While ABO/Rh(D) red blood cells (RBC)-matched transfusions are generally considered as safe, a significant risk of alloimmunization to non-A/B blood group antigens exists; especially in chronically transfused patients. Indeed, alloimmunization to non-A/B antigens can be so severe that RBC transfusion can no longer be safely administered without the risk of a potentially deadly immune haemolytic reaction. Currently, no satisfactory solutions exist either to prevent blood group alloimmunization or to cost-effectively treat patients with severe alloimmunization. To address this problem, we have pioneered the immunocamouflage of donor RBC. The immunocamouflaged (stealth) RBC is manufactured by the covalent grafting of biologically safe polymers to RBC membrane proteins. As a result of the grafted polymer, non-A/B blood group antigens are biophysically and immunologically masked. Of particular interest is the immunocamouflage of the Rh(D) antigen which could be used to improve blood inventory and transfusion safety. The polymer-modified RBCs are morphologically normal and, in mice, exhibit normal in vivo survival at immunoprotective grafting concentration. In this chapter, we explore both the biophysical and immunological consequences of the grafted polymers, explore the conditions in which they might be appropriately used, and describe the technology necessary to manufacture functional transfusable units of these cells within the clinical setting.
- red blood cell
The transfusion of red blood cells (RBC) remains the most common, and best tolerated, form of tissue transplantation. Indeed, an estimated 108 million units of whole blood (~49 million litres) are collected annually worldwide for processing and eventual transfusion . In spite of this massive collection effort, the need for blood constantly exceeds availability due to a combination of collection, manufacturing, storage and, most important clinically, biological (i.e., immunological) issues. The biological challenges facing successful RBC transfusions are vastly underappreciated, largely because of the long history and ubiquity of blood transfusions in modern medicine. Indeed, the RBC is an immunological complex cell with 35 major blood group systems that give rise to over 300 unique antigens capable of eliciting an immune response. Moreover, this immunological complexity is further exacerbated by the finding that the non-A/B (often referred to as minor) blood group antigens exist with varying frequencies among different ethnic and racial groups [2, 3]. Within the non-A/B antigens, Rh(D) deserves special attention.
Among the non-A/B blood groups, the Rh system, and in particular Rh(D), is considered to be the most immunogenic antigen. Indeed, the Rh(D) antigen is highly immunogenic and when Rh(D)+ blood is transfused into an Rh(D)− individual, there is a 50% risk for the development of anti-Rh(D) antibodies resulting in very high risk of a haemolytic transfusion upon a second Rh(D)+ transfusion. Consequent to its immunogenicity, Rh(D) is always determined simultaneously with ABO type and constitutes the ‘±’ found alongside the ABO phenotype. Consequent to its immunogenicity, Rh(D) poses a significant challenge to blood operators since Type O Rh(D)− (O−) blood is the universal donor cell. In Euro-centric populations, 6–7% of the population is O− making the maintenance of an adequate inventory of this universal donor blood problematic but possible. Indeed, in North America and Europe, virtually all blood service providers experience a chronic shortage of Type O− blood. However, in other geographic regions, especially Asia, Rh(D)− individuals are extremely rare. Indeed, in China, only 0.1–0.4% of the population, regardless of ABO type, is Rh(D)−, making the Rh(D)− individual (especially with the increasing influx of European tourists) an at-risk patient . Thus, within transfusion medicine, Rh(D) remains a significant problem in terms of both supply and its clinical risk.
Despite the immunological complexity of the RBC, simple ABO/Rh(D) matching has been, typically, considered sufficient for most acute transfusion needs. However, even when ABO/Rh(D) are appropriately matched, transfusion reactions still occur as mismatched non-A/B antigens do carry some immunological risks to a patient. While the incidence of clinically noteworthy (i.e., significant patient morbidity) transfusion reactions is relatively low (~0.017% of transfused individuals), less severe transfusion reactions (e.g., transient fever, malaise, premature RBC clearance) and alloimmunization are considerably more frequent and increase with the number of transfusions received by an individual [5–9]. Indeed, with approximately 108 million units of whole blood collected worldwide per annum for blood product preparation, the actual numbers of adverse events become quite significant. Of clinical importance, alloimmunization to non-ABO group antigens is significantly exaggerated in individuals (~30%), especially minorities, receiving chronic transfusion therapy as seen in thalassemia and sickle cell anemia. Indeed, alloimmunization to non-ABO blood groups can be so severe that blood transfusion can no longer be safely administered without risk of a potentially deadly immune haemolytic reaction.
