Inhaled Nitric Oxide Attenuates the Adverse Effects of Transfusing Stored Syngeneic Erythrocytes in Mice with Endothelial Dysfunction after Hemorrhagic Shock

• Transfusion of stored (more than 14 days) blood is associated with an increased rate of infection, multiorgan failure, and mortality

• The precise mechanisms involved in these adverse effects remain uncertain

• In a murine model of hemorrhagic shock, endothelial dysfunction (induced by feeding a high-fat diet) exacerbates, whereas breathing nitric oxide ameliorates, the adverse effects of resuscitation with stored erythrocytes

Blood transfusion is a lifesaving treatment for hemorrhagic shock (HS). During ex vivo  storage, erythrocytes undergo numerous biochemical, structural, and functional alterations, which are collectively termed the “storage lesion.”1,2In 2007, the average duration of storage of transfused units was 19.5 days in the United States and 20 ± 11 (mean ± SD) days for trauma patients.3A number of clinical studies have documented that transfusion of blood stored for more than 14 days is associated with an increased rate of infection,4multiorgan failure,5extended length of stay in hospital,6and mortality.7–10In contrast, other clinical studies of selected patient populations, including those undergoing cardiac surgical procedures, trauma victims, and critically ill patients, reported no association between the duration of erythrocyte storage and adverse clinical outcomes.5,11–15 

The precise mechanisms responsible for the adverse clinical effects seen after transfusion of erythrocytes stored for prolonged periods are uncertain. One possibility is that these adverse effects are attributable to release of free hemoglobin from stored erythrocytes, both in solution and as lipid-enclosed microparticles.16Hemoglobin in plasma can scavenge nitric oxide more avidly than erythrocyte-encapsulated hemoglobin17and can cause vasoconstriction,18inflammation, and platelet activation.19Increased plasma hemoglobin levels induce a “nitric oxide deficiency” state, contributing to the complications of various genetic and acquired hemolytic disorders such as sickle cell disease, malaria, and hemolysis-associated smooth muscle dystonia.20 

Our previous studies of infusing hemoglobin-based oxygen carriers or tetrameric hemoglobin indicated that scavenging of endothelium-derived nitric oxide by plasma hemoglobin produced vasoconstriction in both mice and sheep.18We reported that preexisting endothelial dysfunction (a deficiency of vascular nitric oxide availability commonly associated with diabetes and atherosclerosis) dramatically enhanced the susceptibility to hemoglobin-based oxygen carriers or hemoglobin-induced systemic vasoconstriction in mice.21Moreover, we reported that breathing nitric oxide before infusing a hemoglobin-based oxygen carrier or tetrameric hemoglobin in mice prevented the vasoconstriction.18 

Reproducible animal models of erythrocyte storage can provide important insights into the effects of transfusing stored blood in an HS model and aid in the development of methods to prevent stored blood toxicity. Our study had three objectives. First, we measured tissue injury, hemodynamic changes, and survival rate in mice subjected to HS for 90 min and resuscitated with erythrocytes stored for less than 24 h (fresh erythrocytes) or with erythrocytes stored for 2 weeks (stored erythrocytes). Second, we studied the impact of endothelial dysfunction (induced by feeding mice a high-fat diet [HFD-fed] for 4–6 weeks) after resuscitation from HS with fresh or stored erythrocytes. Third, we studied whether nitric oxide inhalation during and for 2 h after resuscitation could prevent or reduce the adverse effects (tissue injury, organ dysfunction, inflammation, oxidative stress, and mortality) of resuscitation with stored erythrocytes in HFD-fed mice. We report that resuscitation with stored erythrocytes induced greater tissue injury, inflammatory response, oxidative stress, mortality, and plasma hemoglobin levels than did resuscitation with fresh erythrocytes. All these adverse effects were exacerbated in HFD-fed mice resuscitated with stored erythrocytes after HS. Inhaled nitric oxide reduced tissue injury, plasma hemoglobin levels, and oxidative stress and improved the survival rate of HFD-fed mice transfused with stored erythrocytes.

