Nitric Oxide System in Needle-InducedTransmyocardial RevascularizationTakayuki Saito, MD, PhD, Marc P. Pelletier, MD, Hani Shennib, MD, andAdel Giaid, PhDDepartments of Pathology and Surgery, The Montreal General Hospital and McGill University, Montreal, Quebec, Canada

Background. Nitric oxide (NO) promotes endothelialproliferation and migration, essential for angiogenesis.The purpose of this study was to determine the cellularexpression of inducible and endothelial nitric oxidesynthases (iNOS and eNOS) in an ischemic cardiomyop-athy animal model of needle-induced transmyocardialrevascularization (TMR).

Methods. Myocardial infarction was created in rats byligating the left coronary artery, and animals were di-vided into two groups: no-TMR group (served as control)and TMR group (underwent concomitant TMR by thecreation of six transmural channels with a 25-gaugeneedle in the ischemic area). Rats were sacrificed atintervals of 1, 2, 4, and 8 weeks. Immunohistochemistryusing specific antisera was performed for iNOS, eNOS,and endothelial cell marker factor VIII. Vascular densityand positive staining area with either iNOS or eNOSwere assessed in the infarcted myocardium.

Results. Vascular density in the infarcted myocardiumwas significantly increased in the TMR group (p < 0.001).The positive staining area for iNOS and the intensity ofiNOS immunoreactivity in cardiomyocytes, vascular en-dothelium, and macrophages were significantly greaterin the TMR group (p < 0.05). However, these differenceswere seen only in the first 2 weeks after TMR. There wasno significant difference in the expression of eNOSbetween groups.

Conclusions. A mechanical injury using needle punc-ture in an ischemic myocardium increased vascular den-sity and is associated with increased expression of myo-cardial iNOS. Increased production of NO derived fromiNOS may contribute to the angiogenic response of TMR.

(Ann Thorac Surg 2001;72:129–36)© 2001 by The Society of Thoracic Surgeons

The possibility of restoring adequate myocardial bloodflow in patients with advanced coronary artery dis-

ease by creating intramyocardial channels is being cur-rently investigated by various groups [1]. Initial trialssuggest that transmyocardial revascularization (TMR)may be effective in revascularizing the myocardiumwhen conventional therapy is not possible. Many inves-tigators have also shown the efficacies of TMR in exper-imental animal models [2–4]. However, the mechanismsby which TMR achieves its therapeutic effects and thechoice of method for creating transmural channels hasnot been fully elucidated. It is generally believed thatTMR induces angiogenesis and improves myocardialcollateral circulation through the processes of tissueinjury and wound healing. During the inflammatory andproliferating phases of wound healing, there is signifi-cant upregulation of various growth factors that promoteangiogenesis and neovascularization. Angiogenesis,which consists of endothelial cell proliferation and mi-gration, remodeling of extracellular matrix, and tubularstructure formation, is a pathophysiological event occur-ring after tissue injury, or in tumor growth and metasta-

sis [5]. This process is tightly regulated by the actions ofangiogenic cytokines such as vascular endothelial growthfactor (VEGF) [6], basic fibroblast growth factor (bFGF)[7], angiopoietin-1 [8], and transforming growth factor-beta (TGFb). It is well known that these angiogeniccytokines are induced in response to ischemia and con-tribute to regulate neovascularization. We have previ-ously demonstrated that angiogenesis, after mechanicalinjury, in an infarcted myocardium is associated with asignificant increase in the expression of VEGF, bFGF, andTGFb [2–4].

It has been well recognized that activation of theinducible form of nitric oxide (NO) synthase (iNOS) byvarious insults including myocardial ischemia [9] re-leases excessive amounts of NO from macrophages andcardiomyocytes. Diverse effects of NO involve not onlyinducing vascular smooth muscle relaxation but alsoinhibiting platelet aggregation [10], leukocyte adherenceto the endothelium [11], and inducing endothelial cellproliferation [12] and migration. Guo and associates [13]demonstrated that the exogenous administration of NOdonor significantly stimulated endothelial proliferationin vitro. Furthermore, production of NO is regulated byseveral angiogenic growth factors such as VEGF andTGFb [14, 15].

