Interendothelial Clefts Research Articles

A model for interpreting the tracer labeling of interendothelial clefts.

We extended the model describing the low molecular weight electron dense tracer wake in the interendothelial cleft and surrounding tissue to describe the time-dependent transport of intermediate size solutes of 1.0-3.5 nm radius by convection and diffusion in an interendothelial cleft containing a fiber matrix. This model provides a quantitative basis on which to reinterpret electron microscopic studies of the distribution of tracers such as horseradish peroxidase (HRP; molecular weight = 40,000; Stokes radius = 3.0 nm) along the interendothelial cell cleft from the lumen to the tissue. For example, we show that, in contrast to our results with low molecular weight tracers, the wake of large molecular weight tracers on the abluminal side of the junctional strand is not likely to be detected, because the concentration of the tracer is predicted to be very low in most experiments. Thus the lack of a tracer such as HRP on the abluminal side of the junctional strand and in the tissue is not as strong evidence against the presence of a cleft pathway as suggested previously. The model does provide the basis for the design of experiments to locate both the principal molecular sieve and breaks in the junctional strand from the standing gradient on the luminal side of the junctional strand. An important experimental variable is the pressure in the vessel lumen which can be varied between 0 and 30 cm H2O to change the contributions of diffusive and convective transport to transcapillary exchange through he interendothelial cleft. This approach will also allow the testing of models for transcapillary pathways for large molecules by measuring the distribution of fluorescent traces across the microvessel wall and in the tissue surrounding the microvessel using confocal microscopy.

Ultrastructural study on the interaction of native and cationized albumin-gold complexes with mouse brain microvascular endothelium

The main objective of this ultrastructural study was to gain insights into the cellular mechanisms responsible for the enhanced brain uptake of blood-borne cationized albumin observed by several authors utilizing quantitative methodology. Mice were injected intravenously or into the common carotid artery (in vivo experiments) or perfused in situ with solutions of native or cationized bovine serum albumin complexed with colloidal gold (BSA-G or cBSA-G respectively). The results indicate that: (1) the main drawbacks of in vivo experiments are very intense phagocytosis of the tracer particles by Kupffer cells located in the liver sinusoids and also escape of the tracer through capillaries of skeletal and heart muscles. This results in a rapid decline of the concentration and disappearance of the circulating tracer particles; (2) BSA-G and cBSA-G both in vivo (up to 30 min circulation) or after perfusion in situ (up to 15 min) do not cross the wall of brain microvessels representing the blood-brain barrier; (3) enhanced entrance of cationized albumin into the brain occurs through fenestrated endothelium of the capillaries located in the examined circumventricular organs (median eminence and neurohypophysis). Although BSA-G is also transported by these fenestrated capillaries, this process is evidently less intense than in the case of cBSA-G; (4) the enhanced passage of cBSA-G across fenestrated capillaries occurs mainly via vesicular transport (adsorptive transcytosis), through transendothelial channels and eventually through interendothelial junctional clefts; (5) the fenestrated capillaries of the choroid plexus appear to be less permeable for both tracers than those located in the other circumventricular organs.

A diffusion wake model for tracer ultrastructure-permeability studies in microvessels.

