INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

INCREASED BLOOD FLOW without increased pressure to the pulmonary arteries is a common scenario for a number of congenital heart defects, including the frequently occurring atrial septal defect (1:1,000 live births; see Ref. 3). Although the pulmonary vasculature is relatively tolerant to increased flow (reviewed in Ref. 18), if such a defect is left unrepaired, chronic exposure of the pulmonary vasculature to increased pulmonary blood flow can result in irreversible pulmonary hypertension and right ventricular failure, usually after the second decade of life (25, 34). Previously, Friedli et al. (12) in 1975 observed that piglets exposed to increased pulmonary blood flow via an arteriovenous fistula and pneumonectomy began to develop thickening of the pulmonary arterial wall. To date, pulmonary vascular remodeling that results from increased pulmonary blood flow alone without the confounding variable of increased pressure has been poorly described, and the mechanisms are unknown. One potential mechanism is a flow-mediated alteration in the regulation of endogenous pulmonary vasodilators.

Nitric oxide (NO) is a potent endogenous vasodilator produced in the lung and other tissues by isoforms of the enzyme NO synthase (NOS). NO is produced in pulmonary endothelial cells from the amino acid L-arginine and molecular oxygen by endothelial NOS (eNOS). NO diffuses from the endothelium to adjacent smooth muscle cells and activates soluble guanylate cyclase to increase cGMP, resulting in vasodilation. The role of NO in the regulation of pulmonary vasculature basal tone and maintenance of low basal pulmonary artery pressures remains controversial (2, 8, 11). Previously, we and others have demonstrated that eNOS is increased in the hypertrophied pulmonary vasculature of rats subjected to chronic hypoxic pulmonary hypertension (22, 30, 32) and in humans with pulmonary hypertension (31). This increase in eNOS precedes and then progresses with the development of vascular remodeling (32). In vitro data suggest that shear stress (i.e., viscous drag) exerted on the endothelial surface by streaming of blood regulates eNOS expression (reviewed in Ref. 1), although the role of increased shear stress versus hypoxia or hypoxia-induced factors in the upregulation of eNOS is not clear. Evidence exists for the regulation of eNOS by flow in the aorta of exercising dogs (29). Recently, NO was also found to be associated with inhibition of smooth muscle cell proliferation and therefore may serve to regulate vascular smooth muscle cell growth (19, 23, 27) during vascular remodeling. Presently, it is unclear whether pulmonary vascular eNOS is also regulated by flow or whether an alteration in eNOS regulation could be responsible for mediating vascular remodeling to increased flow. Therefore, we hypothesize that increased pulmonary blood flow without changes in pressure results in increased eNOS expression, possibly playing a role in pulmonary vascular remodeling.

