Three-dimensional arrangement of the vasa vasorum in explanted segments of the aged human great saphenous veinScanning electron microscopy and three-dimensional morphometry of vascular corrosion casts.код для вставкиСкачать
THE ANATOMICAL RECORD PART A 281A:1372–1382 (2004) Three-Dimensional Arrangement of the Vasa Vasorum in Explanted Segments of the Aged Human Great Saphenous Vein: Scanning Electron Microscopy and Three-Dimensional Morphometry of Vascular Corrosion Casts ALOIS LAMETSCHWANDTNER,1* BERND MINNICH,1 DAVID KACHLIK,2 MAREK SETINA,3 AND JOSEF STINGL2 1 Vascular and Muscle Research Group, Department of Organismic Biology, University of Salzburg, Salzburg, Austria 2 Department of Anatomy, Third Medical Faculty of Charles University, Prague, Czech Republic 3 Department of Cardiovascular Surgery, Regional Hospital, Ceske Budejovice, Czech Republic ABSTRACT The vasa vasorum of skeletonized and nonskeletonized segments of ﬁve human great saphenous veins (GSVs), harvested during coronary bypass grafting, were cannulated, rinsed, and injected (casted) with the polymerizing resin Mercox-Cl-2B. After removal of the dry vascular tissue, the casts were examined using scanning electron microscopy. Stereopaired images (tilt angle, 6°) were taken, imported into a 3D morphometry system, and the 3D architecture of the vasa vasorum (arterial and venous vasa as well as capillaries) was studied qualitatively and quantitatively in terms of vasa diameters, intervascular and interbranching distances, and branching angles. Diameters of parent (d0) and large (d1) and small (d2) daughter vessels of arterial and venous bifurcations served to calculate asymmetry ratios (␣) and area ratios (␤). Additionally, deviations of bifurcations and branching angles from optimal branches were calculated for selected arterial vasa. The arrangement of the vasa vasorum closely followed the longitudinally oriented connective tissue ﬁbers in the adventitia and the circularly arranged smooth muscle cell layers within the outer layers of the media. Venous vasa by far outnumbered arterial vasa. Vasa vasorum changed their course several times in acute angles and revealed numerous circular constrictions, kinks, and outpouchings. Due to their spatial arrangement, the vasa vasorum are prone to tolerate vessel wall distension generated by acute increases in blood pressure or stretching of the vessel without severe impact on vessel functions. Preliminary comparisons of data from the bifurcations of cast arterial vasa vasorum, with calculated optimal bifurcations, do not yet give clear insights into the optimality principle(s) governing the design of arterial vasa vasorum bifurcations of the human GSVs. © 2004 Wiley-Liss, Inc. Key words: man; saphenous vein; vasa vasorum; vascular casts; scanning electron microscopy; 3D morphometry The vessels within the wall of larger blood vessels are known as the vasa vasorum. In normal blood vessels with less than 29 layers of vascular smooth muscle cells (VSMCs) or lamellar units (Wolinsky and Glagov, 1967), vasa vasorum lie in the adventitia. They supply oxygen and nutrients to the outer layers of the vessels; blood circulating in the vessel lumen supplies the inner wall areas. Vasa vasorum also provide an exchange area for delivery to and absorption of substances from the surrounding wall tissue (Williams and Heistad, 1996). In blood vessels with more than 29 lamellar units, adventitial vasa vasorum penetrate into the outer layers of © 2004 WILEY-LISS, INC. *Correspondence to: Alois Lametschwandtner, Organismic Biologie, Vascular and Muscle Res. Group, University of Salzburg, Hellbrunnerstrasse 34, A-5020 Salzburg, Austria. Fax: 43-662-80445698. E-mail: firstname.lastname@example.org Received 16 February 2004; Accepted 25 May 2004 DOI 10.1002/ar.a.20098 Published online 14 October 2004 in Wiley InterScience (www.interscience.wiley.com). VASA VASORUM IN HUMAN GREAT SAPHENOUS VEIN the media (Wolinsky and Glagov, 1967). Vasa vasorum either abut against the proper vessel lumen (vasa vasorum interna), against branches of the vessel, or against arteries next to the vessel (vasa vasorum externa) (Schönenberger and Müller, 1960; Gössl et al., 2003). In the stem of the human great saphenous vein (GSV), only vasa vasorum externa are present, whose stems penetrate the GSV every 5 to 15 mm (Kachlik et al., 2002). In arteries, the vasa vasorum drain into concomitant nearby veins (Gössl et al., 2003); in GSVs, they drain into branches of the vein proper or into concomitant veins (Kachlik et al., 2002). Under altered conditions, the vasa vasorum contribute to the neovascularization of atherosclerotic lesions (plaques) and neointima formation. A direct relationship between the severity of atherosclerosis and the degree of adventitial vasa vasorum is established (Kumamoto et al., 1995). Moreover, hypercholesterolemia has been shown to enhance coronary vasa vasorum adventitial neovascularization (Kwon et al., 1998) and the coexistence of hypercholesterolemia and hypertension is reported to impair adventitial vasa vasorum formation, resulting in insufﬁcient compensation for