Vasoregulation by the beta1 subunit of the calcium-activated potassium channel.

Small arteries exhibit tone, a partially contracted state that is an important determinant of blood pressure. In arterial smooth muscle cells, intracellular calcium paradoxically controls both contraction and relaxation. The mechanisms by which calcium can differentially regulate diverse physiological responses within a single cell remain unresolved. Calcium-dependent relaxation is mediated by local calcium release from the sarcoplasmic reticulum. These 'calcium sparks' activate calcium-dependent potassium (BK) channels comprised of alpha and beta1 subunits. Here we show that targeted deletion of the gene for the beta1 subunit leads to a decrease in the calcium sensitivity of BK channels, a reduction in functional coupling of calcium sparks to BK channel activation, and increases in arterial tone and blood pressure. The beta1 subunit of the BK channel, by tuning the channel's calcium sensitivity, is a key molecular component in translating calcium signals to the central physiological function of vasoregulation.

Natural history of infection with Bartonella bacilliformis in a nonendemic population.

An investigation was performed after an outbreak of bartonellosis in a region of Peru nonendemic for this disorder. Symptoms of acute and chronic bartonellosis were recorded. Serological analysis was performed on 55% of the affected population (554 individuals), 77.5% of whom demonstrated previous infection with Bartonella bacilliformis. The attack rate of Oroya fever was 13.8% (123 cases); the case-fatality rate was 0.7%. The attack rate of verruga peruana was 17.6%. A new specific immunostain was developed and used to confirm the presence of B. bacilliformis in the biopsied skin lesions. Most seropositive individuals (56%) were asymptomatic. The symptoms that were associated with prior infection, as determined by Western blot, included fever (37.2% of the seropositive vs. 17.2% of the seronegative population; P<.001), bone and joint pain (27% vs. 9%; P<.001), headache (27% vs. 12.3%; P <.001), and skin lesions described as verruga peruana (26.8% vs. 4.9%; P<.001). Our findings suggest that infection with B. bacilliformis causes a broad spectrum of disease that is significantly milder in severity than that frequently reported.

Rhinosporidium seeberi: a human pathogen from a novel group of aquatic protistan parasites.

Rhinosporidium seeberi, a microorganism that can infect the mucosal surfaces of humans and animals, has been classified as a fungus on the basis of morphologic and histochemical characteristics. Using consensus polymerase chain reaction (PCR), we amplified a portion of the R. seeberi 18S rRNA gene directly from infected tissue. Analysis of the aligned sequence and inference of phylogenetic relationships showed that R. seeberi is a protist from a novel clade of parasites that infect fish and amphibians. Fluorescence in situ hybridization and R. seeberi- specific PCR showed that this unique 18S rRNA sequence is also present in other tissues infected with R. seeberi. Our data support the R. seeberi phylogeny recently suggested by another group. R. seeberi is not a classic fungus, but rather the first known human pathogen from the DRIPs clade, a novel clade of aquatic protistan parasites (Ichthyosporea).

Tophaceous pseudogout of the mitral valve.

This report describes a 61-year-old patient on chronic hemodialysis with multiple, left-sided, intracardiac masses causing intermittent coronary obstruction. Mitral valve replacement was performed. Massive deposition of calcium pyrophosphate crystals in and around the valve cusps led to the diagnosis of tophaceous pseudogout (tumoral calcinosis) of the mitral valve.

Gamma interferon induces Fas-dependent apoptosis of Peyer's patch T cells in mice following peroral infection with Toxoplasma gondii.

Since we previously observed a remarkable decrease in the numbers of T cells in the Peyer's patches of the small intestines in C57BL/6 mice following peroral infection with Toxoplasma gondii, we performed studies to examine the mechanism(s) whereby this decrease in numbers of the T cells occurs. We found that apoptotic cell death of CD4+ and CD8+ alphabeta T cells occurred in Peyer's patches following infection. Upregulation of Fas expression was observed in these T cells. C57BL/6-background mutant mice which lack functional Fas antigen did not develop apoptosis in their Peyer's patches following infection. Treatment of infected C57BL/6 mice with anti-gamma interferon (IFN-gamma) monoclonal antibodies prevented the upregulation of Fas on their Peyer's patch T cells and inhibited the occurrence of apoptosis of these T cells. These results indicate that IFN-gamma induces Fas-dependent apoptosis in CD4+ and CD8+ alphabeta T cells in Peyer's patches in C57BL/6 mice following peroral infection with T. gondii.