In the present study, an abdominal aorta-to-inferior vena cava shunt was created in rats to mimic the chronic increased pulmonary artery flow state of an atrial septal defect. We demonstrated that increased pulmonary blood flow resulting from the shunt induced medial thickening of the pulmonary vascular wall without evidence of ongoing smooth muscle cell proliferation. In this model, eNOS mRNA, protein levels, and immunohistochemical staining pattern of the pulmonary vasculature were unchanged in the lungs of the shunt animals.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Aortocaval shunt. Forty-two-day old Sprague-Dawley rats (Hilltop) underwent creation of an abdominal aorta-inferior vena caval shunt superior to the renal arteries using the technique described by Garcia and Diebold (14). In brief, animals were anesthetized using pentobarbital sodium, the abdominal aorta and inferior vena cava were isolated, and the fistula was created by passing a 20 (large-shunt)- or 22 (small-shunt)-gauge sterile needle through the inferior vena cava and into the aorta. The needle was slowly withdrawn, allowing a small hematoma to form at the aortic and vena caval puncture. Super Glue (S. P. Richards, Atlanta, GA) was applied as a sealant, and the needle was removed. Subsequently, the abdominal wall was closed, and the animal was allowed to recover. Sham animals were treated similarly without puncturing the aorta or vena cava. To assay for changes in vascular smooth cell proliferation, on the night before and the morning of the end of the study period, both sham and shunt groups received a 10 mg/kg ip injection of the thymidine analog bromodeoxyuridine (BrdU) as a marker of DNA replication (Sigma Chemicals, St. Louis, MO). After 42 days, the animals were anesthetized with pentobarbital sodium and ventilated with a respirator. A catheter was placed in the right carotid artery, and arterial pressures were recorded. The abdomen was exposed, and the aortocaval shunt was inspected for size, the presence of a palpable thrill, and obvious shunting of oxygenated blood into the inferior vena cava. Subsequently, the chest was opened, and pulmonary artery pressure was measured directly by main pulmonary artery puncture. The heart and lungs were rapidly perfused with 20 ml of heparinized (1 U/ml) normal saline, rapidly removed, and weighed. Approximately one-half of each lung was flash-frozen in liquid nitrogen and stored at 70°C for subsequent Northern and Western blot analyses. At the same time, 1- to 2-mm-thick pieces of lung from different lobes were cut and immersed in fixative (4% paraformaldehyde in phosphate-buffered saline) for 90 min on ice. The tissues were subsequently processed as previously described, with 2- to 4-µm sections cut for immunocytochemical analysis (32). After lung removal, the fistula was dissected open and examined under an operating microscope to further confirm patency.

Northern blot. Total RNA from lungs was extracted after the method of Chomczynski (6) using TriReagent (10). Poly(A)⁺ RNA was isolated from total RNA using oligo(dT) cellulose (5 Prime  3 Prime, Boulder, CO). Twenty micrograms of poly(A)⁺ RNA per rat were denatured with glyoxal-DMSO, electrophoresed through a 0.01 M sodium phosphate-buffered agarose gel, and transferred to a nylon membrane (Hybond-Nt; Amersham, Arlington Heights, IL) as previously described (22). The membranes were hybridized with a 4,091-bp eNOS cDNA (a kind gift of Dr. W. C. Sessa) labeled by random-prime labeling (22). Hybridization and washes were at 65°C as previously described (10, 22, 32). To correct for the amount of RNA loading, RNA transfer efficiency, and gene specific expression, all blots were probed simultaneously with the cDNA for the control gene -actin. eNOS mRNA signal was detected and quantified using a phosphorimager and ImageQuant software (Molecular Dynamics, Sunnyvale, CA). Autoradiograms were also obtained by exposure to film (Hyperfilm-MP; Amersham, Arlington Heights, IL).

Western blotting. Western blotting was performed as previously described (22, 32). Briefly, lung homogenates (150 µg/rat) were separated under denaturing conditions in a 7.5% SDS-PAGE gel, followed by blotting of the proteins to nitrocellulose (Bio-Rad, Burlingame, CA). Blots were blocked at room temperature for 1 h in 50 mM Tris · HCl, pH 7.4, 0.15 M NaCl, 2% BSA, and 0.1% Tween 20. Subsequently, blots were incubated with a mouse anti-eNOS monoclonal antibody (dilution 1:500; Transduction Laboratories, Lexington, KY) for 1 h at room temperature. Subsequently, membranes were washed at room temperature and incubated with an anti-mouse IgG conjugated to horseradish peroxidase (Bio-Rad) for 1 h at room temperature. eNOS immunoreactive protein was detected with enhanced chemiluminescence (ECL; Amersham) and exposure to film (Hyperfilm-ECL; Amersham). Signal bands were quantified by densitometry (Personal Densitometer, ImageQuant; Molecular Dynamics).

