Endothelial vesiculo-vacuolar organelles, pockets and multi-layered fenestrated lamellae in the capillaries of the mouse carotid body

ARTICLE IN PRESS Ann Anat 187 (2005) 333—344

www.elsevier.de/aanat

Endothelial vesiculo-vacuolar organelles, pockets and multi-layered fenestrated lamellae in the capillaries of the mouse carotid body Amina B. El-Fadalya,b,, Wolfgang Kummera a

Institute for Anatomy and Cell Biology, Justus-Liebig-University, Aulweg 123, 35385 Giessen, Germany Anatomy Department, Faculty of Medicine, Cairo University, Egypt

b

Received 14 February 2005; accepted 5 April 2005

KEYWORDS Carotid body; Endothelial cells; Plasmalemmal vesicles; Fenestrae; Endothelial pockets; Vacuolar organelles; Ultrastructure

Summary Fenestrated capillaries represent the basic structural unit in the carotid body. They mediate a characteristic hyperpermeability state in this organ. Endothelial fenestrae and plasmalemmal vesicles are of particular importance in this respect. The present electron microscopic study of the capillaries of the mouse carotid body demonstrates prominent endothelial cell structures that are suggested to be closely related to endothelial fenestrae and plasmalemmal vesicles. These structures include: (1) Vesiculo-vacuolar organelles formed by fusion and intercommunication of vesicles and vacuoles of variable dimensions. (2) Pockets in the form of fenestrated membranebound vacuoles that communicate either with the capillary lumen, pericapillary space or both via multiple apertures or fenestrae. (3) Multi-layered fenestrated lamellae where the endothelial cytoplasm is divided into multiple attenuated sheets provided with several fenestrae. The latter two structures are preferentially located in the thick perinuclear region of the endothelial cell. Their fenestrae are always distributed in linear series and show close similarity to the usual chains of fenestrae in the attenuated periphery of the endothelial cells. The individual apertures of the fenestrated vacuoles and multi-layered fenestrated lamellae are closely similar to the stomata of fully opened plasmalemmal vesicles suggesting a relationship between them. Morphological and morphometrical analysis of a series of fenestrae belonging to these structures revealed that they are identical to the usual chains of fenestrae in the attenuated periphery of the endothelial cells. & 2005 Elsevier GmbH. All rights reserved.

Corresponding author. Tel.: +49 641 99 47001; fax: +49 641 99 47009.

E-mail address: [email protected] (A.B. El-Fadaly). 0940-9602/$ - see front matter & 2005 Elsevier GmbH. All rights reserved. doi:10.1016/j.aanat.2005.04.003

ARTICLE IN PRESS 334

Introduction Blood capillaries are lined by a monolayer of highly specialized endothelial cells. These cells mediate continuous exchange of water and solutes between the circulating blood plasma and the interstitial fluid. For this exchange, endothelial fenestrae and plasmalemmal vesicles are of particular importance. In continuous capillaries, the role of plasmalemmal vesicles in the transcellular passage of substances is now well accepted (Ghitescu and Bendayan, 1992; Predescu and Palade, 1993; Schnitzer et al., 1995; Predescu et al., 1997). In fenestrated endothelia, there exist openings or fenestrae that allow free exchange of materials. In addition, numerous vesicles have also been reported to be involved in transendothelial transport in this type of endothelia (Clementi and Palade, 1969). Although there is general agreement as to the role of plasmalemmal vesicles in transport across the endothelium, the issue concerning the mechanism(s) underlying this transport is under debate. Some investigators have suggested a shuttling of plasmalemmal vesicles or caveolae between the luminal and abluminal fronts of the endothelial cells (Palade, 1960; Simionescu et al., 1973; Ghitescu and Bendayan, 1992; Bendayan and Rasio, 1996; Rasio and Bendayan, 2002). In this model, the material is endocytosed on one front of the endothelial cell, and the vesicle is then transported towards the other front where its content is discharged by exocytosis. However, internalisation of caveolae has been questioned (Bundgaard, 1983; Bundgaard et al., 1983; Anderson, 1993; van Deurs et al., 1993). These authors have remarked that, in endothelial cells, most, if not all, caveolae connected to the surface are rather stationary and do not detach and move within the cell. On the other hand, according to Parton et al. (1994), plasmalemmal vesicles can lose their connection to the cell surface; a process being regulated by a kinase activity. Another mechanism involving plasmalemmal vesicles in the process of transport across the endothelium is fusion of such vesicles and possible formation of transendothelial channels connecting the luminal and abluminal fronts of the endothelial cells (Hashimoto, 1972; Bendayan and Rasio, 1996; Rasio and Bendayan, 2002). These channels correspond to a chain of two or more intercommunicating vesicles opening simultaneously on both cell fronts. In addition, fusion and intercommunication between clusters of vesicles and vacuoles with the formation of structures termed vesiculo-vacuolar organelles (VVOs) have also been reported as

A.B. El-Fadaly, W. Kummer participating in transcellular transport of the endothelium (Kohn et al., 1992; Dvorak et al., 1996; Feng et al., 1996; Dvorak and Feng, 2001). Opening of the VVOs on the luminal and abluminal surface membrane and creation of transcellular pathways was confirmed by the examination of ultrathin serial sections and computer-assistedthree-dimensional reconstructions (Feng et al., 1996). The VVOs were considered by Kohn et al. (1992), Dvorak et al. (1996) and Feng et al. (1996) to be characteristics of hyperpermeable endothelium of normal venules and tumour associated microvessels although similar structures have also been detected in capillaries of other microvascular beds such as brain (Hashimoto, 1972) and muscle (Simionescu et al., 1975). In the carotid body microvasculature, VVOs have not been previously reported. The present morphological study was carried out to investigate these organelles and related structures in the fenestrated capillaries of the mouse carotid body. Particular attention was paid to the relationship between these structures and the endothelial fenestrae. In the course of this investigation, the occurrence of a peculiar endothelial structure was noted which we call ‘‘multi-layered fenestrated lamellae’’, that appears to be structurally related to VVOs and that has not been described previously in any vascular bed.

