Capillary Endothelial Cell Swelling Research Articles

Vascular permeability of skeletal muscle microvessels in rat arterial ligation model: in vivo analysis using two-photon laser scanning microscopy.

Peripheral artery disease (PAD) in the lower limb compromises oxygen supply due to arterial occlusion. Ischemic skeletal muscle is accompanied by capillary structural deformation. Therefore, using novel microscopy techniques, we tested the hypothesis that endothelial cell swelling temporally and quantitatively corresponds to enhanced microvascular permeability. Hindlimb ischemia was created in male Wistar rat's by iliac artery ligation (AL). The tibialis anterior (TA) muscle microcirculation was imaged using intravenously infused rhodamine B isothiocyanate dextran fluorescent dye via two-photon laser scanning microscopy (TPLSM) and dye extravasation at 3 and 7 days post-AL quantified to assess microvascular permeability. The TA microvascular endothelial ultrastructure was analyzed by transmission electron microscopy (TEM). Compared with control (0.40 ± 0.15 μm3 × 106), using TPLSM, the volumetrically determined interstitial leakage of fluorescent dye measured at 3 (3.0 ± 0.40 μm3 × 106) and 7 (2.5 ± 0.8 μm3 × 106) days was increased (both P < 0.05). Capillary wall thickness was also elevated at 3 (0.21 ± 0.06 μm) and 7 (0.21 ± 0.08 μm) days versus control (0.11 ± 0.03 μm, both P < 0.05). Capillary endothelial cell swelling was temporally and quantitatively associated with elevated vascular permeability in the AL model of PAD but these changes occurred in the absence of elevations in protein levels of vascular endothelial growth factor (VEGF) its receptor (VEGFR2 which decreased by AL-7 day) or matrix metalloproteinase. The temporal coherence of endothelial cell swelling and increased vascular permeability supports a common upstream mediator. TPLSM, in combination with TEM, provides a sensitive and spatially discrete technique to assess the mechanistic bases for, and efficacy of, therapeutic countermeasures to the pernicious sequelae of compromised peripheral arterial function.

The effect of calpeptin on injury and atrophy of diaphragm under mechanical ventilation in rats

To investigate the effect of calpeptin on diaphragmatic injury and atrophy under controlled mechanical ventilation in rats. A total of 24 SPF Sprague-Dawley (SD) rats were randomly divided into anesthetized control group (CON group), 24-hour controlled mechanical ventilation group (CMV group), and 24-hour CMV + treatment with calpeptin group (CMVC group), with 8 rats in each group. Animals in the CON group received an intraperitoneal injection of pentobarbital sodium without CMV and continuous infusion of pentobarbital sodium. A small-animal ventilator was used for 24 hours in rats of CMV group. Rats of CMVC were treated with a specific calpain inhibitor calpeptin (4 mg/kg). The drug was injected subcutaneously 2 hours before and 8, 15 and 23 hours after mechanical ventilation. Changes in diaphragm ultrastructure, light microscopic picture, and myosin heavy chain (MHC) expression were observed. (1) Alignment of myofilaments and normal Z-band, and the shape of mitochondria were maintained in CON group as revealed by electron microscope. The signs of misalignment of myofibrils, disruption of Z-band and vacuolar mitochondria were found in CMV group, and they were obviously improved in CMVC group. The density of muscle injury (× 10⁻²/μm²) in CMV group was significantly higher than that in control group (36.8 ± 13.7 vs. 6.4 ± 6.3, t=6.373, P=0.001), and that in CMVC group was significantly lowered (17.6 ± 9.1 vs. 36.8 ± 13.7, t=3.694, P=0.002).(2) In CON group, the diaphragm fibers appeared regular in cross section without pathologic change under light microscopy. Fuzzy muscle striations, irregular muscle fibers, centralized nuclei and swelling of capillary endothelial cells were observed in CMV group, while pathological changes in the CMVC group were milder significantly. (3) In CMV group, the density of MHCslow and MHCfast was lower compared with that of CON group, and the gray value was lowered by 61.1% (t=8.138, P=0.001) and 77.1% (t=8.844, P=0.001), respectively, especially in MHCfast. However, the gray values of MHCslow and MHCfast were increased by 1.51 folds (t=4.601, P=0.010), and 1.33 folds (t=2.859, P=0.011), respectively, after treatment with calpeptin, and the elevation was more significantly in MHCslow. Diaphragmatic injury and atrophy were found after CMV for 24 hours. Calpeptin could reverse the detrimental effects of CMV, and it suggested that calpain plays an important role in modulating the ventilator-induced dysfunction of the diaphragm.