Historically, various interventions have been used in an attempt to prevent transfusion reactions arising from alloimmunization. While ABO/Rh(D) typing has been used since the 1940s, the practice of phenotyping some of the more problematic non-A/B/Rh(D) antigens is still uncommon and likely underlies the high frequency of alloimmunization in chronically transfused patients. In studies on β thalassemia, up to 20% of these chronically transfused individuals demonstrate clinical evidence (i.e., mild to severe transfusion reactions) of alloimmunization against non-A/B donor blood group antigens [10, 11]. Rates in patients with sickle cell anemia are even higher (>30%) . As a result, many US National Institutes of Health (NIH) funded Sickle Cell Centres now evaluate a blood recipient for high risk (for alloimmunization) blood group antigens and to prophylactically utilize phenotypically matched blood for transfusion in this cohort of patients. However, even antibody screening does not identify patients alloimmunized to less common RBC antigens or, more importantly, prevent primary alloimmunization to non-tested antigens. To date, the only solutions to prevent alloimmunization, or for individuals with very rare blood type, are to store autologous blood (4°C), maintain an inventory of frozen rare blood group units, keep a blood bank registry of potential donors with rare blood types, utilize extensive RBC phenotyping prior to transfusion and/or encourage minority blood donations [7, 12–27]. While all of these steps are prudent and variably effective, situations still arise where an appropriate (or even satisfactory) blood match cannot be made.
2. Bioengineering the red blood cell
Currently, no satisfactory solutions exist to prevent or cost-effectively treat blood group alloimmunization or to improve the inventory of Rh(D)− blood. To address these unmet needs, the covalent grafting of biocompatible polymers to donor RBC has been proposed to immunocamouflage the allogeneic RBC. The immunocamouflaged (stealth) RBC is manufactured by the covalent grafting of methoxypoly(ethylene glycol) [mPEG; PEGylation], as well as other polymers (e.g., polyoxazolines, POZ; and hyperbranched polyglycerols, HPG), to membrane proteins on the surface of allogeneic donor RBC (Figure 1).
Most commonly, the chemically activated polymers are covalently grafted to proteins at exposed lysine residues. As a result of the grafted polymer, donor blood group antigens are biophysically and immunologically masked while the modified RBC remaining biologically and functionally viable. To date, most studies have focused on mPEG as the polymer of choice due to its superior ability to both sterically and charge camouflage allogeneic RBC and its well-characterized, and safe, pharmacological profile. The basic chemical structure of mPEG is HO-(CH2CH2O)n-CH2CH2CH3. mPEG is of low toxicity and is US FDA approved for oral, intravenous, subcutaneous and intramuscular administration . The mPEG polyether polymer is neutrally charged, available in an extraordinarily wide range of molecular weights, and is highly soluble in aqueous-based solutions making it very suitable for pharmacological use. In contrast, both the POZ (e.g., PEOZ) and HPG polymers are poorly soluble in aqueous solutions and only confer weak charge camouflage.
A large number of biological and biophysical studies have been done to characterize the effects of polymer grafting on immune recognition and in vitro and in vivo viability [28, 30–58]. These studies have demonstrated the significant potential of this immunocamouflage technology in transfusion medicine. Example uses include but are not limited to: (1) derivitize RBC to diminish transfusion reactions arising from mismatched blood or alloimmunization; (2) utilization by clinical blood banks to camouflage the Rh(D) antigen to improve blood inventories and utilization; (3) use of mPEG-modified RBC as a ‘chain-breaker’ (i.e., preventing RBC aggregates arising from abnormal cell-cell interaction) in vascular occlusive diseases such as sickle cell anemia; and (4) prevention of transfusion-associated graft-versus-host disease. Other uses outside of transfusion medicine include the implantation of derivitized cells or cell aggregates (e.g., pancreatic islets) to correct enzyme deficiencies; the induction of tolerance via PEGylated leukocytes; and the prevention of viral infections via antiviral gels.
