- Email: [email protected]
Prostaglandin E2 synthesis and cyclooxygenase expression in abdominal aortic aneurysms
Prostaglandin E 2 synthesis and
cyclooxygenase expression in abdominal aortic aneurysms Dennis R. H o l m e s , M D , William Wester, BA, R o b e r t W. T h o m p s o n , M D , and Jeffrey M. Reilly, M D , St. Louis, Mo.
Purpose: The purpose of this study was to evaluate the expression of prostaglandin E 2 (PGE2) and the two cyclooxygenase isoforms (cox1 and cox2) in human abdominal aortic aneurysm (AAA) tissue. Methods: Ten specimens each of normal aortas and aneurysmal aortas were collected and used for histologic analysis and whole organ culture. An enzyme-linked immunosOrbent assay for PGU, 2 w a s performed on the media from the aortic explant whole organ culture. An immunohistochemical analysis for I~GE2 w a s performed, as was in situ hybridization for cox1 and cox2 on tissue sections. Results: PGE 2 production of AAA specimens was found to be 67,287 -+ 27,303 p g / m l as compared with 1698 -+ 858 p g / m l for normal aortic specimens (p < 0.001). PGE 2 was localized by immunohistochemical analysis to the inflammatory infiltrate in AAAs. Minimal expression was noted in normal aortas. Using in situ hybridization, little expression of cox1 was noted in either the normal or the AAA specimens. Cox2 was expressed by macrophage-like cells within the inflammatory infiltrate of the AAA specimens but was not significantly expressed in the normal aorta. Conclusion: The expression of PGE 2 is associated with the pathogenesis of human AAAs. Its expression is localized to macrophage-like cells within the inflammatory infiltrate and is controlled by the cox2 isoform of cyclooxygenase. Cox2 is, therefore, a potential target for pharmacotherapy ofAAAs. (J Vasc Surg 1997;25:810-5.)
The pathogenesis o f abdominal aortic aneurysms (AAAs) remains poorly understood. The inflammatory response that is associated with AAAs appears to play an important role in the destruction o f the elastin o f the aortic wall that is characteristic o f the disease. 1-4 This inflammatory infiltrate includes B and T lymphocytes, plasma cells, and macrophages. Macrophages in particular have been implicated in this destructive process, and their effects appear to be mediated, at least in part, through the expression o f matrix metalloproteinases (MMPs). s,6 M M P expression in macrophages is dependent on prostaglandin E 2 (PGE2). 7-1° PGE 2 is synthesized from arachidonic acid, and cyclooxygenase is the rate-limiting enzyme in this pathway. There are two isoenzymes o f cyclooxygenase, which are encoded by two independently From the Department of Surgery, Washington UniversitySchool of Medicine, and the John Cochoran Veterans Administration Medical Center. Presented at the Twentieth Annual Meeting of the Midwestern Vascular Surgical Society, St. Louis, Mo., Sep. 27-28, 1996. Reprint requests: JeffreyM. Reilly, MD, Washington University School of Medicine, Department of Surgery, 216 S. Kingshighway, St. Louis, MO 63110. 24/6/79847 810
regulated genes, cyclooxygenase 1 (coxl) and cyclooxygenase 2 (cox2). h a 2 The purpose o f this study was to evaluate the production o f P G E 2 and the differential expression o f coxl and cox2 in AAA tissue and compare this with normal aortic tissue, hypothesizing that P G E 2 and cyclooxygenase are intermediaries involved in the pathogenesis o f AAA. METHODS H u m a n aortic tissues. Full-thiclmcss specimens o f normal infrarenal abdominal aorta that did not have any visible evidence o f atherosclerosis were obtained in the operating r o o m from healthy organ transplantation donors at the time o f organ harvest. Tissue specimens were also obtained from patients who underwent aortic reconstruction procedures for AAAs from the anterior wall o f the mid-infrarenal aorta. All studies were performed with approval by the Washington University School o f Medicine H u man Research Subjects Committee. For each specimen obtained, one portion o f the aortic wall was used for the aortic cultures and an adjacent portion was processed for the histologic studies.
JOURNALOFVASCULARSURGERY Volume25, Number5
Holmes et al.
140 ] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....................................
1001
II
811
[ 7 N ormal
.................
...................................
E ot o. o o o od 141 0 D,.
801 ........................