Therefore, we hypothesized that mechanical injury inthe ischemic myocardium may be associated with in-

Presented at the Thirty-sixth Annual Meeting of The Society of ThoracicSurgeons, Fort Lauderdale, FL, Jan 31–Feb 2, 2000.

Address reprint requests to Dr Giaid, The Montreal General Hospital,McGill University, 1650 Cedar Ave, Suite L3-314, Montreal, Quebec, H3G1A4, Canada; e-mail: adel.giaid@mcgill.ca.

© 2001 by The Society of Thoracic Surgeons 0003-4975/01/$20.00Published by Elsevier Science Inc PII S0003-4975(01)02687-X

duced expression of iNOS, which in conjunction withother factors, may contribute to the process of angiogen-esis associated with TMR.

Material and Methods

AnimalsLewis male rats weighing 250 to 300 g (Charles RiverLaboratory, Wilmington, MA) were used in this study.All animal work was performed in accordance withinstitutional guidelines, which is in compliance with the“Guide for the Care and Use of Laboratory Animals,”published by the National Institutes of Health (NIHpublication 85-23, revised 1985). Animals were dividedinto: no-TMR group (n 5 25), underwent ligation of theleft coronary artery only; TMR group (n 5 28), underwentligation of the left coronary artery along with TMR at thesame operation; sham group (n 5 6), underwent leftthoracotomy only.

Experimental Animal ModelLeft ventricular free-wall myocardial infarction was in-duced as described previously [16] with a slight modifi-cation. In brief, each rat was anesthetized with enflurane,intubated with a 14-gauge intravenous catheter, andmechanically ventilated with room air by use of a smallrodent ventilator (Harvard, South Natick, MA) at a rate of90 cycles per minute and a tidal volume of 1 mL/100 gbody weight. A left thoracotomy was then performed inthe fourth intercostal space. After the pericardium wasincised, the proximal portion of the left coronary arterywas ligated with one suture of 5-0 silk. In the TMR group,transmyocardial punctures were performed with a 25-gauge needle. Six transmural channels were created inthe distribution of the left coronary artery, using cautionto avoid puncturing any large coronary veins. Hemosta-sis was achieved with slight pressure in all cases. In bothgroups, the chest was closed in three layers of 4-0 Vicryl(Ethicon, Somerville, NJ). After surgery, rats were al-lowed to recover and postoperative pain was controlledwith subcutaneous injections of buprinorphine (0.1 to0.5 mg/kg). Both TMR and no-TMR groups were thendivided into four subgroups, which were sacrificed atperiods of 1, 2, 4, and 8 weeks postoperatively. Apartfrom the coronary artery ligation, sham rats underwentan identical procedure and were sacrificed at 1 weekpostoperatively.

Tissue PreparationAt sacrifice, animals were again anesthetized, intubated,and ventilated as described above. After the sternum wasremoved, isolation and cannulation of the aortic root witha 20-gauge intravenous cannula was employed. Beatinghearts were arrested and fixed in diastole with 4% para-formaldehyde, then stored at 4°C. After 24 hours, thehearts were washed and stored in phosphate-bufferedsaline (PBS) solution containing 15% sucrose, again at4°C.

ImmunohistochemistryCryostat sections of tissue were immunostained withantisera to eNOS, and iNOS with a modification of theavidin-biotin-peroxidase methods [17]. Briefly, sectionswere incubated serially with the following solutions: (1)2% hydrogen peroxide for 30 minutes to block endoge-nous peroxide activity; (2) 0.3% Triton-X 100 for 15minutes to permeabilize the membrane; (3) 10% normalgoat serum for 60 minutes to reduce nonspecific bindingof the antiserum; (4) primary antisera for 16 hours at 4°C;(5) biotinylated goat anti-mouse or goat anti-rabbit IgGfor 45 minutes; and (6) avidin-biotinylated horseradishperoxidase complex (Vectastain; Vector Laboratories,Burlingame, CA) for 45 minutes. Immunoreactive siteswere visualized by incubation with 0.025% 3,3-diaminobenzidine and 0.01% hydrogen peroxide for 3minutes. PBS, pH 7.4, was used to dilute each solutionand to wash the sections three times between each step.Antiserum to factor VIII (the endothelial cell marker vonWillebrand factor) was also used.