Influence of intrauterine hypoxia on lung blood vessel development in rats and expression of vascular endothelial growth factor in the lungs

Objective To observe the impact of intrauterine hypoxia on the development of rat lungs and expression of vascular endothelial growth factor (VEGF) in the lungs as the time of hypoxia was extended.Methods To create a model of intrauterine hypoxia,12 pregnant rats were divided into four groups as follows:air-control group,hypoxic 2-day group,hypoxic 6-day group,and hypoxic 10-day group.At birth,we performed pulmonary vascular morphometry in newborn rats with Nis software,and measured pulmonary arterial diameter,wall thickness and wall thickness/pulmonary arterial diameter.We detected expression of VEGF protein by immunohistochemistry and mRNA by real-time polymerase chain reaction.Changes in pulmonary capillary endothelium under electron microscope were observed.One-way analysis of variance and the Student Newman Keuls q (SNK-q) test were applied for statistical analysis.Results As the hypoxic time was extended,wall thickness and wall thickness/pulmonary arterial diameter increased.Compared with the air-control group,pulmonary vascular wall thickness in the hypoxic 10-day group increased [(16.4 ± 5.9) vs (10.8±2.8) μm; q=-8.04,P＜0.05].Wall thickncss/pulmonary artcrial diameter in the hypoxic 10-day group increased compared with that in the air control group,hypoxic 2-day group and hypoxic 6-day group [(31.3±5.1) ％,(22.2±4.9) ％and (23.6±3.9) ％vs (24.1±3.9) ％;q=-7.08,-4.92 and-5.0,all P＜0.05].Expression of VEGF protein in the lungs increased in the hypoxic 6-day group compared with the air-control group [(13.7±3.9) ％ vs (9.3±3.5) ％; q=-6.83,P＜0.05],while the expression was higher in hypoxic 10-day group than in the air-control group and hypoxic 2-day group [(15.2±4.7) ％,(9.3±3.5) ％ vs (11.8 ± 3.3) ％] (q=-9.16 and-5.19,all P＜0.05).Expression of VEGF mRNA in the lungs increased in the hypoxic 6-day group compared with the air-control group [(1.6±0.2)vs (0.8 ±0.2); q=-5.07,P＜0.05],while the expression was higher in the hypoxic 10 day group than in the air-control group and hypoxic 2-day group [(2.2±0.3),(0.8±0.2) vs (1.3±0.2)] (q=-9.54 and-6.42,all P＜0.05).Electron microscopy showed puhnonary capillary endothelial cell swelling as the hypoxic time was extended.In the air-control group,there was no capillary endothelial cell hyperplasia and swelling; in hypoxic 2-day group,there was mild swelling of the capillary endothelial cells and a small amount of hyperplasia; in hypoxic 6-day group,there was moderate swelling of the capillary endothelial cells; and in hypoxic 10-day group:there was significant swelling of the capillary endothelial cells,and pyknosis.Conclusions Intrauterine hypoxia resulted in higher expression of VEGF protcin and mRNA.VEGF in the lungs of newborn rats was involved in the vascular development process. Key words: Fetal hypoxia; Vascular endothelial growth factor A; Lung; Animals, newborn; Rats

Na‐K‐ATPase activity involved in noise induced blood‐labyrinth‐barrier leakage

Na+/K+‐ATPase plays an important role in mediating barrier integrity in the polarized epithelial cells. In the cochlea, we found that Na+/K+‐ATPase subunit alpha 1 was the most abundant protein of the vascular tissue in the blood‐labyrinth‐barrier (BLB) and was expressed predominantly at the luminal side of the endothelium. Na‐K‐ATPase is known to be required for controlling the tight junctions in epithelial cells. In this study, we investigated the involvement of Na+/K+‐ATPase in noise induced BLB leakage. In mice, the expression and activity of Na+/K+‐ATPase were found to be correlated with the changes of expression of tight junction proteins as well as the change in BLB permeability. Increased expression of Na+/K+‐ATPase, at both transcriptional and protein levels, occurred after noise exposure at 120 dB SPL for 3h per day for two days. However, Na+/K+‐ATPase enzymatic activity was significantly decreased. There was a significant decrease in endothelial tight junction membrane contact points. In vitro inhibition of Na+, K+‐ATPase by ouabain corroborated the in vivo studies and showed swelling of capillary endothelial cells and increased ratio between occludin phosphorylation and dephosphorylation. The results support a functional link between Na+, K+‐ATPase and occludin in the BLB as one of the mechanisms for noise induced in the BLB.Support: NIDCD DC 00888, NIDCD DC 00105, NIDCD DC 005983 (p30)

