Glutaraldehyde-Polymerized Hemerythrin: Evaluation of Performance as an Oxygen Carrier in Hemorrhage Models

Hemoglobin-based oxygen carriers (HBOCs) have been proposed and tested for several decades for the treatment of hemorrhage. We have previously proposed replacing hemoglobin (Hb) in HBOC with the oxygen-carrying protein hemerythrin (Hr), from marine worms, showing that Hr-based derivatives can perform at least as well or even better than Hb-based HBOC in a range of in vitro assays involving oxidative and nitrosative stress as well as in top-up animal models, where small amounts of Hr- or Hb-HBOC were injected into rats. Here, these experiments are extended to a hemorrhage experiment, in which Hr polymerized with glutaraldehyde, alone or conjugated with human serum albumin, is administered after a loss of 20–30% blood volume. The performance of these preparations is compared with that of Hb-based HBOC measured under the same conditions. Polymerized Hr is found to decrease the survival rate and can hence cannot be used as an oxygen carrier in transfusions. On the other hand, an Hr-albumin copolymer restores survival rates to 100% and generally yields biochemical and histological parameters similar to those of glutaraldehyde-polymerized bovine hemoglobin, with the exception of an acid-base imbalance. The latter may be solved by employing an allogeneic albumin as opposed to the human albumin employed in the present study.


Introduction
Hemoglobin-based oxygen carriers (HBOCs) have been advocated for use in transfusions due to their potential ability to transport oxygen, as opposed to classical plasma expanders currently employed in emergency medicine or surgery. Hemoglobins from various sources, derivatized chemically by various means (often (co) polymerization with glutaraldehyde, but also reticulation with other agents, or derivatization with biocompatible polymers/oligomers to increase the apparent molecular volume and thus prevent extravasation), have been proposed to this end. Bovine hemoglobin polymerized with glutaraldehyde is currently approved for limited human use in two countries, while no other HBOC is currently approved for clinical use anywhere else, despite constant progress in producing less reactive and more stable Hb derivatives [1][2][3][4][5][6][7][8][9][10][11][12]. Protein-free approaches (oxygen-encapsulating emulsions, fuorinated hydrocarbons, and heme-based dendrimers) have also been reported [9,[13][14][15].
An alternative we previously proposed for Hb in HBOC has involved hemerythrin (Hr), an oxygen-carrying protein originally extracted from marine worms but also available in recombinant form overexpressed in E. coli [15][16][17]. Te active site of Hr contains a nonheme diiron center which compared to hemoglobin heme ofers several interesting diferences besides the obvious diference in color (Hr is almost colorless compared to Hb, thus not interfering with diagnostic tests that rely on the color of hemoglobin) [16][17][18]. Tus, the O 2 molecule when bound to Hb gains a superoxide-like character, which allows it to react rapidly with nitric oxide (since both superoxide and nitric oxide are free radicals) [16][17][18]; this reaction can lead to serious side efects in terms of arterial tension in HBOC-based transfusions [9,[19][20][21]. By contrast, O 2 when bound to Hr gains a peroxide-like character and is essentially inert to nitric oxide [22]. Hr is also, for the same reason, inert to hydrogen peroxide, while the reaction of Hb with peroxides is well-characterized to lead stronglyoxidizinghigh-valent iron and free radicals. Hr is also less reactive to nitrite, compared to Hb [22]. We have reasoned that Hr would have an advantage over Hb if used in HBOC, as being more resistant to oxidative and nitrosative stress agents. In cell culture tests with human cells [23,24], as well as in topup animal models (rats injected with low amounts of HBOC, without prior hemorrhage) [25], some Hr-based HBOCs did, in fact, match or exceed the performance of any of the tested Hb-based HBOC.
We have previously reported an evaluation of a range of physiological parameters in rats injected with small amounts of blood substitute candidates based on hemoglobin, [26] and then also in hemorrhage models. Tese candidates were generally based on bovine and ovine hemoglobin, polymerized with glutaraldehyde alone or together with serum albumin or with a peroxidase, meant to alleviate oxidative stress [1,23,24,[27][28][29]. Here, two Hr-based HBOC are tested under conditions identical to previous library of Hbbased HBOC in hemorrhagic Wistar rats: glutaraldehydepolymerized hemerythrin and a hemerythrin copolymer with human serum albumin.