Animal studies were approved by the Subcommittee on Research Animal Care at Massachusetts General Hospital, Boston, Massachusetts. Eight- to 10-week–old male C57BL/6J mice were obtained from Jackson Laboratory (Bar Harbor, ME). Mice were fed either a standard diet (10% of calories from fat) or a HFD (60% of calories from fat; Research Diets, Inc., New Brunswick, NJ) for 4–6 weeks. Murine blood was collected and stored as previously described.22 

We studied seven groups of C57BL/6 mice. Sham control groups fed the standard diet or HFD-fed mice (n = 8/group) received anesthesia, mechanical ventilation, and vascular cannulation but were not subjected to hemorrhage or resuscitation. We studied the effects of resuscitation with fresh erythrocytes (n = 16–17) and stored erythrocytes (n = 16–18) after HS in both standard diet– and HFD-fed mice. An additional group of HFD-fed mice (n = 16) was resuscitated with stored erythrocytes and breathed 80 parts per million (ppm) nitric oxide commencing 10 min before blood and fluid resuscitation and ending 2 h after fluid resuscitation was completed. Inhaled nitric oxide was delivered via  the mechanical ventilator, as previously described.18 

At 2 h after resuscitation, mice were extubated, and vascular catheters were removed. Cefazoline (400 mg/kg; Sandoz Inc., Princeton, NJ) was administered intraperitoneally at 2 h after resuscitation and readministered every 24 h. In a separate study, additional mice (5–7/group) were euthanized at 2 h after resuscitation to obtain blood and tissue samples.

Mice were anesthetized with an intraperitoneal injection of ketamine (120 mg/kg) and xylazine (4 mg/kg) and mechanically ventilated at a respiratory rate of 120/min and tidal volume of 10 µl/g with an inspired oxygen fraction of 0.21 (Mini Vent 845; Harvard Apparatus, Holliston, MA). Both femoral arteries were cannulated with heparinized polyethylene-10 tubing (BD Diagnostics, Franklin Lakes, NJ). One arterial line was used for continuously measuring and recording arterial blood pressure (Ponemah Physiology Platform, Valley View, OH). The other arterial line was used for blood withdrawal and fluid infusion. Rectal temperature was maintained between 36.5 and 37°C using a heating pad. Anesthetized mice were bled over 10 min to a mean arterial pressure of 40 mmHg. A stable mean arterial pressure of 40 mmHg was maintained for 80 min by either withdrawing additional blood or infusing lactated Ringer’s solution. This is a model of severe blood loss, approximately 40% of the blood volume is acutely shed (table 1). After 90 min of HS, mice were resuscitated over 20 min by infusing a volume of fresh or stored erythrocytes equal to the shed blood volume with an equal volume of lactated Ringer’s solution. We measured the survival rate up to 7 days after HS and resuscitation.

Blood samples were obtained only from mice killed at 2 h after resuscitation by cardiac puncture. Standard base excess, arterial pH, PaO², PaCO², and HCO³⁻were measured using a blood gas analyzer (ABL 800 FLEX; Radiometer Medical, Brønshøj, Denmark). Lactate levels were measured using a Blood Lactate Measuring Meter (Lactate Plus; Nova Biomedical, Waltham, MA). Hematocrit was measured by centrifuging 70 μl of heparinized blood.

At 2 h after resuscitation, oxygen dissociation curves were determined with a Hemox-Analyzer (TCS Scientific Corp., New Hope, PA). P⁵⁰, defined as the partial pressure of oxygen at which hemoglobin is 50% saturated with oxygen, was calculated from the oxygen dissociation curve. Erythrocyte 2,3-diphosphoglycerate (2,3-DPG) concentration was measured with a kit (Roche Diagnostics, Mannheim, Germany).

Two hours after resuscitation with fresh or stored erythrocytes, plasma hemoglobin levels were determined with a QuantiChrom Hemoglobin Assay Kit (BioAssay Systems, Hayward, CA). Haptoglobin and hemopexin levels were measured with murine haptoglobin and hemopexin enzyme-linked immunosorbent assay kits (Life Diagnostics Inc., West Chester, PA).

Plasma aspartate aminotransferase (AST) and creatine phosphokinase (CPK) concentrations, and blood urea nitrogen (BUN) levels, were measured by spectrophotometric analysis with commercial assay kits (BioAssaySystems, Hayward, CA) at 2 h after resuscitation with either fresh or stored erythrocytes.

As a measure of inflammation, plasma interleukin (IL)-6 level and pulmonary myeloperoxidase activity were determined. IL-6 levels were measured using a Duoset^®murine IL-6 enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, MN). Myeloperoxidase activity was evaluated in lung homogenates, as described previously.23 

As an index of in vivo  oxidative stress,24levels of 4-hydroxynonenal (4-HNE)-histidine protein adducts were measured in tissue homogenates using a commercial enzyme-linked immunosorbent assay kit (Cell Biolaboratories, San Diego, CA). To evaluate oxidative damage to DNA, levels of 8-hydroxy-2-deoxy guanosine (8-OHdG) were measured in DNA extracted from lung, liver, and kidney using a commercial enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI).25Total lung protein levels were measured using bicinchoninic acid protein reagent (Thermo Scientific, Rockford, IL).

