Experimental Physiology (1992), 77, 917-920 Printed in Great Britain
TRANSCELLULAR OPENINGS THROUGH MICROVASCULAR WALLS IN ACUTELY INFLAMED FROG MESENTERY C. R. NEAL AND C. C. MICHEL* Department of Physiology & Biophysics, St Mary's Hospital Medical School, Imperial College of Science, Technology & Medicine, Norfolk Place, London W2 JPG (MANUSCRIPT RECEIVED 30 JULY 1992, ACCEPTED 27 AUGUST 1992)
SUMMARY
Openings in the endothelium of frog mesenteric microvessels associated with acute inflammation following mild thermal injury have been examined by reconstruction from serial ultrathin sections. While all the openings lay close to the intercellular junctions only a minority (seven out of thirtyeight) were continuous with the intercellular clefts. The majority of gaps or openings (thirty-one out of thirty-eight) were transcellular and passed through the peripheral cytoplasm of one endothelial cell, separated from the intercellular cleft by a cellular process which was usually less than 2 lim wide. INTRODUCTION
The development of openings or gaps in microvascular endothelium of capillaries and postcapillary venules is widely regarded as the ultrastructural basis of the increased microvascular permeability of acute inflammation (Majno & Palade, 1961). It is believed that the gaps arise from intercellular clefts as a result of the contraction (or shape change) of adjacent endothelial cells (Majno, Shea & Leventhal, 1969). Electron micrographs prepared from perfused microvessels in the acutely inflamed frog mesentery led us to question whether all the gaps were intercellular in this preparation. To define more clearly the relations between the endothelial gaps and the intercellular clefts, we have examined the gaps by reconstructing them from serial ultrathin sections. From this analysis, we now present evidence that the majority of gaps in the endothelium of acutely inflamed frog microvessels pass through the peripheral cytoplasm of a single cell and are distinct from nearby intercellular clefts. METHODS
All experiments were carried out on the exposed mesenteries of frogs (Rana temporaria) whose brains and upper spinal cords had been destroyed by pithing after stunning. In each experiment a single capillary or venule was cannulated with a sharpened micropipette and perfused with frog Ringer solution containing a few human red cells (to act as flow markers) and, in some experiments, bovine serum albumin (50 mg mn- 1 ) or Ficoll 70 (Sigma, UK, 40 mg ml 1 ) or a mixture of bovine serum albumin (10 mg ml- I ) and Ficoll 70 (40 mg ml- l). The hydraulic permeability, Lp, of the perfused vessel and the effective osmotic pressure exerted across its walls were estimated by the method of Michel (1980), first under control conditions when the tissue was being washed by a superfusing Ringer solution at room temperature and then at 1 min intervals after the temperature of the superfusate had been abruptly raised to 35-40 OC. An abrupt rise in tissue temperature to more than 30 OC induces a mild inflammatory response in the tissues *
To whom reprint requests should be sent.
918
C. R. NEAL AND C.C.MICHEL
(Clough, Michel & Phillips, 1988). As soon as an increase in fluid filtration became obvious to the experimenter (usually 5 min after temperature was raised), the flow of superfusate was stopped and the tissues were fixed in situ with ice-cold, 2-5 % glutaraldehyde in 0-1 M cacodylate buffer (pH 7-2) and prepared for electron microscopy. The perfused microvessel was identified in the block and transverse ultrathin sections were cut using a Reichert-Jung Ultracut E. Dimensions of endothelial openings were determined in the transverse plane directly from the electron micrographs (magnification calibrated by using a cross-grating replica of 2160 lines per mm) and, in the direction of the vessel axis, from the thickness and number of serial sections in the reconstruction. Measurements are given as means± 1 S.D. RESULTS
Initial observations from electron micrographs of transverse sections cut at random intervals along the perfused microvessels in sixteen frogs revealed numerous openings in the endothelium. One edge of nearly all these openings lay within 2 gm of a short but intact intercellular cleft (Fig. 1). Because reconstructions based on ultrathin sections made at 1 pm intervals did not support our initial interpretation of these images as representing openings at points where three or more cells met, sequences of serial sections were then made. The serial sections were cut deliberately thick (100 nm) so that an area of microvascular wall completely surrounding an opening could be defined within one sequence. From four microvessels, each prepared from a different animal, thirty-eight openings or gaps were examined and their margins completely described by reconstruction from the serial sections. Seven of these openings were found to be continuous with intercellular clefts. Reconstruction of the remaining thirty-one gaps, however, revealed that although they were all close to intercellular clefts, they were completely surrounded by the cytoplasm of a single endothelial cell (Fig. 2). We refer to these openings as transcellular gaps. Because many intercellular junctions are formed with one endothelial cell overlying its neighbour, some transcellular gaps opened on to the lower part of an intercellular cleft, their exits being partially or completely separated from the basement membrane by the underlying process of the adjoining cell. The intercellular gaps were elongated in the direction of the cleft having a mean length in that direction of 0.32±0.10 pm and a maximum width of 0*12±0*07 pm. The transcellular gaps were larger and more circular with a mean length (parallel to the vessel axis) of 0 79±0.47 pm and a maximum width of 0-73±0.42 pm.