hypoxia in the thickened media (Kai et al., 2002). A blockade of angiogenesis of vasa vasorum is therefore considered a promising new strategy to inhibit vessel restenosis (Fuchs et al., 2001). Recently, it was shown that in an experimental model (common carotid artery of a rabbit) upregulation of hypoxia-induced factor-1␣ (HIF-1␣) and vascular endothelial growth factor (VEGF) expression drive angiogenesis of the vasa vasorum (Bayer et al., 2002) and that formation of adventitial vasa vasorum is induced through the HIF-1␣/ VEGF/Ets-1 (transcription factor) pathway in hypertension (Kuwahara et al., 2002). It is assumed that blockade of this pathway is a possible means to prevent restenosis of grafted or stented vessels. Although there is a long-lasting interest in the vasa vasorum, the ﬁne distribution and the three-dimensional arrangement of the entire vasa vasorum bed within the wall of blood vessels still remain to be elucidated in more details. A promising new technique, microcomputed tomography (-CT) (Jorgensen et al., 1998), enables the study of the spatial arrangement of the vasa vasorum in large specimens (Lerman and Ritman, 1999; Bentley et al., 2002; Wan et al., 2002; Gössl et al., 2003), even in its frozen state (Kantor et al., 2002). In conjunction with semiautomated tree analysis software, it can provide quantitative geometrical data not yet available by other techniques (Gössl et al., 2003). Regrettably, due to the voxel width of ⬃ 20 m used in current -CT, vasa vasorum with a diameter smaller than 40 m cannot be visualized. Thus, -CT images lack the entire capillary bed and pre- and postcapillary vasa vasorum (Kwon et al., 1998; Gössl et al., 2003). In contrast, scanning electron microscopy (SEM) of vascular corrosion casts (Murakami, 1971; Lametschwandtner et al., 1984; Aharinejad and Lametschwandtner, 1992) has a sufﬁciently high power of resolution to visualize the entire vasa vasorum circulatory bed from the feeding arterial stem vessels throughout the capillaries to the draining venous vasa vasorum. The great depth of focus of the SEM, and the possibility to take stereopaired images for subsequent 3D morphometry (Malkusch et al., 1995; Minnich et al., 1999, 2002), enables the measurement of parameters in the casted vasa vaso- 1373 rum required to calculate ﬂow and pressure gradients within the vascular network. Additionally, it can provide the vascular parameters needed to deﬁne which optimality principle(s) (Murray, 1926a, 1926b; Gössl et al., 2003) are applicable to the vasa vasorum of the aged human GSV. As yet, only the vasa vasorum of the rabbit common carotid artery have been studied using SEM of vascular casts in an experimental model (Bayer et al., 2002). The present study is the ﬁrst to analyze the vasa vasorum system in a vein. The human GSV was chosen because it is frequently used as a coronary bypass graft. It is also known that this vein has a high incidence of stenosis that dramatically impairs graft patency. In the current study, SEM of vascular corrosion casts (Murakami, 1971) in combination with 3D morphometry (Minnich et al., 1999) provides detailed qualitative insights into the 3D arrangement of the vasa vasorum, as well as quantitative data on vasa diameters, intervascular and interbranching distances, and branching angles. From these data, it is possible to calculate branching indexes of arterial, capillary, and venular bifurcations (mergings). These data can then be used to calculate optimum vessel diameters and optimum branching angles to ﬁnd out which of the four optimality principles (minimum lumen volume, minimum pumping power, minimum lumen surface, minimum endothelial shear force) (Murray, 1926a, 1926b) is achieved in the vasa vasorum network. MATERIALS AND METHODS Vascular Casting Five explanted segments of human GSV were taken during their dissection for coronary bypass grafts in ﬁve patients who granted written consent and who were aged 67–72 years (three males, two females; risk factors, triplevessel disease, hypertension, previous myocardial infarction, and/or diabetes mellitus). The segments (lengths, 3–5 cm; external diameter, 3– 4 mm) were transported to the laboratory in ice-cooled heparinized saline (1,000 I.E./ml 0.9 sodium chloride) and casted within 6 –10 hr after harvesting. GSV segments were subsequently submerged in saline at room temperature (20°C) and a feeding artery was dissected from the surrounding tissue under a dissecting microscope. It was then cannulated using a mechanical micromanipulator (WPI Instruments, Berlin, Germany) with a ﬁne-tipped glass cannula (outer diameter, 80 –120 m) and rinsed with warm (37°C) saline. Then 4 ml of prepolymerized resin (Mercox-Cl-2B; Ladd Research, Burlington, VA) was gently mixed with 1 ml of monomeric methylmethacrylic acid containing 62.5 mg paste MA (accelerator) and injected using an automatic infusion pump with a ﬂow rate of 7 ml/hr. The injection stopped when the efﬂux from the open-cut vasa vasorum of the vascular segment became highly viscous. The specimens were then left at room temperature (20°C) for 30 min to allow polymerization of the