60
40
20
0
Fig. 1. Raw data for ELISA for PGE2. There were ten specimens in each category. Of note is that PGE z levels in culture media of AAA specimens are in all instances an order of magnitude greater than in normal group. I m m u n o h i s t o c h e m i c a l analysis. Fresh aortic tissue samples were rinsed free o f blood with normal saline solution, fixed in 10% neutral formalin at 4 ° C for 24 hours, processed for routine paraffin embedding, and then sectioned (5 Ixm). The sections were deparaffinized and rehydrated, and endogenous peroxidase activity was quenched by incubation with 0.3% H 2 0 2 at room temperature for 30 minutes. Samples were then incubated with affinity-purified mouse-antihuman PGE 2 (1:500, Monsanto, St. Louis) or 10% horse serum, and immune complexes were detected with immunoperoxidase staining using a Vectastain Elite ldt (Vector Laboratories, Inc., Burlinghame, Calif.). The sections were then counterstained in preparation for the histologic analysis. Aortic o r g a n cultures. Ten specimens each o f normal aorta and aneurysmal aorta were transported to a sterile tissue culture h o o d in cold (4 ° C) Dulbecco's modified Eagle's media supplemerited with 1% bovine serum albumin and antibiotics. Each specimen was divided into 2 mm 2 segments o f full-thickness aortic wall, and these segments were placed into separate wells o f six-tissue culture plates. After 20 minutes to allow adherence to the tissue culture plate, each well was supplemented with 1.5 ml of culture medium, and the plates were incubated at 37 ° C in a humidified 5% CO 2 atmosphere. After 72
hours, the conditioned media was collected and stored at - 2 0 ° C until they were used for enzymelinked immunosorbent assay (ELISA). ELISA. The PGE 2 content in the culture media from aortic explant whole organ culture samples was quantified by competitive-binding indirect ELISA as described. 13 The assay was performed on a total o f 20 specimens (10 normal, 10 AAA). Standards curves were included in each assay using purified human PGE 2. Statistical analysis was performed using a paired t test, with statistical significance being defined as a p value less than 0.05. I n situ hybridization. Probes for in situ hybridization were prepared as described by Stahl-Backdahl et al.l~ Briefly, a 400 base-pair (bp) fragment o f the 3' end o f human cox2 cDNA ( p M O N #23907, Monsanto) and a 240 bp fragment o f the 3' end of human coxl eDNA ( p M O N #23914, Monsanto) were each subcloned in p G E M 5 Z transcription vectors (Promega, Madison, Wis.) and then linearized. The RNA probes were transcribed from the linearized eDNA templates and labeled with digoxigeninl l - U T P under conditions recommended by and with reagents from Boehringer Mannheim (Indianapolis). In situ hybridization was performed as described. 14 Briefly, 5 ixm tissue sections were treated with proteinase K (Sigma Chemical Co., St. Louis)
812
JOURNAL OF VASCULAR SURGERY May 1997
H o l m e s et al.
Fig. 2. Immunohistochemical analysis i:or PGE 2 in AAA tissue. A, Specific staining within macrophage. B, Staining within macrophages and less-intense staining over smooth muscle cells.
and washed in 0.1 m o l / L triethanolamine buffer containing digoxigenin-labeled antisense or sense RNA probes and incubated at 42 ° C for 18 hours in a humidified chamber. After hybridization, the slides were washed under stringent conditions, including treatment with ribonuclease-A. The sections were treated with alkaline phosphatase-conjugated antidigoxigenin antibody and 10% normal sheep serum, followed by 5-bromo-4-chloro-3-indoyl phosphate and nitroblue tetrazolium as chromogens. After development for 4 to 24 hours in the dark, the color reaction was stopped with Milli-Q-water, and the slides were stained with nuclear-fast red for final histologic analysis. RESULTS ELISA performed on the conditioned media from whole organ aortic wall explant cultures AAA showed markedly elevated levels of PGE 2 production as compared with normal aortic specimens (67,287 -+ 27,303 p g / m l vs 1696 _+ 858 p g / m l ; 10 < 0.001). The raw data is depicted in Fig. 1, and this clearly demonstrates both the uniformly low level of PGE 2 production by normal aorta and the markedly elevated levels by aneurysmal aorta. Histo-
logic evaluation demonstrated the inflammatory infiltrate associated with the aortic wall in AAA tissue. Immunohistochemical analysis for PGE 2 demonstrated staining within the inflammatory infiltrate of AAA, with specific staining within macrophage-like cells as well as within some smooth muscle cells (Fig. 2). This staining was similar in all AAA specimens. No PGE 2 antigen staining was detected on the normal aortic specimens. In situ hybridization using a human eDNA cox2 probe showed hybridization within the macrophagelike cells in the inflammatory infiltrate within the aortic wall in all of the AAA specimens (Fig. 3). No qualitative difference was noted between the AAA specimens. In contrast, normal aortic tissue showed little or no hybridization with this probe. There was little or no expression of coxl in either normal or AAA specimens. DISCUSSION
This study documents that AAA tissue in whole organ culture produces significant quantities of PGE 2 and, conversely, shows that normal aortic tissue produces very little PGE 2. In addition, PGE 2 was localized to the inflammatory infiltrate associated
JOURNAL OF VASCULAR SURGERY Volume 25, Number 5
Holmes et aL
8][3
Fig. 3. In situ hybridization for cox2 in AAA tissue. A and B, High-power views of AAAtissue demonstrate the specific staining within macrophage-likecells within inflammatoryinfiltrate.