Total messenger ribonucleic acid (mRNA) was extracted from murine lungs and livers using Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA). Complementary DNA was synthesized by the reverse transcriptase reaction (MMLV-RT; Invitrogen Life Technologies). Real-time amplification of transcripts was performed by the Synergy Brands method using an Eppendorf Mastercycler Realplex (Eppendorf, Hamburg, Germany). The relative expression of target transcripts was normalized to the levels of 18S ribosomal RNA and analyzed using the relative C^Tmethod. The sequences of primers used are listed in table 2.

Standard-diet– and HFD-fed mice were euthanized with pentobarbital (200 mg/kg, intraperitoneal). A second-order mesenteric artery was dissected free, cleared of surrounding fatty tissue, and mounted onto a glass cannula in a pressure myograph (DMT-USA Inc., Ann Arbor, MI). Vessels were perfused with Krebs solution for 45 min at 60 mmHg. Vessels were constricted with phenylephrine (10⁻⁵M). The vasodilator response to increasing concentrations of acetylcholine (10⁻⁹–10⁻⁵M) was measured and calculated, as previously reported.26 

All values are expressed as mean ± SD. We verified that all variables were normally distributed by Shapiro-Wilk test. Data were analyzed using a one-way analysis of variance with a post hoc  Newman-Keuls test (GraphPad Prism; GraphPad Software Inc., La Jolla, CA). Survival rates were compared with a log-rank test. A two-way analysis of variance with repeated measures was performed to determine the effect of diet on mesenteric luminal dilation over a wide range of concentrations of acetylcholine (fig. 1). Relationships between plasma hemoglobin levels versus  plasma AST, CPK, and BUN levels were analyzed by linear regression correlation. Plasma hemoglobin levels versus  tissue 4-HNE and 8-OHdG levels and plasma hemoglobin levels versus  pulmonary myeloperoxidase activity were also analyzed by linear regression correlation. Probability values were two-tailed, and a P  value less than 0.05 was considered significant.

Previous studies from our laboratory reported that diabetic mice exhibited endothelial dysfunction and were sensitized to the vasoconstrictor effects of infusing tetrameric hemoglobin and stored erythrocytes.21,22However, in pilot studies of db/db mice subjected to HS and resuscitation, none of them survived the 90 min of HS. Previous study showed that wild-type mice consuming an HFD for 9 days developed endothelial dysfunction.27Thus, we chose to study HFD-fed mice to learn whether their host factors would contribute to the adverse effects after transfusing stored erythrocytes. To confirm that mice fed an HFD for 4–6 weeks would develop endothelial dysfunction, we compared the ability of acetylcholine to dilate second-order mesenteric arteries obtained from standard-diet– and HFD-fed mice. We found that HFD-fed mice have a significantly impaired vasodilator response to acetylcholine (P = 0.01 by two-way analysis of variance with repeated measures, fig. 1). These results suggest that HFD-fed mice provide an animal model of endothelial dysfunction.

To learn whether resuscitation with stored erythrocytes after HS would cause injury to liver, kidney, or muscle in mice, we measured plasma levels of AST, BUN, and CPK at 2 h after resuscitation. Baseline levels of plasma AST, BUN, and CPK did not significantly differ between standard-diet– and HFD-fed mice. In standard-diet–fed mice, resuscitation with either fresh or stored erythrocytes increased plasma AST, BUN, and CPK levels compared with standard-diet–fed mice that were not challenged with HS (table 3). AST and BUN levels were greater in standard-diet–fed mice resuscitated with stored erythrocytes than in standard-diet–fed mice resuscitated with fresh erythrocytes. In HFD-fed mice, resuscitation with fresh or stored erythrocytes induced higher levels of AST, BUN, and CPK than in HFD-fed mice that were not subjected to HS. AST, BUN, and CPK levels were much greater in HFD-fed mice resuscitated with stored erythrocytes than in HFD-fed mice resuscitated with fresh erythrocytes. Furthermore, AST, BUN, and CPK levels were greater in HFD-fed mice resuscitated with stored erythrocytes than in standard-diet–fed mice resuscitated with stored erythrocytes (P < 0.05 for each comparison, table 3). In HFD-fed mice resuscitated with stored erythrocytes that breathed nitric oxide, the AST, BUN, and CPK levels were less than in HFD-fed mice resuscitated with stored erythrocytes that breathed air without nitric oxide. Our data suggest that resuscitation with stored erythrocytes induced greater liver, kidney, and muscle injury in both standard diet– and HFD-fed mice than did resuscitation with fresh erythrocytes. Tissue injury associated with resuscitating HS with stored or fresh erythrocytes was exacerbated in HFD-fed mice, and the injury caused by transfusing stored erythrocytes in HFD-fed mice was attenuated by breathing nitric oxide.