Materials and Methods
Standard reagents and protein derivatives were of the same sources and stocks as previously described for the top-up and hemorrhagic shock experiments with the very same HBOC as used in the present study (see supporting information for a more detailed description) [26,30]. Te Hr concentrations were 150 μM (calculated per monomer, based on the active UV-vis spectrum of the Hr site as previously described [15,22,25]) in the glutaraldehydepolymerized Hr (pHr) and the glutaraldehyde copolymer of Hr and human serum albumin (pHrHSA) preparations used for treating animals.
Healthy adult male Wistar rats weighing 160 ± 20 g, 24 weeks old, were given free access to standard rat food and water. Rats were kept in a light/temperature controlled room with a light/dark cycle of 12/12 h at 22°C. Animal care and procedures were carried out in accordance with Directive 2010/63/EU and national legislation. Te actual project was approved by the Ethical Committee of Babeş-Bolyai University (IRB no. 2012/03.02.2016). At the end the animals were humanely killed by deep anesthesia isofurane (2%), and they were considered dead when no respiratory and heart activity was recorded. Te irreversibility of the phenomena was ensured by axo-atloidian dislocation.
Te experiment was performed on male Wistar rats randomly divided into 8 groups, 10 animals each, as previously described [30]. Te groups were defned as follows: C (control, not subjected to 30% hemorrhage but with blood extracted for biochemical analyses), H (control hemorrhage, subjected to 30% hemorrhage but not to transfusion), P (subjected to 30% hemorrhage then treated with plasma), pHr (30% hemorrhage, transfused with glutaraldehyde-polymerized hemerythrin in PBS), and pHrHSA (30% hemorrhage, transfused with a copolymer of hemerythrin and human serum albumin). According to previous results and in line with procedures performed by Gutierrez et al. [31] as well as Kowalsky and Brandis (2022) [32] or Hooper and Armstrong (2022) [33], the efusion of 30% of total blood volume has induced a Class II hypovolemic shock that stimulated the rheological behavior of the cardiovascular system with unchanged systolic blood pressure (under physiological conditions).
HBOCs were administered intravenously in a proportional volume with blood efusion during hemorrhageá jeun under deep narcosis as previously described [30]. Hemorrhagic status was induced under narcosis, by blood efusion from the retroorbital plexus until the blood volume was 30% of the total blood volume of the rat as previously described [34]. Te animals were monitored every 10 minutes for the frst two hours after intravenous administration of HBOC, and then every two hours for the next 8 hours, and then again at 24 hours. At the end, the animals were subjected to deep isofurane narcosis and blood was collected and analyzed as previously described [26,30]. Te values for the C, H, and P groups have also been reported in [30].
Te results are presented as mean ± standard deviation of the mean (SD). Biochemical data were subjected to ANOVA followed by Tukey's multiple comparison test when comparing all experimental groups. Histopathological score changes for each tissue were analyzed by using the rank-based nonparametric Kruskal-Wallis test with Dunn's test based on rank for multiple comparisons. In the ANOVA test, p < 0.05 was considered statistically signifcant. Tukey's multiple comparison test was considered statistically signifcant at p < 0.05 and was interpreted as follows: * p < 0.05, * * p < 0.01, and * * * p < 0.001. Signifcant diferences after comparisons across groups were indicated as follow: # p < 0.05, ## p < 0.01, and ### p < 0.001. For each analysis, N (no. of rats or samples) was ten. Statistical analyzes were performed using Graph Pad Prism version 5.0 for Windows, Graph Pad Software, San Diego, CA, USA. Table 1 shows the survival rates for these experiments, along with the arterial tension (AT) values at the time points. Te pHr group shows a reduction in survival rate compared to the untreated hemorrhage group (H), down to 50% from 75%. However, pHrHSA shows full recovery to 100% survival. Te AT values immediately after hemorrhage are similar in all three groups. Treatment with pHr or pHrHSA immediately restores 50% of this gap. However, at 24 hours, the pHr group shows a distinct increase in AT compared to the initial values prior to hemorrhage, while in the pHrHSA group, the ATreturns to values similar to the prehemorrhage state. From these points of view, pHrHSA appears to be a reasonable candidate for HBOC, while pHr does not. Figure 1 shows immunological and clotting parameters after transfusion with pHr or pHrHSA, compared to control groups, not subjected to transfusion (H: loss of 30% blood and C: loss of 5% blood). For IgA, pHr shows levels identical to those of the reference H group, while pHrHSA shows levels identical to those of the control group. For IgM, there are no statistically signifcant diferences between the groups/samples. On the other side, similar immunoglobulin variations in IgA and IgM levels were observed after allicin administration as an immunostimulatory agent, where the IgA and IgM concentration was increased (∼45 mg/dL for IgA and ∼50 mg/dL for IgM) in a dose-dependent manner, and these changes were accepted as benefcial in terms of immunostimulation [35,36] IgG variation after pHr exposure demonstrated that pHr was recognized as a potent antigen that induced a signifcant increase in IgG concentration without other immunological imbalances. Tese reactions also reveal the activation of immunocompetent cells after pHr exposure and a prominent humoral immune response, which is consistent with previous observations according to which IgG amplifcation is a physiological reaction after blood transfusion [35]. Te pHr triggered a strong immune response refected in elevated plasma levels of IgG accompanied by elevated levels of the C3 complement fraction. Tis might be a plausible explanation for the high mortality observed in the pHr group, something not seen in   previous top-up models, where much lower amounts of pHr were injected [25]. Interestingly, no such problem is seen for pHrHSA, suggesting that HSA in this preparation is efcient at blocking the immunogenic sites on Hr. Tis again correlates with the fact that the pHrHSA shows a 100% survival rate, improved both over the untreated group (H) as well as over the pHr group. CRP expectedly did not show statistically signifcant change in any of the groups. For PT, no diferences are seen between the hemorrhage group and the treated groups, while for aPTT, both hemerythrin samples show increases over the control and hemorrhage groups. For fbrinogen, pHrHSA shows a distinct drop compared to all other samples. According to some studies [37,38], high concentrations of human serum albumin form an intravascular sequential adsorption surface for immunoglobulin G, as well as fbrinogen, and their interaction with albumin generates fne fbrils without crosslinking, determining the decrease in measured fbrinogen and IgG in blood. Tables 2 and S2 show histological data collected from liver, lung, and kidney tissues. Hemorrhage has previously shown to induce slight proliferation of Kupfer cells in the liver and infammation in the lungs under these conditions, as hypovolemic shock induces congestion through hemodynamic overload. Candidate Hb-based HBOC were also previously shown to generally show worse diagnoses than the hemorrhage group; exceptions were HBOC based on ovine hemoglobin. Te performance of pHr and pHrHSA in Table 2 appears similar in terms of damage level to the performance of glutaraldehyde-polymerized Hb, i.e., HBOC was previously approved for limited human use [30].