Materials and methods Four male adult 129SvEv x CF1 mice and two adult female FVB mice were sacrificed by inhalation of sevoflurane (Abbott, Wiesbaden, Germany). Immediately upon arrest of respiration, the thorax was opened and a cannula was inserted into the ascending aorta via the left ventricle. The vascular system was flushed with heparin and lidocaine containing rinsing solution (Forssmann et al., 1977) and the animals were fixed by perfusion with 1.5% glutardialdehyde, 2% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The carotid bifurcations of both sides were dissected, stored for an additional 3 h in the same fixative, and then washed in 0.1 M Tris–HCl buffer, osmicated for 1 h in aqueous 1% OsO4, washed for 3
15 min in 0.05 M maleate buffer (pH 5.2), stained en bloc for 1 h in 1% uranyl acetate in maleate buffer at pH 6.0, washed again for 3
15 min in 0.05 M maleate buffer (pH 5.2), and then routinely dehydrated in ethanol/propylene oxide and embedded in Epon. Ultrathin sections were cut with an ultramicrotome (Reichert Ultracut E, Bensheim, Germany). The

ARTICLE IN PRESS Endothelial vesiculo-vacuolar organelles in the capillaries of the mouse carotid body sections were then stained with alkaline lead citrate and viewed with an EM 902 transmission electron microscope (Zeiss, Jena, Germany). The data obtained from size measurements on photographic prints were expressed as mean7standard deviation (S.D.).

Results The descriptive analysis of the structures in the endothelial cells of the fenestrated capillaries was performed in a total of 304 capillary profiles. This was made for three distinct regions: (1) a thick perinuclear region which contained the nucleus, (2) a peripheral attenuated region, and (3) a parajunctional zone of about 0.5 mm for each adjoining cell along the intercellular boundaries (Fig. 1). The peripheral attenuated region showed fenestrated and non-fenestrated segments which alternate at random intervals along the course of the capillary. The endothelial cell thickness at the fenestrated areas was extremely attenuated to 6278 nm (range: 40–92 nm) whereas, at the nonfenestrated segments, it varied considerably from 86 up to 568 nm. The numerical fenestration density of different capillary profiles showed considerable variation with a minimum of 3 and a maximum of 42 fenestrae per capillary cross section. This variation was particularly evident

Figure 1. Cross section profile of a fenestrated capillary showing the thick portion of an endothelial cell containing the nucleus (N) and attenuated parts with fenestrae (F). The arrows point to the junctions between adjacent endothelial cells. A Golgi apparatus (G) is seen in the cytoplasm at the luminal aspect of the nucleus. (Bar ¼ 1 mm.)

335

between capillary profiles having similar diameters. Within each fenestrated segment, the fenestrae were regularly spaced at an average of 204750 nm, center to center (range: 149–297 nm). The individual fenestrae measured about 6778 nm in diameter (range: 56–82 nm, n ¼ 1556). Some of them appeared as cylindrical channels with no detectable substructure other than a homogeneous background of medium electron density. Others exhibited a single-layered membrane or diaphragm, often with a central knob (Fig. 2). A large population of plasmalemmal vesicles was found distributed all over the endothelial cells. The

Figure 2. Substructure of the endothelial fenestrae. (a) Two fenestrae show no visible substructure other than a homogeneous material of medium electron density. (b) A fenestra exhibits a single-layered diaphragm with a central knob (arrow). The stoma of a fully opened vesicle (V) is filled with a material with electron density comparable to that in the adjacent fenestrae. P ¼ pericyte, L ¼ lumen (Bar ¼ 0.25 mm).

ARTICLE IN PRESS 336

A.B. El-Fadaly, W. Kummer

vesicles were generally spherical with an approximate diameter of 50–70 nm (n ¼ 3600). The total number of the vesicles per capillary cross section varied considerably from animal to animal as well as from capillary to capillary in closely adjacent areas. Quantitative evaluation, however, revealed that the number of the vesicles had a tendency to decrease as that of the fenestrae increases (linear correlation coefficient of 0.574, po0:05) (Fig. 3). The vesicles might lie free in the cytoplasmic matrix or contact the surface membrane and open into the capillary lumen or the pericapillary space by vesicular stomata. These stomata might show a single-layered diaphragm with a central knob (Fig. 4). Generally, the stomata of the fully opened vesicles; i.e. those opening directly into the surface without a narrow neck, measured 62714 nm (range: 52–91 nm, n ¼ 1480). They were filled with a homogeneous material of a characteristic texture and of an electron density comparable to that found in the endothelial fenestrae.

