Synergistic toxicity of gram-negative bacteria and free colloidal hemoglobin

Potentiation of hydrogen peroxide toxicity: From catalase inhibition to stable DNA-iron complexes

Hydrogen peroxide (H₂O₂) is unique among general toxins, because it is stable in abiotic environments at ambient temperature and neutral pH, yet rapidly kills any type of cells by producing highly-reactive hydroxyl radicals. This life-specific reactivity follows the distribution of soluble iron, Fe(II) (which combines with H₂O₂ to form the famous Fenton's reagent),Fe(II) is concentrated inside cells, but is virtually absent outside them. Because of the immediate danger of H₂O₂, all cells have powerful H₂O₂ scavengers, the equally famous catalases, which enable cells to survive thousand-fold higher concentrations of H₂O₂ and, in combination with adequate movement of H₂O₂ across membranes, make the killing H₂O₂ concentrations virtually impractical to generate in vivo. And yet, low concentrations of H₂O₂ are somehow used as an efficient biological weapon. Here we review several examples of how cells potentiate H₂O₂ toxicity with other chemicals. At first, these potentiators were thought to simply inhibit catalases, but recent findings with cyanide suggest that potentiators mostly promote the other side of Fenton's reaction, recruiting iron from cell depots into stable DNA-iron complexes that, in the presence of elevated H₂O₂, efficiently break duplex DNA, pulverizing the chromosome. This multifaceted potentiation of H₂O₂ toxicity results in robust and efficient killing.

Globin attenuates the innate immune response to endotoxin

Hemoglobin is an endotoxin (lipopolysaccharide; LPS)-binding protein that synergistically increases the release of proinflammatory cytokines from the innate immune system in response to LPS. It has been suggested that this activity of hemoglobin facilitates the recognition of Gram-negative bacteria in a wound, thereby maximizing immune efficiency. This synergy may be important to the pathogenesis of a broad spectrum of clinical conditions because elevated hemoglobin levels frequently are observed in patients after the transfusion of red cells, trauma, cardiopulmonary bypass surgery, hemolysis, in addition to other disorders. To determine the molecular basis of the specific hemoglobin-LPS synergy, in this article we tested the effects of globin itself on macrophage responses to LPS. Paradoxically, these studies revealed that globin suppressed tumor necrosis factor (TNF) synthesis in LPS-stimulated murine and human macrophage cultures. LPS comigrated with globin on non-denaturing electrophoretic gels, giving direct evidence for binding. Globin specifically inhibited LPS activity in the standard Limulus assay but did not inhibit interleukin-1beta-mediated TNF synthesis. Iron supplementation of macrophage cultures significantly increased interleukin-1beta-induced TNF release. Intraperitoneal administration of globin protected mice against both LPS-induced lethality and experimentally induced bacterial infection. Thus, the heme-iron moiety of hemoglobin, and not the binding of LPS to globin, enhanced macrophage responses to LPS.

Hemoglobin infusion augments the tumor necrosis factor response to bacterial endotoxin (lipopolysaccharide) in mice

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.

Hemoglobin increases mortality from bacterial endotoxin

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.

Safety evaluation of a polymerized hemoglobin solution in a murine infection model

Several investigators have observed that free hemoglobin may increase the mortality rate in experimental Escherichia coli peritonitis in animals. This effect is probably mediated by the heme moiety of hemoglobin, but the mechanism remains controversial. Free hemoglobin might impair neutrophil function, and it might serve as a source of iron, which is necessary for bacterial replication. Several modified hemoglobin solutions, developed as blood substitutes, are currently being tested in clinical studies, but concern exists that these solutions may have the potential to exacerbate a bacterial infection. At the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, a blood substitute based on modified hemoglobin (PolyHbXl) has been developed that has improved oxygen affinity and prolonged vascular retention. In the present study the potential risk of this solution on the promotion of infections has been evaluated. PolyHbXl was intravenously injected into mice in a clinically relevant dose of 1.5 gm/kg body weight 1 hour before intravenous administration of a sublethal number of Listeria monocytogenes, Salmonella typhimurium, E. coli, or Candida albicans organisms. PolyHbXl did not promote the proliferation of any of these microorganisms in the liver and spleen, nor did it lead to an increased mortality rate in the mice. Also, the in vitro proliferation of L. monocytogenes, S. typhimurium, and E. coli was not increased by PolyHbXl. In conclusion, PolyHbXl does not affect the course of infection with various microorganisms in mice, and no indication was found that this new blood substitute compromises the host defense against infections.

Effects of bacterial endotoxin on human cross-linked and native hemoglobins

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.

Current status of injectable oxygen carriers

In this review the current status of what commonly are termed "blood substitutes" is discussed. The term blood substitute is a misnomer because the formulations under development at this time transport respiratory gases but do not perform the metabolic, regulatory, and protective functions of blood. Either hemoglobin or a perfluorochemical form the base to transport oxygen; the advantages and disadvantages of each base are discussed. The availability of a blood substitute in the U.S. will require approval by the Food and Drug Administration (FDA) and, by law, both its efficacy and safety must be demonstrated prior to approval. Showing efficacy of any blood substitute is complicated by the oxygen reserve and the compensatory mechanisms to acute blood loss in man. The challenge is to prove that the administration of these formulations offer clinical advantages compared with replacement of volume alone. Several efficacy models, the most attractive among them being perioperative hemodilution, should provide data that would bring these formulations into clinical practice. When hemoglobin is not within the favorable environment of the red cell, whether the hemoglobin is derived from expression vectors developed through recombinant biotechnology or from lysed human red cells, it acquires a left-shifted oxygen disassociation curve. Further, because the tetramer disassociates when injected intravenously and the resulting dimers are cleared rapidly from the circulation by the kidneys, intravascular dwell time is brief. Hemoglobins have been modified chemically and linked intramolecularly, intermolecularly, and to macromolecules to correct these problems. While these manipulations have normalized the p50 and extended the dwell time significantly, some toxicity problems remain unresolved. The binding of nitric oxide to hemoglobin preparations and the presumably resultant systemic and pulmonary hypertension observed in animals may be the most difficult to overcome, although the implications of these reactions in man is poorly understood. Perfluorochemicals (PFC) provide a fundamentally different and simpler approach to oxygen transport than hemoglobin formulations. Typically, the PFCs used are liquids composed of 8 to 10 carbon atoms that dissolve oxygen and obey Henry's law. Thus, the recipient's inspired oxygen and cardiac output assume importance. Because they are insoluble in water, PFCs are administered as emulsions, that is, as small droplets about 0.1 to 0.2 microns in diameter. In this respect, they are very similar to the lipid emulsions widely used for parenteral nutrition. Egg yolk phospholipid and poloxamers are most commonly used as emulsifiers. PFCs are not metabolized and are excreted unchanged by the lungs, following temporary storage by the monocyte-macrophage system (MMS).(ABSTRACT TRUNCATED AT 400 WORDS)

Toxicity of hemoglobin solutions: hemoglobin is a lipopolysaccharide (LPS) binding protein which enhances LPS biological activity

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.