Altered Flow–Induced Arterial Remodeling in Vimentin-Deficient Mice

Wall shear stress is a major determinant of vascular lumen diameter and, indirectly, of vascular wall mass.^12345678 Arterial structural responses to changes in shear stress operate during physiological adaptations to postnatal development,⁹¹⁰ exercise,¹¹ and pregnancy¹² and in pathological conditions, such as arteriovenous shunting¹³ and arterial occlusive disease.¹⁴¹⁵ They have been suggested to be endothelium dependent¹⁶ and to display similarities to acute shear-induced vasodilatation in response to an increase in arterial blood flow (BF).³¹³¹⁷ Endothelial cells align and change their patterns of gene expression and release of vasoactive mediators in response to a change in shear stress.¹⁸ The endothelial cytoskeleton and its anchoring to the extracellular matrix have been suggested to play a key role in these responses.¹⁹²⁰

The intermediate filament protein vimentin is a component of the cytoskeleton of vascular endothelial cells.²¹ Vimentin-deficient mice (V−/−)²² exhibit a blunted, acute, flow-induced arterial vasodilatation²³²⁴ and an altered balance between endothelin-1 and nitric oxide (NO), 2 endothelium-derived vasoactive factors that may also be involved in arterial remodeling.¹³²⁵²⁶

In the present study, we evaluated whether vimentin plays a pivotal role in the capacity of arterial blood vessels to adjust their lumen diameter and wall mass in response to altered BF. Therefore, we applied unilateral carotid artery ligations²⁷ in wild-type mice (V+/+) and V−/−.²³²⁴ Our findings provide direct support for an important role of vimentin in diameter and wall mass changes during flow-induced arterial remodeling.

Methods

Experimental Animals

All experiments were performed according to institutional ethical guidelines. Vimentin-null mice (V−/−) were obtained by targeted inactivation of the vimentin gene through in-frame insertion of an Escherichia coli β-galactosidase coding sequence into the exon 1 of the vimentin gene.²² Heterozygous mice (V+/−) were originally obtained by crossing a chimeric male (generated from a C57/Bl6 blastocyst injected with the mutated 129 Sv cells) with 129 Sv females. F1 V+/− were then crossed to give V−/−. The V−/− colony was then maintained by brother-sister mating. Male mice of this colony were used in the present study with 129 Sv mice as controls. To identify wild-type mice (V+/+) and homozygous vimentin-deficient mice (V−/−), DNA was extracted from the tails of mice, and the presence of targeted vimentin alleles was evaluated by previously described polymerase chain reaction methods.²⁸

Unilateral Carotid Artery Ligation

At 3 to 4 months of age, V+/+ (n=19) and V−/− (n=19) were anesthetized with ketamine and xylazine (100 and 10 mg/kg SC). Both common carotid arteries were exposed through a midline incision in the neck. In half of both groups of animals (experimental animals), the left common carotid artery (LCA) was ligated with 5.0 surgical suture just proximal to the carotid bifurcation as described by Kumar and Lindner.²⁷ This intervention resulted in cessation of BF in the unbranched occluded vessel. One of the V+/+ (sham) and 4 of the V−/− (2 sham and 2 experimental) died during or after the surgery. The animals were allowed to recover for 4 weeks after sham surgery or unilateral carotid artery ligation. At this stage, they were anesthetized again with ketamine and xylazine.

BF Measurements

During initial surgery and 4 weeks later, body temperature was maintained at 37.5°C by a thermostatically controlled heating platform. BF in the LCA and right carotid artery (RCA) was recorded for V+/+ and V−/− that had undergone either sham surgery or unilateral carotid artery ligation. We used a transit-time ultrasonic flow probe (0.5-mm V series, Transonic Systems) that was mounted on a micromanipulator and positioned halfway between the aortic arch and the carotid artery bifurcation. The pulsatile BF signal was recorded (Figure 1), and mean BF was obtained by averaging during a 5-minute period for the LCA and RCA of each animal. Complete ligation of the LCA resulted in BF recordings that fell below the detection limit of the flowmeter, indicating that there is no net forward BF in these occluded arteries (Table and Figure 1).