An Obturator Reduces Time for Volumetric Measurements of the Foot and Ankle

Increased volume of a body region is often described as swelling, and can be intra- or extra-articular. Swelling in the lower limb can follow immobilization, surgery, or trauma such as an ankle sprain. Assessment and reduction of swelling should be pursued vigorously, especially after trauma, because fibrinous exudation and swelling of capillary endothelial cells can result in scar tissue formation that impedes rehabilitation. Different methods are used to measure volume of the foot and ankle, such as the figure-of-eight method with a tape, and water displacement volumetry. Water displacement provides a direct measure of volume and is the gold standard or criterion reference for other methods. The foot is immersed into a water-filled tank and the volume of displaced water is measured. The advantage over tape methods is that volume can actually be determined, compared with indirect estimates of volume that only correlate with volume change. One significant disadvantage of water displacement volumetry is th...

Authors' reply:

Authors' reply:

Electron microscopic stereology of capillary endothelial cells and cardiomyocytes in artificially arrested canine hearts.

In open heart surgery and transplantation, sufficient structural preservation of the myocardium immediately following cardioplegic arrest is a precondition for overcoming ischemia and for resumption of postischemic function. Therefore, we compared the protective effect of three clinically applied cardioplegic solutions with fibrillating and beating hearts using structural criteria. Left ventricular samples were taken from (1) beating, or (2) fibrillating or arrested hearts following coronary perfu-sion with (3) St. Thomas' Hospital solution, (4) histidine tryptophane ketoglutalate (HTK) (Custodiol), or (5) University of Wisconsin (UW) solution and fixed by immersion. Ultrastructural differences in the swelling of capillary endothelial cells and myocytes were quantitatively evaluated using stereological methods. Endothelial cells were somewhat more swollen after St. Thomas perfusion than those in beating and fibrillating hearts. HTK-arrested hearts showed significantly lower values for cellular edema than beating hearts. UW perfusion resulted in the (significantly) lowest degree of endothelial cell edema. Edematous changes in myocytes were significantly greater in St. Thomas-arrested hearts than in UW- or HTK-arrested hearts. Cardiomyocyte edema in beating and fibrillating hearts was comparable to that in St. Thomas-perfused hearts. Thus, the stereol-ogical analysis revealed significant differences between cardioplegic solutions in structural preservation of myocardial ultrastructure.

Hepatic and Renal Ultrastructural Lesions in Experimental Clostridium perfringens Type A Enterotoxemia in Chickens

Hepatic and renal electron microscopic investigations were carried out in 10 chickens that had experimental intraduodenal infection with Clostridium perfringens Type A. Fourteen days postinfection, there were ultrastructural changes in hepatocytes and renal tubular epithelial cells; these included mitochondrial lesions (swelling, cristolysis, rarefication of the matrix, myelin figures), glycogen loss, and capillary endothelial cell swelling in both organs, as well as thickening of glomerular basement membrane. The findings are discussed with particular reference to the pathogenesis of Clostridium perfringens Type A enterotoxemia.

Mechanisms and implications of capillary endothelial swelling and luminal narrowing in low-flow ischemias.

Decreased functional capillary density (FCD) is a characteristic of low-flow ischemia which often persists on reperfusion of a tissue even though supply blood flow is reestablished. Capillary narrowing is a possible mechanism for the reduced blood reflow through an increase of the hydraulic resistance of the capillary network. We attribute the narrowing to swollen endothelial cells, a condition that could be observed directly and was found to develop with blood acidosis in both hemorrhagic shock and intentional acid infusion. A computer model of blood flow in skeletal muscle was used to predict the impact of a 21% decrease in capillary luminal diameter on flow and leukocyte transit through the network. For hemorrhagic shock, reduction in blood flow due to capillary narrowing and lowered systemic pressure was offset by hemodilution causing a low-flow state. On reperfusion, flow could be restored if the narrowing was rectified by infusion of a hypertonic saline-dextran solution. Reinfusion with conventional Ringer's lactate only partially restored flow because of a persistent capillary narrowing. The rheological contribution of passive leukocytes added only a few percent to network resistance even with shock-narrowed capillaries. With leukocyte activation increased cell cytoplasmic viscosity caused prolonged transit times for leukocytes in the network with a concomitant elevation in resistance. We conclude that capillary narrowing is a mechanism of FCD reduction in reperfusion after low-flow ischemia, either by a direct effect on flow or by enhancing the transient plugging of leukocytes. Therapeutic strategies should aim to ameliorate or prevent capillary endothelial cell swelling, and to minimize the level of activated circulating leukocytes.

