S. COTRAN Medical May
Phagocytosis by vascular endothelium outside the reticuloendothelial system (RES) has been the subject of considerableinvestigation. The light-microscopic studies, based on the localization of colloidal particles or dyes, have been reviewed (Altschul, 1954; Benacerraf et al., 1959; Zweifach, 1961; Majno, 1964). Such studies indicated that, under normal conditions, non-RES vascular endothelium was not phagocytic; however, considerable stimulation or “activation” was said to occur after a variety of stimuli, such as injections of histamine, endotoxin, or antigen-antibody complexes, inflammation, and overloading with colloidal particles (Jansco, 1955; Gozsy and Kato, 1957; Benacerraf et al., 1959; Zweifach, 1961). Recently, electron-microscopic observations have cast doubt on the occurrence of phagocytosis in non-RES endothelium, even under the special circumstancesjust listed. It has been shown, for example, that many of the light-microscopic images allegedly representing endothelial phagocytosis actually corresponded to accumulations of colloidal particles between the endothelial cell and the basementmembrane (Majno and Palade, 1961; Marchesi, 1962; Pappas and Tennyson, 1962; Peterson and Good, 1962; Cochrane, 1963; Movat, 1963; Rowley, 1963). Though phagocytosis was occasionally observed, it was a late and secondary event: the endothelium appeared to phagocytize particles only after they had escapedthrough intercellular gaps and accumulated against the basement membrane (Majno and Palade, 1961). Pinocytic uptake of somecolloidal particles (Palade, 1961; Jennings et al., 1962; Wissig, 1964) did not lead to the formation of phagocytic bodiesin significant number. These various findings also concurred to indicate that the problem of endothelial phagocytosis lies mostly beyond the range of conventional histology (Majno, 1964). In the present communication, the occurrence of endothelial phagocytosis is demonstrated at the level of the electron microscopeafter overloading the RES of rats and mice with colloidal carbon. MATERIALS
The animalswere male white rats of the Sprague-Dawley strain (Holtzman) weighing 270400 gm, and white Swiss mice of both sexesweighing 22-35 gm. They were fed with laboratory chow and water ad lib&m. The suspensionof carbon was a shellacfree, nontoxic, stable suspensionfrom the firm of Gunther Wagner of Hannover, Germany (Batch No. Cll-1431/a), containing approximately 100 mg of carbon per milliliter. The suspensionwas filtered and dialyzed against water before use to remove 1 This work was 06275, and 2G-113).
the phenol. All injections were given into the tail veins with the animals under light ether anesthesia. For histologic study, tissues were fixed in IOy, formalin or in Bouin’s fixative and stained with hematoxylin and eosin. For electron microscopy, tissues were removed while the animal was under deep ether anesthesia or after decapitation. Small pieces of tissue were transferred immediately into cold fixative [ 2% osmium tetroxide in Verona1 buffer at pH 7.3-7.4 (Palade, 1952) with 0.4 M sucrose (Caulfield, 1957)]. After 2-4 minutes the tissues were cut into cubes of approximately 1 mm and transferred to fresh fixative for 2-3 hours at 4°C. Samples of myocardium from the right and left ventricle were processed separately. After dehydration in increasing concentrations of acetone, some pieces were ‘Lprestrained” with potassium permanganate according to the method of Parsons (1961) ; others were dehydrated in graded ethanols without prestaining in permanganate. All specimens were embedded in Epon 812 (Luft 1961). Thin sections were cut on an LKB ultrotome with a diamond knife and placed on uncoated grids or grids coated with Parlodion; they were stained with lead hydroxide according to Karnovsky’s method A ( 1961). Some preparations were stained first with uranyl acetate and then with lead. Most of the micrographs were taken on an RCA 3D electron microscope; some on a Philips EM 200. Thick (0.5 CL) Epon sections were stained for light microscopy with toluidine blue (Trump et al., 1961). The animals were examined after the following treatments:
Dose of carbon (md100 am)
75 dailv 75 twice
Duration treatment (days)
Interval between last injection and sacrifice
to 40 hours to 24 hours
1 min 30 min 1 hour
to 7 days
Controls included untreated animals and animals injected once with 8-16 mg of carbon/100 gm, a dose that is normally cleared from the circulation without phagocytosis by non-RES endothelium (Benacerraf et al., 1959). RESULTS LIGHT-MICROSCOPIC