Pressure-Diameter Curves

After the flow measurements, which were performed at 4 weeks after the initial surgery, the LCA and the RCA were isolated. The proximal end of each vessel was mounted on a steel cannula (diameter, 0.3 mm), after which the distal end was ligated with 7-0 surgical suture, and the arterial segments were incubated in an organ chamber at 37°C in a calcium-free physiological salt solution (KRB) containing 10 μmol/L sodium nitroprusside to inactivate the arterial smooth muscle. The cannulated (“blind sac”) arterial segments were connected to a feedback-controlled pressure source (Living Systems Instrumentation) and were mounted on the stage of an Invertoscope (Nikon TMS) equipped with a video camera (Stemmer) and a digital device (LSI) for the recording of arterial outer diameter. Diameter analysis was limited to the outer diameter because (especially at low distending pressure) mice carotid arteries were not sufficiently translucent to monitor lumen diameter. Intra-arterial pressure was increased in 10 mm Hg steps from 20 to 150 mm Hg. Outer diameter was recorded at 2 minutes after each pressure increment. After the highest pressure was reached, pressure was returned to 100 mm Hg, and 30 minutes later, the arterial preparations were exposed to phosphate-buffered (pH 7.4) formaldehyde (4%).

Outer diameter was expressed as a function of distending pressure. To determine the maximal outer diameter, pressure-diameter (P-D) curves were analyzed by a 4-parametric logistic sigmoidal curve fitting [y=y_min+(y_max−y_min)/(1+10^{(log E50−x) · Hill slope})] of individual curves (Graphpad Prism 2.01). The correlation coefficient of each curve fit exceeded 0.99 in each case.

Morphometry

After overnight fixation in formaldehyde at 100 mm Hg, the arteries were removed from the arteriograph system and stored in 70% ethanol for, at most, a week. To concentrate histological analysis on the central part of the vessels, fixed arteries were cut in half, and both parts were embedded side by side in paraffin. Cross sections (4 μm) were stained with Lawson’s solution (Boom BV), which highlights the elastic laminae. Medial cross-sectional area (CSAm), lumen diameter, and medial thickness were determined as previously described²⁹ with the use of video images generated by a Zeiss Axioscope, a standard CCD camera (Sony), and commercial software (JAVA 1.21, Jandell Scientific).

Data Analysis

Results are shown as mean±SEM. Statistical significance of differences between experimental and sham-operated animals and between V+/+ and V−/− were evaluated by a 2-way ANOVA, followed by a post hoc comparison with a Newman-Keuls significant difference test. A value of P<0.05 was considered significant.

Results

General Characteristics

In sham-operated V+/+ and V−/−, mean BF did not differ significantly between the LCA and RCA (Table). Body weight and BF did not differ significantly between V+/+ and V−/− (Table). In both groups of mice, LCA ligation resulted in an elimination of BF in the occluded vessel and a substantial increase of BF in the RCA (Figure 1 and Table). The extent to which contralateral BF was increased was less marked in V−/− (from 0.33±0.02 to 0.50±0.05 mL/min) than in V+/+ (from 0.34±0.05 to 0.68±0.04 mL/min).

Carotid Arteries of V−/− Compared With V+/+

In both groups of sham-operated mice, the passive P-D curves could be superimposed (Figure 2), and the CSAms were identical for the RCA and LCA (Figure 3).

The outer diameters of LCA and RCA tended to be smaller in V−/− than in V+/+. This was statistically significant for the maximal diameters calculated by curve fitting of the P-D curves (Figures 2 and 4). CSAm was significantly smaller for sham-operated V−/− than V+/+ (Figure 3). Combined, the LCA and RCA diameters of V−/− tended to be smaller and the medial mass was significantly smaller than corresponding values in V+/+. Surprisingly, the smaller medial mass did not seem to alter mechanical properties, such as circumferential compliance (reflected by the tangent to the P-D curve) in the arteries of V−/− (Figure 2), or their distensibility (not shown).