We developed a time-dependent diffusion model for analyzing the concentration profiles of low-molecular-weight tracers in the interendothelial clefts of the capillary wall that takes into account the three-dimensional time-dependent filling of the surrounding tissue space. The model provides a connecting link between two methods to investigate transvascular exchange: electron-microscopic experiments to study the time-dependent wake formed by low-molecular-weight tracers (such as lanthanum nitrate) on the tissue side of the junction strand discontinuities in the interendothelial cleft of frog mesentery capillaries (R. H. Adamson and C. C. Michel. J. Physiol. Lond. 466: 303-327, 1993) and confocal-microscopic experiments to measure the spread of low-molecular-weight fluorescent tracers in the tissue space surrounding these microvessels (R. H. Adamson, J. F. Lenz, and F. E. Curry, Microcirculation 1: 251-265, 1994). We show that the interpretation of the presence of tracer as an all-or-none indication of a pathway across the junctional strand is likely to be incorrect for small solutes. Large-pore pathways, in which the local tracer flux densities are high, reach a threshold concentration for detection and are likely to be detected after relatively short perfusion times, whereas distributed small-pore pathways may not be detected until the tissue concentrations surrounding the entire vessel approach threshold concentrations. The analysis using this approach supports the hypothesis advanced by Fu et al. (J. Biomech. Eng. 116: 502-513, 1994) that the principal pathways for water and solutes of < 1.0 nm diameter across the interendothelial cleft may be different and suggests new experiments to test this hypothesis.

Scanning and transmission electron microscopic studies of microvascular pathology in the osmotically impaired blood-brain barrier

The present investigation focused on the structural events occurring in endothelial cells lining the lumina of brain microvessels in rats subjected to a single intracarotid injection of hypertonic 1.8 M L (+) arabinose solution with or without intravenous injection of horseradish peroxidase. Blood vessels from cerebral cortex and thalamus were evaluated by transmission and scanning electron microscopy. After short-term exposure (10-12 min) there was widespread flooding of peroxidase into the brain neuropil of the ipsilateral hemisphere. Peroxidase tracer was frequently observed within vesiculo-tubular profiles, and occasionally within widened interendothelial junctional clefts. Partially fragmented, necrotic endothelial cells appeared to be in the process of desquamation. Individual endothelial cells appeared to be shrunken with widened interendothelial spaces. Some healthy endothelial cells appeared to be involved in repair processes, manifested by the extension of thin cellular processes towards the area of vessel injury. Other pathological alterations included a conspicuous increase in the number of endothelial cell microvilli, large crater-like invaginations of the endothelial plasma membranes and muscular blood vessels in the process of spasm. We also observed a platelet reaction with or without endothelial cell necrosis and attached microthrombi in some arterial segments.

Differentiation of the microvascular endothelium during early angiogenesis and respiratory onset in the chick chorioallantoic membrane

The present study served to determine the extent of microvascular endothelial differentiation during early stages of morphogenesis (days 4.5–5.5 of the 21-day incubation) in the chick chorioallantoic membrane (CAM). CAM's, which serve as the embryonic lung, were prepared for intravital injections of a graded series of FITC-dextrans and subsequent ultrastructural morphometric analyses of the microvascular units. The precapillary, capillary, and postcapillary microvascular segments presented a continuous endothelium that was substantially thicker than that of adult lung endothelia ( DeFouw, 1988). Further, plasmalemmal vesicles were uniformly sparse, while endothelial vacuoles, of variable diameters, were present continuously in the proliferating microvascular units. Average widths and depths of the interendothelial clefts were uniform and suggested complete structural differentiation from the onset of CAM morphogenesis. Based on our recent estimates of CAM microvascular permeability coefficients ( Rizzo et al., 1995), the observed endothelial ultrastructure was associated with microvascular selectivity comparable to that of adult pulmonary microvessels ( Lanken et al., 1985). Therefore, despite incomplete ultrastructural differentiation of the early CAM microvascular endothelium, these angiogenic microvessels presented adult-like barrier properties. Further they were less permeable than ( Wu et al., 1993; Yuan et al., 1993) and ultrastructurally distinct from ( Kohn et al., 1992) certain tumorigenic microvessels. Thus, angiogenesis is likely not a routinely homogeneous process, and CAM microvascular permeability characteristics may be teleologically significant.