injected resin and then transferred into a water bath at 60°C to temper overnight. The following morning, the specimens were transferred to a solution of 7.5% potassium hydroxide (40°C) to remove all organic material. After maceration, the remaining plastic vascular casts were rinsed several times with distilled water, then submerged for 5–15 min in 5% formic acid. After further rinses with distilled water, the specimens were frozen (in distilled water). Ice-embedded casts 1374 LAMETSCHWANDTNER ET AL. were then freeze-dried (the ice is sublimated and dry vascular casts remain), mounted on specimen stubs, sputtered with gold, and examined in Stereoscan 250 SEM (Cambridge, U.K.) at an accelerating voltage of 10 kV. For further details on cast preparation, see Aharinejad and Lametschwandtner (1992). for an optimality principle based on minimizing lumen surface and total shear force on endothelial tissue (Zamir, 1988a). Finally, the deviation of diameters of daughter vessels and branching angles of several consecutive bifurcations from optimal values were calculated according to Equation 7: Quantiﬁcation For quantiﬁcation of the vasa vasorum, stereopaired images (tilt angle, 6°) were taken with the SEM at an accelerating voltage of 10 kV, a working distance of 10 mm, and magniﬁcations ranging from 20⫻ to 500⫻. Images were imported into the 3D morphometry system (ComServ, Minnich and Muska OEG, Salzburg, Austria) and the following measures were taken: diameters, intervascular and interbranching distances, and branching angles. Data were used to calculate the following: branching indexes ␣ (⫽ asymmetry ratios; d2/d1), diameter ratios of daughter vessels (d1, d2) to parent vessel (d0), and area ratios ␤ of branches to parent vessel [(d12 ⫹ d22)/d02]. For testing the optimality of consecutive arterial and capillary branchings and venous mergings, the two branching diameters d1 (large branch) and d2 (small branch) were normalized in terms of the diameter of the parent vessel (d0) at each bifurcation (merging). Optimally, the diameters of the two branches at a bifurcation would be (Zamir, 1988a, 1988b; Gössl et al., 2003): 1 d1 ⫽ d0 (1 ⫹ ␣3)1/3 (1) d2 ␣ ⫽ d0 (1 ⫹ ␣3)1/3 (2) To ﬁnd out the optimality principle upon which the bifurcations (mergings) of the vasa vasorum of the aged human GSVs are based, the following formulas proposed by Zamir (1988a, 1988b) and Gössl et al. (2003) were used: cos 1 ⫽ (1 ⫹ ␣3)4/3 ⫹ 1 ⫺ ␣4 2(1 ⫹ ␣3)2/3 (3) cos 2 ⫽ (1 ⫹ ␣3)4/3 ⫺ 1 ⫹ ␣4 2␣2(1 ⫹ ␣3)2/3 (4) for an optimality principle based on minimizing lumen volume and pumping power, respectively, and cos 1 ⫽ (1 ⫹ ␣3)2/3 ⫹ 1 ⫺ ␣2 2(1 ⫹ ␣3)1/3 (5) cos 2 ⫽ (1 ⫹ ␣3)2/3 ⫺ 1 ⫹ ␣2 2␣(1 ⫹ ␣3)1/3 (6) Deviation from optimal (%) ⫽ Value(s) gained using data from cast vessels ⫻ 100 Value(s) gained using Equations 1–6 (7) RESULTS Qualitative Analysis The casting of the ﬁve segments resulted in three reasonably good vascular casts of different sizes based on the areas supplied by the cannulated arteries and the length of the segments available. Gross inspection of the casts under the dissecting microscope showed that the wall of the GSV contained a rich supply of the vasa vasorum. The vasa ran predominantly parallel with the longitudinal axis of the vein. SEM of resin casts at low magniﬁcation conﬁrmed this ﬁnding and demonstrated that venous vasa by far outnumbered arterial vasa (Figs. 1– 4). The latter generally had smaller diameters and could be differentiated from the venous vasa by their round proﬁles and the characteristically long endothelial cell nuclei imprints, orientated parallel with the vessel axis (Fig. 2). Feeding arterial vasa ran in close association with the draining venous vasa (Figs. 2– 4). Diameters of feeding arteries were approximately 50 m; those of their venous partners were approximately 100 m. After inﬁltrating the external adventitial layers, the feeding arteries generally followed a longitudinal course (Figs. 2– 4). According to the terminology proposed by Schönenberger and Müller (1960), these vessels were termed ﬁrst-order arteries (1A). These arteries generated side branches, termed secondorder arteries (2A). Second-order arteries ran in a circular formation (Figs. 3 and 4), then became longitudinal and were termed third-order arteries (3A). These branched several times before they changed into capillaries (Figs. 5– 8). Locally, 2As followed a corkscrew-like course (Figs. 6 and 7) and ﬁnally formed capillaries that were arranged in a longitudinal fashion and ran parallel (Fig. 7). These capillaries had (luminal) diameters ranging from 7 to 15 m (number of measurements, n ⫽ 22). Intercapillary distances ranged from 12 to 114 m (n ⫽ 15). After average lengths of 45 to 320 m (n ⫽ 12), the capillaries gradually changed into postcapillary venules. Depending on their location, they were termed fourth- (4V) or thirdorder venules (3V), which either followed a circular (4V) or longitudinal (3V) path (Fig. 8). Locally, arteries became narrow in places with adhering plastic strips imitating myocytes and/or pericytes (Fig. 9). Venules of descending order not only changed from a