Results and Discussion
Supporting information Table S1 shows the distribution of iron deposits in the liver. Both pHr and pHrHSA show behavior similar to the control and hemorrhage groups.

Groups
Liver Lungs  Kidneys  GVD  DG  N  S  K  ED  S  N  GVD  DG  TL  S  M  USD  TD  C  --------------H  ----+  ---------+  P - Furthermore, as shown in supporting information Figure S1, the levels of hematocrit and hemoglobin (measured on samples collected at 24 hours, hence at the same time point as the Table S1) are not statistically diferent across the samples. Te pH of blood, as well as O 2 and CO 2 levels are also not afected ( Figure S2). pHrHSA shows a drop in BE and in bicarbonate compared to all other samples. Tis latter observation is matched by urea and creatinine levels, showing a disruption of the renal function, but also by increased lactate levels (cf. Figures 2 and S3, hence lactic acidosis, accompanied by decreased glucose levels). Te sodium, potassium, and calcium levels show essentially no variations (only slight decreases in sodium and potassium for pHrHSA, cf. Figure S4). Te lactic acidosis seen with pHrHSA may be ascribed to the presence of HAS, as pHr shows no statistically signifcant diferences compared to the control group. Figure 2 shows total protein, transferrin, iron, glucose, and lactate levels. Iron and transferrin levels are slightly afected in Hr samples, but remain at values similar to the control or the untreated hemorrhage group. Te total protein level does not change in a statistically signifcant manner after transfusion with pHr or pHrHSA. Te glucose levels are distinctly lower in the Hr groups and slightly more in pHrHSA compared to pHr; this is mirrored by increase in lactate. Te changes in the pHrHSA group may be linked to the fact, previously commented upon, that serum albumin from other organisms may cause imbalances itself, beyond the efect of the oxygencarrying protein in the HBOC. If so, an Hr copolymer with rat albumin would perform better than Hr-HAS, which was previously observed for Hb-albumin copolymers [30].

Conclusions
Two hemerythrin-basedglutaraldehyde-polymerized HBOCs were analyzed in shock models (30% blood loss, treated by transfusion with polymerized Hr or with a Hr-HAS copolymer). A notable immune response was observed for pHr in terms of IgG (but not IgM or IgA), while no such response was observed for the copolymer pHrHSA. Consistent with this, pHr transfusion led to a decrease in survival rate compared to the untreated hemorrhage group, while pHrHSA restored the survival rate to 100%. Both pHr and pHrHSA induced limited tissular damage to levels similar to those observed for glutaraldehyde-polymerized bovine hemoglobin. Acidosis was detected with both of the Hr-based preparations (especially for pHrHSA), while other physiological parameters remained within normal limits. To conclude, Hr alone cannot form the basis of an efcient HBOC; however, when conjugated with other large molecules (e.g., serum albumin) an acceptable HBOC might be generated if the acid-base problem seen for pHrHSA can be solved.

Data Availability
Supporting information data are available on hematocrit, hemoglobin, acid-base parameters, renal function parameters, blood ion concentrations, iron deposit evaluation. Furthermore, primary data for all tables and fgures in the manuscript are available upon request from the authors.

Conflicts of Interest
Te authors declare that they have no conficts of interest.