with AAA. As would be expected on the basis of the organ culture data, no PGE 2 antigen was detected in normal aortic tissue. In addition, in situ hybridization showed specific staining for the cox2 isozyme of cyclooxygenase in macrophage-like cells within the inflammatory infiltrate in AAA specimens. There was no appreciable staining for cox2 in normal aortic tissue, and coxl was not detected in significant quantifies in either AAA or normal aortic tissue. These results suggest that PGE 2 is involved in AAA pathogenesis and that its expression is dependent on the cox2 isozyme in macrophages. These results are not surprising because other studies have implicated macrophages in the pathogenesis of AAA. Macrophages are lmown to play an important role in regulating the turnover ofextracellular matrix (ECM) in both normal and pathologic conditions through the secretion of proteases, including MMPs, protease inhibitors, and cytoldnes. Levels of MMPs, particularly the 72 kD (MMP-2) and 92 kD (MMP-9) gelatinases, which are potent elastases, 15 are elevated in human AAA tissue and are believed to play a role in the pathogenesis oI'AAA. 5,6 MMP-9 has been localized to macrophages in AAAs,6 as has been urokinase, 13,16 which synergisti-
cally enhances the ability of macrophages to destroy ECM. 17 Levels of plasmin, the end product of uroldnase activity and an activator of MMPs, are also elevated in AAA tissue. 18 These studies strongly support the role of macrophages in the pathogenesis of AAA and suggest that their effects are mediated through a synergistic combination of MMPs, urokinase, and plasmin. This study addresses another aspect of the biologic function ofmacrophages within AAA. In other models, MMP expression in macrophages has been shown to be dependent on PGE2. 7q° PGE 2 is synthesized from arachidonic acid, and cyclooxygenase (cox) is the rate-limiting enzyme in this pathway. There are two isozymes ofwclooxygenase, which are encoded by two independently regulated genes, cyclooxygenase-1 (cox1) and cyclooxygenase-2 (cox2). 11,12 The activation of macrophages has been previously correlated with the induction of COX2,19 and our study documents that cox2 has been induced in macrophages within the inflammatory infiltrate of AAA. These data, coupled with those of previous studies, further implicate macrophages in the pathogenesis of AAA. Although the primary event that initiates the for-
814
Holmes et al.
marion of AAAs is currently not known, therapeutically targeting the biologic events--for example, the expression of matrix metalloproteinases, which result in destruction of the aortic wall--could be a means to attenuate aneurysm growth. The data here suggest that blocking cox activity could decrease the expression of MMPs. This intervention is attractive because nonsteroidal antiinflammatory drugs (NSAIDs) that inhibit both coxl and cox2 activity are already in common use. In fact, we have previously documented that indomethacin, a NSAID, attenuates aneurysm growth in an animal model and is associated with decreased expression of MMP~9. 2° However, NSAID therapy is not benign and is associated with untoward side effects, including peptic ulceration, increased bleeding, and nephrotoxicity. These side effects appear to be mediated by coxl, and cox2 appears to be associated with response to injury and inflammarion.21, 22
A recent study performed in rats using NS-398, a selective-cox2 inhibitor, demonstrated that proinflammatory prostaglandin synthesis could be effectively blocked with this drug without the gastric ulcerogenic side effects that were seen with indomethacin in the same model. 23 If the pathogenesis of AAA is in part dependent on PGE 2 and is mediated by cox2 expression in macrophages (as is suggested by the data presented here), then a selective cox2 inhibitor has the potential to prevent or slow the growth of AAA. This intervention is particularly attractive in the setting of small aneurysms. Currently, the management of small aneurysms (less than 4.0 to 4.5 cm in size) is controversial because the risk of rupture is low, but the natural history of most is continued expansion, and more than 80% of small aneurysms eventually rupture or require surgical intervention. If an effective oral therapy were available, many patients with small aneurysms could avoid surgery, which even in the best of circumstances entails morbidity and death for some patients. CONCLUSION We conclude that PGE 2 is associated with AAA pathogenesis. Its expression is controlled by the cox2 isoform of cyclooxygenase, which is expressed by macrophage-like cells within the inflammatory infiltrate associated with the aortic wall in aneurysm disease. We further suggest that cox2 is a potential target for pharmacotherapy of AAA. We gratefully thank Scott Hausen of Monsanto Corp. for providing c D N A probes for c o x l and cox2, Stephen
JOURNAL OF VASCULARSURGERY May 1997
Mnich o f Searle, Inc., for providing the monoclonal antibody for PGE2, and Alexander Shaffer o f Searle, Inc., for assistance with the PGE 2 ELISA.