Immunocytochemistry. Immunohistochemical analysis was performed as previously described (32) in groups of 10 animals each (4 sham, 6 shunt). Separate lung sections were incubated with one of the following antibodies: 1) monoclonal -smooth muscle actin antibody (dilution 1:500; Sigma; see Ref. 4); 2) a monoclonal antibody against BrdU (dilution 1:200; Vector Laboratories, Burlingame, CA; see Ref. 9); or 3) a monoclonal antibody against proliferating cell nuclear antigen (PCNA, dilution 1:100; Vector; see Ref. 20) or the eNOS antibody (32) described above. Controls were slides incubated with nonspecific mouse IgG (Vector) at the same concentration as the primary antibody. Briefly, sections were deparaffinized with xylene and decreasing concentrations of alcohols and finally washed in PBS (10 mM NaPO₄-0.9% NaCl, pH 7.5) in preparation for immunodetection with the exception of sections for eNOS staining, which were further treated by boiling for 25 min in 10 mM citric acid, pH 6, as recommended by the manufacturer. Sections were blocked for 1 h at room temperature with horse serum followed by incubation for 1 h with the primary antibody. After the sections were washed for 30 min with PBS plus 0.5% Tween 20, immunodetection was observed using the avidin-biotin complex method (Vector Laboratories) with diaminobenzidine (brown) as a color substrate. Sections were not counterstained with the exception of eNOS-stained sections, which were counterstained with Gill's hematoxylin (Sigma). All slides were mounted and examined using an Olympus Vanox AHBS3 microscope for bright-field microscopy.

Pulmonary vascular histology was described according to the method of Meyrick and Reid (24). Pulmonary veins were differentiated from arteries by veins having thinner media, less medial elastic material, and a larger lumen compared with arteries accompanying a bronchiole. Approximately 30 arteries per animal were randomly evaluated in a blinded manner. Using an Olympus-BHS microscope coupled to an MTI color video camera (DAGE-MTI, Michigan City, IN) and I Cube video grabber board, lung images were captured, and arterial diameter and medial wall area were measured using Image Pro Plus software (Media Cybernetics, Silver Spring, MD) after calibration with an Olympus 0.01-mm microscope calibration slide. Artery diameters for two vessel sizes (50-100 and 101-200 µm) were determined by measuring the length of the internal elastic lamina. Medial wall area (µm²) was calculated as the area between the internal elastic lamina and the adventitia.

Statistics. All data are expressed as means ± SE. Statistical comparisons of groups were performed by one-way ANOVA by ranks using the Kruskal-Wallis statistic (Sigmastat; Jandel Scientific, San Rafael, CA). A value of P < 0.05 was considered statistically significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Physiological measurements. Creation of an arteriovenous fistula (shunt) between the abdominal aorta superior to the renal arteries and the inferior vena cava as described by Garcia and Diebold (14) is a well-tolerated surgical procedure with high animal survival. Of the 24 animals that had a shunt created, all but 4 survived. Mortality in all cases was in the first hours of shunt creation. There were no late deaths, with the remaining animals surviving to the end of the study period (6 wk), and both groups demonstrated a similar weight gain (Table 1). Despite the creation of a large systemic arteriovenous shunt, neither systemic nor pulmonary artery blood pressures were altered in the shunted animals. As shown in Table 1, heart weight, heart-to-body weight ratio, and lung-to-body weight ratio were all significantly increased (>2-fold; P < 0.05) in the shunt animal group compared with the sham control. Therefore, increased venous return to the right heart, with a subsequent increase in pulmonary blood flow, resulted in a compensatory increase in cardiac and pulmonary mass without an alteration in systemic or pulmonary blood pressures.