The light-microscopic observations were made on animals that had already cleared of circulating carbon. There was considerablevariation in the amount of carbon present in or around the endothelium of blood vesselseven in animals that received the largest dosesof carbon. The findings, however, were generally similar to those reported in mice by Benacerraf et al. (1959). Carbon deposits were found in the small blood vesselsof the myocardium, stomach, skeletal muscle, and skin; in the glomerular and peritubular capillaries of the kidney; in the lung capillaries; and in the endocardium. In the liver, carbon was present in markedly enlarged Kupffer cells and in the endothelium of blood vesselsin the portal tracts. In the spleen, both red and white pulp contained carbon, and in the adrenal, carbon was present along the sinusoids.There was minimal carbon deposition in the endothelium of the aorta. In the paraffin sections,
it was not possible to resolve whether the carbon was adhering to the endothelium, enclosed in the endothelial cytoplasm, or trapped along the basement membrane. The histologic picture in some lung capillaries and in a few myocardial capillaries suggested carbon thrombi plugging the lumina; however, blackening of endothelial cells without obvious thrombosis was the most common finding in the myocardium. ELECTRON-MICROSCOPIC
The majority of the observations were made on the endothelium of the small myocardial vessels of 14 overloaded animals. Other tissues examined were lung, kidney, aorta, spleen, and liver. Heart. The ultrastructure of the normal myocardial capillaries is well known (Palade, 1961; Majno, 1964). In our material, marginal folds or flaps (Fawcett, 1963), projecting from the luminal surface of the plasma membrane, were present with variable frequency. Control rats and mice receiving single injections showed carbon within the endothelial cytoplasm with extreme rarity.
FE. 1. Capillary from the myocardium of a rat which received four injections of carbon (75 mg/lOO pm) in 2 days and was sacrificed 35 minutes after the last injection. There are four phagocytic carbon bodies in the endothelium; 3 are membrane-bounded. X 18,600. (Scale for all micrographs z 1 CL.) B -1 basement membrane ; C = carbon particles; E = endothelial cell; F = flap; L = lumen; R = red blood cell; S = extravascular space.
In overloaded animals, the amount of carbon deposited within the cytoplasm of endothelial cells varied, as expected from the light-microscopic findings, even in pairs of animals submitted to the same treatment. However, blood vessels with carbon in the endothelium were present in almost all the tissue samples. The carbon particles usually formed clusters (Figs. 1 and 2) sometimes surrounded by a distinct membrane. These phagocytic bodies numbered l-7 in the cross section of any given endothelial cell; they averaged 1 ~1in diameter and were occasionally confluent. They had no con-
FIG. 2. Myocardial capillary (rat: 25 mg/lOO gm of carbon after last injection). This blood vessel contained a single carbon myocardium; N = nucleus. X 31,200.
daily for accumulation
13 days; killed in one section.
I hour My =
sistent position in the cell and they bore no apparent relation to other organelles. Occasionally, within a carbon-containing vacuole, there were clear areas sometimes stippled with fine granular material. Such clear areas have been described in phagocytizing Kupffer cells (Hampton, 1958) and pulmonary endothelium (Collet and Policard, 1962) (Fig. 6). Occasionally, rounded dense bodies, probably fat droplets, were incorporated in the vacuoles containing carbon (Figs. 1 and 3). The lining endothelial cells of the endocardium contained similar carbon accumulations. It was not possible to quantitate the amount of phagocytized carbon in relation to
time of sacrifice after the last injection, However, carbon-containing vacuoles were present in moderate numbers in the animals sacrificed 2 and 7 days after the last injection of carbon, at a time when no free circulating carbon was seen within the lumina of blood vessels (Fig. 3). At no time was carbon seen in the intercellular junctions. Intracellular carbon accumulations were also seen in pericytes and perivascular phagocytes (Fig. 4), but not
FIG. 3. Myocardial last injection). Note
capillary (mouse: 75 mg/lM) gm of carbon for 6 days; the lipid droplet (arrow) in one of the 5 carbon-containing
killed 7 days after bodies. X 25,000.