Molecular Organization and Functional Regulation of Cell to Cell Junctions in the Endothelium

Intercellular junctions are important structural determinants of endothelial permeability. These organelles are formed by a complex network of transmembrane proteins linked to a well developed plasmalemmal undercoat. One of the typical characteristics of endothelial junctions is their dynamic organization. Endothelial cells are able to rapidly change the architecture of the junctions to allow the passage of plasma constituents and circulating cells. This effect, in most of the cases, is quickly reversible and the endothelium is able to disorganize/reorganize the intercellular junctions within minutes. The mechanisms that regulate the opening and the closure of endothelial junctions are still obscure. It is conceivable that inflammatory agents increase permeability by binding to specific receptors generating intracellular signals which in turn cause cytoskeletal reorganization and opening of interendothelial gaps. Endothelial junctions also control leukocyte extravasation. Once leukocytes have adhered to the endothelium, a coordinated opening of interendothelial clefts occurs. The mechanism by which this takes place is unknown, but it might present characteristics similar to that triggered by soluble mediators.

A Junction-Orifice-Fiber Entrance Layer Model for Capillary Permeability: Application to Frog Mesenteric Capillaries

The recent serial section electron microscopic studies by Adamson and Michel (1993) on microves gels of frog mesentery have revealed that the large pores in the junction strand of the interendothelial cleft are widely separated 150 nm wide orifice-like breaks whose gap height 20 nm is the same as the wide part of the cleft. In this paper a modified version of the model in Weinbaum et al. (1992) is first developed in which this orifice structure is explored in combination with a random or ordered fiber matrix layer that is at the luminal surface and/or occupies a fraction of the wide part of the cleft. This basic orifice model predicts that for the measured Lp to be achieved the fiber layer must be confined to a relatively narrow region at the entrance to the cleft where it serves as the primary molecular filter. The model provides a much better fit of the permeability P for intermediate size solutes between 1 and 2 nm radius than the previous model in Weinbaum et al., where the junction strand breaks were treated as finite depth circular or rectangular pores, but like the previous model significantly underestimates P for small ions. However, it is shown that if a small frequent pore of 1.5 nm radius with characteristic spacing comparable to the diameter of the junction proteins or a continuous narrow slit of approximately 1.5 to 2.3 nm gap height is also present in the continuous part of the junction strand, small ion permeability can also be satisfied. The 1.5 nm radius pore does not significantly change Lp, whereas the continuous narrow slit provides a contribution to Lp that is comparable to, or in the case of the 2.3 nm slit greater than, the widely spaced 150 nm orifices. Thus, for the narrow slit the contribution to Lp from the orifices can be as low as 1.0 x 10(-7) cm/s/cm H2O and it is also possible to satisfy the 2.5 fold increase in permeability that occurs when the matrix is enzymatically removed from the luminal side of the cleft, Adamson (1990). The likelihood of each of these cleft structures is discussed.

The pentafraction of hydroxyethyl starch inhibits ischemia-induced compartment syndrome.

Pentafraction (PF), a solution of biodegradable hydroxyethyl starch macromolecules with molecular weights of 10 to 100 x 10(4) daltons, has been shown to minimize tissue edema by sealing interendothelial clefts at the capillary level. The effect of PF on ischemia-reperfusion-induced compartment syndrome was studied. Ten rabbits underwent bilateral femoral artery occlusion following ligation of branches from the terminal aorta to the popliteal artery. After 7 hours of ischemia, reperfusion was established with heparinized polyethylene shunts. Experimental animals (n = 5) received PF and control animals (n = 5) received normal saline (NS) as an intravenous infusion (30 mL/kg) for 1 hour, beginning 10 minutes after shunt placement. During reperfusion, anterior compartment pressure was continuously monitored in the left lower extremity. To quantitate oxidative metabolism, triphenyltetrazolium chloride (TTC) reduction (micrograms of TTC per milligram of protein) of tibialis anterior muscle from the right lower extremity was measured at femoral artery occlusion, 7 hours of ischemia, and 2 hours of reperfusion. In the NS group, anterior compartment pressure significantly increased from the end of the ischemic interval, 10.8 +/- 4.14 to 36.4 +/- 9.9 mmHg and 44.6 +/- 15.4 mmHg, after 1 and 2 hours of reperfusion (p < 0.007) compared with the PF group, which did not change significantly, 10.6 +/- 2.6 to 11.4 +/- 12.9 mmHg and 7.4 +/- 2.8 mmHg, after 1 and 2 hours of reperfusion (p < 0.67).(ABSTRACT TRUNCATED AT 250 WORDS)