circumferential course into a longitudinal course (e.g., from 4V to 3V respectively from 2 V to 1V), but often also changed from a deeper layer into a more superﬁcial layer (Fig. 10). In venous vasa vasorum, tight constrictions were often found (Figs. 11 and 12). Locally, small capillary loops were squeezed between venules, thereby overbridging the venules (Fig. 12). To ﬁt into the narrow intervenular space between two larger VASA VASORUM IN HUMAN GREAT SAPHENOUS VEIN 1375 Fig. 1. Vascular corrosion cast (VCC) of a segment of the proximal portion of the human GSV revealing the vasa vasorum from an adventitial view. Note the prevailing longitudinal arrangement of the vasa. A, artery; V, vein. Arrows mark the feeding and/or draining vessels inﬁltrating the GSV every 0.5–1.5 cm. Fig. 2. VCC of the proximal segment of the GSV. Adventitial view. Longitudinal axis of GSV: left to right. Hierarchy of feeding arterial and draining venous vasa (terminology introduced by Schönenberger and Müller, 1960). Vasa that have a longitudinal arrangement are designed orders 1, 3 and 5 (1A, 3A, 5A for arterial vasa and 1V, 3V, 5V for venous vasa) and orders 2, 4, and 6 (2A, 4A, 6A and 2V, 4V, 6V ) for vasa with a circular arrangement. Note the characteristic close association of sup- plying artery and draining vein and the difference in vessel calibers. Asterisks mark extravasations. Fig. 3. Branching pattern and caliber of feeding artery and draining vein penetrating the adventitia of the GSV every 5–15 mm. Detail from Figure 1 (boxed area). Note that the main branch of the feeding artery (asterisk) does not supply the GSV but the adjacent fatty tissue. Arrows indicate direction of blood ﬂow. Fig. 4. Longitudinal ﬁrst-order artery (1A), circular second-order artery (2A), longitudinal 3A, and circular 4A. Note that the corresponding veins (1V to 3V) have a larger diameter than the arteries. Arrows indicate direction of blood ﬂow. E, evasate. venules, anastomoses followed a hairpin-like course with the two loops close beside each other (Fig. 12). In the external media, venous vasa vasorum with diameters ranging from 24 to 40 m had a circular arrangement occupying empty spaces where vascular smooth muscle cells formerly resided. Capillaries ran a longitudinal 1376 LAMETSCHWANDTNER ET AL. Fig. 5. Branching pattern and course of third-order arteries (3A) and second- and third-order veins (2V and 3V). C, capillary. Arrows mark direction of blood ﬂow. Fig. 6. First-order artery (1A) branching off a second-order artery (2A) in an acute angle. Adventitial view. Note the ﬁrst-order vein (1V) running parallel is much larger than the artery. Capillaries drain either into secondorder (2V) or third-order veins (3V). Arrows indicate direction of blood ﬂow. Fig. 7. Terminal portion of a circular second-order artery (2A) giving rise to capillaries running parallel in a longitudinal arrangement. Detail from Figure 6 (boxed area). Arrows mark direction of blood ﬂow. Fig. 8. Transition of a third-order artery (3A) into a capillary and its merging into a third-order vein (3V). Arrows mark direction of blood ﬂow. course with intercapillary distances from 12 to 43 m (Figs. 13–16). Spacing of circular 4V was around 140 m (Fig. 13). The innermost layer of the vasa vasorum facing the luminal side consisted mostly of capillaries with draining veins at various points (Fig. 14). A close-up view revealed that capillaries formed rectangular or squared meshes with occasionally plastiﬁed pericytes (Hodde, 1981; Castenholz et al., 1982) sticking on them (Figs. 14 and 15). Plastiﬁed means that the pericytes by a hitherto controversially disputed process have been permeated by the casting medium, which polymerized and prevented their structure from digestion (removal) by the maceration solution (7.5% KOH). Postcapillary venules that drained capillaries (e.g., 4V in Fig. 14) ran a circular course over a short distance only to turn into a longitudinal one (e.g., 3V in Fig. 14). Transverse section casts VASA VASORUM IN HUMAN GREAT SAPHENOUS VEIN 1377 Fig. 9. Plastiﬁed myocyte/pericyte (P) embracing and constricting an artery. Fig. 10. Pattern and course of the adventitial vasa vasorum. Overview. Note the rather large venous anastomoses between longitudinal running ﬁrst-order veins (1V) formed by circular running second-order (2V) or third-order veins (3V) that run transversely. Note also the spacing of the origin of second-order arteries (arrowheads). 