Despite the large volume of blood that mice acutely shed to produce severe hypotension and the variation of body weight between groups, there were no significant differences within groups for body weight, estimated blood volume, shed blood volume, and the volume of blood and lactated Ringer’s solution infused during shock and resuscitation (table 1). No significant differences were found in mean arterial pressure between any of the groups before, during, and 2 h after HS and resuscitation (data not shown). No differences in arterial blood gas tensions or arterial pH were noted between standard-diet– and HFD-fed mice at 2 h after resuscitation with either fresh or stored erythrocytes (table 1). However, blood lactate levels at 2 h in mice fed an HFD and resuscitated with stored erythrocytes after HS were greater than those of sham mice, of standard-diet–fed mice resuscitated with either fresh or stored erythrocytes, or of HFD-fed mice transfused with fresh erythrocytes (table 1). The increase of plasma lactate levels measured in HFD-fed mice transfused with stored erythrocytes could be prevented by breathing nitric oxide (table 1).

To investigate the effect of resuscitation with fresh or stored erythrocytes on mortality after HS, we measured the short-term (12 h) and long-term (7 days) survival rates after HS and resuscitation (fig. 2). Short-term and long-term survival rates did not significantly differ in standard-diet–fed mice resuscitated with fresh or stored erythrocytes (fig. 2Aand 2B). However, the short-term survival was greater in HFD-fed mice resuscitated with fresh erythrocytes than in HFD-fed mice resuscitated with stored erythrocytes (P < 0.01, fig. 2C). None of the HFD-fed mice resuscitated with stored erythrocytes after HS survived for 7 days (fig. 2D).

Because nitric oxide attenuated tissue injury, we examined whether breathing 80 ppm nitric oxide could improve the survival rate of HFD-fed mice resuscitated with stored erythrocytes. Nitric oxide inhalation improved the short-term survival rate of HFD-fed mice transfused with stored erythrocytes (80% survival with nitric oxide breathing vs.  30% without nitric oxide, P = 0.014, fig. 2D). However, there was no significant long-term survival benefit produced by breathing nitric oxide for the first 2 h after HFD-fed mice were resuscitated with stored erythrocytes (P = 0.093). Our results suggest that HFD-fed mice are more vulnerable to an early death (within 12 h) after HS and resuscitation with stored erythrocytes, and that the increased early mortality associated with transfusing stored erythrocytes can be prevented by breathing 80 ppm nitric oxide for 2 h.

We have previously reported that levels of P⁵⁰and 2,3-DPG are markedly reduced in murine blood stored for 2 weeks.22To explore whether resuscitation with stored erythrocytes after HS would impair oxygen delivery to the periphery, we measured 2,3-DPG and P⁵⁰levels at 2 h after resuscitation. In HFD-fed mice, erythrocyte 2,3-DPG concentrations were greater after resuscitation with fresh erythrocytes than after resuscitation with stored erythrocytes (4.2 ± 0.3 and 2.4 ± 0.2 mM, respectively). P⁵⁰was greater after transfusion of fresh erythrocytes than after transfusion with stored erythrocytes (44 ± 1 vs.  33 ± 2 mmHg, respectively) after HS. Our results suggest that transfusion of 40% of the blood volume with erythrocytes stored for 2 weeks to resuscitate mice from HS produces markedly decreased 2,3-DPG levels and increased erythrocyte oxygen affinity (lower P⁵⁰).

To investigate whether the inflammatory response to HS is greater after resuscitation with stored erythrocyte than after resuscitation with fresh erythrocyte, plasma IL-6 levels were measured at 2 h after resuscitation. There was no significant difference in IL-6 levels between sham-operated standard-diet–fed mice and sham-operated HFD-fed mice. In standard-diet–fed mice, plasma IL-6 levels were greater in mice resuscitated with fresh erythrocytes than in sham-operated controls. Resuscitation with stored erythrocytes induced a greater increase of plasma IL-6 levels than did resuscitation with fresh erythrocytes (P < 0.01, table 3) in standard diet–fed mice. In HFD-fed mice, resuscitation with either fresh or stored erythrocytes increased IL-6 levels compared with sham-operated HFD-fed mice. IL-6 levels were greater in HFD-fed mice resuscitated with stored erythrocytes than in HFD-fed mice resuscitated with fresh erythrocytes (P < 0.05). In addition, IL-6 levels were greater in HFD-fed mice resuscitated with either fresh or stored erythrocytes than in standard-diet–fed mice resuscitated with fresh or stored erythrocytes, respectively (P < 0.05). Of note, inhaled nitric oxide did not attenuate the increase of IL-6 levels in HFD-fed mice resuscitated with stored erythrocytes. These results suggest that resuscitation from HS with either fresh or stored erythrocytes induces an inflammatory response. Resuscitation with stored erythrocytes induces a greater inflammatory response than does resuscitation with fresh erythrocytes. Furthermore, HFD-fed mice are more sensitive to the inflammatory effects of HS and resuscitation.