Effects of Altered BF in V+/+

In V+/+, LCA ligation resulted in a narrower, stiffer, and thicker vessel (Figure 5). CSAm was significantly increased (Figure 3), whereas the steepness of the P-D curve was markedly reduced (Figure 5), as was the maximal outer diameter (Figure 4).

Despite the doubling of BF, the RCA of V+/+ displayed no significant alteration of wall mechanics and diameter (Figure 5). If anything, medial mass tended to be reduced in these hyperperfused vessels (Figure 3).

Effects of Altered BF in V−/−

The diameter reduction and stiffening of the ligated LCA was considerably less in V−/− than in V+/+ (Figure 5). The maximal outer diameter was reduced by 77 μm in V−/−; this was approximately the double (162 μm) of that in V+/+ (Figure 4). Medial hypertrophy, on the other hand, was more pronounced in the occluded LCA of V−/− (10×10³ μm²) than of V+/+ (5×10³ μm², Figure 3).

In contrast to V+/+, the hyperperfused RCA of V−/− exhibited a statistically significant increase in maximal diameter (Figures 4 and 5) and in CSAm (Figure 3).

Discussion

Acute arterial responses to altered BF have been attributed to shear stress–induced release of endothelium-derived vasoactive mediators.¹⁸ These responses involve the endothelial cytoskeleton and its connections to the basement membrane and extracellular matrix.¹⁹²⁰ The intermediate filament protein vimentin, which is an integral part of the cytoskeleton in vascular endothelium,²¹ may play a role in these responses.²³²⁴ We now demonstrate that the absence of vimentin results in modified structural responses of the mouse carotid artery to chronically altered BF. Inward remodeling in response to BF cessation was blunted, whereas outward remodeling in response to elevated BF was increased. Furthermore, arterial hypertrophic responses to altered BF were potentiated in V−/−.

Kumar and Lindner²⁷ introduced unilateral carotid artery ligation in mice for the study of neointimal formation after BF cessation. Application of this approach in knockout mice revealed a key role of P-selectin³⁰ and of endothelial NO synthase²⁶ in neointimal formation and overall arterial structural changes, respectively, in response to a drastic fall in arterial BF. We reasoned that this experimental approach could also be used to address the role of a cytoskeletal protein, such as vimentin, and that not only arterial structural responses in the absence of BF but also responses to chronically increased BF could be evaluated. We strengthened this experimental approach by combining measurements of BF, mechanics, and structure at the level of the mouse carotid artery.

Vimentin is expressed by various cell types, including vascular endothelial and smooth muscle cells.²¹ This expression is developmentally regulated³¹³² and changes during the course of the repair of arterial and myocardial injuries.^33343536 Yet, vimentin-deficient mice, V−/−, which do not show upregulation of other intermediate filament or cytoskeletal proteins, display a relatively mild phenotype.²² They develop and reproduce normally,²² and their arterial blood pressure does not differ from that of wild-type control mice, V+/+.²³²⁴ It has been reported, however, that V−/− cannot cope with a drastic reduction of renal mass²⁴ and display impaired arterial vasodilatation in response to elevated BF²³²⁴ as a result of an altered balance between the vascular levels of endothelin-1 and NO.²⁴ The evolutionary advantage of the highly conserved vimentin sequence has been proposed to lie not only within the role of the intermediate filament in cellular motility and contractility but also in its possible role in pathological conditions that require vascular adaptations.²⁴

In various previous experimental settings, chronic flow increases and flow decreases have been reported to result in outward and inward arterial remodeling, respectively.^348101317 The structural diameter changes were observed to lead to a normalization of wall shear stress and to be accompanied by a compensatory change in medial mass, which restores circumferential wall stress. Changes in arterial smooth muscle cell size and number have been suggested to participate in these compensatory or adaptive forms of arterial remodeling in response to altered BF.⁴⁸¹⁰