Lung pathology in high-altitude pulmonary edema.

Autopsy findings in 10 cases of high-altitude pulmonary edema have been collected from published articles and personal observations. All cases were males with a mean age of 37 years (22-62). The altitude of occurrence was from 8400 to 17 500 feet. The mean combined lung weight in nine cases was 1682 g (1200-3000 g). Cerebral edema was present in five of eight cases. The most frequency pulmonary findings in addition to diffuse edema consisted of leukocyte infiltrates, alveolar hemorrhages, thrombi in small pulmonary arteries, and alveolar hyaline membranes. Pulmonary infarction was present in only one case. Right ventricular dilatation was commonly present. The left ventricle was normal. No significant coronary disease was present.

Benign lymphangioendothelioma.

We describe a 40-year-old white man with a red-brown, indurated plaque on the proximal aspect of his right thigh. The lesion had been present since birth, and the patient had a 20-year clinical history of recurrent cellulitis in the same area. The histopathologic features of the lesion included permeation of the dermis by flattened, endothelium-lined channels without cellular atypia, hemorrhage, or inflammation. The endothelial cells were stained intensely with monoclonal antibody anti-CD34 (clone MY10). In addition, antibodies to factor VIII antigen, HLA-DR, smooth muscle actin, ICAM-1, and the lectin Ulex europaeus labeled the luminal cells. The basement membrane of the channels stained with anti-type IV collagen and laminin. Desmin-positive cells were abundant adjacent to the channels. Factor XIIIa stained both mononuclear cells and occasional dendritic cells in the perivascular area. Ki-67 immunolabeling could not be demonstrated on fresh or frozen tissue. Electron microscopy revealed the presence of both tight junctions and a well-formed, continuous basement membrane but the absence of Weibel-Palade bodies.

Effectiveness of a purified human hemoglobin as a blood substitute in the perfused rat liver.

The efficacy of purified cross-linked human hemoglobin solution in maintaining the metabolic integrity of perfused livers from fed rats was determined and compared with that of livers recirculated with red blood cells, perfluorocarbon, and Krebs' Ringer bicarbonate buffer media at normal and accelerated flow rates. The data indicate that oxygen utilization was comparable in livers perfused with red blood cell medium (53.5 +/- 4.0 microL.g liver-1.min-1), hemoglobin (45.7 +/- 1.9), and perfluorocarbon (57.2 +/- 6.1) and less in livers perfused with Krebs' Ringer bicarbonate buffer solution at normal (17.4 +/- 1.4) and high (27.7 +/- 1.4) flow rates. Bile flow, the outflow of glucose and lactic acid, and residual glycogen levels were similar when livers were perfused with red blood cells (hematocrit, 19) and hemoglobin solutions containing 7 g hemoglobin/dL at flow rates of 1.1-1.2 mL.g liver-1.min-1. However, livers perfused with perfluorocarbon at 1.1 mL.g-1.min-1 showed a significantly greater (P less than 0.01) decline in bile flow and outflow of glucose and lactic acid. Livers perfused with Krebs' Ringer bicarbonate buffer at normal (1.3 mL.g-1.min-1) and accelerated flow rates (3.0 mL.g-1.min-1) also showed a progressive decrease in bile flow and marked glycogenolysis as well as depletion of adenosine triphosphate content. In addition, morphological studies (light and electron microscopy) showed more vacuoles, membrane alterations, and increased mitochondrial swelling in livers perfused with Krebs' Ringer bicarbonate buffer and perfluorocarbon. These findings suggest that hepatocyte function in livers from fed rats is maintained equally well with hemoglobin solutions and with red blood cell medium, suggesting that cross-linked hemoglobin solution may serve as an effective blood substitute for maintaining adequate oxygenation and metabolic integrity of the isolated perfused rat liver.

The reversibility of canine vein-graft arterialization.

