Invited Symposium: Carbon Monoxide and Cardiovascular Function



Materials & Methods


Discussion & Conclusion



INABIS '98 Home Page Your Session Symposia & Poster Sessions Plenary Sessions Exhibitors' Foyer Personal Itinerary New Search

Heme Oxygenase and CO Actions in Blood Vessels

Contact Person: Kanji Nakatsu (nakatsuk@post.queensu.ca)


Our laboratory has been involved for some time in studies on the mechanism of action of organic nitrate vasodilators, and the relationship between vascular relaxation and the formation of NO. More recently we have turned our attention to carbon monoxide (CO) because it too relaxes smooth muscle; other features of this molecule indicate that it deserves full consideration for a regulatory role in vascular tone. In this presentation we address some of the criteria that must be satisfied in order to consider such a role for CO. Thus we have examined the effect of an antagonist on mechanical activity. For these experiments, we have tested the ability of ODQ, an inhibitor of sCG, on the inhibitory action of CO and NO in rabbit aorta. We have also investigated a number of vascular preparations to determine the presence of heme oxygenase (HO), the enzyme responsible for the formation of CO from heme. Histochemical localization of this enzyme was accomplished with antibodies for HO-1 and HO-2.

Back to the top.

Materials and Methods

Relaxation of the rabbit aorta in response to CO and NO

Rabbit aortic rings (RAR) two to three mm wide, without endothelium were mounted in water-jacketed tissue baths containing Krebs' solution warmed to 37 C and aerated with O2/CO2 (95%-5%). Endothelial removal was confirmed by a contractile response of isolated rabbit aortic rings to acetylcholine (0.01 - 1 uM). Isometric changes in tension were recorded by a computer-based data acquisition system (MacLab®). Tissues were contracted submaximally (60-80%) with 0.1 uM PE; those RAR that could not attain a 60-80% contraction in response to 0.1 uM PE were excluded from the study. After attaining steady-state contraction, the control response to CO was obtained by bubbling the gas directly into the tissue bath at a rate of 2 ml/min until the tension level reached a plateau (calculated final concentration of CO, 30 uM), after which the CO stimulus was removed and tension returned to a submaximal level. Then, DMSO vehicle, 0.1, 1, or 10 uM ODQ was added to the muscle bath for a 15-min incubation period and CO was bubbled (2 ml/min) into the tissue bath until a steady-state relaxation plateau was achieved. Relaxation of RAR to NO was determined in a similar manner except that the different doses of NO were injected via a gas-tight syringe into the tissue bath. In all experiments, ODQ was present in the tissue bath during the addition of vasodilators.

Measurement of HO enzymatic activity in microsomal fractions of bovine pulmonary artery (BPA), vein (BPV) and BPA adventitial and medial layers, and guinea pig placenta.

HO activity in BPA, BPV and BPA adventitial and medial layer, and guinea pig placenta microsomal fractions was determined by measuring the rate of CO formation during the NADPH-dependent degradation of heme. CO in the headspace gas was quantitated by gas chromatography. The amount of CO in the headspace gas was calculated by interpolating the peak area value of the CO chromatographic signal on the linear CO standard curve (10-164 pmol CO).

Data analysis

The data are presented as group means ± SD. Statistical analysis of the data was conducted by randomized-design, two-way analysis of variance. For a significant F-statistic (p less than 0.05), a repeated-measures, one-way analysis of variance followed by Newman-Keuls test or Student's t test for unpaired or paired data (two-tailed) was conducted, depending on which test was statistically appropriate. Two groups of data were considered to be statistically different when p was less than 0.05.