Fig. 1. Electron micrograph of a transverse section of inflamed capillary in the region of the intercellular cleft V, intercellular cleft; E, endothelial cell; G, gap; L, lumen; U, underlying cell process. Scale bar: 500 nm.
MICROVASCULAR PERMEABILITY IN INFLAMMATION
919
LC
_
E -/
E
U
~~G
0
o 00
o
o
F
0
r_
Fig. 2. Diagram of a luminal surface view reconstructed from twenty-four serial sections of inflamed capillary. E, endothelial cell; F, fenestration; G, gap; LC, luminal entrance to the intercellular cleft; U, underlying cell process in the gap. Scale bar: 500 nm.
DISCUSSION
Our finding, that over 80 % of the openings in the endothelium of inflamed frog microvessels are transcellular, is contrary to our initial expectation. It is worth noting, however, that much of the evidence for an intercellular origin of gaps in endothelia is based on relatively low-resolution studies involving light microscopy or scanning electron microscopy. Two exceptions are the studies of Fox, Galey & Wayland (1980) and Braverman & Keh-Yen (1986). These groups used serial ultrathin sections and computer-assisted reconstruction techniques. Fox et al. (1980) examined openings induced by histamine in the venules of the cat mesentery. Although the gaps did communicate with intercellular clefts, the relations were not simple and it was not clear how all of them could have arisen from the opening of the clefts. Braverman & Keh-Yen (1986) examined gaps in human skin microvessels biopsied at the site of lesions in patients with psoriasis and at the site of application of histamine to the skin of normal individuals. Whereas the gaps associated with histamine application were intercellular, both intercellular and transcellular gaps were observed in the vicinity of the psoriatic lesions. A high frequency of transcellular openings of microvascular endothelium has not previously been described in acute inflammation. While we are conscious of the dangers of extrapolating observations on frog to other species and of the possibility that endothelial responses to heating may differ from those to mediators such as histamine, we are impressed by the similar ultrastructural appearances of the endothelial gaps in preparations of frog microvessels and published micrographs of inflamed mammalian microvessels. Just how transcellular openings develop is at present obscure. They could result from retraction of adjacent cells if the peripheral cytoplasm was mechanically weaker than the cell junctions. Alternatively they could arise from rapidly enlarging vesicles or vacuoles in the cell periphery: the presence of enlarged vesicles or vacuoles
920
C. R. NEAL AND C.C.MICHEL
has been described in the periphery of endothelial cells under conditions when permeability has increased (Clough & Michel, 1988). We thank the Wellcome Trust for supporting this project and Dr Geraldine Clough for helpful discussion during the early stages of the work. REFERENCES BRAVERMAN, I. M. & KEH-YEN, A. (1986). Three-dimensional reconstruction of endothelial cell gaps in psoriatic vessels and their morphologic identity with gaps produced by the intradermal injection of histamine. Journal of Investigative Dermatology 86, 577-58 1. CLOUGH, G. & MICHEL, C. C. (1988). The ultrastructure of frog microvessels following perfusion with the ionophore A23187. Quarterly Journal of Experimental Physiology 73, 123-125. CLOUGH, G., MICHEL, C. C. & PHILLIPS, M. E. (1988). Inflammatory changes in permeability and ultrastructure of single vessels in the frog mesenteric microcirculation. Journal of Physiology 395, 99-114. Fox, J., GALEY, F. & WAYLAND, H. (1980). Action of histamine on the mesenteric microvasculature. Microvascular Research 19, 108-126. MICHEL, C. C. (1980). Filtration coefficients and osmotic reflexion coefficients of the walls of single frog mesenteric capillaries. Journal of Physiology 309, 341-355. MAJNO, G. & PALADE, G. E. (1961). Studies on inflammation. I. The effect of histamine and serotonin on vascular permeability: an electron microscopic study. Journal of Biophysical and Biochemical Cytology 11, 571-605. MAJNO, G., SHEA, S. M. & LEVENTHAL, M. (1969). Endothelial contraction induced by histamine-type mediators. An electron microscopic study. Journal of Cell Biology 42, 647-672.