Differential counts of vesicle profiles opening at the two fronts of the endothelial cells showed some variations between the different regions of these cells. In the peripheral attenuated region, the vesicular stomata were almost equally distributed on both cell fronts. In the perinuclear region, they were more numerous at the abluminal front while in the parajunctional zone they were mainly concentrated at the luminal front (Table 1). The plasmalemmal vesicles showed a marked tendency to aggregate in chains or clusters. Profiles suggestive of fusion and communication of vesicles and vacuoles were seen in all the capillaries studied (Fig. 5). The fused vesicles and vacuoles formed

Number of Fenestrae

50 d = 5-10 µm d< 5 µm d >10 µm

40 30

r = -0,574 20 10 0 0

5

10

15

20

25

30

35

40

45

Number of Vesicles

Figure 3. Numbers of endothelial fenestrae versus plasmalemmal vesicles in the attenuated periphery of 20 randomly chosen transverse capillary sections. There is an inverse relation between the numbers of fenestrae and that of vesicles (linear correlation coefficient of 0.574, po0:05). d ¼ capillary diameter.

Figure 4. A peripheral attenuated part of an endothelial cell showing a number of vesicles located within the cytoplasmic matrix. The inset shows one of these vesicles opening into the capillary lumen (L) via a vesicle stoma. This stoma is filled with a homogeneous matrix of medium electron density and displays a single-layered diaphragm with a central knob (K). (Bar ¼ 0.25 mm, bar in inset ¼ 0.1 mm.)

Table 1. Frequency distribution of plasmalemmal vesicles in the three different regions of 40 randomly chosen capillary endothelial cells Region

Perinuclear Peripheral attenuated Parajunctional

Number of vesicles

1900 1220 480

Proportion L (%)

P (%)

C (%)

6.3 26 58.3

21 31 29.2

72.7 43 12.5

L ¼ vesicles opening on the luminal surface of the endothelial cell; P ¼ vesicles opening on the pericapillary surface of the endothelial cell; C ¼ vesicles lying free in the cytoplasmic matrix. As there was no significant difference in regional proportion of vesicles in the endothelial cells studied, counts performed for each region were pooled. In the parajunctional zone, counts of vesicles opening into the intercellular cleft are included in (P).

ARTICLE IN PRESS Endothelial vesiculo-vacuolar organelles in the capillaries of the mouse carotid body channels of extremely heterogeneous morphology. These channels might establish continuity with the surface membrane at the channel stomata. The

337

individual units of the channels were generally spherical and outlined by a triple-layered membrane. They had variable dimensions, usually 4100 nm diameter. The direction of vesicle aggregation and channel orientation varied in the different regions of the endothelial cell. In the perinuclear region, they were mostly oriented parallel to the cell surface (Fig. 5a). In the peripheral attenuated region, they appeared oblique or perpendicular to the cell surface. In such a case, they might communicate with the capillary lumen or the pericapillary space by a single stoma (Fig. 5b). Profiles of vesicles forming channels parallel to the cell surface were also seen in this region, but at a much lower frequency. In such cases, the channels might open at the cell surface by multiple stomata (Fig. 5c). In the parajunctional zone, the vesicles tended to open separately to the cell surface. Intracytoplasmic channels were less frequently seen in this zone. In certain areas, the aggregated vesicles or cytoplasmic channels were spatially related to microtubules or bundles of cytoplasmic filaments and appeared to closely follow their course (Figs. 5b and 6). Membrane-bound vacuoles of variable dimensions, ranging from 0.3 to 2.0 mm were seen within the cytoplasmic matrix of 31 endothelial profiles (Figs. 6 and 7). Of these cases, 26 vacuoles were found in the perinuclear region of the endothelial cells. Those with larger diameter were predominantly located on the abluminal aspect of the nucleus. The vacuoles were characteristically situated immediately (distance of 6773 nm, range: 63–74 nm) beneath the cell surface. They communicated either with the capillary lumen or with the pericapillary space via chains of apertures. In most cases, the vacuoles were found within focal collections of vesicles that might contact their surfaces or open into their lumina. In favourable sections, a chain of apertures belonging to large

Figure 5. Perinuclear (a) and peripheral (b, c) regions in different endothelial cells showing cytoplasmic channels apparently formed by fusion and intercommunication of vesicles and vacuoles. In (a) and (c) the channels are oriented more or less parallel to the cell surface. In (b) they follow an oblique course between the two fronts of the cell. Bundles of cytoplasmic filaments (arrowheads) are shown closely related to these oblique channels. The individual component units of the channels are large and show variable dimensions when compared with the small, more uniform-sized non-fused vesicles (arrows). A trilaminar membrane outlining the channel in (c) is evident. This channel opens at the surface by two stomata (S). (Bar ¼ 0.25 mm.)