Morphological analysis, -smooth muscle actin, PCNA, and BrdU labeling of shunt lungs. Examination of the lungs from the shunt and sham animals revealed that the basic architecture of the lung respiratory apparatus (i.e., alveoli and bronchioles) appeared normal in both study groups. The most striking feature was the development of medial thickening in pulmonary arterioles (Fig. 1). As shown in Fig. 1, the medial layer of pulmonary arterioles was increased relative to the sham animals, whereas the intimal layer was unchanged. This is similar to the Heath-Edwards type 1 classification of vascular arteriopathy (17). -Smooth muscle actin immunostaining demonstrated specific localization of -smooth muscle actin to the pulmonary vascular wall and muscular portion of the airway. Importantly, as shown in Fig. 2, there is an increase in -smooth muscle actin and thus muscularity of very small (<50 µM) arterioles of the shunt lungs compared with the sham controls. The presence of pulmonary artery thickening was confirmed by measuring the arterial medial wall area (Fig. 3). As shown (Fig. 3), there was a significant increase (180%; P < 0.05) in the medial area of arteries 50-100 and 101-200 µm in diameter from the large-shunt group. To determine whether the increase in medial wall area was a result of ongoing smooth muscle hyperplasia, lung sections were stained for PCNA, an endogenous marker for cells entering cell division and for the presence of BrdU incorporation into cells replicating their DNA. As shown in Fig. 4, no PCNA-positive cells were detected in the media or intimal portions of the arteriolar wall. However, PCNA-positive cells in the pulmonary epithelium could be readily identified and served as a positive control for PCNA staining (Fig. 4) in each section. Similarly, BrdU-positive cells could not be detected in the arteriolar wall, indicating a lack of accelerated smooth muscle cell replication in the arteriolar wall. Positive BrdU labeling was identified in pulmonary epithelial cells and served as a positive control in each section (Fig. 4).

View larger version (74K): [in this window] [in a new window]   Fig. 1.   Histological appearance of pulmonary arteries from sham and large-shunt lungs. Representative histological sections of sham and large-shunt lungs are shown demonstrating hypertrophy of the arterial medial wall. A, pulmonary arteriole; B, bronchiole; arrowhead, location of the internal elastic lamina. Scale bar = 25 µm.

View larger version (62K): [in this window] [in a new window]   Fig. 2.   -Smooth muscle actin staining of sham (A) and large-shunt (B) lungs. Representative sections of sham and large-shunt lung are shown, with increased staining for -smooth muscle actin in the small arterioles (arrows) of the shunt compared with the sham lung.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Measurement of pulmonary arterial medial wall area in sham and shunt lungs. Graph demonstrates arterial medial area (µm2) of arteries 50-100 and 101-200 µm in diameter. As shown, the area of the arterial wall medial layer was significantly increased in the large-shunt group, * P < 0.05.

View larger version (93K): [in this window] [in a new window]   Fig. 4.   Proliferating cell nuclear antigen (PCNA) and bromodeoxyuridine (BrdU) staining of large-shunt lungs. As shown in representative arterioles, no PCNA staining or BrdU-labeled proliferating cells could be detected in the pulmonary arteries of the shunt lungs. Arrows indicate PCNA-  or BrdU-positive respiratory epithelial cells, which served as a positive control.

eNOS expression in the shunt lung. To determine whether increased pulmonary blood flow to the lung results in changes in eNOS expression, lung eNOS mRNA and protein levels were determined in shunt and sham animals. As shown in Fig. 5, A and B, Western blot analysis of whole lung homogenates demonstrated no significant differences in eNOS protein levels in the lungs of animals with both small and large shunts compared with the sham controls. Similarly, eNOS mRNA levels (Fig. 6, A and B) as determined by Northern blot analysis were not significantly altered in the shunt lungs, whether shunts were small or large. -Actin hybridization indicated similar RNA loading and transfer. To examine whether there could be local changes in eNOS expression in the lungs of the shunt animals, eNOS immunostaining of lung sections was performed (Fig. 7). As expected, endothelium in shunt and sham lungs was positive for eNOS immunostaining. As shown in Fig. 7, pulmonary artery endothelium was positive for eNOS, although no differences in vascular localization or pattern of eNOS immunostaining were observed. Occasional weak staining of respiratory epithelial cells was also observed but was unchanged between sham and shunt groups.