Effect of torbafylline on muscle blood flow, performance, and capillary supply in ischemic muscles subjected to varying levels of activity

Torbafylline, a novel xanthine derivative, was given to rats by gavage (2 x 25 mg.kg-1.day-1, morning and evening) to study its effect upon fast muscles (tibialis anterior and extensor digitorum longus) made ischemic by unilateral ligation of the common iliac artery and subjected to two types of activity imposed by chronic electrical stimulation at 10 Hz: (i) strenuous, 6 h.day-1 as 3 x 2 h with 90- to 120-min intervals or (ii) mild, 105 min.day-1 as 7 x 10-15 min with 90- or 85-min intervals for 12-14 days. Some of the deleterious effects on ischemic muscles of strenuous activity (reduced blood flow during contractions, less resistance to fatigue) were remedied by torbafylline treatment, and values normalized. Most notably, torbafylline significantly reduced the degree of capillary endothelial cell swelling. In addition, the degree of atrophy was reduced and fatigue resistance improved in muscles contralateral to ischemic, which had been impaired with the strenuous regime. Torbafylline treatment had little further effect on ischemic muscles subjected to mild stimulation, which on its own improved functional hyperemia, resistance to fatigue, and the capillaries per fibre ratio in tibialis anterior, although it did significantly increase the capillary per fibre ratio in extensor digitorum longus. These data indicate a possible role for torbafylline as an adjunct to exercise therapy for chronic muscle ischemia.

Swelling of capillary endothelial cells contributes to traumatic hemorrhagic shock-induced microvascular injury: a morphologic and morphometric analysis.

The endothelial cell (EC) response during the first 2 h after traumatic hemorrhagic shock (THS) was analyzed in the rat mesentery by electron microscopy. Using a computer-assisted image analysis system, we interactively measured THS-induced changes of the area and the mean height of EC as well as the number of swollen EC occluding the capillary lumen. Analysis distinguished between capillaries presenting with the lumen blocked by corpuscular blood cells and capillaries with an open lumen. THS resulted in a significant increase in EC height of capillaries with an open lumen, but not of capillaries with lumen blocked by blood cells when compared with the control group (p < 0.05). This phenomenon was found to be most prominent 60 min after THS. In addition, THS was accompanied by a significantly increased number of swollen EC which occluded capillaries with an open lumen. From these results we conclude that swelling of EC contributes to THS-induced microvascular injury. Occlusion of the capillary lumen by EC swelling may be regarded as the morphological correlate of the THS-induced 'no-reflow' phenomenon.

Amiloride-sensitive Na+ pathways in capillary endothelial cell swelling during hemorrhagic shock.

We recently discovered that the endothelium of skeletal muscle capillaries swells in the low-flow ischemia induced by hemorrhagic shock. The present study was undertaken to determine the Na+ transmembrane pathways involved in this swelling, since hypoxic cell swelling is attributed to an influx of Na+ and water. In an initial series of experiments, amiloride (5 mg/kg body wt), which blocks multiple Na+ pathways, was infused intravenously into anesthetized rabbits 30 min prior to shock (40% single-withdrawal hemorrhage). Intravital microscopy of treated capillaries in the rabbit tenuissimus muscle showed that after a 1-h shock period, there was no endothelial cell swelling, as evidenced by no measurable change in the width of red blood cells traversing the capillary. In contrast, the swollen endothelium of untreated capillaries reduced the luminal diameter by 20-25% with a preserved stationary abluminal membrane. The specific effects of amiloride on Na+ transport were investigated with amiloride analogues. Animal pretreatment with 5-(N,N-hexamethylene)amiloride, a selective inhibitor of Na(+)-H+ activity, in a dose of 0.5 mg/kg did not significantly mitigate shock-induced swelling; however, a dose of 1 mg/kg completely prevented it. Phenamil, a selective inhibitor of Na+ channel conductance, even at a potent dosage of 0.5 mg/kg, did not affect swelling. These results suggest a primary role for Na(+)-H+ exchange in endothelial cell swelling during hemorrhagic shock, possibly as a means to regulate cellular pH, which may become acidic during ischemia. Narrowed capillaries with elevated hydraulic resistances could delay and diminish resumption of microcirculatory flow on shock resuscitation.