ARTICLE IN PRESS 338

Figure 6. (a) A cluster of vesicles aggregated near the pericapillary surface of the endothelial cell. A membrane-bound vacuole (V) is located within this cluster immediately beneath the surface membrane. It opens into the pericapillary space by two apertures (arrows). No morphological difference can be noted between these apertures and the fenestrae located in a usual fenestrated segment (F). Higher magnification (b) of the region indicated by rectangle in (a) shows prominent microtubules (MT) spatially related to the aggregated vesicles. It appears that the direction of vesicle aggregation is guided by the course of these microtubules. L ¼ lumen (Bar ¼ 0.5 mm).

vacuoles, together with stomata of vesicles and smaller vacuoles in their immediate vicinity, were seen in one plane of section. All of them appeared distributed in a linear series at an average spacing of 261743 nm, centre to centre (range: 216–300 nm) (Fig. 7b). These quantitative data and the general ultrastructural characteristics

A.B. El-Fadaly, W. Kummer were identical to those of a chain of fenestrae in the usual fenestrated segments of the attenuated periphery of the endothelial cells. A peculiar structural feature of the capillaries of the mouse carotid body was the division of the endothelial cytoplasm into two attenuated laminae, each with several fenestrae (Fig. 8). This was noticed in nine capillary profiles. The occurrence of multi-layered fenestrated lamellae (MFL) could also be noted in the thick perinuclear region of two additional capillary profiles. Here, the whole thickness of the endothelial cytoplasm was divided into multiple fenestrated sheets with clear spaces in between (Fig. 9). The cytoplasm in the vicinity of the MFL usually contained numerous intercommunicating vesicles and vacuoles of variable dimensions. Moreover, tangential views of the fenestrated laminae were obtained and their fenestrae appeared circular or vesicular with a diameter corresponding to the smallest vesicle diameter measured in the surrounding cytoplasm. Another characteristic and frequently occurring feature of the endothelial cells was the appearance of cell processes projecting into the capillary lumen. The extent of their occurrence showed some variation from animal to animal as well as from capillary to capillary in closely adjacent areas. Their general distribution, however, showed almost a constant pattern along the entire length of the endothelial cell. They were mainly concentrated at the luminal front of the perinuclear and parajunctional areas with extreme paucity at the peripheral attenuated region (Fig. 10). Although it was generally observed that these processes were more numerous when vesicles were infrequent, images showing close association with cytoplasmic vacuoles with heterogeneous morphology were frequently encountered (Fig. 10a and b). The processes were generally not accompanied by deep infoldings of the underlying nuclear membrane.

Discussion Fenestrated capillaries have openings or fenestrae that allow free exchange of materials between blood and tissue fluid. In addition to these fenestrae, the endothelial cells of this type of microvasculature display numerous vesicles that are also involved in transendothelial permeability (Clementi and Palade, 1969). The present quantitative data revealed appreciable variations in the number of the vesicles and fenestrae from animal to animal as well as from capillary to capillary in closely adjacent areas. Despite these variations,

ARTICLE IN PRESS Endothelial vesiculo-vacuolar organelles in the capillaries of the mouse carotid body there was a general tendency of the vesicle number to decrease as that of the fenestrae increases. These observations might reflect a dynamic rela-

339

tionship between both structures. Previous reports indicated that endothelial fenestrae in vitro could be formed from caveolae or caveolae-like vesicles (Chen et al., 2002). This view was challenged by the study of So ¨rensson et al. (2002) due to the lack of caveolin-1 in the membrane adjacent to the fenestrae and by the unaltered ultrastructure of the fenestrae in caveolin-1 deficient mice. Therefore, fenestrae might be formed, at least in part, by caveolin-free vesicles. The present study demonstrates that vesicle stomata and fenestrae share common morphological criteria. The stoma of a fully opened vesicle approximates the diameter of a fenestra. Generally, vesicle stomata and fenestrae appeared as channels filled with a homogeneous material of almost the same texture and electron density. Both apertures might be bridged by a single-layered diaphragm with a central knob. These data are in keeping with the fact that the presence of a diaphragm is a specific feature of fenestrae as well as endothelial caveolae that distinguishes them from caveolae of other cell types (Stan, 2002). Furthermore, both diaphragms seem to share essential structural elements as evidenced by the identification of the protein PV-1 in association with both caveolar and fenestral diaphragms (Stan et al., 1997; Stan et al., 1999a, b; Stan, 2004). Accordingly, fenestrae and vesicle stomata are essentially identical structures. In the present material, the plasmalemmal vesicles of the fenestrated capillaries of the mouse carotid body had a great tendency to aggregate in chains or clusters. In certain areas channels formed of communicating vesicles and vacuoles of variable dimensions were also demonstrated. They morpho-

Figure 7. This figure demonstrates large membranebound vacuoles (V) in the perinuclear region of three different capillaries. The cytoplasm facing the pericapillary surface (a, b) and the capillary lumen (L) (c) is attenuated to a thin sheet provided with several fenestrae. Some vesicles (small arrows) are seen opening into the lumen of the cavities in (a) and (b). Other vesicle and smaller vacuoles (large arrows) are also seen in the immediate vicinity of the large vacuole in (b). They open into the pericapillary space by individual stomata. These stomata together with the fenestrae belonging to the large vacuole are arranged in a linear series at almost regular intervals. No obvious differences can be noted between the different apertures within this series. A regularly structured basal lamina (arrowheads) covers the pericapillary surface of the endothelial cells and continues over the fenestrae and the vesicle stomata. Note the well-developed Golgi apparatus (G) on the luminal side of the nucleus (N) immediately beneath the vacuole in (c). (Bar ¼ 0.5 mm in (a, b); 1 mm in (c).)