Hypoxia induced disruption of the cardiac endothelial glycocalyx: implications for capillary permeability

The aim was to determine the effect of hypoxia on the ultrastructure of the endothelial glycocalyx of cardiac capillaries. Isolated rat hearts were perfused with oxygenated or hypoxic Krebs solution for 30 min after equilibration with oxygenated medium. They were then perfused with a cationic marker (ruthenium red, lanthanum nitrate, or cationised ferritin) to delineate the cardiac endothelial glycocalyx in the electron microscope. With ruthenium red and lanthanum, perfusions were carried out both in the presence and absence of bovine serum albumin. Ferritin perfused hearts were used to quantify changes in the glycocalyx as a result of hypoxia and to measure the cross sectional area of the endothelial cells. In all the hearts perfused with well oxygenated solution, all three markers showed an even, electron dense layer on the luminal surface of the capillaries. With ruthenium red and lanthanum (but not with ferritin), the marker was occasionally observed throughout the length of the interendothelial clefts and on the albuminal surface. After 30 min hypoxic perfusion, both ruthenium red and lanthanum showed disruption and irregular clumping of the glycocalyx, with or without albumin. Ferritin, however, showed a sparse and uneven layer. Measurements of endothelial cell area showed that some cells from hypoxic hearts were swollen when compared with controls. Measurements of the percentage of luminal membrane covered by ferritin molecules showed a significant loss of glycocalyx in hypoxic hearts. There was, however, no correlation between loss of glycocalyx and endothelial cell swelling. The endothelial cell glycocalyx of continuous capillaries is sensitive to changes in PO2. The disruption of this surface coat may explain the reported increase in capillary permeability in hypoxia.

Hapten‐tagged plasma proteins as immunocytochemical probes for the study of vascular permeability

Bovine serum albumin and transferrin were covalently coupled with fluorescein isothiocyanate and digoxigenin, respectively, and intravenously co-injected in equal amounts in mouse. The derivation of the two proteins induces minor alterations of their physicochemical properties as well as of their physiological functions. The two tracers were revealed within vascular and extravascular compartments of diaphragm by quantitative postembedding immunocytochemistry, using antibodies against each of the haptens in conjunction with the protein AG-gold complexes. The influence of different fixatives and embedding protocols on the immunodetectability of the hapten-tagged proteins was assessed. Both resist reasonably well to osmication and embedding in Epon. None of the haptens reacted with the heterologous antibody. At 30 minutes after injection, the tracers were detected in blood plasma, interstitium, and endothelial plasmalemmal vesicles. The presence of both proteins within the interendothelial clefts was inconspicuous. The ratios between the labeling densities found over endothelium, interstitial space, and vascular lumen were similar for both tracers. This suggests that the endothelium of mouse diaphragm capillaries might exhibit comparable permeabilities towards serum albumin and transferrin which are similar in size and charge. The study shows that hapten-tagged polypeptides are close to the corresponding native macromolecules, and represent interesting tools for the morphological study of dynamic processes such as transcytosis.