We assessed the reversibility of functional and morphological changes of arterialized vein segments by returning them to the venous circulation. Thirteen dogs underwent right carotid and femoral veno-arterial grafting. After 12 weeks, veno-arterial grafts were removed for contractility (norepinephrine [NE] and 5-hydroxytryptamine [5-HT]), luminal prostacyclin (PGI2), and morphometric analyses; the remaining segments were used as left jugular and femoral veno-venous grafts. After another 12 weeks, the veno-venous grafts were harvested. To NE, veno-arterial grafts (ED50, 5.4 +/- 0.1 [-log M]) were less sensitive than control veins (ED50, 6.0 +/- 0.2) or veno-venous grafts (ED50, 6.4 +/- 0.2) but were more sensitive than control arteries (ED50, 4.0 +/- 0.1); the maximum tension of veno-arterial grafts (6.2 +/- 0.6 g) was greater than that of veins, less than that of arteries (9.8 +/- 1.0 g), and comparable with that of veno-venous grafts (5.1 +/- 1.1 g). To 5-HT, veno-arterial (ED50, 7.5 +/- 0.1) and veno-venous (ED50, 7.3 +/- 0.2) grafts were more sensitive than arteries (ED50, 6.0 +/- 0.3), while the vein was unresponsive; the maximum tension of veno-arterial grafts (5.0 +/- 0.7 g) was less than that of arteries (6.9 +/- 0.9 g) and greater than that of veno-venous grafts (1.4 +/- 0.3 g). PGI2 production in veins (3.6 +/- 0.8 ng/ml), veno-arterial grafts (3.9 +/- 0.8 ng/ml), and veno-venous grafts (3.3 +/- 0.9 ng/ml) was comparable and less than that of arteries (6.4 +/- 0.9 ng/ml). Veno-arterial graft intimal thickness (127 +/- 8 microns) and intimal area (15.6 +/- 1.8 x 10(3) microns 2) tended to be greater than that in the veno-venous graft (113 +/- 9 microns and 12.4 +/- 1.8 x 10(3) microns 2); also, the veno-arterial graft medial area (103.0 +/- 7.3 x 10(3) microns 2) was greater than that of the veno-venous graft (80.3 +/- 6.9 x 10(3) microns 2), thereby resulting in a similar relative intimal area (13 +/- 1%). Therefore, some changes associated with arterialization, for example, adrenergic sensitivity, maximum tension to 5-HT, medial thickening, and perhaps intimal hyperplasia, reverted toward venous values when replaced in the venous environment, possibly due to variations in pressure, flow, shear stress, and/or graft preparation techniques. Luminal PGI2 was unchanged in the grafts, implying that graft contractility was not modulated by luminal PGI2.

Immunohistochemical localization of low density lipoprotein receptors in adrenal gland, liver, and intestine.

The localization of LDL receptors in adrenal gland, liver, and intestine was studied using immunohistochemistry. The anti-LDL receptor antibody used was shown to be monospecific and did not react with striated muscle, a tissue which has a very low level of LDL receptors. Similarly, cerebral cortex showed only faint reactivity and that was to an area previously demonstrated to have LDL receptors. Adrenal gland was intensely reactive with the zona fasciculata, having a greater density of receptors than the zona reticularis. In normal liver, LDL receptors were present on the sinusoidal membranes and were sparse in the areas of hepatocyte-to-hepatocyte contact without an obvious portal to central gradient. LDL receptors were present throughout the intestine. In jejunum, staining was most intense at the base of the villus and extended up toward the villus tip. At the base of the villus, the receptor was primarily at the basal lateral membrane, but toward the villus tip, there was appreciable intracellular staining. Staining in crypts was more faint; in duodenum, staining in crypts equaled that in the villus region in intensity. In colon, there was intense staining throughout the epithelial cells. These results provide new information about the cellular and subcellular localization of LDL receptors and raise the interesting possibility that there is a role for LDL-derived cholesterol in new lipoprotein formation.

Mechanisms responsible for inhibition of vein-graft arteriosclerosis by fish oil.