ARTICLE IN PRESS 340

A.B. El-Fadaly, W. Kummer

Figure 9. Multi-layered fenestrated lamellae. The perinuclear cytoplasm is divided into lamellar fenestrated sheets (arrows) with large clear spaces in between. Numerous intercommunicating vesicles and vacuoles of variable dimensions are seen on the luminal side of the nucleus in the vicinity of these fenestrated lamellae. Owing to the obliquity of the section in certain regions of the fenestrated lamellae, some of their fenestrae (arrowheads) appear rather circular or vesicular with dimensions approximating those of the smallest vesicles in the surrounding cytoplasm. L ¼ lumen (Bar ¼ 1 mm).

Figure 8. In this capillary, the region indicated by the rectangle shows a peripheral attenuated segment of the endothelial cell divided into two lamellae. Higher magnification of this segment shows that these lamellae have several fenestrae. L ¼ lumen (Bar ¼ 1 mm).

logically resemble the VVOs previously identified in the hyperpermeable endothelia of normal venules, tumour associated microvasculature and venules associated with allergic inflammation (Kohn et al., 1992; Dvorak et al., 1994; Feng et al., 1996). It is suggested that these channels represent the VVOs in the fenestrated capillaries of the carotid body. The morphological data obtained from the present study along with those reported by these authors indicated that these organelles are formed, at least in part, by fusion of plasmalemmal vesicles. Their individual units are very similar to the non-fused vesicles being spherical and outlined by a smooth triple-layered membrane. This outlining membrane might establish continuity with the plasmalemmal

surface membrane at the channel stomata. The dimensions of the individual units of the channels were variable and usually had a larger diameter (4100 nm) than the more uniform-sized, non-fused vesicles (approximate diameter of 50–70 nm). The large size of the units of the VVOs may be the result of fusion of smaller vesicles (Feng et al., 1999). This assumption was supported by the present morphological observation that the VVOs were usually located within focal collections of vesicles with the usual diameter of 62714 nm. These vesicles often contacted or communicated with the peripheral units of the VVOs. These morphological data are corroborated by the immunohistochemical study of Vasile et al. (1999) that reveals the presence of caveolin-1 in both caveolae and VVOs. The orientation of the VVOs varied in the different regions of the endothelial cell. In the peripheral attenuated region, they were commonly oriented perpendicular to the cell surface and might communicate with the capillary lumen or the pericapillary space by a single stoma. In the perinuclear region, they were mostly oriented parallel to the cell surface and might open into the cell surface by multiple stomata. In certain

ARTICLE IN PRESS Endothelial vesiculo-vacuolar organelles in the capillaries of the mouse carotid body areas, cytoplasmic channels or VVOs and focal collections of plasmalemmal vesicles were seen in close spatial relationships to microtubules or

341

bundles of cytoplasmic filaments and appeared to closely follow their course. This arrangement suggests that the cytoskeletal elements act as a guidance mechanism or tracks that direct the traffic and fusion of the vesicles. At the molecular level, an interaction is assumed between caveolae and filamin, an actin binding protein, thereby attaching the caveolin-1 coated vesicles to the cortical actin cytoskeleton (Stahlhut and van Deurs, 2000). The present study also demonstrates membranebound vacuoles that appear to be generated by similar fusion events as the VVOs. Although these vacuoles show extreme variation in dimensions (ranging from 0.3 to 2 mm), they share common morphological criteria in that they are situated immediately beneath the cell surface and communicate with the extracellular aspect via multiple apertures or fenestrae. These fenestrae always occur in chains within attenuated cytoplasmic sheets. Morphological and morphometrical data reveal that the individual apertures of such attenuated cytoplasmic sheets are closely similar to the stomata of the fully opened plasmalemmal vesicles. It appears highly probable that the observed fenestrated vacuoles are created by fusion and sequential communication of vesicles aggregated in chains or clusters parallel to the cell surface. Accordingly, their individual apertures would merely represent the stomata of these vesicles at the cell surface. An additional morphological criterion in support of this concept is the frequent location of these vacuoles within collections of vesicles that might contact their surfaces or communicate with their lumina. Furthermore, in favourable sections, the chain of apertures or fenestrae belonging to the vacuoles together with stomata of vesicles and smaller vacuoles in their immediate vicinity was seen in one plane of section. All of them appeared distributed in a linear series at regular intervals, centre to centre. Morphological and morphometrical evaluation indicated that this series is identical to a chain of fenestrae in a usual fenestrated segment of the same capillary with no clear-cut differentiation

Figure 10. Finger-like projections (arrows) of the luminal front of the perinuclear regions (a, b) and the parajunctional zone (c). In the perinuclear region, the cytoplasm underlying these projections shows intercommunicating vesicles and vacuoles of extreme heterogeneous morphology. One of these vacuoles () is fenestrated. An excessive amount of Golgi areas (G) are also demonstrated in the perinuclear region in (a). (Bar in (a–c) ¼ 1 mm.)