View larger version (30K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   Fig. 5.   Western blot analysis of endothelial nitric oxide synthase (eNOS) protein expression in sham and shunt lungs. A: representative Western blot of 150 µg of whole lung homogenate protein from sham and small- and large-shunt animals was analyzed. Monoclonal antibody detected eNOS protein as a single 135-kDa band. B: densitometric analysis of eNOS protein signal from Western blots of sham and shunt lung homogenates. Results are means ± SE.

View larger version (65K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Fig. 6.   Northern blot analysis of eNOS mRNA levels in sham and shunt lungs. A: Northern blot showing 20 µg poly(A)+ mRNA from sham and small- and large-shunt lungs and bovine arterial endothelial cells (BAEC, positive control) analyzed for eNOS mRNA content. B: densitometric analysis of eNOS mRNA signal from Northern blots of RNA from sham and shunt lungs. Results are expressed as means ± SE. As shown, eNOS mRNA levels were not significantly different in the sham and shunt groups.

View larger version (129K): [in this window] [in a new window]   Fig. 7.   eNOS immunostaining of sham and shunt pulmonary arteries. Sham (B), small-shunt (C), and large-shunt (D) lung eNOS immunostaining was positive in the endothelium (arrow) of the pulmonary arterioles. A: representative control section (no primary antibody) from large-shunt animal. eNOS immunostaining was unchanged in the shunt vs. sham lung arterioles.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The pulmonary arterial bed is tolerant to increases in pulmonary blood flow (reviewed in Ref. 15). This is clearly demonstrated in the case of humans with atrial septal defects in which high levels of elevated pulmonary blood flow are well tolerated for decades (16, 25, 34). However, if unrepaired, this high flow state can eventually result in pulmonary hypertension and right heart failure (16, 34). The mechanisms whereby increased pulmonary blood flow alone results in an eventual detrimental remodeling of the pulmonary vasculature are unknown. The hemodynamic consequences of an atrial septal defect can be mimicked by the creation of systemic arteriovenous connections (i.e., aortocaval shunts). As a consequence of the reduced vascular resistance in the systemic venous system compared with the arterial system, a substantial left-to-right shunt occurs, with increased volume load on the right ventricle and increased pulmonary blood flow at low pressures (5, 13-15, 28). The development of a rodent model of cardiac and pulmonary flow overload by the creation of an aortocaval shunt has allowed the determination of the fate of prolonged volume overload on heart and lung growth and physiology (5, 14, 28).

The present study demonstrates that, in the rat, the pulmonary arterial bed also has features of flow tolerance. As demonstrated, even with a large shunt that results in increased heart and lung weights, pulmonary artery pressures were unchanged. Increased pulmonary blood flow alone, without an associated increase in pressure, appeared to be a significant stimulus for medial thickening of the pulmonary vascular wall and muscularization of small arterioles. These changes in the pulmonary vasculature are similar to the changes described by Heath and Edwards (17) as a type 1 lesion. Similarly, Friedli et al. (12) reported that piglets with an arteriovenous (a-v) fistula and a late pneumonectomy similarly developed occasional hypertrophy of small arterioles, whereas an a-v fistula alone was insufficient to alter the pulmonary architecture. The addition of the pneumonectomy also resulted in an increase in pulmonary artery pressure. Unfortunately, this prior study was composed of a small number of animals, and it is unclear whether significant vascular hypertrophy could have developed in the a-v fistula group alone. In addition, age may also play a factor because the piglets in that study were newborn, whereas the rats in our study were 42 days of age at the time of shunting. Significant elevations in pulmonary artery pressures were not seen in our study, suggesting that pulmonary vascular resistance was not increased. Similarly, Fullerton et al. (13) did not observe an increase in pulmonary artery pressure in dogs 5 mo after creation of a femoral artery-vein fistula, with a sustained 3:1 left-to-right shunt. Importantly, lambs with an a-v fistula for 6 wk and a shunt of 2.4:1 did not have increased pulmonary vascular resistance or altered vascular compliance. Combined, these studies indicate that at least 6 wk of increased pulmonary blood flow has little effect on pulmonary vascular resistance. Whether vascular shear forces are also altered in this model is unclear and requires further study.