Biophysical and biological characterization of the stealth RBC: The immunocamouflage of cells is a function of the biophysical and biochemical nature of the grafted polymer (Figure 1). Biophysically, the grafted polymers confer its immunoprotective effects via both steric hindrance and charge camouflage (Figure 2). The efficacy of membrane immunocamouflage is dependent upon both the density (i.e., how much) and depth (i.e., thickness; polymer molecular weight) of the polymer layer. As shown in Figure 1, steric hindrance arises from either the rapid mobility arising from intra-molecular flexibility of the polymer (mPEG and PEOZ) and/or polymer density itself (HPG). Perhaps of even more importance is the ability of the polymer to obscure the surface charge (charge camouflage) of the cell. Biophysically, charge camouflage arises from polymer-mediated extension of the shear plane (SP) thereby decreasing the apparent surface charge (Figure 2A and B). The grafted polymers can give rise to both DIRECT (direct binding to the antigen in question) and INDIRECT (binding to sites other than the antigen but conferring indirect camouflage) immunocamouflage of blood group antigens (Figure 2C). The importance of indirect immunocamouflage is significant as blood group antigen proteins do not exist in isolation but are most commonly part of complex protein clusters. As shown, the Rh(D) protein is deeply buried within a large protein complex on the surface of the RBC making the direct camouflage of Rh(D) difficult. However, the indirect immunocamouflage of Rh(D) arising from polymers grafted to surrounding proteins (e.g., Kell protein, Band 3) results in the highly efficient immunocamouflage of RhD [28, 52, 55, 59].
Indeed, one of the most promising prospects of RBC immunocamouflage is in both diminishing the risk of Rh(D) alloimmunization and safely increasing blood inventory during emergency situations or in circumstances where Rh(D)- blood is unavailable. As shown in Figure 3A, immune recognition and phagocytosis of anti-D (RhoGAM®; Rho(D) Immune Globulin (Human) RhoGAM Ultra-Filtered PLUS; Ortho Clinical Diagnostics)-opsonized Rh(D)+ RBC are blocked in a grafting concentration-dependent manner by the grafted mPEG polymer. Importantly, RhoGAM® is a highly purified and concentrated human-derived anti-D IgG antibody that is highly effective at RBC opsonization yielding Monocyte Index (MI) scores in the monocyte-monolayer assay (MMA) in the range of 60–100%. The RhoGAM® antibody is used clinically for the prevention of Rh immunization, including during and after pregnancy and other obstetrical conditions or incompatible transfusion of Rh-positive blood. However, RhoGAM® does not fully reflect the biological/clinical heterogeneity of anti-D alloantibodies arising in alloimmunized individuals. To assess the potential utility of mPEG-RBC in alloimmunized individuals, human-sourced anti-D alloantibodies (N = 8) were assessed using the MMA. Of note, the naturally occurring heterogeneity of anti-D responses is demonstrated by the wide range in MI values when used to opsonize Rh(D)+ RBC (Figure 3B). Indeed, only five of eight of the alloantibodies exhibited MI values of >5% (i.e., potential for acute haemolytic transfusion reaction). However, as predicted by the RhoGAM findings, grafting of the 20 kDa polymer (2 mM) yielded significant reductions (p < 0.01) in the MI values of all samples. Of the five plasma samples with MI values >5%, PEGylation resulted in four having MI values of <5% while the remaining sample had an MI of 5.5 ± 0.9. Moreover, the anti-phagocytic effect of polymer-mediated immunocamouflage of Rh(D)+ RBC was observed regardless of the IgG subclass of the alloantibody. It is also important to note that the serological score (ranging from W+ to 4+) did not correlate significantly with the MI value . These findings are in line with previous studies that demonstrated that the antiglobulin test does not accurately reflect the amount of IgG bound and is, at best, only a weak predictor of RBC phagocytosis [60–63]. In aggregate, these findings suggest that immunocamouflaged Rh(D)+ RBC could be safely transfused into Rh(D)− patients in an emergency situation.