with the same frequency as in the endothelium proper. In some preparations, subendocardial carbon-laden mononuclear cells were conspicuous, and occasionally one was seen separating the lining endocardial cells (Fig. 5). Free carbon particles were rarely present between the external endothelial plasma membrane and the basement membrane (Fig. 6). The lumina of small vessels sometimes contained mononuclear cells with one or more membrane-bounded carbon deposits (Fig. 7). These cells were always distinct from the
4. Myocardium from clumps; the adjoining
same animal as in Fig. 3. A perivascular phagocyte contains several blood vessel is free of carbon. H = histiocyte; P = pericyte. X 10,200.
FIG. 5. Endocardium (right ventricle) of an overloaded rat 35 minutes after the last carbon injection. A mononuclear cell (MO) lies between a lining endothelial cell (E) and its basement membrane (B). Note the free fibrin (f) and carbon particles, and the carbon within the mononuclear cell. Cf = collagen fibrils. X 5200.
endothelium; they contained the usual cell organelles and a small number of vesicles. They were present even in the animal sacrificed 7 days after the last injection of carbon. They were especially numerous in the blood vessels of the lung (Fig. 8) and on the endocardial surface of the right ventricle, but they were also present in the left ventricle (Fig. 7). Although the origin of these free circulating carbon-laden cells cannot be ascertained, other studies suggest that the majority are circulating phagocytic cells detached from other locations, principally from the liver (Nicol and Bilbey, 1958).
FIG. 6. Lung capillary (mouse: 75 mg/lOO gm of carbon daily for 6 days, sacrificed 2 days after the last injection). There are carbon particles in an endothelial vacuole, as well as free between the endothelium and the basement membrane (arrows) ; none are present in the epithelium (Ep). Note the clear areas within the phagocytic vacuole. x 25,000.
Thrombi, containing carbon and fibrin, were occasionally seen in small vessels or adherent to the endocardium. In a rare vessel with a thrombus the endothelium was grossly disrupted (Fig. 9). Intraluminal thrombi may have accounted for some of the black deposits seen by light microscopy; in the thick Epon sections stained with toluidine blue, they could be distinguished from intramural carbon. The presence of phagocytic endothelium, however, was unrelated topographically to the presence of thrombi.
Evidence for participation of endothelial flaps in the process of phagocytosis was equivocal. Because of the variability in the number of flaps in control animals, it was not possible to show whether the number of flaps in overloaded animals was increased. However, in animals which were killed while some carbon was still circulating, some images suggested (Fig. 10) that the flaps may be involved in the initial steps of phagocytosis. Other tissues. There was considerable evidence of phagocytosis by the vascular endothelium of the lung (Fig. 11). The phagocytizing cells were similar to those found in
FIG. 7. A small myocardial 7 days after the last injection. x 11,300.