The vascular lesion induced by the increased pulmonary blood flow in this study was medial thickening, as intimal expansion was not observed. Ongoing hyperplasia was not obvious in the thickened arteries, as confirmed by PCNA staining and BrdU labeling of proliferating cells, both of which were negative in vascular smooth muscle cells. This confirms the findings of Heath and Edwards (17) that the initial lesion in the spectrum of pulmonary vascular obstructive disease is medial hypertrophy alone. Furthermore, it also explains why a type 1 lesion is able to regress after surgical repair of a shunt lesion (17, 25).

Because the mechanisms by which flow regulates pulmonary vascular remodeling are largely unknown, we examined the potential role of eNOS in this process. As shown in the present study, eNOS protein, mRNA, and immunohistochemical localization were not changed in the lungs of the shunt animals. eNOS could play two potential roles in modulating the pulmonary vascular response to increased pulmonary blood flow: first, increased expression in small arteries to facilitate vascular recruitment and maintain flow tolerance (i.e., increased flow without alteration in pulmonary vascular resistance) or second, increased expression to regulate smooth muscle cell growth (19, 23, 27). However, the findings of the present study suggest that, in this model, increased blood flow alone is not a significant regulator of pulmonary eNOS expression and that alterations in eNOS are not responsible for mediating the arterial remodeling observed. This hypothesis is supported by the recent report that a twofold increase in pulmonary artery blood flow as a result of contralateral pulmonary artery stenosis also does not alter lung eNOS expression (21). There are, however, data to suggest that blood flow regulation of eNOS is different between the aorta and pulmonary arteries. The eNOS gene does appear to be regulated by increases in blood flow in the aorta of the dog where Sessa et al. (29) demonstrated that exercise for 10 days resulted in increased aortic endothelial eNOS mRNA and increased NOS activity in epicardial coronary arteries. Similarly, Nadaud et al. (26) demonstrated that aortic eNOS mRNA and protein were increased in rats after creation of an a-v shunt. It is unclear whether this represents tissue-specific regulation of eNOS or whether this effect is specific to large-conduit arteries. Support for pulmonary-specific regulation of vasodilators is suggested by Fullerton et al. (13) where increased pulmonary blood flow resulted in decreased endothelium-dependent cGMP-mediated relaxation of pulmonary artery rings with acetylcholine after 2 wk and further after 5 mo of increased flow. The lack of endothelium-dependent cGMP-mediated relaxation may be related to progressive thickening of the arterial wall, rendering the pulmonary artery less responsive. Unfortunately, eNOS mRNA, protein, or vasculature morphology was not assessed in their study. In addition, it is unclear as to what effect the increased pulmonary arterial oxygen content produced by the creation of the shunt could have on eNOS regulation, as previously we have demonstrated that chronic hypoxia can increase eNOS protein and gene expression in the rat lung (22, 32).

The current study in combination with the report by Le Cras et al. (21) strongly suggests that the upregulation of eNOS in the chronic hypoxia model is related to hypoxia per se and not increased flow. Collectively, these data would suggest that pulmonary and peripheral arterial eNOS are regulated differently by flow. It is difficult to implicate eNOS in the regulation of the smooth muscle cell hypertrophy in the current model because eNOS levels were unchanged.

In summary, chronic increased pulmonary blood flow in the shunted rat model results in pulmonary arteriole medial thickening without changes in eNOS protein or mRNA expression. Therefore, we speculate that eNOS does not play a significant role in mediating pulmonary vascular remodeling in this model of increased pulmonary blood flow.