Importantly, at immunologically protective grafting concentrations, the grafted polymer does not affect RBC structure or function as evidenced by normal morphology, O2 uptake and delivery, cellular deformability or ion transport [30, 33, 34, 37, 41, 52, 55]. While virtually indistinguishable from unmodified cells in most aspects, one interesting difference was noted between unmodified and mPEG-modified RBC. Consequent to the charge camouflage of the RBC, the cell:cell interactions necessary for Rouleaux formation were abrogated (Figure 4). Because the grafted polymer camouflages the charge of the cell necessary for cell:cell interaction, Rouleaux formation is inhibited resulting in attenuation of RBC sedimentation and, physiologically, decreased low-shear viscosity (Figure 4). Importantly, the decrease in low-shear viscosity may make the use of stealth RBC highly suitable for patients with diseases characterized by RBC-vaso-occlusive events (e.g., sickle cell). Importantly, RBC PEGylation inhibits both non-antibody (e.g., sickle cell self-aggregation) and antibody-mediated aggregation events and donor mPEG-RBC can serve as an efficient chain-breaker in pro-aggregation states .
The grafting of immunologically ‘inert’ polymers to the membrane of allogeneic RBC effectively camouflages multiple non-ABO antigens from immune recognition. The combined actions of both steric and charge camouflage underlie the ability of the grafted polymer to camouflage allogeneic blood group antigens (immunocamouflage) from the recipient’s immune system. These immunocamouflaged (i.e., stealth) RBC may be an effective tool in both preventing and treating alloimmunization in the chronically transfused patient; the transfusion of individual patients with rare blood phenotypes; emergency situation or geographic locations (e.g., China) where Rh(D)− blood is unavailable. Moreover, immunocamouflaged RBCs are inexpensively and easily manufactured in the clinical setting.
3. Manufacturing the ‘stealth RBC’
The immunocamouflaged RBC is unlikely to be a mass-market product but rather a boutique blood product for the discriminating customer. Hence, on demand manufacturing of the stealth RBC will be the most likely scenario facing the clinician and blood provider. Fortunately, the manufacturing process of the stealth RBC is rapid and straightforward and requires no specialized equipment on part of the blood provider, hospital or clinic. Moreover, current laboratory tests exist that can, relatively rapidly (~48 h), evaluate the potential clinical utility of the stealth RBC in the at-risk alloimmunized patient. Crucial to the clinical use of the stealth RBC is the manufacturing process. The key tenet of the manufacturing process is the maintenance of a constant polymer:cell ratio in order to achieve a homogenous grafting of the polymers to the individual cells within donor RBC unit (Figure 5). If cells are under-PEGylated, they retain significant immunogenicity/antigenicity; if cells are over-PEGylated, the in vivo viability (i.e., circulation time) of the cell is compromised.
To achieve the homogeneity necessary for a clinical mPEG-RBC unit (Figure 5), two scalable devices (Figures 6 and 7) utilizing micro-mixing chambers (alternatively Y-connectors to induce turbulence and mixing) have been designed, constructed and validated to semi-automate the RBC derivatization process and minimize the risk of contamination. Both approaches can be done aseptically using modifications of existing blood bags and sterile docking devices. The Syringe Pump Method (Figure 6) has an advantage in better controlling the hydrolysis rate (i.e., inactivation rate; t1/2 of 33 min for succinimidyl valerate-activated methoxypoly(ethylene glycol) [SVAmPEG]) of the activated mPEG as the activated polymer is prepared in smaller batches. The Peristaltic Pump Method (Figure 7) has the advantage of being a continuous flow device with fewer points of possible disruptions in blood sterility. As demonstrated in Figure 6, the described manufacturing devices are capable of producing a transfusable unit of PEGylated RBC.
RBC manufacturing and product quality control: Quality assurance of the mPEG-RBC to document the reproducibility and homogeneity of SVAmPEG derivatization can be assessed using a two-phase (PEG-Dextran) partitioning system [37, 41]. The partitioning of PEGylated RBC in this system is governed by the ratio of PEG:Dextran, as well as by the density and size (m.w.; number of ethoxy units) of the grafted mPEG. In this simple assay, mPEG-RBC are added to an immiscible PEG:Dextran solution, rapidly mixed and allowed to separate. As shown, unmodified RBC preferentially partition to the Dextran or interface region while the PEGylated RBC preferentially partition to the PEG-rich phase (Figure 8). Importantly, both the automated derivatization and the purification methodologies are highly scalable and can be applied to existing blood bank devices and workflows.
