Blood transfusions save lives every day and have become an indispensable part of clinical practice in modern medicine. However, there have always been numerous limitations and problems with their use, namely dependence on donor readiness, short shelf life , risk of infection , transfusion reactions in the case of non-blood group-specific transfusion , immunization with the formation of antibodies , and an overall high personnel and financial expense .

For decades, science has been trying to overcome these problems . Although erythrocytes also pose as a CO2 transporter and native hemoglobin as a buffer system , research is being focused on their function as oxygen transporters. The search is on for a laboratory-produced oxygen carrier that is relatively simple and inexpensive to produce, can be stored for a long time and administered universally, is safe to use, and has few to none adverse drug effects.

One approach associated with great hopes are hemoglobin-based oxygen carriers (HBOCs). Initially, however, serious side effects were encountered during development, namely nitrogen monoxide scavenging and associated hypertensive crises, massive renal damage due to tubular reabsorption of hemoglobin (Hb), decay into dimers , and oxidative stress . Various approaches of intra- as well as intermolecular modifications of HBOCs have been devised. Crosslinking of different Hb chains, polymerization of Hb molecules , surface modification , and techniques for encapsulation already brought us somewhat closer towards a safe use of HBOCs . However, a persistent problem is the severely limited retention time of oxygen carriers in the circulatory system. Erythrocytes and the Hb they contain have a mean survival time of about 120 days within the human body . HBOCs, in contrast, have been eliminated within minutes to a few hours in previous experiments . Much of this process appears to occur in the liver . The question of the mechanisms by which HBOCs are sequestered remains partly unclear though.

Possible degradation pathways include haptoglobin (Hp), which, depending on the size and surface properties of HBOCs, could bind its physiological target protein hemoglobin . CD163, the receptor for Hp–Hb complexes also shows some affinity for Hb . The corresponding binding site at the receptor appears to be the same as for the binding of Hp–Hb complexes, according to Schaer et al. ; whereas within Hb, the binding site for direct interaction with CD163 is probably within the β chain of Hb (binding of Hb to Hp via a binding site within the Hb α chain) . Furthermore, not only cells of the monocyte/macrophage lineage appear to be involved in sequestering Hb, but also hepatocytes . It can be speculated that the same could be true for HBOCs. This fits with a case report by Drieghe et al., which suggests that hemopexin (Hpx) may also play a role in the elimination of HBOCs . Hpx was depleted before a change in Hp levels could be observed when catecholamine-requiring patients were treated with an HBOC for intentional NO scavenging and consecutive increase in peripheral vascular resistance. Whether binding between cell surface proteins and Hb/HBOC can occur, and how high the corresponding affinity is, probably depends on modifications made to Hb. Intramolecular crosslinking has an impact, depending on whether the binding site within the Hb α chain is freely accessible (exclusively β-crosslinked Hb vs α-crosslinked Hb) . Intermolecular modifications changing the molecular or polymer size of HBOCs are relevant for protein binding as well. Hemoglobin sub-micron particles (HbMPs) obtained via coprecipitation–crosslinking–dissolution (CCD) are promising as HBOCs. CCD provides particles that are malleable and show a consistent morphology and narrow size distribution, as well as a negative zeta potential . It could be shown that neither NO scavenging nor vasoconstriction can be detected when using HbMPs as oxygen carriers . In addition to transporting oxygen, HbMPs can also be used as drug carriers. However, in a pharmacokinetic study with HbMPs, accumulation of the particles in the sinusoids of the liver, where the Kupffer cells are located, was observed . The mechanisms of the interaction of liver macrophages with HbMPs have not been systematically investigated yet. Since HbMPs are composed of Hb, the elimination via Hp and Hpx seems likely. Hp binds freed Hb, Hpx binds freed heme. The resulting complexes are then bound by the respective receptors, namely CD163 for Hp and CD91 for Hpx, and taken up by phagocytes (e.g., Kupffer cells in liver sinoids), where Hb or heme are subsequently degraded. If this mechanism cannot be bypassed, Hp and Hpx must be fully saturated to achieve and maintain the effect of the HBOCs.

Here is an example calculation with commercially available HBOC 201 (Hemopure®; hemoglobin glutamer-250 (bovine); Hemopure, HbO2 Therapeutics LLC, Souderton PA 18964, USA): A dose of 60 g HBOC 201 in 5 L of blood results in an HBOC concentration of 2.9 × 1019 HBOCs/mL. Hp serum concentrations amount in average to about 1.5 g/L; this corresponds to 1 × 1015 molecules/mL. The average Hpx serum concentration of 0.6 g/L corresponds to 4.5 × 1015 molecules/mL. Thus, the dose of 60 g HBOC 201 is sufficient to bind the available Hp and Hpx, allowing subsequent doses of HBOCs to remain in circulation until these proteins are replenished. In case of infusion with HbMPs of 20% (v/v) in reference to a blood volume of 5 L, this would correspond to a concentration of 2.3 × 1011 HbMPs/mL. In this case, Hp and Hpx would still be present in excess and could thus bind and eliminate all HbMPs. The need arises to verify whether it is indeed Hp and/or Hpx that are responsible for the elimination of HbMPs. A direct interaction with the monocyte receptor CD163, as already described in the literature for other Hb derivatives, is also conceivable. Another possibility would be an elimination of HbMPs independent of the Hb content, which is instead influenced by particle size or other physical properties of the particles.