Favorable changes in lipoproteins, inhibition of platelet aggregation, reduction of serum thromboxane (TX), altered plasma-membrane fluidity, and reduced production of growth factors (mitogens) have all been implicated as possibly being involved in the inhibition of arteriosclerosis by fish oil (FO), which is rich in omega 3 fatty acids; however, causal relations are mostly lacking. Several putative mechanisms responsible for the salutary effects of FO were investigated in a canine model of accelerated vein-graft arteriosclerosis. Venoarterial autografts (N = 192) were implanted in 48 hypercholesterolemic dogs divided into six groups: group A, control; B, FO (as MaxEPA, 200 mg/kg/day eicosapentaenoic acid); C, aspirin (ASA, 50 mg/kg/day); D, TX synthetase inhibitor (TXSI [CGS-12970], 10 mg/kg/day); E, FO + ASA; and F, FO + TXSI. At sacrifice 3 months later, there was no significant difference in plasma lipoproteins, hepatic low density lipoprotein-receptor concentration, red blood cell fragility, bleeding time, or platelet count compared with controls; the decrease in platelet aggregation (30 +/- 5% [mean +/- SEM]) was similar in all treatment groups. Arterialized vein-graft intimal thickening was significantly inhibited by FO (with or without ASA), while ASA alone was ineffective. Conversely, serum TX was significantly lower only in the ASA and FO + ASA groups. Serum mitogenic activity was higher at 3 months in the control group versus all treatment groups. Compared with baseline values, serum mitogenic activity rose significantly over time in the control and the TXSI groups, and an increase or rising trend was present in all other treatment groups except for the FO-treated animals. Thus, the salutary biologic effect of FO in this hypercholesterolemic model of arterialized vein grafts may have been more related to in vivo inhibition of platelet-mitogen growth factor release than to changes in lipoproteins, low density lipoprotein receptors, platelet function, or eicosanoid metabolism. These observations underscore the need for further studies to clarify the interactions between FO (omega 3 fatty acids) and paracrine cellular mitogenic factors in the context of atherosclerosis prevention.

Effects of fish oil on arteriosclerosis in the Japanese quail.

The effects of fish oil on the development of arteriosclerosis were assessed using a special susceptible strain (SEA) of Japanese quail (Coturnix coturnix japonica). Sixty four quail were randomly divided into two groups and placed on isocaloric and approximately isocholesterolic (2% by weight) diets. Group A (control) was supplemented with 10% beef tallow oil, while group B received 10% Menhaden fish oil. The birds were sacrificed at 10 weeks (early) and 15-16 weeks (late). Based on semiquantitative histological grading of the arteriosclerotic lesions in the proximal aorta and brachiocephalic arteries, a score from 1 (no lesion) to 5 (severe, diffuse lesions) was assigned. A total of 57 quail were evaluated (seven died prior to scheduled sacrifice). At the early period, the mean arteriosclerosis scores for group A (n = 8) and group B (n = 8) were 3.3 (SD 1.0) and 1.9(1.0) respectively (p less than 0.017); 63% of the quail in group A and 13% of those in group B had a score greater than or equal to 3 (p less than 0.25, NS). At the late period, the scores for group A (n = 20) and group B (n = 21) were 3.8(0.6) and 2.6(0.9), respectively (p less than 0.001); 95% of the birds in group A and 43% of those in group B had a score greater than or equal to 3 (p less than 0.005). Histopathological examination of the arteriosclerotic lesions revealed disruption of the innermost elastic lamina, increased proteoglycan deposition in the medial interlamellar spaces, and the distinct involvement of macrophage like cells. Compared to human disease, arteriosclerosis in the quail is marked by distinct similarities, as well as differences. The SEA strain of Japanese quail appears to be a practical model for the study of arteriosclerosis; fish oil reduces the severity of disease in these birds when fed a high cholesterol diet.

Effects of aspirin, dipyridamole, and cod liver oil on accelerated myointimal proliferation in canine veno-arterial allografts.

The effects of the administration of aspirin (ASA), dipyridamole (DPM), and cod liver oil (CLO) on graft patency rate and degree of intimal hyperplasia were investigated in a canine, hypercholesterolemic veno-arterial allograft model in an attempt to modify this immunologically mediated vascular injury. The drug regimens were ASA 1 mg/kg/day, DPM 10 mg/kg/day, combined ASA and DPM (ASA + DPM), and CLO (1.8 g/day eicosapentanoic acid [EPA] and 1.2 g/day docosahexanoic acid [DHA]), and control. The early angiographic patency rate (1-3 weeks) was 81% +/- 10% (+/- 70% confidence limits); the 90-day overall patency rate was 60% +/- 4% (87/144), with no statistically significant differences among the groups (range 46 +/- 10-71 +/- 9%). Qualitatively, there was no difference in luminal thrombus, intimal hemorrhage, or lesion eccentricity. Considering the relatively short time of graft implantation, an extensive amount of microscopic disease was observed; quantitatively, the mean intimal thickness was 515 +/- 17 microgram overall but was not statistically different between the groups. The fraction of potential lumenal area occupied by intimal thickening was 0.37 +/- 0.01 but again did not differ significantly between the groups. These doses of ASA, DPM, ASA + DPM, and CLO did not alter graft occlusion or retard the marked degree of subintimal myointimal cell hyperplasia that was generated in this hypercholesterolemic canine veno-arterial allograft preparation. Possible explanations for these negative findings include inadequate dosage or form of omega-3 fatty acids and the antiplatelet drugs administered, excessive variability in graft response due to uncharacterized immunologic histocompatibility, and the possible influence of non-platelet-mediated mechanisms. Nevertheless, this preparation is attractive as a reproducible model of accelerated (immunologically mediated) experimental arteriosclerosis.

Carcinogen bioassay of isoflurane in mice.

A carcinogen bioassay of isoflurane was performed in groups of Swiss/Webster mice exposed to either air (n = 181), 0.1% isoflurane (n = 167), or 0.4% isoflurane (n = 165), for 4 h per day, 5 days per week. After 78 weeks of exposure, mice were left untreated for 3 weeks and were then killed. Mice killed at this time when they were 86 weeks of age, and those killed or dying at other times during the study were subjected to complete gross and microscopic examination. Throughout most of the study, mean body weights of mice exposed to 0.1% isoflurane and 0.4% isoflurane were less by 1-5% and 5-8%, respectively, than that of mice exposed to air alone. Otherwise, no gross toxic treatment effects were noted. The first neoplastic lesion was detected 23 weeks after starting treatment and, by the end of the study, 190 tumors had been detected in 179 mice. However, there were no statistical differences among the groups in the number of mice with a particular tumor at a specific site, the ratio of benign to malignant tumors, or the time to tumor appearance. It was concluded that isoflurane is unlikely to have carcinogenic potential and is a remarkably non-toxic anesthetic in mice.

Endoproteolytic cleavage of gp160 is required for the activation of human immunodeficiency virus.

The envelope protein of human immunodeficiency virus (HIV) is synthesized as a polyprotein (gp160) and cleaved intracellularly to a gp120-gp41 heterodimer. In this study, the tryptic-like endoproteolytic cleavage site was removed by site-directed mutagenesis and replaced with a chymotryptic-like site. The resultant mutant, RIP7/mut10, was found to be indistinguishable from wild-type HIV when analyzed at the level of proviral replication, RNA processing, protein expression, and viral assembly. However, the gp160 polyprotein was not cleaved and the mutated virions were biologically inactive, until and unless they were exposed to limiting concentrations of chymotrypsin. As is the case for other enveloped mammalian viruses, endoproteolytic cleavage of the HIV envelope protein and release of a unique hydrophobic domain appear to be necessary for the full expression of viral infectivity.

Inhibition of vein graft intimal thickening by eicosapentanoic acid: reduced thromboxane production without change in lipoprotein levels or low-density lipoprotein receptor density.

Marine lipids containing omega-3 fatty acids (chiefly, eicosapentanoic acid [EPA] and docosahexanoic acid [DHA]) may inhibit the development of atherosclerotic vascular disease, but the mechanisms responsible for this putative beneficial effect are unknown. We investigated the effects of EPA and DHA in a canine model of accelerated vein graft arteriosclerosis during a 3-month period. Twenty-five dogs were divided into three dietary groups: group I (control), group II (2.5% cholesterol), and group III (2.5% cholesterol plus 2 gm EPA/day [as MaxEPA]). The effects of EPA on vein graft intimal thickening, platelet and vascular prostaglandin metabolism, lipid and lipoprotein receptor metabolism, and hematologic parameters were assessed. Cholesterol feeding caused a significant 54% increase in graft intimal thickness compared with control animals (124.9 +/- 50.4 vs 81.2 +/- 32.4 micron; p = 0.013), which was prevented by supplementation with EPA in group III (56.9 +/- 30.0 micron; p = 0.001 vs group II). Intimal thickness in group III was not significantly different from that of control. EPA supplementation was also associated with a 38% decline in serum thromboxane levels from 457.0 +/- 129.3 pg/0.1 ml in group II to 283.5 +/- 96.9 pg/0.1 ml in group III (p = 0.007). The alterations in lipoprotein metabolism associated with cholesterol feeding were not affected by EPA: in both groups II and III, serum cholesterol and high-density lipoproteins and liver cholesterol content were elevated and hepatic low-density lipoproteins (LDL) receptor content was reduced. There were no differences between the three groups in terms of vein graft or native vessel prostacyclin production, hematocrit, platelet count, or coagulation parameters. In this canine model, dietary supplementation with marine omega-3 fatty acids reduced the extent and magnitude of accelerated vein graft intimal thickening induced by hypercholesterolemia; moreover, this beneficial effect was associated with lower serum thromboxane production and appeared to be independent of alterations in lipoprotein metabolism or LDL receptor density.

Cardiac cryolesions as an experimental model of myocardial wound healing.

The standard coronary ligation model for experimental myocardial infarction results in variable areas and patterns of necrosis; therefore, the healing of such infarctions is also variable. The authors developed an experimental myocardial injury model using simple cryoinjury, which allows standardization of the size, depth, and location of the wound. Thirty-eight left ventricular cryolesions were created in 19 dogs, which were then killed from 3 to 35 days after injury. A consistent decrease in the depth of scar (p less than 0.005) and accumulation of collagen (p less than 0.0001) over time characterized this healing myocardial wound. Histologic examination revealed that the cellular pattern of healing myocardial cryolesions is similar to that of a healing myocardial infarction but with less variability. The authors advocate the use of cardiac cryolesions as a model of experimental myocardial wound healing.

Incomplete biochemical adaptation of vein grafts to the arterial environment in terms of prostacyclin production.

Biochemical (or functional) adaptation of venoarterial grafts has been demonstrated recently. We reexamined one aspect of this biochemical "arterialization" process: prostacyclin (PGI2) production by canine venoarterial autologous grafts and the responsiveness of this biosynthetic pathway to maximal stimulation with substrate enhancement. Four reversed autologous grafts (femoral vein) were interposed into both carotid and femoral arteries in eight dogs. After 12 weeks, the grafts were removed, and radioimmunoassay was used to determine luminal surface production of 6-keto-PGF1 alpha (the stable metabolite of PGI2) in both the basal and stimulated (27 mumol/L arachidonic acid [AA]) states. PGI2 production by the venous autologous grafts was compared with that of control native artery and vein. We confirmed that PGI2 production (measured in nanograms per milliliter) by control artery was greater than vein under both basal conditions (5.8 +/- 0.4 [+/- SEM] vs. 2.7 +/- 0.5, p less than 0.001) and stimulated conditions (8.8 +/- 0.8 vs. 5.5 +/- 0.4, p = 0.002); moreover, AA stimulation significantly increased PGI2 production in both native artery and vein compared with basal PGI2 production. Under basal conditions, graft PGI2 production (6.3 +/- 1.6 ng/ml) was not significantly different than basal arterial levels (p = 0.8) but was higher than basal venous levels (p = 0.05). However, in marked contrast to both native artery and vein, the vein graft flow surface showed no significant response to substrate enhancement with AA: basal (6.3 +/- 1.6 ng/ml) vs. stimulated (5.9 +/- 0.9 ng/ml) (p = 0.8). These observations confirm that canine venoarterial autologous grafts undergo biochemical "arterialization"; however, this process appears to be an incomplete one.(ABSTRACT TRUNCATED AT 250 WORDS)

Halothane hepatotoxicity in Fischer 344 rats pretreated with isoniazid.

Male Fischer 344 rats were used to investigate the hepatic effects of exposure to halothane under normoxic conditions (FIO2 = 0.21) in isoniazid-treated rats. Animals were treated with saline or isoniazid (50 mg/kg) for 7 days and then were exposed to either 1% halothane or air for 2 hr. One-half of the rats from each treatment and exposure group were killed 24 hr postexposure; the remaining were killed 4 days postexposure. Twenty-four hours following halothane exposure, serum transaminase levels were significantly elevated in isoniazid- compared with saline-treated rats (i.e., aspartate aminotransferase = twofold; alanine aminotransferase = seven-fold). Cholesterol levels were significantly depressed by halothane exposure in both saline- and isoniazid-treated rats. Other serum parameters indicative of hepatic and renal function were not different: alkaline phosphatase, total protein, total bilirubin, hematocrit, uric acid, creatinine, urea nitrogen, Na+, K+, Ca2+, and inorganic phosphate. Neither saline-treated nor isoniazid-treated rats exposed to air exhibited histologic evidence of hepatic damage. Halothane-exposed rats, however, showed a circumscribed disruption of cellular morphology. The most severe lesions were observed with isoniazid-treated animals with extensive pericentral hepatocellular necrosis and infiltration by leucocytes and Kupffer cells. Serum concentrations of two products of the oxidative metabolism of halothane, trifluoroacetic acid and bromide, were significantly elevated in isoniazid- compared with saline-treated rats. Serum levels of fluoride, a product of reductive metabolism, were not different. These results strongly suggest that hepatic injury following halothane administration can be produced by intermediates of oxidative metabolism.