The PEGylation Recipe
PEGylation Buffer: While buffer flexibility exists, the chemistry of PEGylation via activated mPEG requires alkaline conditions [28, 30–58]. The formulation of the mPEG-buffer used by our laboratory for the PEGylation of RBC is as follows: 50 mM K2HPO4, 105 mM NaCl, pH 8.0.
Activated mPEG: Following extensive analyses of multiple polymer molecular weights and linker chemistries, 20 kDa succinimidyl valerate-activated methoxypoly(ethylene glycol) [SVAmPEG] has been selected as our primary polymer species. While a wide range of polymer molecular weights (e.g., 2–40 kDa) have been tested, in vitro (human and mouse) and in vivo (mouse) studies demonstrate that the 20 kDa polymer provides the optimal immunological camouflage of the allogeneic cell while maintaining normal in vivo survival. Linker chemistry selection was based on the commercial availability of the polymer, the relatively long half-life of the activated polymer and its excellent in vitro and in vivo findings. Clinically compliant SVAmPEG can be purchased from Laysan Bio Inc. (
4. Evaluating the potential clinical utility of the stealth RBC
While multiple studies have demonstrated that RBC immunocamouflage can effectively block immune recognition of multiple blood group antigens, the diversity of alloantibodies produced to a single blood group antigen by humans is staggering. Hence, a crucial step in the clinical use of the stealth RBC should be evaluating the potential efficacy of the stealth cell in the individual patient. While one might assume that standard serological testing techniques would suffice, this is not the case for a variety of reasons. Primary among these reasons, and as shown in Figure 3C, is that the laboratory serological score does not correlate significantly with the MI value or the risk of an acute transfusion reaction. Indeed, multiple studies have demonstrated that the antiglobulin test poorly reflects the amount of IgG bound and is, at best, a very weak predictor of RBC phagocytosis [60–63]. This confounding finding is actually by design. Serological testing is ‘overly’ sensitive in order to detect miniscule amounts of bound antibody to assure the appropriate typing of an individual or to detect the presence of a potentially dangerous alloantibody. Another potential complication is that a large number of commercial testing protocols employ PEG (as either a listed or unlisted ingredient) as a component of the testing reagents. The reagent PEG will cause the mPEG-RBC to segregate as ‘PEG likes PEG’ (Figure 8B). Hence, other predictors of the potential clinical utility of the stealth RBC for an individual patient are needed.
Perhaps the most definitive testing approach for the potential clinical value of the stealth RBC is the monocyte-monolayer assay (MMA). The MMA assesses FcγR-mediated adherence and phagocytosis of alloantibody-opsonized donor RBC by monocytes and has been clinically correlated with in vivo transfusion safety (i.e., prevention of acute transfusion reactions) and efficacy (24 h RBC survival) [60–62, 65–68]. The in vitro MMA is reliable, reproducible and, within the transfusion medicine community, is considered to be the best assay currently available for the evaluation of FcγR-mediated phagocytosis of antibody-coated human red cells, having more than 20 years of proven validity for the comparison of in vitro phagocytosis to in vivo clinical relevance.
Schematically, the MMA is described in Figure 9 (and experimentally demonstrated in Figure 3). The MMA examines FcγR-mediated phagocytosis in vitro using adherence-purified monocytes isolated from peripheral blood mononuclear cells obtained from normal volunteers. Blood group antigen positive RBCs are incubated with buffer (negative control) or sera or plasma from alloimmunized patients for opsonization, washed and overlaid on the monocyte monolayer. Both donor-obtained and reagent RBC are suitable for use in the MMA. The MMA uses a visual readout whereby the numbers of adherent and phagocytized control and opsonized RBC are enumerated per 100 monocytes (MI). For validation purposes, consistent positive and negative controls should be used, most commonly anti-Rh(D)-opsonized Rh(D)+ (positive control) and Rh(D)− (negative control) human RBC (see Figure 3). The visual inspection is simplified and enhanced by the use of phase contrast microscopy. Using the MMA, MI values of ≤5% indicate that the donor cells can be given without risk of an overt haemolytic reaction. However, it is worth noting that the MMA is most predictive of acute haemolytic transfusion reactions and is less predictive of long-term survival of donor RBC.