Skeletal Muscle Adaptations to Ischemia and Severe Exercise

The influence of peripheral arterial insufficiency and severe exercise on muscle mass, capillarity and mitochondrial content was analyzed in rat extensor digitorum (EDL) muscles. Animals (n = 4) were assigned to three groups: ligation of the common iliac artery, chronic electrical stimulation (10 Hz, 6 h/day) of the peroneal nerve, and a combination of both. All interventions were carried out unilaterally, animals were treated for 10–12 days. The effects of the experimental procedures were judged by a paired analysis of treated muscles to contralateral untreated muscles of the same animals. Ligation of the common iliac artery alone had no effect on the quantitative structural composition of the EDL muscle. If ischemia was combined with electrical stimulation, a massive reduction in muscle mass occurred. However, as a consequence of an increase in mitochondrial and capillary density, the total mitochondrial volume and the total capillary length was maintained in these muscles. Stimulated ischemic muscles thus had improved conditions for aerobic energy supply for the remaining reduced myofibrillar apparatus. Extensive swelling of capillary endothelial cells was observed in stimulated as well as in contralateral muscles of all animals that received electrical stimulation. It is concluded that severe exercise may lead to extensive damage in ischemic skeletal muscles.

Pulmonary lesions induced by 3-methylindole and bovine respiratory syncytial virus in calves

SUMMARY Our objectives were to describe the ultrastructural morphogenesis of pulmonary lesions induced by 3-methylindole in 30- to 45-day-old Holstein calves and to determine whether toxic exposure to 3-methylindole exacerbates pulmonary lesions induced by bovine respiratory syncytial virus. Administration of 3-methylindole (0.25 g/kg) to calves resulted in interstitial edema and ultrastructural swelling of type-I alveolar epithelial cells and nonciliated bronchiolar epithelial cells as early as 4 to 6 hours after intraruminal administration. More severe alveolar edema containing protein was associated with swelling of capillary endothelial cells at 2 days after administration. Proliferation of type-II alveolar epithelial cells was first observed at 2 days after 3-methylindole administration, and marked hyperplasia of type-II epithelial cells and nonciliated bronchiolar epithelial cells was evident by 4 days after administration. Pulmonary cytochrome P-450 monooxygenase concentrations decreased significantly (P &lt; 0.001) by 12 hours after administration and did not increase significantly again by 8 days after administration. Calves were inoculated with bovine respiratory syncytial virus 3 days after administration of 3-methylindole, and pulmonary lesions were assessed 5 days after viral inoculation. Viral replication was demonstrated by fluorescence microscopy for viral antigen or by transmission electron microscropy in ciliated and nonciliated airway epithelial cells. Viral antigen was identified infrequently in alveolar macrophages and in type-II alveolar epithelial cells. 3-Methylindole exposure in calves did not result in more widespread distribution of viral antigen in alveolar tissue of respiratory syncytial virus-inoculated calves or in significant enhancement of viral pneumonia. The results indicated that young calves are susceptible to 3-methylindole-induced pulmonary injury and that nonciliated bronchiolar epithelial cells, type-I alveolar epithelial cells, and pulmonary endothelial cells are most susceptible to injury. The results failed to indicate that 3-methylindole pulmonary toxicosis significantly enhances respiratory disease induced by bovine respiratory syncytial virus.

The ultrastructural effects of acute decompression on the lung of rats: the influence of frusemide.

Rats were placed in a decompression chamber and the chamber and the pressure reduced to 265 mm Hg over a period of 1 hr. Other rats were subjected to the same treatment after having been given an injection of frusemide. An ultrastructural study of the lungs of these rats showed that acute decompression of this magnitude can cause swelling and disintegration of type I alveolar epithelial cells. Although capillary endothelial cells showed only minor cytoplasmic damage many capillaries were ruptured with diapedesis of red cells. This is believed to be caused by gross capillary dilatation associated with pulmonary congestion. Pretreatment with frusemide prevented capillary rupture and reduced the degree of epithelial degeneration. However, frusemide induced swelling of capillary endothelial cells which was unassociated with acute decompression.

Progressive perfusion impairment during prolonged low flow myocardial ischemia in dogs.