ARTICLE IN PRESS 342 between them. These observations provide more supportive evidence for the concept that, in the fenestrated endothelia, fenestrae and vesicle stomata are essentially identical structures. The fenestrated vacuoles demonstrated here show specific locations within the endothelial cell. They were mostly found in the thick perinuclear cytoplasm. Profiles showing a large single vacuole with multiple apertures are more common on the abluminal aspect of the nucleus. On the other hand, a frequent occurrence of multiple smaller vacuoles with individual apertures was noticed on the luminal side. It appears that, on the luminal side, smaller vacuoles open early and their walls become integrated into the surface membrane before larger ones are formed. In support of this interpretation, a quantitative estimation of the number of individual vesicle stomata opening on the luminal versus abluminal front revealed conspicuous asymmetry favouring the abluminal front. Also, irregularities or finger-like processes or projections of the cell membrane are frequently associated with such smaller vacuoles. Such processes showed a striking preferential location on the luminal front of the perinuclear area in most of the cases studied. It should be stressed that, in many of these cases, no deep infoldings of the nuclear membrane were seen, excluding the possibility that the cell membrane projections were the result of cellular contraction. It seems more probable that these irregularities or fingerlike processes represent fully opened vacuoles that lost their stomata and became integrated into the surface membrane. Endothelial pockets first described in murine renal peritubular capillaries (Milici et al., 1986) were found to be a characteristic structural feature in the present study. In many cases, the endothelial cytoplasm was divided into two fenestrated sheets enclosing a clear space in between. This arrangement appears to be a multifold version of the fenestrated vacuoles described above. In addition to these pockets, the present study also demonstrated a complete division of the cytoplasm in the perinuclear region into MFL. Here, the entire thickness of the perinuclear cytoplasm was transformed into multilamellar fenestrated sheets; a finding which has not been referred to in any other vascular bed. It is suggested that endothelial pockets and MFL result from progressive steps of the formation of the VVOs as evidenced by their same predominant location and a similar relationship to the vesicles in the surrounding cytoplasm. It is increasingly becoming clear that a dynamic transport exists between trans-Golgi cisternae and plasmalemmal vesicles with possible interaction

A.B. El-Fadaly, W. Kummer between the endocytotic and secretory systems (Kurzchalia et al., 1994; Luetterforst et al., 1999; Schlegel and Lisanti, 2001; Vetterlein et al., 2002). It should be expected that such interaction is more prominent in the perinuclear region where the Golgi apparatus and most of the cell organelles are concentrated. Consequently, it could be suggested that the structures described here, being more frequently observed in the perinuclear region, might represent structural manifestations of this interaction. The question of whether the transendothelial channels and VVOs are static or dynamic structures has been debated for many years. The extreme heterogeneity of such structures shown here as well as in previous reports favours the dynamic concept. Furthermore, as already discussed, there are accumulating data suggesting their formation, at least in part, by fusion of caveolae which already have been proved to be dynamic structures (Rothberg et al., 1992; Parton et al., 1994; Chen et al., 2002). However, the exact machinery controlling this dynamic process has so far not been identified. The studies of Kohn et al. (1992) and Feng et al. (1996, 1999, 2000) indicated that certain permeabilizing mediators such as vascular permeability factor/vascular endothelial growth factor (VPF/ VEGF), histamine and serotonin bind to the VVOs and enhance transcellular leakage via these organelles. In this regard, it should be emphasized that a large number of mast cells have been found in the mouse carotid body (Bo ¨ck, 1973) and that serotonin has been detected in glomus cells of rabbits (Mo ¨llmann et al., 1972a, b), cats (Chiocchio et al., 1971) and humans (Hamberger et al., 1966). Recently, large amounts of histamine were also demonstrated in the rat carotid body (Koerner et al., 2004). Moreover, VPF/VEGF was reported to be induced in the carotid body by hypoxia and has been suggested to contribute to the increased vascularity in this organ in hypoxia (Schweiki et al., 1992; Chen et al., 1997) and in carotid body tumours (Jyung et al., 2000). The role of VPF/ VEGF in the formation of endothelial fenestrations is still controversial. Some investigators have reported that it increases fenestration in microvascular endothelia (Roberts and Palade, 1995, 1997; Esser et al., 1998; Yokomori et al., 2003), while others have shown that it enhances vascular permeability but does not increase endothelial fenestration (Kohn et al., 1992; Feng et al., 2000). In conclusion, the capillary endothelia of the mouse carotid body display multifold structural manifestations of vascular permeability, extending from simple fenestrations over fenestrated vacuoles, pockets and VVOs to fenestrated multi-

ARTICLE IN PRESS Endothelial vesiculo-vacuolar organelles in the capillaries of the mouse carotid body layered sheets or lamellae (MFL) that have not been reported previously in any other vascular bed. These latter structures appear to be formed by the same mechanisms that cause the genesis of the VVOs which still have to be elucidated in detail.

Acknowledgements The authors thank Ms. K. Michael, Mr. G. Kripp and Mr. G. Magdowski for their skilful technical assistance.