By using the MMA, the potential efficacy of RBC PEGylation can be done on an individualized basis. The probable success of the stealth RBC transfusion can be further enhanced by serologically selecting (or even better MMA testing) the best possible matches from the donor RBC inventory so as to minimize the risks of additional complications. Once the serological or MMA testing of the donor blood unit has been done, the unit can be PEGylated as described then tested against the individual’s alloantibodies via the MMA prior to transfusion into the recipient. However, the identification of the donor unit(s), PEGylation and MMA testing does require 48–96 h lead time. Hence, identification of potential patients should be done as early as possible to assure the availability to source appropriate polymer stock, prepare the PEGylation device and identify and test possible donor units. Concurrent with the evaluation of the potential clinical value of the stealth RBC, the physician/transfusion service must also receive institutional and governmental approval for their use.
5. Institutional and governmental approval for patient use
Prior to the actual clinical use of the stealth RBC in a seriously ill patient, compassionate use approval must be obtained from both the hospital Research Ethics Board (REB; or equivalent) and the appropriate governmental agencies (e.g., in Canada, Health Canada). This is likely to be a physician-driven process done in conjunction with the hospital's transfusion service and/or blood provider. Key to these requests is the need to clearly cite the lack of, or very limited availability, of suitable donor RBC. Once institutional approval has been obtained, the hospital REB would likely lead the interaction with the appropriate governmental agency (e.g., Health Canada) regarding an Investigative New Drug (IND) submission. For compassionate use in a single patient who would be likely to die in the absence of a transfusion, a formal IND submission may or may not be necessary. These steps will, obviously, change from country to country.
One question likely to be raised by the REB is whether the proposed mPEG-dosing is safe. The answer to this important question is, at least in part, addressed by recent Phase I–III clinical trials of PEGylated human haemoglobin (PEG-Hb; Sangart, San Diego, CA, USA) [69–73]. These clinical trials have infused humans, at the highest dosing schedule, with up to 8.33 ml/kg of PEG-Hb. At this dosing, the typical male volunteers (180 lbs/81.8 kg) received 680 ml of the PEG-Hb solution as a single dose, an infusion of ~25 g of PEG. Importantly, no adverse effects were noted in any of the human volunteers receiving this dose. Furthermore, in animal studies (e.g., rats), PEG-Hb was safely infused at a single adjusted dose exposure of to 46 g of PEG-Hb .
Grafting of immunologically ‘inert’ polymers to the membrane of allogeneic RBC can effectively camouflage non-ABO antigens from immune recognition. These immunocamouflaged (i.e., stealth) RBCs may be an effective tool in both preventing and treating alloimmunization in the chronically transfused patient; the transfusion of individual patients with rare blood phenotypes; for emergency situation or geographic locations (e.g., China) where RhD-negative blood is unavailable. Importantly, several characteristics of the immunocamouflaged RBC may also make them highly suitable in patients/diseases characterized by RBC-mediated vaso-occlusive events (e.g., sickle cell) consequent to the polymer-mediated reduction in low-shear viscosity. For the hospital or clinic, the immunocamouflaged RBCs are inexpensively and easily manufactured using commonly available equipment and existing blood bags. Moreover, the potential clinical utility of the stealth RBC can be evaluated for the individual patient using the clinically validated monocyte-monolayer assay in which antigen-mismatched RBCs are PEGylated and then opsonized with the patient’s own alloantibody.
The authors would like to thank the past laboratory members whose work as graduate students and postdoctoral researchers have been cited in this chapter: Dr Duncheng Wang, Dr Yevgeniya Le, Dr Dana Kyluik-Price, Dr Li Li, Dr Amanda Bradley, Dr Audrey Chen and Dr Kari Murad. This study was supported by grants from the Canadian Institutes of Health Research (Grant No. 123317; MDS), Canadian Blood Services (MDS) and Health Canada (MDS). The views expressed herein do not necessarily represent the view of the federal government of Canada. We thank the Canada Foundation for Innovation and the Michael Smith Foundation for Health Research for infrastructure funding at the University of British Columbia Centre for Blood Research. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.