Interendothelial junctions during blood-brain barrier development in the rat: morphological changes at the level of individual tight junctional contacts

The endothelium of brain capillaries represents the structural basis for the blood-brain barrier in vertebrates. Individual endothelial cells are linked by a continuous belt of complex tight junctions (zonulae occludents). Hydrophilic solutes and macromolecules are believed to cross the barrier through specific carrier mechanisms. Unspecific paracellular ionic leak is thought to be very low. In rats the blood-brain barrier is not fully developed until postnatal day 24. We investigated the ultrastructure of the developing blood-brain barrier at 5 developmental stages between embryonic day 17 and young adults. The use of high power goniometric tilting of ultrathin sections allows one to gather information about the exact relationship between two opposing membranes throughout the entire length of the cleft. Our results suggest that the maturation of blood-brain barrier interendothelial clefts is accompanied by the establishment of a characteristic ratio of ‘narrow zone’ (complex tight junctions) to ‘wide zone’ (15–20 nm), and of a typical cleft length. Membrane separations larger than 20 nm disappear and individual tight junctional contacts undergo structural changes.

The interendothelial junction in myocardial capillaries: evidence for the existence of regularly spaced, cleft-spanning structures.

Water and hydrophilic solutes cross the endothelium of continuous capillaries via the paracellular cleft and possibly other routes. This pathway shows a selectivity to molecule size and charge. However, it is not yet known which systems confer this selectivity. Isolated rat hearts were perfusion-fixed through the coronary circulation, stained with lanthanum or tannic acid, and further processed for transmission electron microscopy. Thin sections viewed at x 160,000 magnification revealed regularly spaced, cleft-spanning structures in the wider zone of a small percentage of clefts in addition to at least one zonula occludens. Goniometric tilting of the specimen in steps of 5 degrees perpendicular to the plane of the wide zone showed that such "linkers" can be revealed in at least 40% of all clefts. They become visible at some tilt angles, although the same area of the cleft is featureless at other angles. Single linker spacing measurements were obtained using a computerized image analysis system, and compiled in a frequency distribution chart. On the basis of these data, two models of a regular linker distribution within the cleft are illustrated. Our results provide evidence for the presence of regularly spaced, cleft-spanning structures within the interendothelial cleft which may have implications for endothelial cell-cell adhesion and permeability.

Transcapillary transport of solute by the endothelial vesicular system: Evidence from thin serial section analysis

Upon crossing the walls of continuous capillaries, perfused electron-dense solutes often appear to become lodged in the attenuated space bounded by pericytes and the ablumenal capillary surface. In single thin sections, it is difficult to determine whether such accumulations result from transcellular transport via the endothelial vesicular system or whether tracer has filled these spaces retrogradely after paracellular transport through interendothelial clefts. We have taken continuous thin serial sections through patches of terbium trapped between pericytes and the endothelial cells of continuous capillaries in the rete mirabile of the eel. Sampling these areas in three dimensions has revealed that the terbium deposits are bounded and limited to regions of the capillary wall devoid of interendothelial associations. Dense interstitial terbium deposits were also continuous with ablumenal endothelial caveolae and could also be seen in clusters of fused vesicles. These observations strongly implicate transcellular transport by the endothelial vesicular system as the route by which accumulated terbium had crossed the capillary wall.

Blood-brain barrier permeability to micromolecules and edema formation in the early phase of incomplete continuous ischemia

The distribution patterns of ionic Lanthanum (La3+; mol. wt. 139) were evaluated after 15, 30 and 60 min of middle cerebral artery occlusion in perfused-fixed rats. Blood-brain barrier (BBB) permeability to Evans blue (EB) and horseradish peroxidase (HRP; mol. wt. 40,000) in vivo was also evaluated. Brain tissue specific gravity was measured. An increase in brain water content was found as early as 30 min following occlusion. HRP and EB extravasation was not observed. La3+ crossed the interendothelial clefts of venules and capillaries at 30 and 60 min and was seen in both extracellular and intracellular brain compartments at 60 min. La3+ extravasation was seen in nonedematous areas bordering the regions of water accumulation. Our findings suggest that the early phase of incomplete continuous ischemia is accompanied by changes in BBB permeability and the interendothelial clefts of venules and capillaries seem to represent one of the early sites of ischemic damage.

Permeability of frog mesenteric capillaries after partial pronase digestion of the endothelial glycocalyx.