Recent studies have shown that after total coronary artery occlusion, there is impaired "reflow" of blood accompanied by myocardial and capillary endothelial cell swelling. To investigate the effect of prolonged low flow myocardial ischemia on coronary vascular resistance, regional hypoperfusion of the distal left anterior descending coronary artery was studied in 31 autonomically blocked dogs on right heart bypass. Heart rate, aortic pressure, and, during ischemia, left ventricular end-diastolic pressure were held constant. The distal left anterior descending coronary artery was perfused at a substantially reduced perfusion pressure which resulted in an antegrade coronary blood flow that usually was between 3% and 7% (0.5-1 ml/min) of control. When relative hypothermia (33-34 degrees C) was induced in nine dogs, left anterior descending coronary artery vascular resistance did not change during 2.5-3 hours of low flow ischemia. Under euthermic conditions (37-40 degrees C) in 17 dogs there was a consistent progressive increase in distal left anterior descending coronary artery vascular resistance starting at 90 minutes (median) after onset of ischemia. By 110-140 minutes ischemic antegrade flow decreased by 35 +/- 4% (SEM) (P less than 0.01). Directionally similar flow changes were observed in six euthermic experiments using the krypton-85 washout technique. Light microscopy did not reveal hemorrhage as a cause of the increased vascular resistance. The perfusion impairment did not occur in two euthermic, nonischemic hearts. In five dogs elevation of serum osmolality by 23 +/- 11 mOsmol/liter with mannitol attenuated the progressive decrease in flow. Thus, a progressive perfusion defect exists in the ischemic low flow state in the heart which presumably contributes to the extent of eventual necrosis.

Effects of mannitol on cardiac ultrastructure and microcirculation following anoxia.

Electron microscopic and microcirculatory effects of hyperosmolal mannitol were evaluated in the isolated perfused isovolumic rat heart. Specimens for ultrastructural examination were obtained in 26 experiemnts after 15 min of sequential aerobic, anoxic, and reoxygenated perfusion using an isosmolal perfusate of Krebs-Ringer-Henseleit bicarbonate buffer (KRB) (osmolality equals 290 mosmol/kg) vs. a hyperosmolal solution of KRB + mannitol (equals 350 mosmol/kg). No significant changes were noted during aerobic perfusion. Anoxic hearts perfused with isosmolal KRB demonstrated the most severe ultrastructural alterations including: mitochondrial swelling with disruption of cristae, myofibrillar fusion and contraction bands, and subsarcolemmal edema and vacuolization. These subcellular changes were not only partially reversed by oxygenated isosmolal perfusion but were significantly reversed during both the anoxic and reoxygenation perfusion periods with mannitol added. Following silicone rubber injection of the microcirculation, only focal capillary endothelial cell swelling was noted, and no difference in arteriolar or capillary filling was observed with either perfusate. Thus, mannitol significantly reversed the postanoxic ultrastructural changes consistently observed in the absence of increased osmolality. No gross effect on vascular patency could be demonstrated.

The effects of hyperosmolal coronary perfusion on the haemodynamic, metabolic and ultrastructural changes of myocardial anoxia.

Summary Recovery from anoxia has been evaluated in the isovolumic non-recirculating paced perfused rat heart. Seventy studies were performed consisting of 15 min of aerobic perfusion (AP); AP+15 min anoxic perfusion; and AP+15 min anoxic perfusion+15 min reoxygenation (recovery). Krebs-Ringer-bicarbonate+5 mmol glucose (KRB) (290 mmol) was compared to KRB+mannitol (350 mmol). Mannitol decreased myocardial water content. It improved recovery of haemodynamic function after reoxygenation. With KRB alone left ventricular systolic peak pressure (LVSP) decreased 32% and maximum dP/dt by 50%. With mannitol added LVSP decreased 18% and dP/dt 21% (P&amp;lt;0·01). KRB and mannitol did not differentially affect total coronary flow, lactate and glucose extraction, tissue glycogen, creatine phosphate or adenine nucleotide concentrations. No difference in submicroscopic appearance was noted with either perfusate during aerobic perfusion. Anoxic hearts perfused with isosmolal KRB demonstrated the most severe ultrastructural alterations including mitochondrial swelling with disruption of cristae and extraction of matrix components, myofibrillar fusion and contraction bands, and subsarcolemmal oedema and vacuolization. These changes were only partially reversed during reoxygenated perfusion. However, cellular changes were reversed or markedly improved during both the anoxic and reoxygenation perfusion periods with hyperosmolal solutions. When studied by silicone rubber injection of the microcirculation, only focal capillary endothelial cell swelling was noted and no difference in arteriolar or capillary filling was observed with either perfusate. Mannitol appears to improve LV function by direct myocardial osmotic action unrelated to enhanced energy production.