Morphometric StudiesAngiogenesis was assessed by counting the number ofvessels per high-power field (HPF; 3400). Vessels weredefined as round structures with a certain lumen that islined by cells staining positively to factor VIII [2]. Proteinexpression was quantified by measuring the area oftissue sections positively stained for iNOS or eNOS ineach HPF (3400). Measurements were performed using10 sampling sites per infarcted zone. Results were quan-titated as square millimeters per square millimeter ofmyocardial tissue [2]. To compare the immunoreactivityfor iNOS and eNOS in myocytes, macrophages, endocar-dium, and vascular endothelium of the infarcted myocar-dium, positive stained dots in HPF (3400) were high-lighted and the density of dots was calculated afterencircling these cells on the computer screen. The aver-age of 10 samples from each cell type in the infarctedmyocardium was quantitated as dots per square micro-meter. It is important to mention that although most ofthe animals used in this study were included in ourprevious study [2], we have taken different slices of themyocardium and increased the number of fields exam-ined in each section (n 5 10). This explains the differencein vascular density in both studies.

Statistical AnalysisAll data were presented as mean 6 SD. Two-way repeat-ed-measures analysis of variance and Student’s t testwere used. A correlation between the positive stainingarea with iNOS and vascular density was assessed bylinear regression analysis. Values of p less than 0.05 wereconsidered significant.

Results

Vascular DensityThe number of capillaries per HPF was greater in theinfarcted area of both TMR and no-TMR groups when

130 SAITO ET AL Ann Thorac SurgiNOS IN NEEDLE TMR 2001;72:129–36

compared with sham group (Fig 1). When comparedbetween experimental groups, vascular density was sig-nificantly increased by TMR treatment, particularly at 1and 2 weeks postoperatively (p , 0.001). After 4 weeks,there was no significant difference in the number ofvessels between the two experimental groups. In bothTMR and no-TMR groups, the number of vessels wasgreatest at 1 week and decreased progressively until theeighth week.

Inducible Nitric Oxide SynthaseIn the sham group, weak immunoreactivity with iNOSwas seen in the cardiomyocytes, endocardium, and en-dothelium. In the experimental myocardial infarctionmodel, the localization of iNOS immunoreactivity wasseen in the cardiomyocytes, endocardium, and endothe-lium of intramyocardial vessels in both infarcted andnoninfarcted myocardium, and macrophages in the in-farcted myocardium. The expression of iNOS was moreprominent in the infarcted myocardium of the TMR andthe no-TMR groups when compared with the shamgroup (Figs 2A–F and 3A–D). Intensity of iNOS immuno-reactivity in cardiomyocytes, endothelium, and macro-phages was significantly higher in the TMR groups ateither 1 or 2 weeks postoperatively than that of theno-TMR group (p , 0.05). The positive immunoreactivearea for iNOS in the infarcted myocardium was signifi-cantly greater in the TMR than in the no-TMR group at 1and 2 weeks postoperatively (Fig 4A, p , 0.05). However,analysis of variance (ANOVA) revealed no significantdifference in the time course of both positive stainingarea and intensity of immunoreactivity between TMRand no-TMR groups. No iNOS immunoreactivity wasseen in the negative control sections.