REFERENCES 1. Tilson MD. Histochemistry of aortic elastin in patients with nonspecific abdominal aortic aneurysmal disease. Arch Surg 1988;123:503-5. 2. Camps JS, Greenhalgh RM, Posell JT. Elastin degradation in abdominal aortic aneurysms. Atherosclerosis 1987;65:13-21. 3. Rizzo RJ, McCarthy WJ, Dixit SN, Lilly MP, Shively VP, Flima WR" Yao JST. Collagen types and matrix protein content in human abdominal aortic aneurysms. J Vasc Surg 1989; 10:365-73. 4. Brophy CM, Reilly JM, Smith W, Tilson MD. The role of inflammation in nonspecific aneurysm disease. Ann Vasc Surg 1991;5:229-33. 5. Herron GS, Umemori I, Wong M, et al, Connective tissue proteinases and inhibitors in abdominal aortic aneurysms. Arterioscler Thromb Vasc Biol 1991;11:1667-77. 6. Thompson RW, Holmes DR, Mertens RA, et al. Production and localization of92-kilodalton gelatinase in abdominal aortic aneurysms: an elastolytic metalloproteinase expressed by aneurysm-infiltrating macrophages. J Clin Invest 1995;96: 318-26. 7. Corcoran ML, Stetler-Stevenson WG, Brown PD, Wahl LM. Interleuldn 4 inhibition ofprostaglandin E 2 synthesis blocks interstitial collagenase and 92-kDa type IV collagenase/gelatinase production by human monocytes. J Biol Chem 1992; 267:515-9. 8. Penfland AP, Shapiro SD, Welgus HG. Agonist-induced expression of tissue inhibitor of metalloproteinases and metalloproteinases by human macrophage is regulated by endogenous prostaglandin E2 synthesis. J Invest Dermatol 1995; 104:52-7. 9. Shapiro SD, Kobayashi DK, Penfland AP, Welgus HG. Induction of macrophage metalloproteinases by extracellular matrix. J Biol Chem 1993;268:8170-5. 10. Arias-Negrete S, Kelley K, Chadee K. Proinflammatory cytokines regulate cyclooxygenase-2 mRNA expression in human macrophages. Biochem Biophys Res Commun 1995;208: 582-9. 11. Lee SH, Soyotla E, Chanmugan P, et al. Selective expression of mitogen-inducible cyclooxygenase in macrophages stimulated with lipopolysacchatide. J Biol Chem 1992;267: 25934-8. 12. Hla T, Nielson K. Human cyclooxygenase-2 cDNA. Proc Natl Acad Sci U S A 1992;89:7384-8. 13. Reilly JM, Sicard GA, Lucore CL. Abnormal expression of plasminogen activators in aortic aneurysmal and occlusive disease. J Vasc Surg 1994;19:865-72. 14. Stahle-Backdahl M, Inoue M, Giudice GJ, Parks WC. 92 kDa gelatinase is produced by eosinophils at the site of blister formation in bullous permphidoid and cleaves the extracellulax domain of the 180 kDa bullous permphidoid autoantigen (type XVII collagen). J Clin Invest 1"994;93:2022-30. 15. Senior RM, Griffin GL, Flizar CJ, et al. Human 92-kilodalton type IV collagenases are elastases. J Biol Chem 1991;266: 7870-5. I6. Schneiderman J, Bordin GM, Engelberg I, et al. Expression offibrinolytic genes in atherosclerotic abdominal aortic aneurysm wall. J Clin Invest 96:639-45.
JOURNALOF VASCULARSURGERY Volume25, Number 5
17. Werb Z, Banda M~, Jones PA. Degradation of connective tissue matrices by macrophages. J Exp Med 1980;152:1537. 18. Jean-Claude J, Newman ICM,Li H, Tilson MD. Possible key role for plasmin in the pathogenesis of abdominal aortic aneurysms. Surgery 1994;116:472-8. 19. Arias-Negrete S, KelleyK, Chadee IC Proinflammatory cytokines regulate cyclooxygenase-2mRNA expression in human macrophages. Biochem Biophys Res Commun 1995;208: 582-9. 20. Holmes DR, Petrinec D, Wester W, Thompson RW, Rdlly JM. Indomethacin prevents elastase-induced abdominal aortic aneurysmsin the rat. J Surg Res 1996;63:305~9. 21. Fletcher BS, Kujubu DA, Perrin DM, Herschman HR. Structure of the mitogen-inducible TIS 10 gene and demonstration
DISCUSSION Dr. William H. Pearce (Chicago, Ill.). I would like to compliment Dr. Holmes on his presentation and this group in St. Louis who have focused for so many years on the pathogenesis ofaneurysms and have contributed significantly to our literature. Inflammation is a part of atherosclerosis and a part of aneurysm disease, and therefore it is not surprising that one would find these mediators of inflammation in the aneurysm specimens. What is interesting and surprising is the magnitude of the difference. The levels were almost 1700 p g / m l in the normal group and more than 67,000 p g / m l in the aneurysm group. What I would like to lmow is what was the level in the group of normal samples? I think none of us will be surprised that a 20-year-old who has not a speck of atherosclerosis would probably have different levels than a 63-year-old who has an aneurysm and atherosclerosis. In a similar argument, do you have any patients that did have atherosclerosis with whom you could make a comparison? Every time I perform these studies and compare the results with those of the atherosclerotic group, I find that much of the significance will disappear. In sum, I think this is a very interesting and provocative study that
Hohag~es et al.
815
that the TIS10-encoded protein is a functional prostaglandin G / H synthase. J Biol Chem 1992;267:4338-44. 22. Xie WL, Chipman JG, Robertson DL, et al. Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing. Proc Natl Acad Sci U S A 1991;88:2692-6. 23. Masferrer JL, ZweifelBS, Manning PT, et al. Selectiveinhibition of inducible wclooxygenase2 in vivo is antiinflammatory and nonulcerogenic. Proc Natl Acad Sci U S A 1994;91: 3228-32.
Submitted Sep. 30, 1996; accepted Dec. 6, 1996.
gives us an insight into a system that can be manipulated, and wouldn't it be interesting in the future if we could treat our patients with aneurysms with nonsteroidal antiinflammatory drugs? I have two questions for you. Could you better characterize the normal group of your patients? And second, in any of your experiments have you performed tissue cultures where you have given the nonsteroidal antiinflammatory drugs and demonstrated a change? Dr. Dennis R. Holmes. The group of normal samples was obtained from organ transplant donors. All of the specimens were examined, and we selected those that displayed no visible evidence of atherosclerosis. Because of the limited samples, we did not perform a complete study ofatheroocclusive samples. However, we did examine two occlusive specimens and found that the level of PGE 2 was elevated in those samples. The level was riot as high as that in the aneurysm group, but it was significantly higher than in normal tissues. In addition, we did culture aneurysm tissue in the presence of indomethacin, and we found that the PGE 2 production was reduced to almost undetectable levels.
cyclooxygenase expression in abdominal aortic aneurysms Dennis R. H o l m e s , M D , William Wester, BA, R o b e r t W. T h o m p s o n , M D , and Jeffrey M. Reilly, M D , St. Louis, Mo.