Acute pulmonary inflammation is characterized by neutrophil infiltration, which can be quantified by measuring pulmonary myeloperoxidase activity.28To learn whether resuscitation with stored erythrocytes after HS induced an inflammatory response in the lung and whether HFD-fed mice are more sensitive to transfusing stored erythrocytes than are standard-diet–fed mice are, pulmonary myeloperoxidase activity was measured at 2 h after resuscitation (table 3). In standard-diet–fed mice, resuscitation with fresh and stored erythrocytes increased pulmonary myeloperoxidase activity, and the activity of myeloperoxidase was greater in the group resuscitated with stored erythrocytes than in the group resuscitated with fresh erythrocytes. Pulmonary myeloperoxidase activity did not significantly differ in HFD-fed mice subjected to HS and resuscitated with fresh erythrocytes and in sham-operated HFD-fed mice. In contrast, resuscitation with stored erythrocytes increased pulmonary myeloperoxidase activity in HFD-fed mice. Furthermore, pulmonary myeloperoxidase activity after resuscitation with stored erythrocytes was greater in HFD-fed mice than in standard-diet–fed mice. Inhaled nitric oxide reduced the increased pulmonary myeloperoxidase levels after resuscitation with stored erythrocytes in HFD-fed mice. These data suggest that resuscitation with stored erythrocytes induced pulmonary inflammation in both standard-diet– and HFD-fed mice, and that HFD-fed mice are more sensitive to transfusing stored erythrocytes than are standard-diet–fed mice.

To evaluate the prothrombotic response after HS and resuscitation, tissue factor (TF) mRNA levels were measured in lung and liver at 2 h after resuscitation (fig. 3). In standard-diet–fed mice subjected to HS, resuscitation with neither fresh nor stored erythrocytes altered TF mRNA levels in lung and liver. In HFD-fed mice, resuscitation with both fresh and stored erythrocytes increased pulmonary TF mRNA levels, but only resuscitation with stored erythrocytes increased hepatic TF mRNA levels. In addition, pulmonary TF mRNA levels were higher in HFD-fed mice resuscitated with fresh or stored erythrocytes than that in standard-diet–fed mice resuscitated with fresh or stored erythrocytes. Hepatic TF mRNA levels were greater in HFD-fed mice receiving stored erythrocytes after HS than levels in standard-diet–fed mice resuscitated with stored erythrocytes. Inhaled nitric oxide did not reduce the increase of pulmonary or hepatic TF mRNA levels in HFD-fed mice resuscitated with stored erythrocytes. These data suggest that resuscitation with stored erythrocytes after HS induces a prothrombotic response in HFD-fed mice.

To test whether resuscitation from HS with fresh or stored erythrocytes induces oxidative lipid peroxidation and DNA damage, the levels of 4-HNE histidine protein adducts and 8-OHdG were measured in lung, liver, and kidney. In standard-diet–fed mice, resuscitation with either fresh or stored erythrocytes did not increase tissue levels of 4-HNE adducts or 8-OHdG (table 3). In HFD-fed mice, resuscitation with fresh erythrocytes did not alter the tissue levels of 4-HNE adducts or 8-OHdG. In contrast, resuscitation of HFD-fed mice with stored erythrocytes increased the levels of both 4-HNE adducts and 8-OHdG in the lung, liver, and kidney. It is noteworthy that tissue 4-HNE and 8-OHdG levels were greater in HFD-fed mice resuscitated with stored erythrocytes than the levels in standard-diet–fed mice resuscitated with stored erythrocytes. Breathing nitric oxide prevented the increase of the levels of 4-HNE adducts and 8-OHdG induced by transfusing stored erythrocytes after HS in HFD-fed mice. These data indicate that HFD-fed mice are more vulnerable to oxidative damage, including lipid peroxidation and oxidative DNA damage, induced by resuscitation with stored erythrocytes than are standard-diet–fed mice and that breathing nitric oxide can prevent this oxidative stress.