Endothelial Nitric Oxide SynthaseIn the sham group, immunoreactivity with eNOS wasseen in the endocardium and endothelium, and to someextent in the cardiomyocytes. In the experimental myo-cardial infarction model, localization of eNOS immuno-reactivity was seen in the cardiomyocytes, endocardium,and endothelium of intramyocardial vessels in both in-farcted and noninfarcted myocardium. There was nosignificant difference in eNOS immunoreactivity in car-

diomyocytes, endothelium, and endocardium among allgroups (Figs 2G–H and 3E–H). Positive immunoreactivearea for eNOS was significantly reduced in the infarctedmyocardium of both TMR and no-TMR groups comparedwith the sham group (Fig 4B). ANOVA revealed therewas no significant difference in the time course of bothpositive staining area and intensity of immunoreactivitybetween TMR and no-TMR groups. No eNOS immuno-staining was seen in the negative control sections.

Correlation StudiesFigure 5 shows the relationship between the positiveimmunoreactive area for iNOS and vascular density inthe infarcted myocardium of both TMR and no-TMRgroups. Vascular density was lineally correlated withiNOS immunoreactive area in both experimental groups(TMR group: R2 5 0.89, p , 0.0001, no-TMR group; R2 50.732, p , 0.0001).

Comment

In this study, we have demonstrated that myocardialischemia promotes neovascularization and this is en-hanced by the additional mechanical injury in this area.These results suggest that some component of the nee-dle-induced TMR procedure leads to the formation ofnew microvessels. However, the effect of TMR on neo-vascularization was seen only within the first 2 weekspostprocedure. Our current observations are somewhatconsistent with recent clinical studies showing no im-provement in myocardial blood perfusion at intermedi-ate term after TMR [18, 19]. Decreasing vascular densityover time may be explained by the well-known process ofwound healing. With time, the inflammatory processmay lead to increased scar tissue formation via fibroblastproliferation and collagen deposition. Dense scar tissuemay allow fewer vessels to survive. Moreover, becausethe surface area of surviving cardiomyocytes is less than10% over total infarcted myocardium of the same animalmodel (our unpublished data), fewer numbers of vesselsmay be required in this area to supply blood to theremaining cardiomyocytes.

Induction of NOS in cells of the cardiovascular systemin response to a variety of stimuli, and its possiblepathologic role, have been well documented. NO hasspecifically been shown to have angiogenic actions invitro [20]. In the present study, we have demonstratedthat myocardial ischemia significantly induced iNOSexpression particularly in cardiomyocytes, endocardialand vascular endothelium, and macrophages. Interest-ingly, both positive immunoreactive area for iNOS andintensity of the immunostaining were significantly in-creased in the infarcted myocardium of the TMR group.The most significant difference in the expression of iNOSwas seen in only first 2 weeks after the TMR procedure.Also, the number of vessels in the infarcted myocardiumwas lineally correlated with the positive immunoreactivearea for iNOS in the infarcted myocardium. This phe-nomenon was true for both TMR and no-TMR groups.Because we show that myocardial ischemia induced

Fig 1. Vascular density was expressed as number of vessels per HPF(3400) in the myocardium of sham and the two experimentalgroups (no-TMR and TMR groups); *p , 0.001.

131Ann Thorac Surg SAITO ET AL2001;72:129–36 iNOS IN NEEDLE TMR

Fig 2. Immunostaining for iNOS (A–E) and eNOS (F–H) in the myocardium. (A) Strong expression of iNOS in the infarcted myocardium of TMRgroup at 1 week after TMR procedure. Arrow indicates endocardial endothelium. (B) Higher magnification of the heart in A showing intense expres-sion of iNOS in the vascular endothelium (arrows) and macrophages (arrowheads). (C) Also a higher magnification of the heart in A and showsabundant expression of iNOS in the cardiomyocytes (arrowheads) and endocardial endothelium (arrow). (D) Little expression of iNOS in the in-farcted myocardium of the TMR group at 8 weeks after the procedure. Arrow indicates endocardial endothelium. (E) Little expression of iNOS in thenoninfarcted myocardium of the sham group. (F) Immunostaining with eNOS in the infarcted myocardium of the no-TMR group at 1 week after sur-gery. (G) Higher magnification of the heart in F and shows moderate staining with eNOS in the cardiomyocytes (arrowheads) and endocardial endothelium(arrow). (H) Diffuse immunoreactivity for eNOS in the myocytes of the sham group. Magnifications: A, D, and F (3200); B, C, E, G, and H (3400).