Purpose: The purpose of this study was to evaluate the expression of prostaglandin E 2 (PGE2) and the two cyclooxygenase isoforms (cox1 and cox2) in human abdominal aortic aneurysm (AAA) tissue. Methods: Ten specimens each of normal aortas and aneurysmal aortas were collected and used for histologic analysis and whole organ culture. An enzyme-linked immunosOrbent assay for PGU, 2 w a s performed on the media from the aortic explant whole organ culture. An immunohistochemical analysis for I~GE2 w a s performed, as was in situ hybridization for cox1 and cox2 on tissue sections. Results: PGE 2 production of AAA specimens was found to be 67,287 -+ 27,303 p g / m l as compared with 1698 -+ 858 p g / m l for normal aortic specimens (p < 0.001). PGE 2 was localized by immunohistochemical analysis to the inflammatory infiltrate in AAAs. Minimal expression was noted in normal aortas. Using in situ hybridization, little expression of cox1 was noted in either the normal or the AAA specimens. Cox2 was expressed by macrophage-like cells within the inflammatory infiltrate of the AAA specimens but was not significantly expressed in the normal aorta. Conclusion: The expression of PGE 2 is associated with the pathogenesis of human AAAs. Its expression is localized to macrophage-like cells within the inflammatory infiltrate and is controlled by the cox2 isoform of cyclooxygenase. Cox2 is, therefore, a potential target for pharmacotherapy ofAAAs. (J Vasc Surg 1997;25:810-5.)
The pathogenesis o f abdominal aortic aneurysms (AAAs) remains poorly understood. The inflammatory response that is associated with AAAs appears to play an important role in the destruction o f the elastin o f the aortic wall that is characteristic o f the disease. 1-4 This inflammatory infiltrate includes B and T lymphocytes, plasma cells, and macrophages. Macrophages in particular have been implicated in this destructive process, and their effects appear to be mediated, at least in part, through the expression o f matrix metalloproteinases (MMPs). s,6 M M P expression in macrophages is dependent on prostaglandin E 2 (PGE2). 7-1° PGE 2 is synthesized from arachidonic acid, and cyclooxygenase is the rate-limiting enzyme in this pathway. There are two isoenzymes o f cyclooxygenase, which are encoded by two independently From the Department of Surgery, Washington UniversitySchool of Medicine, and the John Cochoran Veterans Administration Medical Center. Presented at the Twentieth Annual Meeting of the Midwestern Vascular Surgical Society, St. Louis, Mo., Sep. 27-28, 1996. Reprint requests: JeffreyM. Reilly, MD, Washington University School of Medicine, Department of Surgery, 216 S. Kingshighway, St. Louis, MO 63110. 24/6/79847 810
regulated genes, cyclooxygenase 1 (coxl) and cyclooxygenase 2 (cox2). h a 2 The purpose o f this study was to evaluate the production o f P G E 2 and the differential expression o f coxl and cox2 in AAA tissue and compare this with normal aortic tissue, hypothesizing that P G E 2 and cyclooxygenase are intermediaries involved in the pathogenesis o f AAA. METHODS H u m a n aortic tissues. Full-thiclmcss specimens o f normal infrarenal abdominal aorta that did not have any visible evidence o f atherosclerosis were obtained in the operating r o o m from healthy organ transplantation donors at the time o f organ harvest. Tissue specimens were also obtained from patients who underwent aortic reconstruction procedures for AAAs from the anterior wall o f the mid-infrarenal aorta. All studies were performed with approval by the Washington University School o f Medicine H u man Research Subjects Committee. For each specimen obtained, one portion o f the aortic wall was used for the aortic cultures and an adjacent portion was processed for the histologic studies.
JOURNALOFVASCULARSURGERY Volume25, Number5
Holmes et al.
140 ] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....................................
1001
II
811
[ 7 N ormal
.................
...................................
E ot o. o o o od 141 0 D,.
801 ........................