1. The proteolytic enzyme pronase, which degrades the endothelial cell glycocalyx, was perfused through single capillaries of frog mesentery. Hydraulic conductivity (Lp) of each vessel was determined before and after pronase perfusion. In three vessels in which Lp increased, the ultrastructure of interendothelial clefts was examined. In a separate group of frogs the effect of pronase on the endothelial glycocalyx was assessed by using cationized ferritin to label the capillary luminal surface. 2. Control Lp was 2.0 x 10(-7) cm s-1 cmH2O-1 (10 mg ml-1 bovine serum albumin, BSA, in frog Ringer solution). Vessels were then perfused with a solution containing 0.1 mg ml-1 pronase and 10 mg ml-1 BSA for 1 min. Lp measured in the same eleven vessels increased to 4.9 x 10(-7) cm s-1 cmH2O-1 (P less than 0.005). 3. Transverse sections of three of these vessels were examined by transmission electron microscopy at eight sites along each vessel. In these sections a total of 156 interendothelial cell clefts were found and photographed. No morphological features, such as fenestrations, transendothelial channels, or intercellular gaps associated with inflammation, were found which might account for the increases in Lp. 4. Measurement of cleft dimensions yielded a harmonic mean cleft depth (delta x) of 0.32 microns and an arithmetic mean cleft depth of 0.64 microns. Mean width (w) of the clefts outside the tight regions was 0.012 microns and the cleft length per unit area (L) was 1330 cm-1. The mean fractional pore area of vessel wall per unit cleft depth, Ap/delta x, calculated as Lw/delta x, was 48.7 cm-1. 5. There was less cationic ferritin (CF) labelling of the luminal glycocalyx in pronase-perfused than in control capillaries. On average, the proportion of the luminal surface covered by CF was 85% in controls and 42% in pronase-treated capillaries (P less than 0.01). In some vessels the CF pattern was greatly disrupted, indicating large changes in the glycocalyx structure. 6. It is concluded that the moderate increases in Lp induced by pronase perfusion are associated with partial digestion of the endothelial glycocalyx but are not accompanied by changes in the dimensions of the intercellular cleft. These observations support the fibre matrix hypothesis of capillary permeability and suggest that the endothelial glycocalyx contributes as much as 60% of the hydraulic resistance of the capillary wall.

Interfacial Transport in Large and Small Blood Vessels

In this paper we shall review some recent mathematical models which have led to new conceptual views of the ultrastructural pathways by which water, solutes and large molecules cross the endothelial interface between tissue and blood. In particular, we shall show how a sequence of models for the endothelium and underlying tissue in large arteries have finally led to the experimental discovery of the large pore via which LDL and other large molecules enter the artery wall and how a new three-dimensional model for the interendothelial cleft in capillaries might reconcile the several long standing paradoxes relating to the measured filtration and solute permeability coefficients in the transcapillary exchange of water and hydrophilic solutes in the microcirculation.

Ultrastructural distribution of terbium across capillary endothelium: detection by electron spectroscopic imaging and electron energy loss spectroscopy.

We used terbium as an intravital tracer of permeability pathways across the walls of capillaries in the rete mirabile of the eel swimbladder and in frog mesentery. Terbium was detected in unstained ultra-thin sections by electron density using electron spectroscopic imaging (ESI) and by electron energy loss spectroscopy (EELS). Enhancement of intrinsic contrast in zero loss images (elastically scattered electrons) permitted imaging of membrane-bound compartments and terbium within them which might otherwise have been undetected in counterstained sections. Element-selective imaging with EELS indicated that terbium was associated with heavy electron-dense deposits, but the terbium mass:volume of sections in areas of lighter deposition was insufficient to obtain a terbium signal. In the rete capillaries, terbium was deposited on the luminal surface, throughout vesicular profiles, and in the interstitium, but could not be traced through interendothelial junctions. Fine terbium deposits were detectable throughout apparent vesicular connections across the endothelium. In the frog mesentery, terbium penetrated some but not all interendothelial clefts, and was detectable in small quantities within luminal and abluminal vesicular profiles and in the interstitium. The results indicate that in the rete capillaries, terbium permeates the capillary via a transcellular route. This route may be provided by transient fusions of luminal and abluminal vesicular compartments.