Previously, we and others have reported that murine erythrocytes stored for 2 weeks undergo hemolysis during storage and after transfusion.22,29To assess alterations of hematologic variables after HS and resuscitation, we measured plasma hemoglobin levels at 2 h after resuscitation with fresh or stored erythrocytes. Plasma hemoglobin levels did not differ in sham-operated standard-diet–fed mice and sham-operated HFD-fed mice. Plasma hemoglobin levels were greater in standard-diet–fed mice subjected to HS and resuscitated with fresh erythrocytes than in sham-operated mice. Plasma hemoglobin levels were greater in standard-diet–fed mice resuscitated with stored erythrocytes than in similar mice resuscitated with fresh erythrocytes (P < 0.01, table 3). In HFD-fed mice, resuscitation with fresh erythrocytes did not significantly increase plasma hemoglobin levels above those measured in sham-operated mice. Plasma hemoglobin levels were greater in HFD-fed mice resuscitated with stored erythrocytes than in HFD-fed mice resuscitated with fresh erythrocytes (P < 0.01). Furthermore, plasma hemoglobin levels were much greater in HFD-fed mice receiving stored erythrocytes than the levels in standard-diet–fed mice resuscitated with stored erythrocytes. To study the effect of breathing nitric oxide on hemolysis of erythrocytes after transfusion, we examined whether nitric oxide breathing would attenuate the increase of plasma hemoglobin levels in HFD-fed mice resuscitated with stored erythrocytes. We found that inhaled nitric oxide reduced the increase of plasma hemoglobin levels in HFD-fed mice resuscitated with stored erythrocytes (P < 0.01, table 3). In addition, plasma hemoglobin levels were correlated with many indices of tissue injury in many organs (e.g. , plasma BUN R  ²= 0.3, P < 0.0001), pulmonary myeloperoxidase activity (R  ²= 0.4, P < 0.0001), and pulmonary and hepatic oxidative stress as measured by 4-HNE (R  ²= 0.3, P = 0.0008) and 8-OHdG (R  ²= 0.4, P < 0.0001). These results suggest that increased plasma hemoglobin levels likely contribute to the adverse effects induced by HS and resuscitation with stored erythrocytes.

Mammals produce proteins to neutralize plasma hemoglobin (e.g ., haptoglobin) and heme (e.g ., hemopexin) and, thereby, prevent inflammatory damage and vasoconstriction.30,31There was no significant difference in plasma hemopexin levels between sham-operated standard diet–fed mice and sham-operated HFD-fed mice. In contrast, greater baseline levels of haptoglobin were measured in HFD-fed mice compared with the haptoglobin baseline of standard-diet–fed mice. Hemopexin levels were less in standard-diet–fed mice resuscitated with fresh erythrocytes than in sham-operated mice, whereas haptoglobin levels did not significantly differ (table 3). Both haptoglobin and hemopexin levels were less in standard-diet–fed mice resuscitated with stored erythrocytes than in mice resuscitated with fresh erythrocytes or in sham-operated mice (P < 0.01 for both). In HFD-fed mice, resuscitation with either fresh or stored erythrocytes reduced plasma haptoglobin and hemopexin levels compared with sham-operated mice. There were no significant differences of haptoglobin levels between standard-diet– or HFD-fed mice resuscitated with either fresh or stored erythrocytes. Similarly, no significant differences were measured of hemopexin levels in standard diet– or HFD-fed mice resuscitated with fresh or stored erythrocytes. Breathing nitric oxide did not attenuate the decrease of haptoglobin and hemopexin levels after resuscitation with stored erythrocytes. Our data suggest that resuscitation with stored erythrocytes released greater amounts of hemoglobin into the plasma and, thereby, lead to increased plasma hemoglobin and heme levels, which were bound by haptoglobin and hemopexin, respectively.