vessel from the left ventricular myocardium of an overloaded Free in the lumen is a mononuclear cell containing 2 carbon
the heart and to those described (Collet and Policard, 1962) after single injections of carbon in rats. In the kidneys, both glomerular and peritubular capillaries (Fig. 12) showed phagocytic endothelium. The aortic endothelium contained intracytoplasmic accumulations of carbon without evidence of endothelial damage or intimal thrombosis. DISCUSSION The results of this study show that considerable phagocytosis can occur in the nonRES vascular endothelium when the RES is overloaded with colloidal particles. The findings thus confirm-with the electron microscope-some of the light-microscopic
observations of Benacerraf et al. (1959) on the localization of colloidal particles in the endothelium. The conditions under which phagocytosis has been demonstrated, i.e., after repeated injections of colloidal carbon, deserve some comment. First, this study could not differentiate between two events: stimulation of endothelial phagocytosis or the demonstration of a normal low level of phagocytosis made more apparent by repeated injections. Furthermore, it is possible that the large doses of carbon injected could have produced a variety of secondary effects in addition to
FIG. 8. A phagocytic after overloading). The free cell, compared with
cell (MO) free in the lumen of a pulmonary capillary (mouse, endothelial cells contain no carbon. Note the sparsity of vesicles the large number of vesicles in the endothelium. x 13,000.
2 days in the
those of overloading per se. It has also been claimed that carbon preparations may contain variable amounts of endotoxin (Hopps and Dent, 1962). On the basis of the ultrastructural findings, two relevant points can be made: (1) the phagocytic endothelial cells were morphologically within normal limits, and (2) intercellular leaks were absent; the latter would have pointed to complicating factors such as release of vasoactive amines, inflammation, or antigen-antibody reactions. Despite the demonstration of phagocytosis, the impression gained from the study of
light- and electron-microscopic preparations was that such phagocytosis did not occur readiIy. Thus some of the most overloaded animals failed to show significant carbon accumulations, One can presume that their RES, stimulated by the repeated injections, was able to cope efficiently with the large amounts of circulating carbon.
FIG. 9. A myocardial blood vessel from an overloaded rat 40 minutes after the last injection. it contains carbon and fibrin. The endothelium is There is a conspicuous thrombus in the lumen; disrupted in several places (arrows). The round dense structures (X) are probably fragments of red blood cells. X 13,000.
The fate of the phagocytized carbon was partly clarified by this study. Some carbon remains sequestrated in the endothelium, at. least for 7 days (Fig. 3). Some may cross the endothelial cytoplasm and the basement membrane to be engulfed by pericytes or perivascular cells. It is also possible that some of the carbon may find its way back into the bloodstream, either free or in detached endothelial cells; convincing evidence for either of these latter events was not gathered in these experiments. In
this regard, Collet and Policard (1962) found that the amount of carbon present in the pulmonary endothelium 17 hours after carbon injection was appreciably reduced at 48 hours. The observations reported in this paper were made primarily on the endothelium of the heart vessels,which is considered to be clearly not a component of the RES. It is likely that non-RES endothelium in other locations may have different phagocytic
FIG. 10. A myocardial blood vessel 35 minutes after last carbon injection (overloaded rat). There is still some free-circulating carbon in the lumen (C). The clump of carbon at (X), while still in the lumen, is apparently bein g enveloped by endothelial flaps (arrows). The phagocytic body Z is intracellular. The carbon-containing vacuoles Y and V, although they appear intracellular, may also be located in depressions of the endothelium. X 31,200.
properties. The situation in the lung deserves some emphasis: Collet and Policard (1962) have shown unequivocal phagocytosis by the endothelium of the rat lung after single dosesof carbon that would not be normally phagocytized by endothelium of myocardial vessels. Phagocytosis by glomerular endothelium has been discussedby Farquhar et al. (1961). In our experiments, phagocytosis was also accomplishedby capillaries of the kidney. Phagocytosis by aortic endothe fenestrated peritubular
thelium confirms Buck’s observations (1958) on the aorta of the rabbit after a single large intravenous injection of thorotrast. Moore et aE. (1961a, b) studied, by light and electron microscopy, the endothelial uptake of saccharated iron oxide. They found saccharated iron oxide (in order of decreasing frequency, for any dose given) in the sinusoidal lining cells of the liver, spleen , glomerular endothelium, pulmonary capillaries, arteries, arterioles, and capillaries of tissues examined.