132 SAITO ET AL Ann Thorac SurgiNOS IN NEEDLE TMR 2001;72:129–36

Fig 3. The intensity of iNOS (A–D) and eNOS immunoreactivity (E–G) in the cardiomyocytes, endocardium, endothelium, and macrophagesin the infarcted myocardium of no-TMR and TMR groups and in the noninfarcted myocardium of sham group; *p , 0.05.

133Ann Thorac Surg SAITO ET AL2001;72:129–36 iNOS IN NEEDLE TMR

iNOS expression, and we know from previous literaturethat the resultant increase in NO production can haveangiogenic effects, it is reasonable to speculate that NOcontributes to the increased vascular density after TMR.

Another isoform of NOS, eNOS, is constitutively pro-duced in endothelial cells as well as other cell types, andplays an important role in maintaining normal vasculartone. Murohara and associates have recently demon-strated deteriorated angiogenic response to hindlimbischemic model in eNOS gene knockout mice [21]. There-fore, eNOS has been considered to be critical for angio-genesis. However, we did not detect an increase in eNOSexpression in the infarcted myocardium. In other words,neither myocardial ischemia nor additional mechanicalinjury could influence the pattern of eNOS expression.On the contrary, we observed a reduction in the area ofeNOS immunostaining in the TMR and no-TMR groups.This might be explained by loss of immunoreactivemyocytes in the infarcted myocardium. Our results arealso supported by the findings of Wildhirt and associates,who showed that iNOS but not eNOS activity is signifi-cantly increased in the infarcted myocardium [22]. Takeninto consideration that iNOS produces high levels of NO,which may have serious cytotoxic and negative inotropiceffects on the myocardium [23], and that eNOS producesphysiological levels of NO, future studies should includethe use of selective iNOS inhibitor, and the use ofcombined TMR and eNOS gene transfer therapy.

We have previously reported that neovessel formationafter TMR is associated with the increased expression ofseveral growth factors such as VEGF, bFGF, and TGFb[2–4]. Among these cytokines, VEGF is an importantregulator of endothelial cell proliferation, migration, andpermeability, and is secreted from tumor and hypoxiccells [24]. Myocardial ischemia has previously beenshown to induce expression of VEGF mRNA in cardiacmyocytes and vascular smooth muscle cells [25]. VEGFbinds to high-affinity tyrosine kinase receptors [26] onvascular endothelial cells, which leads to the release ofNO through the activation of both eNOS and iNOS [14].Interestingly, a recent study has shown that nonselectiveinhibition of NOS prevents VEGF-mediated angiogene-sis [27]. Moreover, we have recently shown that expres-sion of VEGF was highest at 1 week after coronary arteryligation, which is consistent with the current findings ofincreased vascular density and iNOS expression at thesame time. In addition to VEGF, TGFb also promotesproliferation of endothelial cells in vitro, and induces theformation of capillary-like tubes of bovine microvascularendothelial cells [28]. TGFb has a potent chemoattractanteffect for monocytes and fibroblasts [29, 30]. These cellscan then secret angiogenic molecules such as VEGF andbFGF, which act directly on endothelial cells. Consistentwith our present finding, it has recently been shown that

Fig 4. Positive immunoreactive area for iNOS (A) and eNOS (B) inthe infarcted myocardium of the TMR and the non-TMR groups andin the noninfarcted myocardium of the sham group; *p , 0.05.

Fig. 5. Correlation between iNOS immunoreactive area and vasculardensity in the infarcted myocardium of the TMR (A) and no-TMRgroups (B).