Complexes of hemoglobin–haptoglobin or heme–hemopexin are transported to monocytes/macrophages and/or hepatic parenchymal cells, where the heme is metabolized by heme oxygenase-1 (HO-1). Our recent study has shown that transfusion with stored erythrocytes increased hepatic HO-1 mRNA levels.22In the current study, we noted that pulmonary and hepatic HO-1 mRNA levels were not different in sham-operated mice between standard-diet– and HFD-fed mice (fig. 4). In standard-diet–fed mice, pulmonary and hepatic HO-1 mRNA levels were higher in mice resuscitated with fresh erythrocytes than in sham-operated mice (P < 0.01). At 2 h after transfusion, pulmonary and hepatic HO-1 mRNA levels were greater in standard-diet–fed mice that were resuscitated with stored erythrocytes than in the sham group, and hepatic HO-1 mRNA levels were much higher in mice resuscitated with stored erythrocytes than in mice resuscitated with fresh erythrocytes. In HFD-fed mice, HO-1 mRNA levels in lung and liver were higher in mice resuscitated with fresh and stored erythrocytes than in sham-operated mice. Resuscitation with stored erythrocytes induced HO-1 gene expression to a greater extent in lung and liver than did resuscitation with fresh erythrocytes. In addition, pulmonary HO-1 gene expression was greater in HFD-fed mice resuscitated with stored erythrocytes than in standard-diet–fed mice resuscitated with stored erythrocytes (P < 0.05). Breathing nitric oxide during resuscitation with stored erythrocytes in HFD-fed mice did not alter the increase of HO-1 mRNA levels in lung and liver. Our results indicate that the increase of pulmonary and hepatic HO-1 mRNA levels may occur via  other pathways than the reduction of nitric oxide bioavailability. These data suggest that hemolysis after resuscitation with stored erythrocytes induced greater levels of pulmonary and hepatic HO-1 gene expression in both standard-diet– and HFD-fed mice.

In this study, we examined resuscitation with either fresh or stored erythrocytes in a murine model of severe HS. We report that resuscitation of HS with stored erythrocytes is worse than resuscitation with fresh erythrocytes. HFD-fed mice (with endothelial dysfunction) are more sensitive to transfusion than are standard-diet–fed mice. Inhaled nitric oxide reduces some, but not all, of the injuries associated with transfusing stored blood into HFD-fed mice. We also report that recipient factors, and the duration of erythrocyte storage, can play major roles modulating the adverse effects of transfusing stored blood.

The toxicity of transfusing stored blood is likely related to the amount of blood transfused and its duration of storage. A recent clinical study reported that the risk of acute kidney injury was in direct proportion to the number of units of erythrocytes transfused in patients undergoing cardiac surgery.32In our murine model of severe HS (90 min at 40 mmHg mean arterial pressure), we transfused approximately 40% of total blood volume with either fresh or stored syngeneic erythrocytes. We report that after HS resuscitation with either fresh or stored erythrocytes induced tissue injury in all of the resuscitated groups. Tissue injury induced by resuscitation with either fresh or stored erythrocytes was greater in HFD-fed mice than in standard-diet–fed mice.

In this study, we measured a decrease of both circulating 2,3-DPG and P⁵⁰levels in HFD-fed mice resuscitated with stored erythrocytes, but there was no change of these variables from normal levels in HFD-fed mice resuscitated with fresh erythrocytes. Reduced P⁵⁰impairs the ability of blood to provide oxygen to tissues. Others have found that transfusing blood stored for 28 days fails to improve intestinal microvascular oxygenation in a rat HS model.33It is conceivable that the tissue injury and early mortality seen in HFD-fed mice resuscitated with stored erythrocytes may be partly attributable to this reduction in P⁵⁰.

Gladwin and Kim-Shapiro34have proposed that several of the adverse events (vasoconstriction, inflammation, and thrombosis) due to stored erythrocyte transfusion are attributable to a reduction of nitric oxide bioavailability, due to an imbalance between nitric oxide destruction and production. When human erythrocytes are stored for 42 days, hemolysis increases extracellular hemoglobin levels to 25 μM.35Recently, Berra et al .36transfused 1 unit of autologous erythrocytes stored for 40 days into healthy volunteers. They reported that plasma hemoglobin levels reached 25 µM at 2 h after transfusion. In the current study, we found that plasma hemoglobin levels were greater in HFD-fed mice subjected to HS and resuscitated with stored erythrocytes than in standard-diet–fed mice resuscitated with stored erythrocytes. Furthermore, the increased levels of extracellular hemoglobin positively correlated with tissue injury, inflammation, and oxidative stress. Our data also suggest this as a possible mechanism, the reduced nitric oxide bioavailability produced by plasma hemoglobin scavenging is likely to contribute to the adverse effects we measured after transfusion of murine erythrocytes stored for prolonged periods.

Haptoglobin and hemopexin are two major protective proteins against the toxicities produced by free hemoglobin and heme, respectively.37,38In this study, consistent with the increased plasma hemoglobin and heme levels after transfusion, we measured a decrease of haptoglobin and hemopexin levels in all groups resuscitated with stored erythrocytes. Haptoglobin and hemopexin may attenuate the toxicity associated with increased plasma levels of hemoglobin or heme. Baek et al .39have recently demonstrated that coinfusion of haptoglobin attenuates the intravascular hemolysis, acute hypertension, vascular injury, and kidney dysfunction induced by transfusing 28-day-old blood into guinea pigs. In addition, HO-1, a stress-inducible protein with potential antiinflammatory effect, can be induced by increased heme levels.40,41We consistently measured an increase of pulmonary and hepatic HO-1 mRNA levels in all groups resuscitated with fresh or stored erythrocytes.