Frc. 11. Pulmonary carbon. Note endothelial (Ep). x 17,200.
capillary from phagocytosis;
an overloaded mouse, 2 days after the last injection no carbon is present in the lumen or in the epithelial
The relation of the phagocytosis here described to the micropinocytosis attributed to the vesicles of non-RES vascular endothelium is unclear. Studies with tracer particles have suggested that ferritin, thorotrast, and saccharated iron oxide (Jennings et al., 1962) can be carried by vesicles across the endothelial cell: thus the end result of such pinocytosis would be transport across the cell wall, probably in both directions (Wissig, 1964) without intracellular sequestration (Wissig, 1964). It is likely that the process of phagocytosis reported here is unrelated to transport, although the biochemical and morphologic events in pinocytosis and phagocytosis may be comparable (Karnovsky, 1962 ) .
In conclusion it can be stated that the present study confirms the ability of endothelial cells to perform phagocytosis. While it is true that the endothelium does not respond primarily with phagocytosis to a number of important stimuli, as has been suggested (Jansco, 1955; Gozsy and Kato, 1957), there is little doubt that the endothelial cell is capable of considerable phagocytosis under proper circumstances. In
FIG. 12. A fenestrated peritubular capillary from the kidney cortex of a mouse 7 days after carbon overloading. Note the intracellular carbon clumps in the perinuclear area of the endothelial cell. Ep is the basal portion of a tubular epithelial cell. X 8400.
the present experiments we have demonstrated endothelial phagocytosis as induced by systemic overloading; in experiments to be described elsewhere, it will be shown that extensive phagocytosis takes place after intercellular leakage of colloidal particles induced by histamine-a form of “reverse” phagocytosis by non-RES vascular endothelium. SUMM.4RY The RES of rats and mice was “overloaded” by repeated injections of colloidal carbon, and the non-RES vascular endothelium was examined by electron microscopy for evidence of phagocytosis. Phagocytosis of carbon was demonstrated in the endothelium of small myocardial vessels, in the endocardium, in pulmonary capillary endothelium, in the aorta, and in glomerular and peritubular capillary endothelium. Some carbon remained in the endothelium of the heart vessels for at least 7 days after overloading. Carbon particles were also present in circulating mononuclear cells and in perivascular phagocytes. This thelium
study thus confirms-at the level of the electron outside the RES to perform phagocytosis.
I wish to thank Dr. Guido Majno for his invaluable help and advice, Miss M. Manuel LaGattuta for technical assistance, and Mr. E. Garriga for help with the photography.
REFERENCES ALTSCHUL, R. (1954). “Endothelium: Its Development, Morphology, Function and Pathology.” Macmillan, New York. BENACERRAF, B., MCCLUSKEY, R., and PATRAS, D. (1959). Localization of colloidal substances in vascular endothelium. A mechanism of tissue damage. 1. Factors causing the pathologic deposition of colloidal carbon. Am J. Pathol. 36, 75-92. BUCK, R. C. (1958). The fine structure of endothelium of large arteries. J. Biophys. Biochem. Cytol. 4, 187-190. CAULFIELD, J. B. (1957). Effects of varying the vehicle for 0~0~ in tissue fixation. J. Biophys. Biochem. Cytol. 3, 827-829. COCHRANE, C. G. (1963). Studies on the localization of circulating antigen-antibody complexes and other macromolecules in vessels: 1. Structural studies. J. Exfitl. Med. 118, 489-503. COLLET, A., and POLICARD, A. (1962). Essai de localization infrastructurelle dans le poumon des elements du system reticula endotheliel. [email protected]