Immunohistochemical analysis of BPA and BPV and guinea pig placenta

BPA and BPV were dissected from bovine lungs and immediately fixed with 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS; pH 7.2). Vessels were fixed for at least 24 h and processed for paraffin embedding. Guinea pig placenta were processed similarly. An indirect immunohistochemical method was used to detect HO-1 and HO-2 using a commercially available avidin-biotin peroxidase (ABC-HRP) immunostaining kit (Pierce Chemical Co. Rockford, IL). Polyclonal antibodies to HO-1 and HO-2 were obtained from StressGene (Victoria, B.C.). Antibodies were diluted 1:150(HO-1) and 1:100(HO-2) with 3% (v/v) goat serum in PBS and applied to the sections for 2 h at room temperature in a humidified chamber prior to staining. Immunoreactive sites were visualized by incubating sections with 0.05% (w/v) diaminobenzidine (DAB; Sigma Chemical Co.), 0.003% (v/v) H2O2 in 50 mM Tris-HCl (pH 7.6) for 5 min. Sections were then counterstained with hematoxylin, dehydrated in ethanol, mounted with Permount®, coverslipped and examined under a light microscope. The specificity of staining was assessed by the following procedures: (a) elimination of primary antibody from the reaction; and (b) staining sections with a polyclonal antibody to atrial natriuretic peptide (ANP) as per HO-1 and HO-2, respectively. The investigator conducting the immunohistochemical analysis was not aware of the experimental conditions of the individual placental sections.

Back to the top.


Effects of ODQ on CO- and NO-induced relaxation of RAR (1)

In the control response in which ODQ was absent, CO relaxed RAR 47.4 ± 6.1% and this relaxation was significantly reduced to 26.0 ± 4.8% by pretreatment with 0.1 uM ODQ. This relaxation was further attenuated to 7.4 ± 3.2% by 1 uM ODQ and completely blocked by 10 uM ODQ. Injection of NO (2.1, 4.2, 8.3, 16.6, or 20.8 nmol) in the tissue bath resulted in a concentration-dependent relaxation of RAR; the maximum relaxation was 85% of the PE-induced contraction and the ED50 was 10 nmol. NO caused an immediate decrease in tension followed by an increase to the submaximal tension levels. This most likely reflects loss of NO to the atmosphere and inactivation of NO in the tissue bath or the tissues. While the relaxation to CO was completely blocked by 10 uM ODQ, NO-induced relaxation of RAR was only partially attenuated by 10 uM ODQ (by about 40-60%).

Measurement of HO enzymatic activity in microsomal fractions of BPA, BPV and BPA adventitial and medial layers (2)

HO activity in the BPA and BPV microsomal fractions was 0.22 ± 0.07 and 0.24 ± 0.06 nmol CO/mg protein/h, respectively; these HO activity values were not different. HO activity in both BPA and BPV was inhibited (p less than 0.05) by 50 uM chromium mesoporphyrin (CrMP) to the extent of 55 ± 11% and 63 ± 19%, respectively, and by 50 uM zinc protoporphyrin ZnPP to the extent of 35 ± 13% and 51 ± 11%, respectively. The HO activity of the microsomal fraction of the BPA adventitial layer (0.22 ± 0.04 nmol CO/mg protein/h) was significantly greater (p less than 0.05) than that of the BPA medial layer (0.14 ± 0.04 nmol CO/mg protein/h). HO activity in both the BPA adventitial and medial layers was inhibited (p < 0.05) by 50 uM CrMP to the extent of 64 ± 2.0% and 75 ± 24%, respectively.

HO enzymatic activity in guinea pig placenta during gestation (3)

Placental HO activity at GD (gestational day) 34, 50, 62 and full-term was 0.28 ± 0.10, 0.34 ± 0.14, 0.51 ± 0.10 and 0.22 ± 0.08 pmol/microgram protein/15 min, respectively. Thus, at GD 50 and GD 62, the magnitude of HO activity in the placenta was about 40% of the enzymatic activity in the cerebral cortex, hippocampus and cerebellum of the fetal brain.

HO immunohistochemical analysis of BPA and BPV(2)

Apparent HO-1 and HO-2 immunoreactivity was localized in the smooth muscle cells of the medial layer and in the adventitia of BPA, predominantly associated with nerves and vasa vasorum. In the endothelial cells, there was staining of HO, associated with apparent HO-2 immunoreactivity and only trace HO-1 immunoreactivity. Staining of HO in BPV was similar. Specificity of staining was assessed using an ANP polyclonal antibody. Neither dilution of ANP antibody produced any noticeable stain in BPA.