In the current study, we measured an increased plasma IL-6 concentration and pulmonary myeloperoxidase activity in standard-diet–fed mice resuscitated with stored erythrocytes than in mice receiving fresh erythrocytes, with an exacerbated inflammatory response in HFD-fed mice. The increased plasma IL-6 levels measured in HFD-fed mice in the current HS model are consistent with our recent study, where transfusion with stored erythrocytes increased plasma IL-6 levels in HFD-fed mice.22Furthermore, Hod et al .42have demonstrated that transfusion of murine erythrocytes stored for 14 days led to increased plasma proinflammatory cytokine levels. In addition, in HFD-fed mice, we measured increased prothrombotic effects after HS and resuscitation with stored erythrocytes. Together, our data suggest that HS and resuscitation with stored erythrocytes activate the inflammatory and coagulation cascades and contribute to the development of tissue injury and multiorgan dysfunction.

The oxidation of extracellular hemoglobin to methemoglobin and the production of free heme can initiate oxidative damage to lipids, nucleic acids, and proteins,43all of which could contribute to the tissue injury associated with increased plasma hemoglobin levels.44Measurements of 4-HNE modified protein adducts and 8-OHdG are reliable markers of in vivo  oxidative stress.25,45In our study, resuscitating HFD-fed mice with stored erythrocytes produced markedly increased levels of both 4-HNE adducts and 8-OHdG. These data suggest that oxidative stress after resuscitation from HS contributes to the adverse effects after transfusion of stored erythrocytes.

Exposing circulating extracellular hemoglobin (in the absence of methemoglobin reductase outside of erythrocytes) to a high concentration of inhaled nitric oxide rapidly converts it to ferric hemoglobin species, which do not scavenge nitric oxide.46Breathing nitric oxide leads to a rapid accumulation of a variety of nitric oxide metabolites, which are transported to the periphery where they can be converted back to nitric oxide.47Increased tissue levels of nitric oxide may reduce tissue injury, inflammation, thrombosis, oxidative stress, and metabolic acidosis (reduced lactate levels). Furthermore, inhaled nitric oxide interacts with leukocytes and platelets as they transit the lungs impairing their ability to be activated in the periphery. Inhaled nitric oxide reduced plasma hemoglobin levels; this may be due to increased renal excretion of hemoglobin during nitric oxide breathing.48The antiinflammatory effects and vascular dilation function of nitric oxide may have contributed to the improved short-term survival rate of HFD-fed mice resuscitated with stored erythrocytes. However, breathing nitric oxide did not attenuate all the adverse effects. It is conceivable that transfusion with stored erythrocytes may induce inflammation via  other mechanisms, such as induction by lipid or other components of transfused erythrocytes other than extracellular hemoglobin.

Host factors can sensitize individuals to transfusion with blood stored for prolonged periods. Recently, Vlaar et al .49reported that transfusion of erythrocytes stored for 14 days induced mild lung injury in healthy rats, and to a greater extent in lipopolysaccharide-pretreated rats. We have shown that HFD-fed mice exhibit endothelial dysfunction. Deleterious events after transfusion of stored erythrocytes in healthy patients seem to be exacerbated in recipients compromised by preexisting disorders.11If our findings in mice can be extrapolated to patients with endothelial dysfunction, then the complications after stored blood transfusion in patients with HS who suffer from diabetes or atherosclerosis may be similar to the effect of transfusing stored erythrocytes in mice with endothelial dysfunction after HS. Care should be taken when transfusing stored erythrocytes into patients with endothelial dysfunction. Transfusion of “younger” erythrocytes should be considered when they are available, especially for patients in HS who have endothelial dysfunction. Breathing nitric oxide during the transfusion of “older” erythrocytes should be studied as it may prevent some of the deleterious effects of transfusing stored erythrocytes. However, because murine blood differs markedly from human blood (e.g ., shorter erythrocyte half-life, reduced oxygen affinity, lower plasma haptoglobin levels), the conclusions drawn from our murine studies warrant further clinical study in humans.

In conclusion, transfusion of syngeneic murine blood stored for prolonged periods adversely impacts outcome in a mouse model of HS. Endothelial dysfunction sensitizes mice to the adverse effects of transfusing stored erythrocytes. Inhaled nitric oxide can ameliorate many but not all of the adverse effects of resuscitating mice after HS with stored erythrocytes.