60
40
20
0
Fig. 1. Raw data for ELISA for PGE2. There were ten specimens in each category. Of note is that PGE z levels in culture media of AAA specimens are in all instances an order of magnitude greater than in normal group. I m m u n o h i s t o c h e m i c a l analysis. Fresh aortic tissue samples were rinsed free o f blood with normal saline solution, fixed in 10% neutral formalin at 4 ° C for 24 hours, processed for routine paraffin embedding, and then sectioned (5 Ixm). The sections were deparaffinized and rehydrated, and endogenous peroxidase activity was quenched by incubation with 0.3% H 2 0 2 at room temperature for 30 minutes. Samples were then incubated with affinity-purified mouse-antihuman PGE 2 (1:500, Monsanto, St. Louis) or 10% horse serum, and immune complexes were detected with immunoperoxidase staining using a Vectastain Elite ldt (Vector Laboratories, Inc., Burlinghame, Calif.). The sections were then counterstained in preparation for the histologic analysis. Aortic o r g a n cultures. Ten specimens each o f normal aorta and aneurysmal aorta were transported to a sterile tissue culture h o o d in cold (4 ° C) Dulbecco's modified Eagle's media supplemerited with 1% bovine serum albumin and antibiotics. Each specimen was divided into 2 mm 2 segments o f full-thickness aortic wall, and these segments were placed into separate wells o f six-tissue culture plates. After 20 minutes to allow adherence to the tissue culture plate, each well was supplemented with 1.5 ml of culture medium, and the plates were incubated at 37 ° C in a humidified 5% CO 2 atmosphere. After 72
hours, the conditioned media was collected and stored at - 2 0 ° C until they were used for enzymelinked immunosorbent assay (ELISA). ELISA. The PGE 2 content in the culture media from aortic explant whole organ culture samples was quantified by competitive-binding indirect ELISA as described. 13 The assay was performed on a total o f 20 specimens (10 normal, 10 AAA). Standards curves were included in each assay using purified human PGE 2. Statistical analysis was performed using a paired t test, with statistical significance being defined as a p value less than 0.05. I n situ hybridization. Probes for in situ hybridization were prepared as described by Stahl-Backdahl et al.l~ Briefly, a 400 base-pair (bp) fragment o f the 3' end o f human cox2 cDNA ( p M O N #23907, Monsanto) and a 240 bp fragment o f the 3' end of human coxl eDNA ( p M O N #23914, Monsanto) were each subcloned in p G E M 5 Z transcription vectors (Promega, Madison, Wis.) and then linearized. The RNA probes were transcribed from the linearized eDNA templates and labeled with digoxigeninl l - U T P under conditions recommended by and with reagents from Boehringer Mannheim (Indianapolis). In situ hybridization was performed as described. 14 Briefly, 5 ixm tissue sections were treated with proteinase K (Sigma Chemical Co., St. Louis)
812
JOURNAL OF VASCULAR SURGERY May 1997
H o l m e s et al.
Fig. 2. Immunohistochemical analysis i:or PGE 2 in AAA tissue. A, Specific staining within macrophage. B, Staining within macrophages and less-intense staining over smooth muscle cells.
and washed in 0.1 m o l / L triethanolamine buffer containing digoxigenin-labeled antisense or sense RNA probes and incubated at 42 ° C for 18 hours in a humidified chamber. After hybridization, the slides were washed under stringent conditions, including treatment with ribonuclease-A. The sections were treated with alkaline phosphatase-conjugated antidigoxigenin antibody and 10% normal sheep serum, followed by 5-bromo-4-chloro-3-indoyl phosphate and nitroblue tetrazolium as chromogens. After development for 4 to 24 hours in the dark, the color reaction was stopped with Milli-Q-water, and the slides were stained with nuclear-fast red for final histologic analysis. RESULTS ELISA performed on the conditioned media from whole organ aortic wall explant cultures AAA showed markedly elevated levels of PGE 2 production as compared with normal aortic specimens (67,287 -+ 27,303 p g / m l vs 1696 _+ 858 p g / m l ; 10 < 0.001). The raw data is depicted in Fig. 1, and this clearly demonstrates both the uniformly low level of PGE 2 production by normal aorta and the markedly elevated levels by aneurysmal aorta. Histo-
logic evaluation demonstrated the inflammatory infiltrate associated with the aortic wall in AAA tissue. Immunohistochemical analysis for PGE 2 demonstrated staining within the inflammatory infiltrate of AAA, with specific staining within macrophage-like cells as well as within some smooth muscle cells (Fig. 2). This staining was similar in all AAA specimens. No PGE 2 antigen staining was detected on the normal aortic specimens. In situ hybridization using a human eDNA cox2 probe showed hybridization within the macrophagelike cells in the inflammatory infiltrate within the aortic wall in all of the AAA specimens (Fig. 3). No qualitative difference was noted between the AAA specimens. In contrast, normal aortic tissue showed little or no hybridization with this probe. There was little or no expression of coxl in either normal or AAA specimens. DISCUSSION
This study documents that AAA tissue in whole organ culture produces significant quantities of PGE 2 and, conversely, shows that normal aortic tissue produces very little PGE 2. In addition, PGE 2 was localized to the inflammatory infiltrate associated
JOURNAL OF VASCULAR SURGERY Volume 25, Number 5
Holmes et aL
8][3
Fig. 3. In situ hybridization for cox2 in AAA tissue. A and B, High-power views of AAAtissue demonstrate the specific staining within macrophage-likecells within inflammatoryinfiltrate.