134 SAITO ET AL Ann Thorac SurgiNOS IN NEEDLE TMR 2001;72:129–36

myocardial mRNA and protein levels for iNOS and TGFbwere increased in the postischemic myocardium [31].TGFb has been shown to inhibit iNOS expression in vitroand to be upregulated as a negative feedback response toelevated levels of NO [15, 32, 33]. Therefore, an increaseof NO may lead to an abundant production of TGFb andmay consequently accelerate vessel formation. Interest-ingly, the highest expression of both iNOS and TGFb wasalso seen during the first week after ligation, particularlyin the TMR group [2]. These results suggest a possibleinteractive pathway between iNOS, VEGF, and TGFb inthe ischemic myocardium.

The present study has several methodological limita-tions. First, our experimental TMR procedure involvedneedle punctures for an acute ischemic myocardium,which is no relation to the clinical scenario. In addition,although we have previously compared the efficacy ofneovascularization by a needle and a laser [3], in thisstudy, we did not use a laser TMR. To mimic clinicalsituations, therefore, mechanical punctures by either aneedle or a laser have to be carried out in the nonscararea adjacent to the infarction area. Second, in ourpresent study, we measured only immunoreactivity forboth iNOS and eNOS in the infarcted myocardium.Because availability of substrates for these enzymes andmicroenvironment of the ischemic myocardium such ascalcium concentration and pH could influence the enzy-matic activity and NO production, direct measurementsof NOS activity and NO concentration might be signifi-cantly important to support our present results. Third,although we compared the number of vessels in theinfarcted myocardium and showed an augmented angio-genesis by needle punctures, we have not measuredmyocardial blood flow, which might be more importantand relevant to clinical outcome. Finally, we have notshown direct interaction between iNOS and angiogene-sis. Therefore, it is important to use a selective iNOSinhibitor to elucidate the role of this enzyme on angio-genesis in vivo. If our hypothesis is pathophysiologicalycorrect, the selective iNOS inhibitor may theoreticallydiminish the efficacy of the mechanical puncture onangiogenesis. To overcome these limitations may provideus with a more detailed mechanism of this procedure onangiogenesis.

In conclusion, a mechanical injury in the ischemicmyocardium, such as that induced by needle puncture,promotes transient vessel formation. Associated in-creased production of NO derived from iNOS, shown inthis study, may significantly contribute to the process ofneovessel formation seen after needle-induced TMR.

This work was supported in part by a grant from MedicalResearch Council of Canada. Dr Giaid is a recipient of an FRSQscholarship. The animal model used in this study was done incollaboration with Dr R.C.J. Chiu.

References

1. Burkhoff D, Schmidt S, Schulman SP, et al. Transmyocardiallaser revascularization compared with continued medical

therapy for treatment of refractory angina pectoris: a pro-spective randomized trial. ATLANTIC Investigators. AnginaTreatments-Laser and Normal Therapies in Comparison.Lancet 1999;354:885–90.

2. Pelletier MP, Giaid A, Sivaraman S, et al. Angiogenesis andgrowth factor expression in a model of transmyocardialrevascularization. Ann Thorac Surg 1998;66:12–8.

3. Chu VF, Giaid A, Kuang JQ, et al. Angiogenesis in transmyo-cardial revascularization: comparison of laser versus me-chanical punctures. Ann Thorac Surg 1999;68:301–8.

4. Chu V, Kuang JQ, McGinn A, Giaid A, Korkola S, Chiu RCJ.Angiogenic response induced by mechanical transmyocar-dial revascularization. J Thorac Cardiovasc Surg 1999;118:849–56.

5. Folkman J. Angiogenesis in cancer, vascular, rheumatoidand other disease. Nature Med 1995;1:27–30.

6. Senger DR, Ledbetter SR, Claffey KP, Papadopoulos-SergiouA, Perruzzi CA, Detmar M. Stimulation of endothelial cellmigration by vascular permeability factor/vascular endothe-lial growth factor through cooperative mechanisms involv-ing the aÃb3 integrin, osteopontin, and thrombin. Am JPathol 1996;149:293–305.