In this study, we screened several monocytic surface receptors for a possible influence on the uptake of HbMPs by monocytes, which are precursor cells of macrophages. We chose to screen for CD14- as well as CD33-dependent HbMP uptake by monocytes since both proteins are specific for monocytes. CD163 was tested because of its direct affinity to Hb. Also, we tested the effect of blocking CD204 (scavenger receptor A/SR-A). SR-A is a membrane protein occurring in the monocyte/macrophage lineage. Playing an important role in host defense, it exhibits a long list of ligands including modified serum albumin, which might also occur within our HBMPs . The interaction of HbMPs (with different surface modifications) with Hp, as well as with anti-Hb antibodies has already been studied by Prapan and co-workers . With this work, the investigation is extended to the role of CD14, CD33, CD163, and CD204 in the uptake of HbMPs by monocytes. We decided to not research the effect of Hpx on HbMPs in this study, since there is no indication for heme to be accessible within our HbMPs. We hypothesize that CD163 plays an important role in the uptake of our HbMPs into monocytes, while there might be additional other mechanisms at play. The aim of this work is to contribute a further step towards the development of HbMPs as a complete, safe blood substitute for volume expansion and oxygen distribution.

0.9% sodium chloride was purchased from B. Braun SE, Melsungen, Germany. Phosphate-buffered saline (PBS) 10-fold solution was purchased from Fisher scientific, Fair Lawn, New Jersey, USA. Human serum albumin (Plasbumin®20) was purchased from Grifols Deutschland GmbH, Frankfurt am Main, Germany. Trypan blue 0.4% solution was purchased from PAN-Biotech GmbH, Aidenbach, Germany. Manganese chloride, sodium hydroxide, sodium borohydride, sodium carbonate, ethylenediaminetetraacetic acid (EDTA), glutaraldehyde solution, grade II, 25% in H2O, as well as Triton™ X-100 were purchased from Sigma-Aldrich, Steinheim, Germany. Pronase was purchased from Roche diagnostics GmbH, Mannheim, Germany. Bovine hemoglobin (Actoheme®) was provided by Biophyll GmbH, Dietersburg, Germany.

PHAGOTEST™ and PHAGOBURST™ test kits were purchased from Glycotope Biotechnology GmbH, Berlin, Germany. Mouse-derived monoclonal anti-human CD14 antibody (IgG2a, Κ), conjugated with fluorescein isothiocyanate (FITC) was purchased from BD Biosciences, Heidelberg, Germany. Mouse-derived monoclonal anti-human CD14 antibody (IgG2a, Κ) was purchased from BioLegend, San Diego, USA. Mouse-derived monoclonal anti-human CD33 antibody (IgG1, K) was purchased from BioLegend, San Diego, USA. Mouse-derived monoclonal anti-human CD163-antibody (IgG1, clone: GHI61), conjugated with allophycocyanin (APC) was purchased from antibodies-online GmbH, Aachen, Germany. Mouse-derived anti-human CD204-antibody (IgG2a, K) was purchased from BioLegend, San Diego, USA.

Peripheral venous blood was collected from healthy donors and anticoagulated with lithium heparin according to the requirements of the German law regulating transfusion. Written informed consent was obtained from all donors, as well as a positive vote of the Ethics Committee of the Charité - Universitätsmedizin Berlin (EA4/023/22). Blood samples were processed immediately after blood collection, and cell metabolism was maximally slowed down by storage at 0 °C.

For the preparation of hemoglobin-based sub-micron particles (HbMPs), the protocol of particle preparation using the CCD technique previously described has been followed: 0.5% bovine hemoglobin solution was mixed with 0.125 M MnCl2 solution. While stirring rapidly, 0.125 M Na2CO3 was added and stirred further for 30 s to let Hb–MnCO3 particles form. 20% human serum albumin (HSA) solution was added and allowed to incubate with the particles for 5 min. After washing in 0.9% NaCl solution, particles were centrifuged (3000g, 3 min), and the supernatant was decanted. Particles were then resuspended in 0.9% NaCl. For crosslinking, 0.02% glutaraldehyde (GA) was added to the particle suspension, which was then incubated for 1 h at room temperature under stirring. After another washing in aqua dest. the MnCO3 matrix was dissolved using 0.25 M EDTA. After 30 min incubation time, 0.2 mg/mL NaBH4 in 0.1 M NaOH was added to prevent oxidation of Hb to met-Hb. Incubation for another 30 min followed. After triplicate washing in a washing solution of 0.9% NaCl containing 0.2% HSA, particles were resuspended in 0.9% NaCl, checked via light microscopy for aggregation, and stored at 4 °C.