Rend. Sot. Biol. 166, 991-995. FARQUHAR, M. G., WISSIG, S. L., and PALADE, G. E. (1961). Glomerular permeabili:y. 1. Ferritin transfer across the normal glomerular capillary. J. Ezptl. Med. 113, 47-66. FAWCETT, D. (1963) Comparative observations on the fine structure of blood capillaries. In “The Peripheral Blood Vessels” (International Academy of Pathology Monograph No. 4), pp. 1744. Williams & Wilkins, Baltimore, Maryland. “Studies on Phagocytic Stimulation.” Univ. of Montreal Press, G~ZSY, B., and KATZ, L. (1957). Montreal, Canada. HAMPTON, J. C. (1958). An electron microscopic study of the hepatic uptake and excretion of submicroscopic particles injected into the blood stream and bile duct, Acta Amt. 32, 262-291. HOPPS, H. C., and DENT, T. E. (1962). India ink and reticuloendothelial blockade. Arch. Pathol. 74, 285-291. JANSCO, N. (1955). “Stoffanreicherung in Retikuloendothel und in der Niere.” Budapest Academia, Kiado, Hungary. JENNINGS, M. A., MARCHESI, V. T., and FLOREY, H. (1962). The transport of particles across the walls of small blood vessels. Proc. Roy. Sot. (London), Ser. B 156, 14-19. KARNOVSKY, M. J. (1961). Simple methods for “staining with lead” at high pH in electron microscopy. J. Biophys. Biochem. Cytol. 11, 729-732. KARNOVSKY, M. L. (1962). The metabolic basis of phagocytic activity. Phys. Rev. 42, 143-168. LUFT, J. H. (1961). Improvements in epoxy resin embedding methods. J. Biophys. Biochem. Cytol. 9, 409-414. Studies on inflammation. 1. The effect of histamine and MAJNO, G., and PALADE, G. (1961). serotonin on vascular permeability: .4n electron microscopic study. J. Biophys. Biochem. Cytol. 11, 571-605. MAJNO, G. (1964). Ultrastructure of the vascular membrane. In “The Handbook of Physiology,” Section 2, Vol. III. (In press). MARCHESI, V. T. (1962). The passage of colloidal carbon through inflamed endothelium. Proc. Roy. Sot. (London), Ser. B 166, 550-552. MOORE, R. D., RUPP, J., MUMAW, V., and SHOENRERC, M. (1961a). The reticuloendothelial system in the rabbit-Phagocytosis of saccharated iron oxide. Arc-h. Pathol. 72, 51-60. MOORE, R., MUMAW, V., and SHOENBERG, M. (1961). The transport and distribution of colloidal iron and its relation to the ultrastructure of the cell. J. Ultrastruct. Res. 5, 244-256. MOVAT, HENRY, Z. (1963). Acute inflammation. The earliest fine structural changes at the bloodtissue-barrier. Lab. Zwest. 12, 895. NICOL, T., and BILBEY, D. L. (1958). Elimination of the macrophage cells of the RES by way of the bronchial tree. Nature 182, 192. PALADE, G. E. (1952). A study of fixation for electron microscopy. J. Ewptl. Med. 95, 283-298.
PALADE, G. (1961). Blood capillaries of the heart and other organs. Circulalion 24, 368-384. PAPPAS, G., and TENNYSON, V. (1962). An electron microscopic study of the passage of colloidal particles from the blood vessels of the ciliary processes and choroid plexus of the rabbit. 1. Cell Biol. 15, 227-239. PARSONS, D. F. (1961). A simple method for obtaining increased contrast in Araldite sections by using post-fixation staining of tissues with potassium permanganate. J. Biophys. Biorhem. Cytol. 11, 492-494. PETERSON, R., and GOOD, R. A. (1962). Morphology of vascular permeability. 1. Passive cutaneous anaphylaxis. Lab. Invest. 11, 507-513. ROWLEY, D. A. (1963). Mast-cell damage and vascular injury in the rat: An electron microscopic study of a reaction produced by thorotrast. Brit. 1. Exptl. Pathal. 44, 284-290. TRUMP, G. F., SMUCKLER, E. A., and BENDITT, E. P. (1961). A simple method for staining epoxy sections for electron microscopy. J. Ufrastruct. Res. 5, 343-348. WISSIC, S. (1964). The transport by vesicles of protein across the endothclium of muscle capillaries. Anat. Record 148, 411. ZWEIFACH, B. (1961). “Functional Behavior of the Microcirculation.” Thomas, Springfield, Illinois.