with AAA. As would be expected on the basis of the organ culture data, no PGE 2 antigen was detected in normal aortic tissue. In addition, in situ hybridization showed specific staining for the cox2 isozyme of cyclooxygenase in macrophage-like cells within the inflammatory infiltrate in AAA specimens. There was no appreciable staining for cox2 in normal aortic tissue, and coxl was not detected in significant quantifies in either AAA or normal aortic tissue. These results suggest that PGE 2 is involved in AAA pathogenesis and that its expression is dependent on the cox2 isozyme in macrophages. These results are not surprising because other studies have implicated macrophages in the pathogenesis of AAA. Macrophages are lmown to play an important role in regulating the turnover ofextracellular matrix (ECM) in both normal and pathologic conditions through the secretion of proteases, including MMPs, protease inhibitors, and cytoldnes. Levels of MMPs, particularly the 72 kD (MMP-2) and 92 kD (MMP-9) gelatinases, which are potent elastases, 15 are elevated in human AAA tissue and are believed to play a role in the pathogenesis oI'AAA. 5,6 MMP-9 has been localized to macrophages in AAAs,6 as has been urokinase, 13,16 which synergisti-
cally enhances the ability of macrophages to destroy ECM. 17 Levels of plasmin, the end product of uroldnase activity and an activator of MMPs, are also elevated in AAA tissue. 18 These studies strongly support the role of macrophages in the pathogenesis of AAA and suggest that their effects are mediated through a synergistic combination of MMPs, urokinase, and plasmin. This study addresses another aspect of the biologic function ofmacrophages within AAA. In other models, MMP expression in macrophages has been shown to be dependent on PGE2. 7q° PGE 2 is synthesized from arachidonic acid, and cyclooxygenase (cox) is the rate-limiting enzyme in this pathway. There are two isozymes ofwclooxygenase, which are encoded by two independently regulated genes, cyclooxygenase-1 (cox1) and cyclooxygenase-2 (cox2). 11,12 The activation of macrophages has been previously correlated with the induction of COX2,19 and our study documents that cox2 has been induced in macrophages within the inflammatory infiltrate of AAA. These data, coupled with those of previous studies, further implicate macrophages in the pathogenesis of AAA. Although the primary event that initiates the for-
814
Holmes et al.
marion of AAAs is currently not known, therapeutically targeting the biologic events--for example, the expression of matrix metalloproteinases, which result in destruction of the aortic wall--could be a means to attenuate aneurysm growth. The data here suggest that blocking cox activity could decrease the expression of MMPs. This intervention is attractive because nonsteroidal antiinflammatory drugs (NSAIDs) that inhibit both coxl and cox2 activity are already in common use. In fact, we have previously documented that indomethacin, a NSAID, attenuates aneurysm growth in an animal model and is associated with decreased expression of MMP~9. 2° However, NSAID therapy is not benign and is associated with untoward side effects, including peptic ulceration, increased bleeding, and nephrotoxicity. These side effects appear to be mediated by coxl, and cox2 appears to be associated with response to injury and inflammarion.21, 22
A recent study performed in rats using NS-398, a selective-cox2 inhibitor, demonstrated that proinflammatory prostaglandin synthesis could be effectively blocked with this drug without the gastric ulcerogenic side effects that were seen with indomethacin in the same model. 23 If the pathogenesis of AAA is in part dependent on PGE 2 and is mediated by cox2 expression in macrophages (as is suggested by the data presented here), then a selective cox2 inhibitor has the potential to prevent or slow the growth of AAA. This intervention is particularly attractive in the setting of small aneurysms. Currently, the management of small aneurysms (less than 4.0 to 4.5 cm in size) is controversial because the risk of rupture is low, but the natural history of most is continued expansion, and more than 80% of small aneurysms eventually rupture or require surgical intervention. If an effective oral therapy were available, many patients with small aneurysms could avoid surgery, which even in the best of circumstances entails morbidity and death for some patients. CONCLUSION We conclude that PGE 2 is associated with AAA pathogenesis. Its expression is controlled by the cox2 isoform of cyclooxygenase, which is expressed by macrophage-like cells within the inflammatory infiltrate associated with the aortic wall in aneurysm disease. We further suggest that cox2 is a potential target for pharmacotherapy of AAA. We gratefully thank Scott Hausen of Monsanto Corp. for providing c D N A probes for c o x l and cox2, Stephen
JOURNAL OF VASCULARSURGERY May 1997
Mnich o f Searle, Inc., for providing the monoclonal antibody for PGE2, and Alexander Shaffer o f Searle, Inc., for assistance with the PGE 2 ELISA.