TRANSCELLULAR OPENINGS THROUGH MICROVASCULAR WALLS IN ACUTELY INFLAMED FROG MESENTERY C. R. NEAL AND C. C. MICHEL* Department of Physiology & Biophysics, St Mary's Hospital Medical School, Imperial College of Science, Technology & Medicine, Norfolk Place, London W2 JPG (MANUSCRIPT RECEIVED 30 JULY 1992, ACCEPTED 27 AUGUST 1992)
SUMMARY
Openings in the endothelium of frog mesenteric microvessels associated with acute inflammation following mild thermal injury have been examined by reconstruction from serial ultrathin sections. While all the openings lay close to the intercellular junctions only a minority (seven out of thirtyeight) were continuous with the intercellular clefts. The majority of gaps or openings (thirty-one out of thirty-eight) were transcellular and passed through the peripheral cytoplasm of one endothelial cell, separated from the intercellular cleft by a cellular process which was usually less than 2 lim wide. INTRODUCTION
The development of openings or gaps in microvascular endothelium of capillaries and postcapillary venules is widely regarded as the ultrastructural basis of the increased microvascular permeability of acute inflammation (Majno & Palade, 1961). It is believed that the gaps arise from intercellular clefts as a result of the contraction (or shape change) of adjacent endothelial cells (Majno, Shea & Leventhal, 1969). Electron micrographs prepared from perfused microvessels in the acutely inflamed frog mesentery led us to question whether all the gaps were intercellular in this preparation. To define more clearly the relations between the endothelial gaps and the intercellular clefts, we have examined the gaps by reconstructing them from serial ultrathin sections. From this analysis, we now present evidence that the majority of gaps in the endothelium of acutely inflamed frog microvessels pass through the peripheral cytoplasm of a single cell and are distinct from nearby intercellular clefts. METHODS
All experiments were carried out on the exposed mesenteries of frogs (Rana temporaria) whose brains and upper spinal cords had been destroyed by pithing after stunning. In each experiment a single capillary or venule was cannulated with a sharpened micropipette and perfused with frog Ringer solution containing a few human red cells (to act as flow markers) and, in some experiments, bovine serum albumin (50 mg mn- 1 ) or Ficoll 70 (Sigma, UK, 40 mg ml 1 ) or a mixture of bovine serum albumin (10 mg ml- I ) and Ficoll 70 (40 mg ml- l). The hydraulic permeability, Lp, of the perfused vessel and the effective osmotic pressure exerted across its walls were estimated by the method of Michel (1980), first under control conditions when the tissue was being washed by a superfusing Ringer solution at room temperature and then at 1 min intervals after the temperature of the superfusate had been abruptly raised to 35-40 OC. An abrupt rise in tissue temperature to more than 30 OC induces a mild inflammatory response in the tissues *
To whom reprint requests should be sent.
918
C. R. NEAL AND C.C.MICHEL
(Clough, Michel & Phillips, 1988). As soon as an increase in fluid filtration became obvious to the experimenter (usually 5 min after temperature was raised), the flow of superfusate was stopped and the tissues were fixed in situ with ice-cold, 2-5 % glutaraldehyde in 0-1 M cacodylate buffer (pH 7-2) and prepared for electron microscopy. The perfused microvessel was identified in the block and transverse ultrathin sections were cut using a Reichert-Jung Ultracut E. Dimensions of endothelial openings were determined in the transverse plane directly from the electron micrographs (magnification calibrated by using a cross-grating replica of 2160 lines per mm) and, in the direction of the vessel axis, from the thickness and number of serial sections in the reconstruction. Measurements are given as means± 1 S.D. RESULTS
Initial observations from electron micrographs of transverse sections cut at random intervals along the perfused microvessels in sixteen frogs revealed numerous openings in the endothelium. One edge of nearly all these openings lay within 2 gm of a short but intact intercellular cleft (Fig. 1). Because reconstructions based on ultrathin sections made at 1 pm intervals did not support our initial interpretation of these images as representing openings at points where three or more cells met, sequences of serial sections were then made. The serial sections were cut deliberately thick (100 nm) so that an area of microvascular wall completely surrounding an opening could be defined within one sequence. From four microvessels, each prepared from a different animal, thirty-eight openings or gaps were examined and their margins completely described by reconstruction from the serial sections. Seven of these openings were found to be continuous with intercellular clefts. Reconstruction of the remaining thirty-one gaps, however, revealed that although they were all close to intercellular clefts, they were completely surrounded by the cytoplasm of a single endothelial cell (Fig. 2). We refer to these openings as transcellular gaps. Because many intercellular junctions are formed with one endothelial cell overlying its neighbour, some transcellular gaps opened on to the lower part of an intercellular cleft, their exits being partially or completely separated from the basement membrane by the underlying process of the adjoining cell. The intercellular gaps were elongated in the direction of the cleft having a mean length in that direction of 0.32±0.10 pm and a maximum width of 0*12±0*07 pm. The transcellular gaps were larger and more circular with a mean length (parallel to the vessel axis) of 0 79±0.47 pm and a maximum width of 0-73±0.42 pm.