1A and 2A denote ﬁrst- and second-order arteries. Arrowheads mark origin of secondorder arteries. Arrows indicate direction of blood ﬂow. Fig. 11. Drainage of longitudinal postcapillary venules (pv) into ﬁfth(5V), fourth- (4V), and third-order veins (3V). Note the local indentations on 3 V (arrowheads). Cl, capillary loop. Arrows mark direction of blood ﬂow. Fig. 12. Third-order veins (3V) with deep indentations (arrowheads). Capillaries interposed between the veins loop and form an anastomosis between the veins. Arrows indicate direction of blood ﬂow. clearly revealed that terminal arterial vasa extended toward the innermost layer of the vasa vasorum where they capillarized (Fig. 15). The spacing of the vasa vasorum can be seen in Figure 16. Branched structures were found on capillaries facing the luminal side (Fig. 16). On higher magniﬁcation, these structures clearly identiﬁed themselves as plas- tiﬁed pericytes with their processes embracing the capillary (Figs. 16 and 17). Quantitative Analysis Three-dimensional morphometry on GSV vasa vasorum revealed that bifurcation indexes (␣) ranged from 0.4 to 1378 LAMETSCHWANDTNER ET AL. Fig. 13. Arrangement and course of the venous vasa vasorum (5V to V). Note the transition of the longitudinal ﬁfth-order vein (5V) into the circular fourth-order vein (4V) and the transition into the third-order vein (3V). At the transition sites, the veins loop and ascend (5V to 4V) or descend (4V to 3V). Note the parallel arrangement of the capillaries and the space below (asterisks) devoid of vessels. Arrows mark direction of blood ﬂow. pv, postcapillary venule. Fig. 14. Vasa vasorum facing the lumen. Luminal aspect. Capillaries form a rectangular meshwork and drain into circular or longitudinal veins. Arrows indicate direction of blood ﬂow. Fig. 15. Spatial arrangement of the vasa vasorum. Transverse-sectioned vascular cast. Note the transition of the terminal artery into the capillaries of the innermost layer. Arrows mark the direction of blood ﬂow. Fig. 16. Spacing of the vasa vasorum as shown in a transversesectioned GSV vascular cast. Note the plastiﬁed pericytes (arrowheads) embracing luminal capillaries. Asterisks mark transverse-cut vessels. Arrows indicate direction of blood ﬂow. 0.9 in arterioles, 0.7 to 1.0 in capillaries, and 0.4 to 1.0 in venules. Interestingly, in venular bifurcations (mergings), the ratio of diameters of large (d1) and small daughter vessels (d2) to the diameters of their parent vessels (d0), i.e., d1/d0 respectively d2/d0, was the same (mean, 0.6). In arterioles, this ratio was 0.9 (mean). While larger daugh- ter vessels showed the same diameter ratio to the parent vessels in both arterioles and venules (mean, 0.6), this ratio was different in smaller daughter vessels (mean, 0.9 vs. 0.6; Table 1). Analysis of the preliminary data from eight consecutive bifurcations in a single arterial vas using Equations 1– 6 3 VASA VASORUM IN HUMAN GREAT SAPHENOUS VEIN Fig. 17. Plastiﬁed pericyte embracing a capillary. Arrowheads mark lateral processes of the pericyte. Note the internal structure of the resin cast exposed by the cutting (asterisk). revealed a maximum deviation from the theoretical optimum of 20% in branch diameters; a deviation of 301.3% in branching angles 1 if for an optimality principle of minimum vessel volume and minimum pumping power is tested; a deviation of 138.8% if for an optimality principle of minimum surface and minimum drag force is tested; a deviation of ⫺81.2% in branching angles 2 if for an optimality principle of minimum vessel volume and minimum pumping power is tested; and a deviation of 84.2% if for an optimality principle of minimum surface and minimum drag force is tested (Table 2). DISCUSSION Previously unpublished studies by the authors have shown that the vascular bed of organs and tissues can be successfully cast with polymerizing resins for up to 12 hr after excision. The results from this study conﬁrm that a delicate vascular system such as the vasa vasorum of the human GSV can also be cast within this period. Since the injections were carried out via a single perforating artery, the size and quality of the vascular casts varied between injections depending on the area of the wall supplied by the artery. Because the feeding arteries were rather small in diameter (⬍ 100 m), the resin was initially injected retrogradely via the much larger draining vein. Venous injections, however, failed in all cases; only the larger venous vasa vasorum ﬁlled. Capillaries failed to ﬁll. Injection was therefore continued orthogradely via one perforating artery, exposed, and cannulated under the dissecting microscope. Recently, SEM of resin corrosion casts was used in an experimental rabbit model to give an overview of the amount of neovascularization of the vasa vasorum occurring between the rabbit carotid artery cuffed by a segment of its contralateral partner (Bayer et al., 2002). This study 1379 gave no detailed account of the 3D arrangement of the vasa vasorum within the wall of the artery. However, the technique proved valuable as it allowed the authors to demonstrate that some of the newly formed vasa vasorum branched off the lumen of the carotid artery and showed the characteristic arterial endothelial cell nuclei imprint patterns ﬁrst described on the surface of vascular casts by Miodonski et al. (1976). The present study allowed, for the ﬁrst time, a thorough analysis of the entire microvascular bed and its 3D arrangement of the vasa vasorum of a blood vessel in general and of the human GSV in particular. The present work has provided the ﬁrst quantitative data on vasa parameters and the general geometry of the vasa vasorum network within the wall of a vein. Previous work used careful dissections and/or India ink injections to visualize the