Immunohistochemical localization of HO-1, HO-2 and NOS III in Guinea pig placenta during gestation and in full-term human placenta (3)

Faint staining for HO-1 was observed in the adventitial layer of larger fetal blood vessels at GD 34. The intensity of this staining was greater at GD 50 and GD 62, and decreased at full-term. Staining for HO-1 also was evident in the endothelium. Serial sections of placental tissues obtained at GD 50 and GD 62 exhibited lower staining intensity in the fetal adventitial layer when incubated with anti-HO-2 antiserum. Tissue sections incubated with normal rabbit serum diluted to the same concentration as the anti-HO-1 and anti-HO-2 antisera had no apparent staining. The intensity of staining for NOS III appeared to be greatest at mid-gestation, decreased thereafter to term, and was localized primarily to trophoblast lining maternal channels. Staining for NOS III was abolished by preincubating the antibody with the NOS III peptide immunogen used to generate the antibody.

In a preliminary study of a full-term human placenta, there was intense staining for HO-1 in the adventitial layer of fetal blood vessels of stem chorionic villi. A control serial section of this placenta incubated with normal rabbit serum diluted to the same concentration as the anti-HO-1 antiserum did not exhibit staining.

Back to the top.

Discussion and Conclusion

Given that CO, like NO, is proposed to act at the same intracellular receptor, i.e. sGC; it was anticipated that ODQ would interfere with the CO-induced relaxation of RAR. ODQ (0.1 - 10 uM) was capable of attenuating relaxations of RAR to CO in a concentration-dependent manner. These results are interpreted as supporting the hypothesis that CO-induced vascular smooth muscle relaxation is mediated by a mechanism similar to that for NO, viz. activation of sGC.

On the other hand, reports from several laboratories suggest that this concept is too simplistic and merits elaboration. Accordingly, in a purified bovine pulmonary sGC preparation, others have found that millimolar concentrations of CO bind to its heme moiety, but CO is a weak activator of purified sGC; this suggests that CO is not a likely physiological modulator of this enzyme. One explanation for this discrepancy could be that highly purified sGC has lost a significant part of its regulatory properties. In the present study, a lower concentration of CO (30 uM) produced relaxation of rabbit aortic rings; this was sensitive to inhibition by ODQ. The differences between whole cell preparations and broken cell systems may reflect differing ability of sGC to remain stable in its active conformation. This suggestion is based in part on the observation that YC-1, a stimulator of sGC activity independent of NO or CO, is thought to stabilize the active enzyme; others have found that in the presence of YC-1, CO could stimulate sGC activity to levels similar to those achieved with NO.

When compared with relaxation by NO gas, CO-induced relaxation of RAR is much more sensitive to inhibition by ODQ. This is evident as 10 uM ODQ completely blocked CO-mediated relaxation, whereas ODQ-treated RAR could still demonstrate residual relaxations to NO gas. Residual NO relaxation in the presence of ODQ may indicate a sGC-independent component of relaxation.

In BPA and BPV, there was measurable HO activity, determined as the rate of NADPH-dependent, heme-derived CO formation, of similar magnitude in the artery and vein. Furthermore, HO protein was detected by immunohistochemical analysis using HO-1 and HO-2 antibodies. Inhibition of CO formation by incubation with the two HO inhibitors, viz., CrMP and ZnPP, verified the conclusion that there was HO activity in BPA and BPV. HO activity was present in the adventitial and medial layers of BPA; the enzymatic activity was higher in the adventitial layer and was inhibited in both layers by chromium mesoporphyrin. HO activity was not determined in the intimal layer because of the small amount of endothelium. HO protein was identified in the adventitial, medial and endothelial layers, using the HO-1 and HO-2 antibodies in the immunohistochemical analysis. In the adventitia, there was apparent positive staining with the HO-1 and HO-2 antibodies that was associated predominantly with nerves and vasa vasorum. There also was HO protein in the smooth muscle cells of the medial layer that was identified by staining with both HO-1 and HO-2 antibodies. Immunohistochemical analysis of BPV demonstrated localization of HO protein in the adventitial, medial and intimal layers; however, HO activity was not determined experimentally for each layer, because of the difficulty in mechanically separating the adventitial and medial layers and the very small amount of endothelium. The relationship between the localization of HO protein and the presence of HO enzymatic activity in the adventitial, medial and intimal layers of BPV requires further elucidation.