Effect of hypoxia on endothelial morphology and interendothelial junctions in the isolated perfused rat heart

Isolated perfused rat hearts were used to assess the effect of 30 or 60 mins of hypoxia on the ultrastructure of the capillary endothelium and particularly on the interendothelial junctions. Perfusions were carried out both in the presence and absence of albumin. Albumin had no effect on ultrastructure or membrane spacing in the interendothelial clefts, neither in oxygenated controls nor in hypoxic hearts. After 30 and 60 mins of hypoxia some capillaries showed endothelial swelling while after 60 mins the endothelium of others was attenuated. The wide regions of the intercellular cleft were not affected by hypoxia but the “narrow zone” gap between membranes became significantly smaller. We conclude that factors within the clefts other than the “tightness” of the narrow zones are responsible for changes in permeability in hypoxia and with albumin.

A NEW MODEL FOR CAPILLARY FILTRATION BASED ON RECENT ELECTRON MICROSCOPIC STUDIES OF ENDOTHELIAL JUNCTIONS

Existing models to predict the filtration and permeability of hydrophilic solutes in the capillary interendothelial cleft have been based largely on one-dimensional models of channels with constrictions that are represented by long rectangular slits of 6 to 8 nm gap width that run along the length of the junction and do not interact with one another [1,7–10]. It has been commonly assumed based on these simple models that the water and solute fluxes are proportional to the area of the junction that is open. Recent electron microscopic studies [7, 11] using freeze fracture and ultrathin serial sectioning techniques indicate that the pores in the tight junctions might be numerous much narrower disruptions in the protein strands of 10 to 20 nm length. A new model is proposed in which the pores are short periodically distributed equivalent cylindrical holes in the junctional strands that have dimensions comparable to individual missing proteins. Since these pores have spacings which are small compared to the d...

Transcapillary adenosine transport and interstitial adenosine concentration in guinea pig hearts.

We used the multiple-indicator-dilution technique to observe the capillary transport of adenosine in isolated Krebs-Henseleit-perfused guinea pig hearts. Tracer concentrations of radiolabeled albumin, sucrose, and adenosine were injected into the coronary inflow; outflow samples were collected for 10-25 s and analyzed by high-performance liquid chromatography (HPLC) and by gamma- and beta-counting. The albumin data define the intravascular transport characteristics; the sucrose data define permeation through interendothelial clefts and dilution in interstitial fluid (ISF). Parameters calculated from adenosine data include permeability-surface area products for endothelial cell uptake at the luminal and abluminal membranes and intraendothelial metabolism. We found that in situ endothelial cells avidly take up and metabolize adenosine. Tracer adenosine in the capillary lumen is twice as likely to enter an endothelial cell as it is to permeate the clefts. There was no adenosine in the arterial perfusate. Under control conditions, the steady-state venous adenosine concentration was 3.6 +/- 0.8 nM, which from the flow and the parameters estimated from the tracer data gave a calculated ISF concentration of 6.8 +/- 1.5 nM. During dipyridamole infusion (10 microM) at constant pressure, the cell permeabilities went essentially to zero, whereas the venous adenosine concentration increased to 44.0 +/- 12.6 nM, giving an estimated ISF concentration of 191 +/- 53 nM. With constant flow perfusion, venous concentration during dipyridamole infusion was 30.9 +/- 6.3 nM, and estimated ISF concentration was 88 +/- 20 mM. We conclude that in this preparation, at rest, the ISF adenosine concentration is about twice the venous concentration and the ISF adenosine concentration increases with dipyridamole administration.