vascular patterns. The experienced anatomist can positively identify all types of vessels in India ink-injected specimens by careful light microscopical inspection (Kachlik et al., 2002). However, only SEM of microvascular corrosion casts allows reliable information to be gained on the hierarchy of vessels, their exact branching patterns and branching angles, as well as the relationships between parent and daughter vessels in arterial branchings and in venous mergings. As the vascular corrosion casts were devoid of any tissue, no direct statement can be made regarding the proximity of the capillaries to the lumen of the saphenous vein. According to our previous ﬁndings, the proximity is between 250 and 660 m to the lumen of the saphenous vein (Kachlik et al., 2002). If these results are compared with those of Bayer et al. (2002), who were the ﬁrst to demonstrate the adventitial vasa vasorum network of the rabbit common carotid artery, it should be noted that the studies deal with anatomically and physiologically different kinds of blood vessels (artery versus vein), different species (rabbit vs. human), differently sized vessels (1 vs. 4 mm), different aged vessels (months vs. 69 –73 years), different casting conditions (in vivo casting with vessel under normal pressure vs. ex vivo with vessel under zero intravasal pressure), and different injection routes (common carotid artery vs. single feeding artery of the vasa vasorum). Both studies, however, clearly show the rich vascularization of the vessel walls by the vasa vasorum [compare Fig. 5A in Bayer et al. (2002) with Fig. 1 here]. Interestingly, the adventitial vasa vasorum are randomly arranged in the rabbit common carotid artery, while those in the human saphenous vein follow a clear longitudinal orientation (Fig. 1). In the experimental model, the newly formed vasa vasorum are located between the intact common carotid artery and the cuff formed by a segment of the contralateral vessel. This area was occupied by connective tissue that had no obvious characteristic arrangement that could serve as a support structure for the newly grown vessels. In the wall of the saphenous vein, the vasa vasorum, however, followed the course of the longitudinal connective tissue ﬁbers and the circular vascular smooth muscle cells in the outer media layers. Moreover, in the human great saphenous vein, the typical mammalian pattern of blood supply with a small artery running closely aside a much larger vein (clearly seen in Fig. 1) is found, whereas in the cuffed rabbit common carotid, a large undulating artery arose from the 1380 LAMETSCHWANDTNER ET AL. TABLE 1. Ranges and means of branching indexes ⌬, ratios of diameters of larger daughter vessel (d1) and smaller daughter vessel (d2), area ratios ␤, sum of branching angles between daughter vessels (Q1 ⴙ Q2) of selected arterial, capillary and venular vasa vasorum of the GSV* Arterioles Range Mean Capillaries Range Mean Venules Range Mean ␣ d1/d0 d2/d0 ␤ Q1 ⫹ Q2 Interbranching distance (m) 0.4–0.9 0.7 0.7–1.0 0.9 0.4–0.9 0.6 0.8–1.7 1.2 90–150 119 44–131 88 0.7–1.0 0.8 0.7–1.0 0.9 0.6–0.9 0.7 0.8–1.7 1.3 77–154 116 14–117 39 0.4–1.0 0.7 0.3–1.0 0.6 0.4–0.9 0.6 0.6–1.6 1.0 57–199 134 14–111 60 *Note that (merging) venules are similar in dimension while in arterial bifurcations these ratios are different. TABLE 2. Characterization of eight consecutive branchings in a single arterial vasa vasorum. Branching level d0 (m) 1 28.52 2 29.07 3 19.97 4 18.47 5 15.51 6 15.16 7 10.63 8 11.80 d1 (m) d2 (m) d1/d0 measured d1/d0 optimal Deviation from optimal (%) d2/d0 measured d2/d0 optimal Deviation from optimal (%) ␣ 1 (°) measuredb 1 optimalb 1 optimalc Deviation from optimal (%)b,c 2 (°) measured 2 optimalb 2 optimalc Deviation from optimal (%) 28.15 26.06 0.99 0.83 23.57 20.69 0.81 0.84 17.96 12.72 0.90 0.91 17.34 16.23 0.94 0.82 18.63 14.46 15.07 11.72 0.99 0.88 11.11 7.75 11.35 8.44 0.96 0.89 19.2 0.91 0.76 ⫺1.2 0.71 0.73 1.1 0.64 0.63 19.5 0.88 0.76 d 12.5 0.77 0.68 d 0.63 19.7 0.93 49.08 34.10 48.21 ⫺2.7 0.88 29.86 32.23 46.33 1.6 0.71 75.55 23.37 38.87 15.8 0.94 40.32 34.58 48.60 0.78 69.93 26.82 42.00 13.2 0.78 67.14 26.89 42.07 0.70 91.63 22.83 38.37 9.1 0.74 62.53 25.17 40.52 43.9/1.8 31.67 40.86 53.65 ⫺3.1/⫺35.5 13.39 43.20 55.48 223.3/94.4 9.84 52.29 62.39 16.6/⫺17.0 23.08 40.38 53.27 160.8/66.5 20.11 48.52 59.55 149.6/59.5 23.23 48.43 59.48 301.3/138.8 53.52 52.90 62.85 148.4/54.3 26.75 50.30 60.90 1.8/⫺14.8 ⫺46.8/⫺56.1 a 0.88 a 0.68 d ⫺22.5/⫺41.0 ⫺69.0/⫺79.0 ⫺81.2/⫹84.2 ⫺42.8/⫺56.7 ⫺58.6/⫺66.2 ⫺52.0/⫺60.9 a 0.91 a d 7.9 0.72 0.66 a Atypical branching with one daughter vessel larger than the parent vessel. Principle of minimum vessel volume and minimum pumping power. c Principle of minimum surface and minimum drag force (Zamir, 1988a, 1988b). d No calculation done because ofa. b lumen of the carotid artery with no vein aside [see Fig. 5B in Bayer et al. (2002)]. The vein ran at some distance away from the artery and was only slightly wider. This difference most likely resulted from the experimental situation (cufﬁng) with subsequent neovascularization within the space