7. Folkman J, Shing Y. Angiogenesis. J Biol Chem 1992;267:10931–4.

8. Suri C, Jones PF, Patan S, et al. Requisite role for angiopoi-etin-1, a ligand for the TIE2 receptor, during embryonicangiogenesis. Cell 1996;87:1171–80.

9. Akiyama K, Kimura A, Suzuki H, et al. Production ofoxidative nitric oxide in infarcted human heart. J Am CollCardiol 1998;32:373–9.

10. Radomski MW, Palmer RMJ, Moncada S. An L-arginine/nitric oxide pathway present in human platelets regulatesaggragation. Proc Natl Acad Sci USA 1990;87:5193–7.

11. Kubes P, Suzuki M, Granger DN. Nitric oxide: an endoge-nous modulator of leukocyte-adhesion. Proc Natl Acad SciUSA 1991;88:4651–5.

12. Schmidt HH, Walter. NO at work. Cell 1994;78:919–25.13. Guo J-P, Murohara T, Panday MM, Lefer AM. Nitric oxide

promotes endothelial proliferation: role in inhibiting reste-nosis [Abstract]. Circulation 1996;92:I–750.

14. Kroll J, Waltenberger J. VEGF-A induces expression ofeNOS and iNOS in endothelial cells via VEGF receptor-2(KDR). Biochem Biophys Res Commun 1998;252:743–6.

15. Pinsky DJ, Cai B, Yang X, Rodriguez C, Sciacca RR, CannonPJ. The lethal effects of cytokine-induced nitric oxide oncardiac myocytes are blocked by nitric oxide synthase an-tagonism or transforming growth factor beta. J Clin Invest1995;95:677–85.

16. Saito T, Rodger IW, Hu F, Shennib H, Giaid A. Inhibition ofcycloxygenase-2 improves cardiac function in myocardialinfarction. Biochem Biophys Res Commun 2000;273:772–5.

17. Giaid A, Saleh D. Reduced expression of endothilial nitricoxide synthase in the lungs of patients with pulmonaryhypertention. N Eng J Med 1995;333:214–21.

18. Landolfo CK, Landolfo KP, Hughes GC, Coleman ER,Coleman RB, Lowe JE. Intermediate-term clinical outcomefollowing transmyocardial laser revascularization in patientswith refractory angina pectoris. Circulation 1999;100(19 Sup-pl):II128–33.

19. Al-Sheikh T, Allen KB, Straka SP, et al. Cardiac sympatheticdenervation after transmyocardial laser revascularization.Circulation 1999;100:135–40.

20. RayChaudhury A, Fischer H, Malik AB. Inhibition of endo-thelial cell proliferation and bFGF-induced phenotypic mod-ulation by nitric oxide. J Cell Biochem 1996;63:125–34.

21. Murohara T, Asahara T, Silver M, et al. Nitric oxide synthasemodulates angiogenesis in response to tissue ischemia.J Clin Invest 1998;101:2567–78.

22. Wildhirt SM, Suzuki H, Horstman D, et al. Selective modu-lation of inducible nitric oxide synthase isozyme in myocar-dial infarction. Circulation 1997;96:1616–23.

135Ann Thorac Surg SAITO ET AL2001;72:129–36 iNOS IN NEEDLE TMR

23. Balligand JL, Ungureanu D, Kelly RA, et al. Abnormalcontractile function due to induction of nitric oxide synthasein rat cardiac myocytes follows exposure to activated mac-rophage-conditioned medium. J Clin Invest 1993;91:2314–9.

24. Dhweiki D, Itin A, Soffen D, Keshet E. Vascular endothelialgrowth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 992;359:843–5.

25. Ladoux A, Frelin C. Hypoxia is a strong inducer of vascularendothelial growth factor mRNA expression in the heart.Biochem Biophys Res Commun 1993;165:1005–10.

26. De Vries C, Escobedo JA, Ueno H, Houck K, Rerrara N,Williams LT. The fms-like tyrosine kinase, a receptor forvascular endothelial growth factor. Science 1992;255:989–91.