In mice, however, the BF changes resulting from unilateral carotid artery ligation did not result in compensatory adaptive remodeling. In the occluded vessel, the diameter decreased, but medial mass increased (Reference 27²⁷ and the present study), undoubtedly entailing a decrease in circumferential wall stress. Moreover, a significant neointima develops in the distal part of these vessels.²⁷³⁰ Even more surprising is the lack of structural changes in the contralateral carotid arteries. Although we previously described compensatory outward hypertrophic remodeling in rat mesenteric muscular arteries exposed to a doubling of BF,⁸ the diameter and medial mass of mouse carotid arteries seemed to be refractory to a comparable intervention. We cannot firmly exclude that the contralateral hyperperfused vessel regulated wall shear stress primarily by removal of vasomotor tone (flow-induced vasodilatation). Because shear stress is directly proportional to flow and inversely proportional to the third power of vessel radius,³ maintenance of wall shear stress during doubling of BF would, however, require a 26% increase in arterial diameter. This exceeds by far the maximal dilatation that can be induced in carotid arteries of rodents,³⁷ including mice (D.H., unpublished observations, 1997). We have currently no firm explanation for the differences between mouse carotid artery structural responses to altered BF. These differences may involve age,³⁸³⁹ interspecies and regional differences regarding the pulsatility of the local hemodynamic factors, and the abundance of smooth muscle cells, extracellular matrix components, and cytoskeletal features.¹⁸⁴⁰

In the present study, we focused our attention on the central part of the vessels where BF measurements were performed and where alterations in vessel mechanics, CSAm, and vessel diameter occurred in the absence of significant neointimal formation. We constructed P-D curves in vitro to ensure that differences in diameter between groups were not due to differences in transmural pressure or arterial smooth muscle tone.

Occluded carotid arteries of V−/− displayed a less marked reduction of arterial diameter and a more pronounced medial hypertrophy than occluded carotid arteries of V+/+. This finding displays similarities to previous results in endothelial cell NO synthase–deficient mice in which unilateral carotid artery ligation did not alter arterial diameter but resulted in exaggerated media hypertrophy.²⁶ As proposed with respect to acute flow-induced vasodilatation,²³²⁴ the presence of vimentin may be required to translate altered arterial shear stress into a chronic alteration of endothelial cell NO synthase activity.

The contralateral hyperperfused carotid artery of V−/− displayed significant increases in diameter and medial mass, whereas the arteries of V+/+, which experienced a more marked increase in BF, showed no significant structural changes. The blunted contralateral hyperemia seen in V−/− during unilateral carotid artery ligation may find its origin in impaired flow–induced vasodilatation²³²⁴ of the vessels that interconnect the areas perfused by both carotid arteries.

Chronic structural responses of arteries to altered BF have been considered to reflect the summation of many short-term vasomotor events.⁴¹⁴² Findings of the present study indicate, however, that flow-induced vasodilatation and flow-induced remodeling are not directly related. Impaired flow–induced dilatation has been observed in mesenteric²³ and renal²⁴ arteries of V−/−. In the assumption that this also applies to the carotid arteries of V−/−, their outward flow-induced remodeling suggests that remodeling does not develop as a consequence of vasomotor changes but rather in conditions in which tone cannot provide for optimal regulation.

Despite increased diameter and medial mass responses to an imposed increase in BF, carotid arteries of sham-operated V−/− tended to be narrower and were equipped with a considerably smaller media than those of V+/+ littermates. This indicates that imposed flow-induced arterial remodeling in the adult does not represent the recapitulation of processes that operate during the development of the arterial system. Changes in carotid artery BF during early development are most likely much more gradual than the ones that we imposed by arterial occlusion, and the effects of endothelium-derived factors may differ between an immature and a fully differentiated vessel wall.⁴³

In conclusion, the present study provides the first evidence that vimentin modulates arterial structural responses to altered BF. It expands previous conclusions regarding the role of vimentin in the mechanotransduction of shear stress from acute vasomotor responses to chronic arterial remodeling, although the structural changes do not seem to be the direct consequence of altered arterial contractility.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes