Transfusion of soluble hemoglobin.
Explore the source record for details and available documents.
Biomedical subjects
Publications and source records attributed to R I Roth.
Explore the source record for details and available documents.
Bacterial endotoxin (lipopolysaccharide) has affinity for a number of cations, including iron. Previous investigations have demonstrated that lipopolysaccharide can affect the oxidation rate of iron; heme-bound ferrous iron in hemoglobin is oxidized to ferric iron when hemoglobin binds lipopolysaccharide. In the present study, we directly examined the interaction between lipopolysaccharide and iron. Lipopolysaccharide caused a concentration-dependent increase in the rate of iron oxidation, with up to a 23-fold increase in oxidation in the presence of 200 microg/ml Escherichia coli lipopolysaccharide. This effect was seen both with several carbohydrate-rich smooth lipopolysaccharides and also with carbohydrate-poor rough lipopolysaccharide. Extensively deacylated rough lipopolysaccharide had no effect, suggesting a role of the fatty acid components of lipopolysaccharide in this process. Purified lipid A produced inconsistent results: some preparations stimulated iron oxidation and others did not. A series of sugars, starches and a preparation of purified O-chain polysaccharide (the carbohydrate portion of the lipopolysaccharide macro-molecule) had no effect on the rate of iron oxidation, whereas phospholipid-enriched brain tissue extracts (similar to the lipid A component of lipopolysaccharide) stimulated oxidation. We conclude that the lipid moiety of bacterial lipopolysaccharide is responsible for the stimulation of iron oxidation. This process may contribute to the ability of lipopolysaccharide to cause oxidation of heme-bound iron in hemoglobin.
OBJECTIVE: To determine whether cell-free hemoglobin augments the inflammatory cascade, as detected by production of tumor necrosis factor (TNF) elicited by bacterial endotoxin (lipopolysaccharide [LPS]). DESIGN: In vivo and ex vivo study, using a mouse model of sepsis. SETTING: Animal research facility SUBJECTS: Female Swiss Webster mice. INTERVENTIONS: For the in vivo experiments, an LD50 dose (500 microg) of Escherichia coli LPS was injected intraperitoneally into mice. Cell-free crosslinked hemoglobin (60 mg/mouse) or saline was administered intravenously 10 hrs before or coincident with LPS. For the ex vivo experiments, hemoglobin (60 mg/mouse) or saline was administered intravenously to mice, and, 10 hrs later, hepatic Kupffer cells, peripheral blood mononuclear cells, or peritoneal macrophages were isolated. MEASUREMENTS AND MAIN RESULTS: Intravenous infusion of hemoglobin either 10 hrs before or coincident with intraperitoneal LPS resulted in a peak of plasma TNF that was greater than in control mice administered LPS only. Cultured Kupffer cells, isolated from mice that had received hemoglobin in vivo 10 hrs before cell collection, produced more TNF in response to LPS in vitro than cells from normal mice. A trend toward greater TNF production in vitro by peripheral blood mononuclear cells obtained from hemoglobin-treated mice also was observed. Enhanced sensitivity to LPS was not observed with cultured peritoneal macrophages from mice that had received hemoglobin. CONCLUSIONS: Intravenous hemoglobin increased the sensitivity of hepatic macrophages to subsequent stimulation by LPS. This effect may contribute to the increased mortality that we have observed in animals that have received both LPS and hemoglobin.
Limulus antilipopolysaccharide factor (LALF) can neutralize bacterial endotoxin, but its ability to prevent mortality following prolonged endotoxemia is unknown. Mice were challenged with an LD50 dose of intraperitoneal E. coli lipopolysaccharide (LPS) and then received LALF at various times after administration of LPS. Survival at 72 h was significantly improved by the administration of LALF at 4, 10, and even 24 h after LPS (73%, 78%, and 65% survival, respectively, vs. 15% survival in controls). Following intravenous administration of LALF at either 10 or 24 h after LPS, plasma levels of biologically active LPS abruptly fell (> 1000-fold lower than pre-LALF levels). Plasma LALF concentrations fell much more gradually in LPS-treated mice (t1/2 = 120 min) than in control mice (t1/2 = 2.5 min). In conclusion, LALF markedly decreased plasma concentrations of biologically active LPS and protected mice from lethality even when LALF was not administered until long after the onset of continuous endotoxemia.
Cell-free hemoglobin (Hb) is being developed as an erythrocyte substitute. We have previously demonstrated that cell-free Hb is an endotoxin-binding protein which disaggregates endotoxin and subsequently increases the biological activity of endotoxin in several in vitro assays. Because much of the morbidity and mortality associated with gram-negative bacterial infection is the result of pathophysiologic responses to bacterial lipopolysaccharide (LPS; endotoxin), we studied the effect of Hb on LPS-mediated mortality. Hb infused intravenously into mice before, coincident with, or after intraperitoneal LPS injection substantially increased LPS-related mortality from <5% to 50 to 70% 24 h after administration of LPS and from 50% to 60 to 90% at 48 h. Enhanced mortality was observed over a range of doses of injected LPS. At a given LPS dose, enhancement of mortality was shown to be dependent on the dose of Hb administered. Unmodified native human Hb, alpha-alpha-cross-linked human Hb, and beta-beta-cross-linked human or bovine Hb all were shown to enhance LPS-mediated mortality. Depressed reticuloendothelial cell function may have contributed to the enhanced mortality from LPS in the presence of Hb. Therefore, Hb-based blood substitutes, which are currently undergoing clinical trials, may intensify the potentially fatal effects of the sepsis syndrome in patients with trauma, infection, or hypotension who receive Hb for erythrocyte replacement.
(1-->3)-beta-D-glucan is a ubiquitous constituent of fungi, and elevated plasma glucan levels are commonly present in patients with deep mycosis or fungemia. The pharmacokinetics, biologic effects, and distribution in blood and organs of iodine 125-labeled (1--> 3)-beta-D-glucan purified from Candida albicans organisms were analyzed in rabbits during the 24-hour period after intravenous administration of this constituent. The intravascular half-life of beta-glucan was 1.8 minutes in the low-dose group (9.3 micrograms/kg, n = 3) and 1.4 minutes in the high-dose group (222 micrograms/kg, n = 3), and the total body clearance was 1.12 +/- 0.30 ml/min and 1.17 +/- 0.16 ml/min (mean +/- SD), respectively (not significantly different). The serum concentration of (1-->3)-beta-D-glucan was also biologically determined by a test using coagulation factor G of the Japanese horse-shoe crab (G test). There was good correlation between the clearance of beta-glucan measured biologically and isotopically. During the 24-hour period of observation the rabbits remained well and beta-glucan failed to alter blood cell counts, tumor necrosis factor levels, or lipid metabolism. 125I-labeled beta-glucan associated with the blood cellular compartment initially was less than 3% (the majority in the platelets) and decreased further during the following 2-hour period. Over 97% of circulating 125I-labeled beta-glucan was associated with the cell-free plasma, and the majority of this glucan in plasma appeared not to be associated with lipoproteins. The liver contained more than 80% of the 125I-labeled beta-glucan detected in the six major organs analyzed.
Although previous studies have documented a wide variety of derangements in laboratory measurements of blood coagulation and platelets during cardiopulmonary bypass, limited data are available concerning the magnitude of these changes and any association with excessive bleeding. To determine whether abnormalities in commonly available laboratory tests for the evaluation of coagulation, fibrinolysis and hemostasis correlate with postoperative blood loss and transfusion requirements as measures of clinical outcome, 47 consecutive patients undergoing coronary artery bypass grafting with hypothermic cardiopulmonary bypass (CPB) were studied prospectively at 12 time points before, during, and following CPB. Routine blood coagulation tests, coagulation factor levels (fibrinogen, V, VII, VIII, and IX) and fibrinolysis (FDP) became abnormal within 15 minutes after patients were placed on CPB, remained abnormal for the duration of CPB, and recovered at varying rates after discontinuation of CPB. Mean factor V levels declined by the greatest percentage, to 15% of normal, followed by factor VIII which decreased to 30%. Platelet counts declined to below 100 x 10(9)/L after the initiation of CPB and remained low in the postoperative period. Twenty-eight percent of patients had mediastinal output > or = 100 mL per hour during the immediate postoperative period, and were considered to be "bleeders." There were no clinically relevant differences in any of the laboratory measurements between patients with normal postoperative blood loss and those defined as bleeders. Thus, the absence of significant correlations between various laboratory measurements of hemostasis and actual postoperative bleeding indicates that these laboratory derangements are transient, are not predictive of clinically important hemostatic abnormalities, and should not be used in isolation to guide the use of blood components in these patients. Furthermore, although bleeders received more blood components, there was surprisingly little effect on the coagulation factor levels measured.
Human endothelial cells, when incubated with bacterial endotoxin (lipopolysaccharide, LPS), modify their surface in association with prominent production of procoagulant tissue factor (TF) activity. This deleterious biological effect of LPS has been shown previously to be enhanced approximately 10-fold by the presence of hemoglobin (Hb), a recently recognized LPS binding protein that causes disaggregation of LPS and increases the biological activity of LPS in a number of in vitro assays. The present study was performed to test the hypothesis that Hb enhances the LPS-induced procoagulant activity of human umbilical vein endothelial cells (HUVEC) by increasing LPS binding to the cells. The binding of 3H-LPS to HUVEC was determined in the absence or presence of Hb or two other known LPS-binding proteins, human serum albumin (HSA) and IgG. LPS binding was substantially increased in the presence of Hb, in a Hb concentration-dependent manner, but was not increased by HSA or IgG. Hb enhancement of LPS binding was observed in serum-free medium, indicating that there was no additional requirement for any of the serum factors known to participate in the interaction of LPS with cells (e.g., lipopolysaccharide (LPS)-binding protein (LBP) and soluble CD 14 (sCD14)). Hb enhancement of LPS binding also was observed in the more physiologic condition of 100% plasma. LPS-induced TF activity was stimulated by Hb, but not by HSA or IgG. In serum-free medium, TF activity was not stimulated under any of the conditions tested. Ultrafiltration of LPS was dramatically increased after incubation with Hb but not with HSA or IgG, suggesting that LPS disaggregation by Hb was responsible for the enhanced binding of LPS to HUVEC and the subsequent stimulation of TF activity.
Previous investigations have demonstrated that hemoglobin (Hb) is a binding protein for bacterial endotoxin (lipopolysaccharide, LPS) and that the structure and biological activity of LPS are altered in the presence of Hb. In the present study, the influence of LPS on the structure of native human HbA0 and covalently cross-linked Hb (alpha alpha Hb) was studied by analyzing the absorption and circular dichroic spectra of Hb in the wavelength region of 200-650 nm. Incubation of oxyHb with each of several LPSs resulted in a decrease in the intensity of the major Soret band at 414 nm with a shift in the maximum peak to 410 nm, decreases in the intensities of the major visible region peaks at 541 and 577 nm, and the appearance of increased absorbance in the visible region in the range of 630 nm. The resultant spectra are characteristic of methemoglobin formation. These spectral changes were time-dependent and LPS-concentration-dependent. Production of methemoglobin was prominent with chemically modified, partially deacetylated rough LPS, and was observed to a lesser extent both with native, complete rough and with native smooth LPSs. The influence of LPS on the absorption spectrum of methemoglobin also was directly tested. The conversion of methemoglobin to hemichrome in the presence of LPS was demonstrated and was shown to be reversible. Analysis of circular dichroic spectra of Hb demonstrated LPS-induced spectral changes in the visible and Soret regions consistent with the production of a substantial quantity of metHb, but did not demonstrate any alteration in the far-UV region (210-240 nm). Moreover, Hb oxygen affinity was only slightly altered after incubation with any of several LPSs. In conclusion, analyses of absorption and circular dichroic spectra reveal the potential of LPS to produce a facilitated oxidation of both alpha alpha-cross-linked human Hb and native human HbA0, without substantial changes in the secondary structure of the globin.
Cell-free hemoglobin (Hb) and bacterial endotoxin (LPS) synergistically produce toxicity. To elucidate possible mechanisms, three groups of rabbits received LPS alone, LPS plus human serum albumin (HSA), or LPS plus Hb (Hb group). The intravascular retention of injected iodine 125-labeled LPS during the 30-minute period analyzed was significantly longer in the Hb group than in the LPS alone or HSA groups (p = 0.0007 and p = 0.03, respectively), especially during the initial 10 minutes. The intravascular half-life of LPS in the LPS control, HSA, and Hb groups was 2.8, 4.0, and 4.9 minutes, respectively; the area under the curve was 1369 +/- 483, 1594 +/- 360, and 1731 +/- 481, respectively (ng/ml x minutes, mean +/- SD); and the total body clearance was 24.7 +/- 9.2, 20.1 +/- 5.4, and 18.9 +/- 6.0 (ml/min, mean +/- SD), respectively. The proportion of LPS associated with blood cells was very small at the initial 1-minute time period, and this decreased even further during the 30-minute period analyzed (p = 0.0001). Over 96% of injected LPS was associated with the cell-free plasma, with 51% to 54% of LPS in the apoprotein fraction at the initial time point and 35% to 37% in the HDL fraction. The proportion of LPS increased significantly in the HDL fraction and decreased significantly in apoproteins during the 30-minute period analyzed (p = 0.0006 and p = 0.002, respectively). However, there were no differences between the three groups. The liver was the main distribution site (74%) of injected LPS among the six organs evaluated. In the Hb group the accumulation of 125I-labeled LPS in the spleen was significantly lower than that in the HSA group (p = 0.05). The synergism of the in vivo toxicity reported for LPS and Hb may be due, in part, to the decreased rate of intravascular clearance of endotoxin.
Cell-free hemoglobin (Hb) is a purified preparation of human hemoglobin that is being developed as a resuscitation fluid. In vivo administration of hemoglobin has resulted in significant toxicity, due in part to contamination with bacterial endotoxin (lipopolysaccharide (LPS)). To better understand this toxicity, we have studied the interaction between Hb and LPS. Mixtures of each of three different Hb preparations (cross-linked alpha alpha Hb, cross-linked carbon monoxy-alpha alpha HbCO, and non-cross-linked (native) HbAo) and LPS (Escherichia coli O26:B6 or Proteus mirabilis S1959) were examined by several independent methods for evidence of Hb.LPS complex formation. Binding assays in microtiter plates demonstrated saturable binding of LPS to immobilized Hb, with a kD of 3.1 x 10(-8) M. Binding of LPS to Hb also was demonstrated wiht a radiolabeled LPS photoaffinity probe. Ultrafiltration of Hb/LPS mixtures by 300- and 100-kDa cut-off membranes showed that the majority of LPS in these mixtures (87-97 and 64-72%, respectively) was detected in the filtrates, in contrast to the lack of filterability of LPS in the absence of Hb. Density centrifugation demonstrated that LPS co-migrated with each of the three Hbs, whereas unbound LPS had a distinctly greater sedimentation velocity than Hb or Hb.LPS complexes. Nondenaturing polyacrylamide gel electrophoresis demonstrated that in the presence of Hb, LPS migrated into the gel and co-electrophoresed with Hb, whereas LPS alone did not appreciably enter the gel. Finally, precipitation by ethanol of each of the three Hb preparations was increased in the presence of LPS compared with precipitation in the absence of LPS. Interaction of LPS with each of the three Hb preparations was also associated with altered biological activity of LPS, as shown by enhancement of LPS activation of Limulus amebocyte lysate. Therefore, our data provide several lines of independent evidence for Hb-LPS complex formation and indicated that LPS exhibited altered physical characteristics and enhanced biological activity in the presence of Hb.
Human endothelial cells respond to bacterial endotoxin (lipopolysaccharide [LPS]) with changes that transform the endothelium into a surface with prominent procoagulant properties. Production of tissue factor (TF) in response to LPS is a major alteration that favors coagulation. Biologic activities of LPS have previously been shown to be enhanced by the presence of hemoglobin. Therefore, the ability of human hemoglobin (Hb) to modulate TF production by cultured human umbilical vein endothelial cells (HUVEC) was investigated. Cell-free Hb (10 mg/mL), either purified native (HbAo) or chemically cross-linked (alpha alpha Hb), was incubated with LPS (0.1 microgram/mL), and the mixtures then were added to HUVEC in culture. TF activity was quantified with a clotting assay and TF protein was measured with an enzyme-linked immunosorbent assay. Hb preparations greatly enhanced the production of TF activity (11- to 25-fold greater than TF produced by HUVEC alone) compared with minimal TF activity generated by LPS alone (only twofold greater than HUVEC alone). The enhancement of LPS-induced TF activity was Hb concentration-dependent over a range of 1 to 100 mg/mL. Cross-linked alpha alpha Hb also greatly enhanced the production of TF protein compared with TF protein generated by LPS alone (12-fold greater v 3.5-fold greater than HUVEC alone, respectively). The enhancement of LPS-induced TF protein was Hb concentration-dependent over a range of 0.1 to 2 mg/mL. Enhancement of TF activity by Hb required new protein synthesis. These results show that human Hb can augment the ability of LPS to induce endothelial cell TF and suggest that hemolysis associated with disseminated intravascular coagulation during sepsis may further stimulate coagulation. In addition, these results suggest a potential mechanism for generalized thrombosis in animals that has been associated with the infusion of cell-free Hb for resuscitation.
Explore the source record for details and available documents.
Administration of alpha alpha-crosslinked stroma-free hemoglobin (SFH) as a cell-free resuscitation fluid is associated with multiple organ toxicities. Many of these toxicities are characteristic of the pathophysiological effects of bacterial endotoxins (lipopolysaccharide, LPS). To better understand the potential role of LPS in the observed in vivo toxicities of SFH, we examined mixtures of SFH and E. coli LPS for evidence of LPS-SFH complex formation. LPS-SFH complexes were demonstrated by three techniques: ultrafiltration through 300 kDa cut-off membranes, which distinguished LPS in complexes (87-89% < 300 kDa) from LPS alone (90% > 300 kDa); density centrifugation through 5% sucrose, which distinguished denser LPS alone from LPS-SFH complexes; and precipitation by 67% ethanol, which demonstrated 2-3 fold increased precipitability of complexes compared to SFH alone. Interaction of LPS with SFH was also associated with markedly increased biological activity of LPS, as manifested by enhancement of LPS activation of Limulus amebocyte lysate (LAL), increased release of human mononuclear cell tissue factor, and enhanced production of cultured human endothelial cell tissue factor. These results demonstrated that hemoglobin can serve as an endotoxin binding protein, and that this interaction results in the alteration of several LPS physical characteristics and enhancement of LPS biological activities.
Administration of purified hemoglobin (Hb) as a cell-free resuscitation fluid is associated with multiple organ toxicities. Many of these toxicities are characteristic of the pathophysiological effects of bacterial endotoxins (lipopolysaccharide, LPS). To better understand the potential role of LPS in the observed in vivo toxicities of Hb, we examined mixtures of Hb and LPS for evidence of LPS-Hb complex formation. LPS-Hb complexes were demonstrated by three techniques: ultrafiltration through 300 kDa cut-off membranes, which distinguished LPS in complexes (87-89% < 300 kDa) from LPS alone (90% > 300 kDa); density centrifugation through sucrose, which distinguished denser LPS alone from LPS-Hb complexes; and precipitation by 67% ethanol, which demonstrated 2-3 fold increased precipitability of Hb in complexes compared to Hb alone. Interaction of LPS with Hb was also associated with markedly increased biological activity of LPS, as manifested by enhancement of LPS activation of Limulus amebocyte lysate (LAL), increased release of human mononuclear cell tissue factor, and enhanced production of human endothelial cell tissue factor. These results demonstrated that hemoglobin can serve as an endotoxin binding protein, and that this interaction results in the alteration of several of the physical characteristics of LPS and enhancement of the biological activities of LPS. These findings suggest that a mechanism for the toxicity of infused Hb in vivo may involve potentiation of the biological effects of LPS. In addition, these observations suggest a mechanism by which LPS-related morbidity during sepsis could be enhanced by erythrocyte hemolysis.
Limulus amebocyte lysate, obtained from horseshoe crab (Limulus polyphemus) blood cells, contains a coagulation system which is activated by bacterial lipopolysaccharide (LPS). A chromatographic fraction of Limulus lysate, containing the endotoxin-sensitive factor(s) which initiates the coagulation cascade, was studied. We utilized a photoreactive, cleavable, radiolabeled derivative of Salmonella minnesota LPS, LPS-(p-azidosalicylamido)-1,3'-dithiopropionamide (LPS-ASD), to identify LPS-binding proteins. The lysate fraction was incubated with LPS-ASD, and LPS-binding proteins were identified by autoradiography of sodium dodecyl sulfate-polyacrylamide gels. An 82-kDa protein, a major protein component of this fraction from Limulus lysate, was identified as a LPS-binding protein in a majority of lysates. Incubation of whole Limulus lysate with antiserum to this protein resulted in enhanced sensitivity of the lysate to LPS, suggesting that this 82-kDa protein is a negative regulator of coagulation. A minor 50-kDa protein component of lysate also was identified as a LPS-binding protein and is a candidate for the LPS-sensitive coagulation protein in L. polyphemus.
Bacterial endotoxin (lipopolysaccharide [LPS]) is known to interact with numerous components of blood, including erythrocytes, mononuclear cells, platelets, neutrophils, lipoproteins, and plasma proteins. The relative affinities of LPS for these elements, and the distribution of LPS between them, are unknown. Cross-linked stroma-free hemoglobin (SFH), a potential substitute for erythrocyte transfusion, produces in vivo toxicity in animals consistent with significant LPS contamination. Therefore, we studied the distribution of LPS in human and rabbit blood and examined whether the presence of SFH altered LPS distribution. In either the presence or absence of SFH, LPS was associated predominantly with high-density lipoproteins and apoproteins. There was lesser binding to low- and very-low-density lipoproteins. Examination of the apoprotein pool by column chromatography and density centrifugation demonstrated that LPS in this fraction was predominantly protein bound. Binding of LPS to SFH resulted in dissociation of a portion of the LPS into low-molecular-weight complexes. Cell-bound LPS was only 2 to 16% of the total and was unaffected by SFH. The distribution among blood cells demonstrated predominant binding to platelets in human blood but predominant binding to erythrocytes in rabbit blood. Cellular distribution was not significantly altered by SFH.
Horseshoe crabs (Limulus polyphemus and Tachypleus tridentatus) possess a proteolytic blood coagulation system within their amebocytes that, after release and endotoxin activation, generates a polymerized insoluble coagulin clot. Clotting enzyme from horseshoe crab amebocyte lysate is the protease that activates the clottable protein (coagulogen) which then forms the coagulin clot. Comparison of the previously published descriptions of this enzyme has revealed significantly discordant biochemical characteristics. We purified a 60-kDa proclotting enzyme from L. polyphemus amebocyte lysate to a single band on polyacrylamide gel electrophoresis. After electrophoresis, evaluation of enzymatic activity of this protein within gels demonstrated that the band of purified protein corresponded to enzymatic activity, as detected by amidolytic activity for chromogenic substrates and by gelation of coagulogen applied to the gel. The enzymatic activity was inhibited by serine protease inhibitors. The purified proclotting enzyme had a molecular weight and amino acid composition different from the previously published characterizations of proclotting enzymes from both L. polyphemus and T. tridentatus.