References Anderson, R.G.W., 1993. Potocytosis of small molecules and ions by caveolae. Trends Cell Biol. 3, 69–72. Bendayan, M., Rasio, E.A., 1996. Transport of insulin and albumin by the microvascular endothelium of the rete mirabile. J. Cell Sci. 109, 1857–1864. Bo ¨ck, P., 1973. Das Glomus caroticum der Maus. Adv. Anat. Embryol. Cell Biol. 48, 7–84. Bundgaard, M., 1983. Vesicular transport in capillary endothelium: does it occur? Fed. Proc. 42, 2425–2430. Bundgaard, M., Hagman, P., Crone, C., 1983. Threedimensional organisation of plasmalemmal vesicular profiles in the endothelium of rat heart capillaries. Microvasc. Res. 25, 358–368. Chen, J., Dinger, B., Stensaas, L., Fidone, S., 1997. Upregulation of vascular endothelial growth factor (VEGF) and VEGF receptor (Flk-1) in chronically hypoxic rat carotid body. Abstr. Soc. Neurosci. 23, 433. Chen, J., Braet, F., Brodskz, S., Weinstein, MS., 2002. VEGF induced mobilization of caveolae and increase in permeability of endothelial cells. Am. J. Physiol. Cell Physiol. 282, 1053–1063. Chiocchio, S.R., King, M.P., Carballo, L., Angelakos, E.T., 1971. Monoamines in the carotid body of the cat. J. Histochem. Cytochem. 19, 621–626. Clementi, F., Palade, G.E., 1969. Intestinal capillaries. I. Permeability to peroxidase and ferritin. J. Cell Biol. 41, 33–58. Dvorak, A.M., Feng, D., 2001. The vesiculo-vacuolar organelle (VVO): a new endothelial cell permeability organelle. J. Histochem. Cytochem. 49, 419–431. Dvorak, A.M., Tepper, R.I., Weller, P.F., Morgan, E.S., Estrella, P., Monaham-Early, R.A., Galli, S.J., 1994. Piecemeal degranulation of mast cells in the inflammatory eyelid lesions of interleukin-4 transgenic mice. Evidence of mast cell histamine release in vivo by diamine oxidase-gold enzyme-affinity ultrastructural cytochemistry. Blood 83, 3600–3612. Dvorak, A.M., Kohn, S., Morgan, E.S., Fox, P., Nagy, J.A., Dvorak, H.F., 1996. The vesiculo-vacuolar organelle (VVO): a distinct endothelial cell structure that provides a transcellular pathway for macromolecular extravasation. J. Leukocyte Biol. 59, 100–115.

343

Esser, S., Wolburg, K., Wolburg, H., Breier, G., Kurzchalia, T., Risau, W., 1998. Vascular endothelial growth factor induces endothelial fenestrations in vitro. J. Cell Biol. 140, 947–959. Feng, D., Nagy, J.A., Hipp, J., Dvorak, H.F., Dvorak, A.M., 1996. Vesiculo-vacuolar organelles and the regulation of venule permeability to macromolecules by vascular permeability factor, histamine and serotonin. J. Exp. Med. 183, 1981–1986. Feng, D., Nagy, J.A., Pyne, K., Hammel, I., Dvorak, H.F., Dvorak, A.M., 1999. Pathways of macromolecular extravasation across microvascular endothelium in response to VPF/VEGF and other vasoactive mediators. Microcirculation 6, 23–44. Feng, D., Nagy, J.A., Dvorak, A.M., Dvorak, H.F., 2000. Different pathways of macromolecule extravasation from hyperpermeable tumour vessels. Microvasc. Res. 59, 24–37. Forssmann, W.G., Ito, S., Weihe, E., Aoki, A., Dym, M., Fawcett, D.W., 1977. An improved perfusion fixation method for the testis. Anat. Rec. 188, 307–314. Ghitescu, L., Bendayan, M., 1992. Transendothelial transport of serum albumin: a quantitative immunocytochemical study. J. Cell Biol. 117, 745–755. Hamberger, B., Ritzen, M., Wersa ¨ll, J., 1966. Demonstration of catecholamines and 5-hydroxytryptamine in the human carotid body. J. Pharmacol. Exp. Ther. 152, 197–201. Hashimoto, P., 1972. Intracellular channels as a route for protein passage in the capillary endothelium of the shark brain. Am. J. Anat. 134, 41–57. Jyung, R.W., LeClair, E.E., Bernat, R.A., Kang, T.S., Ung, F., McKenna, M.J., Tuan, R.S., 2000. Expression of angiogenic growth factors in paragangliomas. Laryngoscope 110, 161–167. Koerner, P., Hesslinger, C., Schaefermeyer, A., Prinz, C., Gratzl, M., 2004. Evidence for histamine as a transmitter in rat carotid body sensor cells. J. Neurochem. 91, 493–500. Kohn, S., Nagy, J.A., Dvorak, H.F., Dvorak, A.M., 1992. Pathways of macromolecular tracer transport across venules and small veins: structural basis for the hyperpermeability of tumour blood vessels. Lab. Invest. 67, 596–607. Kurzchalia, T.V., Dupree, P., Monir, S., 1994. VIP21caveolin, a protein of the trans-Golgi network and caveolae. FEBS Lett. 346, 88–91. Luetterforst, R., Stang, E., Zorzi, N., Carozzi, A., Way, M., Parton, R.G., 1999. Molecular characterisation of a cis-Golgi targeting domain in the caveolin molecule. J. Cell Biol. 145, 1443–1459. Milici, A.J., Peters, K.R., Palade, G.E., 1986. The endothelial pockets. A new structure in fenestrated endothelia. Cell Tissue Res. 244, 493–499. Mo ¨llmann, H., Knoche, H., Niemeyer, D.H., Alfes, H., Kienecker, E.W., Decker, S., 1972a. Experimenteller Beitrag zur Kenntnis der biogenen Amine im Glomus caroticum des Kaninchens. Elektronen- und fluoreszenzmikroskopische Untersuchungen nach Reserpinund PCPA-Applikation. Z. Zellforsch. 124, 238–246.