between the cuff and the native vessel. A comparison with the ﬁndings gained by -CT (Kwon et al., 1998; Wilson et al., 2002; Gössl et al., 2003) clearly shows that -CT lacks all capillaries and pre- and postcapillary vessels smaller than ⬃ 40 m. Here, SEM of corrosion casts is superior to the previous technique, as it enables the detailed study of each microcirculatory unit of the vasa vasorum and thus allows the origin and merging of individual vessels to be clearly deﬁned. As vascular casts made from Mercox-CL-2B tend to be very brittle, superﬁcial vessels hiding vessels underneath can be removed with the help of a ﬁne tipped insect pin and thus visualization of successive layers of vessels becomes possible. For more details on how to dissect microvascular corrosion casts, see Aharinejad and Lametschwandtner (1992). One of the advantages of -CT, namely, the visualization of the vascular bed within a rather large specimen, can also be gained by removing individual vascular layers until the entire cast is analyzed. As all tissue was removed during maceration, individual vasa vasorum territories with deﬁned layers of the wall of the saphenous vein can be attributed only by correlation with light microscopy of tissue sections done previously (Kachlik et al., 2002). A partial maceration of resin-injected specimens would be possible (Castenholz, 1983), but would allow the correlation of exposed cast blood vessels with the tissue surface only, and deeper layers would be still obscured from investigation. Because of its richness, the vasa vasorum are a unique source of neoangiogenesis (via sprouting) under pathological conditions (e.g., varicosis, intimal hyperplasia, atherosclerosis). We hypothesize that the sites that are most likely to give origin to vascular sprouts are the areas where venules change their course at an acute angle. Here the venous endothelium will be exposed to a higher shear VASA VASORUM IN HUMAN GREAT SAPHENOUS VEIN rate that in turn will initiate vessel sprouting. The lack of sprouts in our specimens is most likely due to the age and health of the specimens from which the major part was used for coronary bypass surgery. With respect to the kinks of the venous vasa vasorum casts in the current study, these structures can be attributed to the age of the donors rather than to pathological changes of the vasa vasorum. If the given spatial arrangement of the vasa vasorum is considered, it can be concluded that the GSV is probably less vulnerable to dimensional changes of the vessel wall emerging from changing intravasal volumes and/or pressures resulting from muscular contractions or stretching of the vein. The corkscrew-like course of circular-running second-order arteries (2A) in the current specimens most likely resulted from the lack of intraluminal pressure during the casting procedure. On the other hand, this course enables a distension of the wall of the vein without severely impairing blood ﬂow of the vein. Optimality in Vasa Vasorum Quantitative analysis of the longitudinally ﬁrst-order arteries already revealed that the diameters of the main stem of these vessels remained almost the same over several branching levels and thus most likely serves as a distributing vessel (Zamir, 1988a, 1988b). The external vasa vasorum of the porcine coronary artery is also reported to have a distributing function (Gössl et al., 2003). Normalized ratios of branch diameters that ranged from 0.81 to 0.99 (d1/d0) respectively 0.64 to 0.91 (d2/d0) in six consecutive bifurcations in a single ﬁrst-order arteriole in the GSV casts (Table 2) support a delivering function of this vessel. A delivering function of arteries is indicated by a signiﬁcant decrease of vessel diameters at each bifurcation level. In the porcine coronary artery, the internal vasa vasorum are reported to have a delivering function (Gössl et al., 2003). Interestingly, branching angles varied to a much greater degree around theoretical optima than diameters of daughter vessels. These preliminary data corroborate the ﬁndings of Gössl et al. (2003), where the diameters in the arterial vasa vasorum were scattered around much closer to theoretical optima. This scattering of branching angles has been reported in the vasculature in general (Zamir and Brown, 1982). More data, however, are needed to evaluate these ﬁndings statistically. ACKNOWLEDGMENTS The authors thank Dr. W.D. Krautgartner for technical support in SEM, Ms. Synnöve Tholo for assistance of the SEM work, Mr. Andreas Zankl for image processing, Ms. Birgit Peukert and Ms. Gudrun Thallinger for 3D data acquisition, and T.R. Koll for programming Equations 1-6, making calculations much easier. This work was supported by grants from the Charles University Prague (GA UK 57/2002/C to D.K. and J.S.) and the Stiftungs- und Förderungsgesellschaft der Paris-Lodron-Universität Salzburg (to A.L.). LITERATURE CITED Aharinejad SH, Lametschwandtner A. 1992. Microvascular corrosion casting. New York: Springer-Verlag. Bayer IM, Caniggia I, Adamson SL, Langille BL. 2002. Experimental angiogenesis of arterial vasa vasorum. Cell Tissue Res 307:303– 313. 