To ensure that the subsequent experiments would be carried out on particles that also met the quality requirements for potential use as oxygen transporters in vivo, the particles were characterized in different ways. Thus, comparability was achieved between different particle batches in different experiments. All experiments within this work have been performed with HbMPs from the same batch.

HbMP suspension was placed in glass capillaries for hematocrit determination and then centrifuged at 15,000g for 10 min. The amount of particle sediment was determined manually, and the remaining suspension was diluted with 0.9% NaCl to a concentration of 2% HbMP.

After dilution of the particle suspension to 0.13% (V/V) with NaCl, the average size and conductivity, as well as the zeta potential of the particles were determined using a zetasizer (zetasizer Nano ZS, Malvern Instruments, Malvern, United Kingdom) with each measurement in triplicate.

To determine the HbMPs’ hemoglobin content, the modified alkaline hematin-D method (AHD method) was used, as described in detail elsewhere . Pronase solution (0.5 mg/mL) was added to the particle suspension and incubated at 45 °C for 30 min. The AHD reagent (25 mg/mL Triton™ X-100 in 0.1 M NaOH) was added to the particle sample (final concentration: 12.5 mg/mL), mixed, and incubated at room temperature for 15 min while stirring. After centrifugation (20,000g; 10 min) the supernatant was taken off and, immediately, the absorbance at a wavelength of 574 nm was determined photometrically in three single measurements (cytation 3 imaging reader, BioTek Instruments GmbH, Bad Friedrichshall, Germany). The hemoglobin concentration of the particles was then calculated from the measured values according to the following formula, as published by Smuda et al. :

(A = absorbance at 574 nm; ƒ = dilution factor (2.1); molar mass of Hb (tetrameric) = 64.5; (monomer) molar extinction coefficient (ε) = 6.945).

Each hemoglobin molecule contains an Fe2+ ion on which the ability of Hb to bind oxygen is based. Methemoglobin has undergone an oxidation process. The central iron ion is trivalent, and the Hb derivative has lost its ability to bind oxygen. Therefore, only bivalent Hb is functional. The “oxygen release method” was used to determine the percentage of functional hemoglobin, as previously described in detail by Kloypan and co-workers . In brief, the oxygen content of the suspension was measured in three individual measurements in 500 µL each of the particle suspension diluted to a total Hb content of 0.5 mg/mL under steady, gentle stirring (Microx 4, PreSens, Regensburg, Germany). After reaching a stable initial value, 50 µL of 10% potassium ferricyanide (K3[Fe(CN)6]) was added, and the stirring speed was increased. The oxygen bound to Hb was thus released, and the peak value of O2 within the suspension was determined. From the difference between initial and peak values, the proportion of functional Hb was calculated.

Table 1: Samples for indirect phagocytosis test no. 1, role of CD163.

aIn a second step, all indirect phagotest samples were also incubated with FITC-tagged E. coli for 10 min at 37 °C; EC = E. coli; AB = antibody.

Table 2: Samples for indirect phagocytosis test no. 2, role of CD14, CD33, and CD204.

aIn a second step, all indirect phagotest samples were also incubated with FITC-tagged E. coli for 10 min, at 37 °C; EC = E. coli; AB = antibody.

Table 3: Samples for exclusion of confounders and identification of leucocyte subpopulations.

awb = (human) whole blood.

A HbMP suspension was initially adjusted to a concentration of 2% (V/V) HbMPs as described. The hemoglobin content in this suspension was 5.4 ± 0.02 mg/mL. Analysis by zetasizer showed an average particle size of 781 ± 7 nm, a conductivity of 1.34 ± 0.07 mS/cm, and a zeta potential of −28.0 ± 0.5 mV. The oxygen release method measurement showed a percentage of 81.4% ± 2.13% functional Hb. Light microscopy showed no aggregation within the particle suspension.

Figure 1: Leukocyte subpopulations, DNA-stained, FITC-tagged anti-CD14 AB (for better readability, all captions of the graphic have been enlarged. The dot plot itself has not been changed in any way).

Figure 2: (a) Reduction of normalized MFI (%) by pre-feeding cells with unlabeled E. coli; antibody blockade of CD163 does not obstruct the uptake of untagged E. coli; N = 3. (b) Monocyte-MFI (normalized) after uptake of FITC-tagged E. coli; reduced MFI (normalized) after pre-feeding with HbMP. Antibody blockade of CD163 does obstruct the uptake of HbMP; N = 3.

Figure 3: Effect of blockade of CD14, CD33, CD163, and CD204 on HbMP uptake by monocytes; N = 3.

Figure 4: Non-interference of utilized antibodies with the uptake of FITC-tagged E. coli; N = 3.

Table 4: Overview of phagotest no. 2 samples, normalized MFI values.

aBlocking monocytic CD14, CD33, CD163, and CD204 reduces HbMP uptake by monocytes (MFI is inversely proportional to the extent of uptake in pre-feeding step); N = 3.