ARTICLE IN PRESS 344 Mo ¨llmann, H., Niemeyer, D.H., Alfes, H., Knoche, H., 1972b. Mikrospektrofluorometrische Untersuchungen der biogenen Amine im Glomus caroticum des Kaninchens nach Resserpin- und PCPA-Applikation. Z. Zellforsch. 126, 104–115. Palade, G.E., 1960. Transport in quanta across the endothelium of blood capillaries. Anat. Rec. 136, 254 (Abstr.). Parton, R.G., Joggerst, B., Simons, K., 1994. Regulated internalization of caveolae. J. Cell Biol. 127, 1199–1215. Predescu, D.N., Palade, G.E., 1993. Plasmalemmal vesicles represent the large pore system of continuous microvascular endothelium. Am. J. Physiol. 265, 725–733. Predescu, S.A., Predescu, D.N., Palade, G.E., 1997. Plasmalemmal vesicles function as transcytotic carriers for small protein in the continuous endothelium. Am. J. Physiol. 272, 937–949. Rasio, E.A., Bendayan, M., 2002. Sequential morphological and permeability changes in the rete capillaries during hyperglycaemia. Microsc. Res. Tech. 57, 327–349. Roberts, W.G., Palade, G.E., 1995. Increased microvascular permeability and endothelial fenestration induced by vascular endothelial growth factor. J. Cell Sci. 108, 2369–2379. Roberts, W.G., Palade, G.E., 1997. Neovasculature induced by vascular endothelial growth factor is fenestrated. Cancer Res. 57, 764–772. Rothberg, K.G., Heuser, J.E., Donzell, W.C., Ying, Y.-S., Glenney, J.R., Anderson, R.G.W., 1992. Caveolin, a component of caveolae membrane coats. Cell 68, 673–682. Schlegel, A., Lisanti, M.P., 2001. Caveolae and their coat proteins, the caveolins: from electron microscopic novelity to biological launching pad. J. Cell Physiol. 186, 329–337. Schnitzer, J.E., Jenny, A., Phil, O.H., 1995. NEM inhibits transcytosis, endocytosis and capillary permeability: implication of caveolae fusion in endothelia. Am. J. Physiol. 268, 48–55. Schweiki, D., Itin, A., Soffer, D., Keshet, E., 1992. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359, 843–845. Simionescu, N., Simionescu, M., Palade, G.E., 1973. Permeability of muscle capillaries to exogenous myoglobin. J. Cell Biol. 57, 424–452.

A.B. El-Fadaly, W. Kummer Simionescu, N., Simionescu, M., Palade, G.E., 1975. Permeability of muscle capillaries to small hemepeptides. Evidence for the existence of patent transendothelial channels. J. Cell Biol. 64, 586–607. So ¨rensson, J., Fierlbeck, W., Heider, T., Schwarz, K., Park, D., Lisanti, M., Bellermann, B.J., 2002. Glomerular endothelial fenestrae in vivo. J. Am. Soc. Nephrol. 13, 2639–2647. Stahlhut, M., van Deurs, B., 2000. Identification of filamin as a novel ligand for caveolin-1: evidence for the organization of caveolin-1 associated-membrane domains by the actin cytoskeleton. Mol. Biol. Cell 11, 325–337. Stan, R.-V., 2002. Structure and function of endothelial caveolae. Microsc. Res. Tech. 57, 350–364. Stan, R.-V., 2004. Multiple PV1 dimers reside in the same stomatal or fenestral diaphragm. Am. J. Physiol. Heart Circ. Physiol. 286, 1347–1353. Stan, R.-V., Roberts, W.G., Predescu, D., Ihida, K., Saucan, L., Ghitescu, L., Palade, G.E., 1997. Immunoisolation and partial characterization of endothelial plasmalemmal vesicles (caveolae). Mol. Biol. Cell 8, 595–605. Stan, R.-V., Kubitza, M., Palade, G.E., 1999a. PV-1 is a component of stomatal and fenestral diaphragms in fenestrated endothelia. Proc. Natl. Acad. Sci. USA 96, 13203–13207. Stan, R.-V., Ghitescu, L., Jacobson, B.S., Palade, G.E., 1999b. Isolation, cloning and localization of rat PV-1, a novel endothelial caveolar protein. J. Cell Biol. 145, 1189–1198. van Deurs, B., Holm, P.K., Sandvig, K., Hansen, S.H., 1993. Are caveolae involved in endocytosis? Trends Cell Biol. 3, 249–251. Vasile, E., Qu-Hong, H.F., Dvorak, A.M., 1999. Caveolae and vesiculo-vacuolar organelles in bovine capillary endothelial cells cultured with VPF/VEGF on floating matrigel-collagen gels. J. Histochem. Cytochem. 47, 159–167. Vetterlein, M., Ellinger, A., Neumu ¨ller, J., 2002. Golgi apparatus and TGN during endocytosis. Histochem. Cell Biol. 117, 143–150. Yokomori, H., Oda, M., Yoshimura, K., Nagai, T., Ogi, M., Nomura, M., Ishii, H., 2003. Vascular endothelial growth factor increases fenestral permeability in hepatic sinusoidal endothelial cells. Liver Int. 23, 467–475.