In view of the presence of HO enzymatic activity in artery and vein of the bovine lung, it would appear that there is a multi-faceted role for the enzyme in the regulation of heme content in vascular tissues and in the formation of CO with consequent regulation of vascular tone. Heme content is regulated by the rate-limiting enzyme of heme biosynthesis, delta-aminolevulinic acid synthetase, and by the heme degradative enzyme, HO. Heme is required for the synthesis of hemoproteins in all cell types, which includes cells in the adventitial, medial and endothelial cells of blood vessels. Additional roles have been proposed for HO including the conversion of excess heme, a prooxidant, to biliverdin and ultimately bilirubin, two antioxidants, and CO. Furthermore, HO-1 is considered to be a stress protein in many tissues. It is conceivable that HO plays a similar role to that of nitric oxide synthase in blood vessels for physiologic and pathophysiologic conditions, whereby CO, like NO, produces vasodilation via stimulation of sGC and increased cGMP formation.

In addition to our studies on bovine blood vessels, we have conducted HO experiments using different blood vessels from a different species. Thus, the objectives of this part of our work were to determine the localization of HO-1 and HO-2 protein (as well as NOS III protein) and to determine the HO activity in guinea pig placenta during gestation, and at birth. HO-1 and HO-2 positive staining was associated primarily with the adventitial layer of larger fetal blood vessels of the placenta, with staining also evident in the endothelium. There was greater intensity of staining for HO-1 protein compared with HO-2 protein, and the magnitude of staining for HO-1 appeared to be greater at GD 50 and GD 62 compared with mid-gestation (GD 34) and full-term. In contrast, NOS III was localized predominantly to trophoblast lining maternal channels, and the intensity of staining was greatest at mid-gestation (GD 34) and decreased thereafter to term. This gestational profile of NOS III protein expression in the guinea pig placenta is similar to that reported for the EDRF-dependent vasorelaxant response of human umbilical artery to histamine, in which there is decreasing response with increasing gestational age. The data of our study clearly demonstrate that the gestational profile of placental HO-1 and HO-2 protein expression is distinctly different from that of NOS III. In a preliminary immunohistochemical investigation of one full-term human placenta, HO-1 protein localization was similar to that in the guinea pig placenta at term, with intense staining in the adventitial region of fetal blood vessels of stem chorionic villi.

It may be that HO plays a role akin to that of NOS in vascular tissue, in which CO, like NO, is an endogenous vasorelaxant metabolite that produces vasodilation via stimulation of sGC and increased cGMP synthesis. The maintenance of uteroplacental and fetoplacental blood flow is important for optimal maternal-to-fetal transfer of oxygen and nutrients. Impaired placental blood flow will decrease nutrient supply and oxygenation of the fetus, thereby leading to intrauterine growth restriction. The NOS-catalysed synthesis of NO and HO-catalysed production of CO may serve as novel gaseous messenger pathways in the maintenance of optimal placental blood flow, and oxygen and nutrient supply to the fetus. In view of the distinct localization and gestational profiles of expression of HO (HO-1 and HO-2) and NOS III proteins in the placenta, it is plausible that these two enzymatic pathways play different, but complementary, roles in the regulation of vascular tone and hemodynamics in the placenta throughout pregnancy. It is noteworthy that for exposure of cultured arterial smooth muscle cells to hypoxia, it was cell-derived CO, not NO, that increased cGMP content.

In summary, we have determined that different blood vessels obtained from cattle and guinea pigs possess the metabolic machinery to produce CO. In all the vasculature reported herein, HO was found in the adventitia, and in some blood vessels it was also found in the media and intima. The vasorelaxant effects of CO were antagonized by the sGC inhibitor, ODQ. In view of the vasorelaxant action of both CO and NO, it is conceivable that HO and NOS pathways play key complementary roles in the regulation of placental hemodynamics during gestation.

Back to the top.


This presentation represents a distillation of observations from the following papers:

1. Can. J. Physiol.Pharmacol. 75:1034-1037, 1997.

2. J. Cardiovasc. Pharmacol. 30: 1-6, 1997.

3. Placenta 19: 509-516, 1998.

Back to the top.

| Discussion Board | Previous Page | Your Symposium |
Nakatsu, K.; Brien, J.F.; Graham, C.H.; McLaughlin, B.E.; Hussain, A.S.; Odrcich, M.J.; Marks, G.S.; (1998). Heme Oxygenase and CO Actions in Blood Vessels. Presented at INABIS '98 - 5th Internet World Congress on Biomedical Sciences at McMaster University, Canada, Dec 7-16th. Invited Symposium. Available at URL http://www.mcmaster.ca/inabis98/wang/nakatsu0681/index.html
© 1998 Author(s) Hold Copyright