REFERENCES 1. Tilson MD. Histochemistry of aortic elastin in patients with nonspecific abdominal aortic aneurysmal disease. Arch Surg 1988;123:503-5. 2. Camps JS, Greenhalgh RM, Posell JT. Elastin degradation in abdominal aortic aneurysms. Atherosclerosis 1987;65:13-21. 3. Rizzo RJ, McCarthy WJ, Dixit SN, Lilly MP, Shively VP, Flima WR" Yao JST. Collagen types and matrix protein content in human abdominal aortic aneurysms. J Vasc Surg 1989; 10:365-73. 4. Brophy CM, Reilly JM, Smith W, Tilson MD. The role of inflammation in nonspecific aneurysm disease. Ann Vasc Surg 1991;5:229-33. 5. Herron GS, Umemori I, Wong M, et al, Connective tissue proteinases and inhibitors in abdominal aortic aneurysms. Arterioscler Thromb Vasc Biol 1991;11:1667-77. 6. Thompson RW, Holmes DR, Mertens RA, et al. Production and localization of92-kilodalton gelatinase in abdominal aortic aneurysms: an elastolytic metalloproteinase expressed by aneurysm-infiltrating macrophages. J Clin Invest 1995;96: 318-26. 7. Corcoran ML, Stetler-Stevenson WG, Brown PD, Wahl LM. Interleuldn 4 inhibition ofprostaglandin E 2 synthesis blocks interstitial collagenase and 92-kDa type IV collagenase/gelatinase production by human monocytes. J Biol Chem 1992; 267:515-9. 8. Penfland AP, Shapiro SD, Welgus HG. Agonist-induced expression of tissue inhibitor of metalloproteinases and metalloproteinases by human macrophage is regulated by endogenous prostaglandin E2 synthesis. J Invest Dermatol 1995; 104:52-7. 9. Shapiro SD, Kobayashi DK, Penfland AP, Welgus HG. Induction of macrophage metalloproteinases by extracellular matrix. J Biol Chem 1993;268:8170-5. 10. Arias-Negrete S, Kelley K, Chadee K. Proinflammatory cytokines regulate cyclooxygenase-2 mRNA expression in human macrophages. Biochem Biophys Res Commun 1995;208: 582-9. 11. Lee SH, Soyotla E, Chanmugan P, et al. Selective expression of mitogen-inducible cyclooxygenase in macrophages stimulated with lipopolysacchatide. J Biol Chem 1992;267: 25934-8. 12. Hla T, Nielson K. Human cyclooxygenase-2 cDNA. Proc Natl Acad Sci U S A 1992;89:7384-8. 13. Reilly JM, Sicard GA, Lucore CL. Abnormal expression of plasminogen activators in aortic aneurysmal and occlusive disease. J Vasc Surg 1994;19:865-72. 14. Stahle-Backdahl M, Inoue M, Giudice GJ, Parks WC. 92 kDa gelatinase is produced by eosinophils at the site of blister formation in bullous permphidoid and cleaves the extracellulax domain of the 180 kDa bullous permphidoid autoantigen (type XVII collagen). J Clin Invest 1"994;93:2022-30. 15. Senior RM, Griffin GL, Flizar CJ, et al. Human 92-kilodalton type IV collagenases are elastases. J Biol Chem 1991;266: 7870-5. I6. Schneiderman J, Bordin GM, Engelberg I, et al. Expression offibrinolytic genes in atherosclerotic abdominal aortic aneurysm wall. J Clin Invest 96:639-45.
JOURNALOF VASCULARSURGERY Volume25, Number 5
17. Werb Z, Banda M~, Jones PA. Degradation of connective tissue matrices by macrophages. J Exp Med 1980;152:1537. 18. Jean-Claude J, Newman ICM,Li H, Tilson MD. Possible key role for plasmin in the pathogenesis of abdominal aortic aneurysms. Surgery 1994;116:472-8. 19. Arias-Negrete S, KelleyK, Chadee IC Proinflammatory cytokines regulate cyclooxygenase-2mRNA expression in human macrophages. Biochem Biophys Res Commun 1995;208: 582-9. 20. Holmes DR, Petrinec D, Wester W, Thompson RW, Rdlly JM. Indomethacin prevents elastase-induced abdominal aortic aneurysmsin the rat. J Surg Res 1996;63:305~9. 21. Fletcher BS, Kujubu DA, Perrin DM, Herschman HR. Structure of the mitogen-inducible TIS 10 gene and demonstration
DISCUSSION Dr. William H. Pearce (Chicago, Ill.). I would like to compliment Dr. Holmes on his presentation and this group in St. Louis who have focused for so many years on the pathogenesis ofaneurysms and have contributed significantly to our literature. Inflammation is a part of atherosclerosis and a part of aneurysm disease, and therefore it is not surprising that one would find these mediators of inflammation in the aneurysm specimens. What is interesting and surprising is the magnitude of the difference. The levels were almost 1700 p g / m l in the normal group and more than 67,000 p g / m l in the aneurysm group. What I would like to lmow is what was the level in the group of normal samples? I think none of us will be surprised that a 20-year-old who has not a speck of atherosclerosis would probably have different levels than a 63-year-old who has an aneurysm and atherosclerosis. In a similar argument, do you have any patients that did have atherosclerosis with whom you could make a comparison? Every time I perform these studies and compare the results with those of the atherosclerotic group, I find that much of the significance will disappear. In sum, I think this is a very interesting and provocative study that
Hohag~es et al.
815
that the TIS10-encoded protein is a functional prostaglandin G / H synthase. J Biol Chem 1992;267:4338-44. 22. Xie WL, Chipman JG, Robertson DL, et al. Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing. Proc Natl Acad Sci U S A 1991;88:2692-6. 23. Masferrer JL, ZweifelBS, Manning PT, et al. Selectiveinhibition of inducible wclooxygenase2 in vivo is antiinflammatory and nonulcerogenic. Proc Natl Acad Sci U S A 1994;91: 3228-32.
Submitted Sep. 30, 1996; accepted Dec. 6, 1996.
gives us an insight into a system that can be manipulated, and wouldn't it be interesting in the future if we could treat our patients with aneurysms with nonsteroidal antiinflammatory drugs? I have two questions for you. Could you better characterize the normal group of your patients? And second, in any of your experiments have you performed tissue cultures where you have given the nonsteroidal antiinflammatory drugs and demonstrated a change? Dr. Dennis R. Holmes. The group of normal samples was obtained from organ transplant donors. All of the specimens were examined, and we selected those that displayed no visible evidence of atherosclerosis. Because of the limited samples, we did not perform a complete study ofatheroocclusive samples. However, we did examine two occlusive specimens and found that the level of PGE 2 was elevated in those samples. The level was riot as high as that in the aneurysm group, but it was significantly higher than in normal tissues. In addition, we did culture aneurysm tissue in the presence of indomethacin, and we found that the PGE 2 production was reduced to almost undetectable levels.