1381 Bentley MD, Ortiz MC, Ritman EL, Romero JC. 2002. The use of microcomputed tomography to study microvasculature in small rodents. Am J Physiol Integr Comp Physiol 282:R1267–R1279. Castenholz A, Zöltzer H, Erhardt H. 1982. Structures imitating myocytes and pericytes in corrosion casts of terminal blood vessels: a methodical approach to the phenomenon of “plastic strips” in SEM. Mikroskopie 39:95–106. Castenholz A. 1983. The outer surface morphology of blood vessels as revealed in scanning electron microscopy in resin cast, non-corroded tissue specimens. Scan Electron Microsc 4:1955–1962. Fuchs S, Kornowski R, Leon MB, Epstein SE. 2001. Antiangiogenesis: a new potential strategy to inhibit restenosis. Int J Cardiovasc Interv 4:3– 6. Gössl M, Rosol M, Malyar NM, Fitzpatrick LA, Beighley PE, Zamir M, Ritman EL. 2003. Functional anatomy and hemodynamic characteristics of vasa vasorum in the wall of porcine coronary arteries. Anat Rec 272A:526 –537. Hodde KC. 1981. Cephalic vascular patterns in the rat. Thesis. Amsterdam: University of Amsterdam. Jorgensen SM, Demirkaya O, Ritman EL. 1998. Three-dimensional imaging of vasculature and parenchyma in intact rodent organs with X-ray micro-CT. Am J Physiol 275:H1103–H1114. Kachlik D, Lametschwandtner A, Rejmontova J, Stingl J, Vanek I. 2002. Vasa vasorum of the human great saphenous vein. Surg Radiol Anat 24:377–381. Kai H, Kuwahara F, Tokuda K, Shibata R, Kusaba K, Niijama H, Tahara N, Nagata T, Imaizumi T. 2002. Coexistence of hypercholesterolemia and hypertension impairs adventitial vascularization. Hypertension 39:455– 459. Kantor B, Jorgensen SM, Lund PE, Chmelik MS, Reyes DA, Ritman EL. 2002. Cryostatic micro-computed tomography imaging of arterial wall perfusion. Scanning 24:186 –190. Kumamoto M, Nakashima Y, Sueishi K. 1995. Intimal neovascularization in human atherosclerosis: its origin and pathophysiological signiﬁcance. Hum Pathol 26:450 – 456. Kuwahara F, Kai H, Tokuda K, Shibata R, Kusaba K, Tahara N, Niiyama H, Nagata T, Imaizumi T. 2002. Hypoxia-inducible factor1alpha/vascular endothelial growth factor pathway for adventitial vasa vasorum formation in hypertensive rat aorta. Hypertension 39:46 –50. Kwon HM, Sangiorgi G, Ritman EL, McKenna C, Holmes DR, Schwartz RS, Lerman A. 1998. Enhanced coronary vasa vasorum neovascularization in experimental hypercholesterolemia. J Clin Invest 101:1551–1556. Lametschwandtner A, Lametschwandtner U, Weiger T. 1984. Scanning electron microscopy of vascular corrosion casts technique and applications. Scanning Electron Microsc 1984:663– 695. Lerman A, Ritman EL. 1999. Evaluation of microvascular anatomy by micro-CT. Herz 24:531–533. Malkusch W, Konerding MA, Klapthor B, Bruch J. 1995. A simple and accurate method for 3-D measurements in microcorrosion casts illustrated with tumor vascularization. Anal Cell Pathol 9:69 – 81. Minnich B, Leeb H, Bernroider EWN, Lametschwandtner A. 1999. Three-dimensional morphometry in scanning electron microscopy: a technique for accurate dimensional and angular measurements of microstructures using stereopaired images and digital image analysis. J Microsc 195:23–33. Minnich B, Bartel H, Lametschwandtner A. 2002. How a highly complex three dimensional network of blood vessels regresses: the gill blood vascular system of tadpoles of Xenopus during metamorphosis—a SEM study on microvascular corrosion casts. Microvasc Res 64:425– 437. Miodonski A, Hodde KC, Bakker C. 1976. Scanning electron microscopy of plastic-corrosion-casts: morphological differences between arteries and veins. Berichte Direktabb Oberﬂ 9:435– 442. Murakami T. 1971. The application of the scanning electron microscope to the study of the ﬁne distribution of the blood vessels. Arch Histol Jap 32:445– 454. Murray CD. 1926a. The physiological principle of minimum work applied to the angle of branching of arteries. J Gen Physiol 9:835–841. 1382 LAMETSCHWANDTNER ET AL. Murray CD. 1926b. The physiological principle of minimum work: 1, the vascular system and the cost of blood volume. Proc Natl Acad Sci USA 12:207–214. Schönenberger F, Müller A. 1960. On the vascularization of the wall of the bovine aorta. Helvet Physiol Pharmacol Acta 18:136 –150. Wan SY, Ritman EL, Higgins WE. 2002. Multi-generational analysis and visualization of the vascular tree in 3D micro CT images. Comput Biol Med 32:55–71. Williams JK, Heistad DD. 1996. The vasa vasorum of the arteries. J Mal Vasc 21(Suppl C):266 –269. Wilson SH, Herrmann J, Lerman LO, Holmes DR, Napoli C, Ritman EL, Lerman A. 2002. Simvastatin preserves the structure of coronary adventitial vasa vasorum in experimental hypercholesterolemia independent of lipid lowering. Circulation 105:415– 418. Wolinsky H, Glagov S. 1967. Nature of species differences in the medial distribution of aortic vasa vasorum in mammals. Circ Res 20:409 – 421. Zamir M, Brown N. 1982. Arterial branching in various parts of the cardiovascular system. Am J Anat 163:295–307. Zamir M. 1988a. The branching structure of arterial trees. Comments Theor Biol 1:15–37. Zamir M. 1988b. Distributing and delivering vessels of the human heart. J Gen Physiol 91:725–735.