Fig. 1. Electron micrograph of a transverse section of inflamed capillary in the region of the intercellular cleft V, intercellular cleft; E, endothelial cell; G, gap; L, lumen; U, underlying cell process. Scale bar: 500 nm.
MICROVASCULAR PERMEABILITY IN INFLAMMATION
919
LC
_
E -/
E
U
~~G
0
o 00
o
o
F
0
r_
Fig. 2. Diagram of a luminal surface view reconstructed from twenty-four serial sections of inflamed capillary. E, endothelial cell; F, fenestration; G, gap; LC, luminal entrance to the intercellular cleft; U, underlying cell process in the gap. Scale bar: 500 nm.
DISCUSSION
Our finding, that over 80 % of the openings in the endothelium of inflamed frog microvessels are transcellular, is contrary to our initial expectation. It is worth noting, however, that much of the evidence for an intercellular origin of gaps in endothelia is based on relatively low-resolution studies involving light microscopy or scanning electron microscopy. Two exceptions are the studies of Fox, Galey & Wayland (1980) and Braverman & Keh-Yen (1986). These groups used serial ultrathin sections and computer-assisted reconstruction techniques. Fox et al. (1980) examined openings induced by histamine in the venules of the cat mesentery. Although the gaps did communicate with intercellular clefts, the relations were not simple and it was not clear how all of them could have arisen from the opening of the clefts. Braverman & Keh-Yen (1986) examined gaps in human skin microvessels biopsied at the site of lesions in patients with psoriasis and at the site of application of histamine to the skin of normal individuals. Whereas the gaps associated with histamine application were intercellular, both intercellular and transcellular gaps were observed in the vicinity of the psoriatic lesions. A high frequency of transcellular openings of microvascular endothelium has not previously been described in acute inflammation. While we are conscious of the dangers of extrapolating observations on frog to other species and of the possibility that endothelial responses to heating may differ from those to mediators such as histamine, we are impressed by the similar ultrastructural appearances of the endothelial gaps in preparations of frog microvessels and published micrographs of inflamed mammalian microvessels. Just how transcellular openings develop is at present obscure. They could result from retraction of adjacent cells if the peripheral cytoplasm was mechanically weaker than the cell junctions. Alternatively they could arise from rapidly enlarging vesicles or vacuoles in the cell periphery: the presence of enlarged vesicles or vacuoles
920
C. R. NEAL AND C.C.MICHEL
has been described in the periphery of endothelial cells under conditions when permeability has increased (Clough & Michel, 1988). We thank the Wellcome Trust for supporting this project and Dr Geraldine Clough for helpful discussion during the early stages of the work. REFERENCES BRAVERMAN, I. M. & KEH-YEN, A. (1986). Three-dimensional reconstruction of endothelial cell gaps in psoriatic vessels and their morphologic identity with gaps produced by the intradermal injection of histamine. Journal of Investigative Dermatology 86, 577-58 1. CLOUGH, G. & MICHEL, C. C. (1988). The ultrastructure of frog microvessels following perfusion with the ionophore A23187. Quarterly Journal of Experimental Physiology 73, 123-125. CLOUGH, G., MICHEL, C. C. & PHILLIPS, M. E. (1988). Inflammatory changes in permeability and ultrastructure of single vessels in the frog mesenteric microcirculation. Journal of Physiology 395, 99-114. Fox, J., GALEY, F. & WAYLAND, H. (1980). Action of histamine on the mesenteric microvasculature. Microvascular Research 19, 108-126. MICHEL, C. C. (1980). Filtration coefficients and osmotic reflexion coefficients of the walls of single frog mesenteric capillaries. Journal of Physiology 309, 341-355. MAJNO, G. & PALADE, G. E. (1961). Studies on inflammation. I. The effect of histamine and serotonin on vascular permeability: an electron microscopic study. Journal of Biophysical and Biochemical Cytology 11, 571-605. MAJNO, G., SHEA, S. M. & LEVENTHAL, M. (1969). Endothelial contraction induced by histamine-type mediators. An electron microscopic study. Journal of Cell Biology 42, 647-672.