27. Kroll J, Waltenberger J. A novel function of VEGF receptor-2(KDR): rapid release of nitric oxide in response to VEGF-Astimulation in endothelial cells. Biochem Biophys Res Com-mun 1999;265:636–9.

28. Kuzuya M, Kinsella JL. Reorganization of endothelial cord-like structures on basement membrane complex (Matrigel):

involvement of transforming growth factor beta 1. J CellPhysiol 1994;161:267–76.

29. Wahl SM, Hunt DA. Transforming growth factor beta in-duces monocyte chemotaxis and growth factor production.Proc Natl Acad Sci USA 1987;84:5788–92.

30. Poslethwaite AI, Keski-Oja J. Stimulation of the chemotacticmigration of human fibroblasts by transforming growthfactor b. J Exp Med 1987;165:251–6.

31. Chandrasekar B, Streiyman JE, Colston JT, Freeman GL.Inhibition of nuclear factor kB attenuates proinflammatorycytokines and inducible nitric-oxide synthase expression inpostischemic myocardium. Biochim Biophys Acta 1998;1406:91–106.

32. Piez, KA, Sporn M. Transforming growth factor-bs. Chem-istry, biology and therapeutics. Ann NY Acad Sci 1990;593:1–379.

33. Border WA, Ruoslahti E. Transforming growth factor-beta indisease: the dark side of tissue repair. J Clin Invest 1992;90:1–7.

DISCUSSION

DR. RAINALD SEITELBERGER (Vienna, Austria): Congratula-tions for your nice paper. Although I had some trouble under-standing, you prove that you really did induce angiogenesis. Youdid show some increase in vascular density after 1 week;however, there was no vascular density increase at all in com-parison to the infarction group after 2 weeks, and then again,after 4 weeks, there was a little increase in vascular density. So,my question is, is this just something that is induced by themethod of measuring vascular density, because theoreticallyyou should have a consistent increase in vascular density overtime? You may have a consistent decrease again over time, butyou have some sort of jumping changes over time. Have you anexplanation for that?

DR SAITO: Yes. As you suggested, at the first week, it mighthave been due to no specific reaction to the tissue injury thatenhanced neovascularization. We found significant differencenot only at the first week but also at 4 weeks. At that time, itmight be due to a significantly different level of expression ofangiogenic markers, including NO and VGF. However, there areseveral articles in which vascular density has been progressivelydecreased after transmyocardial needle revascularization.

DR. SEITELBERGER: But that is not logical, is it? I mean, if youhave an increase in vascular density, it at least should be stableover time, because why should it vanish again under stableconditions?

DR LOUIS P. PERRAULT (Montreal, Quebec, Canada): A veryinteresting study, however, you have not shown really the exact

role of iNOS in the exact process here, because your stainingstudies alone do not substantiate it is one good thing butfunction studies would be even better, so I am sure you areplanning experiments with L-NAME or aminoguanidine tosubstantiate the interesting hypothesis that you bring out withthis work. Do you have any comments? Thank you.

DR. SAITO: I did not understand.

DR. FULLERTON: Have you given any agents to block nitricoxide?

DR SAITO: We did not use any drugs; for example, no selectiveinhibitor of iNOS. But it has been reported that no selectiveiNOS inhibitor prevents VGF-mediated angiogenesis.

DR GIAID: If I could answer the first and second question, too.I am the senior author on this paper. What we are actually doingnow is we are using the specific inhibitor for iNOS in the sameanimal model and we are also measuring the left ventricularend-diastolic pressure in the same animal model.

As to the question related to the increased vascularity duringthe first week, as we all know, when you cause injury, the firstphase of it involves angiogenesis, which explains the increasednumber of vessels seen in the first week, however, this processis followed by the healing and the scar formation, which ex-plained the reduction of the number of vessels during the2-week and the 4-week period.

136 SAITO ET AL Ann Thorac SurgiNOS IN NEEDLE TMR 2001;72:129–36