Future generations of red blood cell substitutes



Chang TMS (Artificial Cells & Organs Research Center, McGill University, Montreal, Quebec, Canada). Future generations of red blood cell substitutes (Minisymposium). J Intern Med 2003; 253: 527–535.

Polyhaemoglobins (PolyHb) and perfluorochemicals are in advanced phase III clinical trials and conjugated haemoglobins in phase II clinical trial. New recombinant human haemoglobin with no vasoactivity is being developed. A soluble macromolecule of PolyHb–catalase–superoxide dismutase is being studied as an oxygen carrier with antioxidant properties. New artificial red blood cells that are more like RBC are being developed. One is based on haemoglobin lipid vesicles. A more recent one is based on nano-dimension artificial red blood cells containing haemoglobin and RBC enzymes with membrane formed from composite copolymer of polyethylene glycol–polylactic acid. Their circulation time is double that of PolyHb.


Fluids for volume replacement have been in routine clinical use for many years. When haematocrit falls below the critical level and fluid replacement alone is not enough, red blood cells or whole blood is now the only next step. The present effort in the development of blood substitutes is to produce a substitute for red blood cells. Concentrated efforts to develop blood substitutes for public use were only seriously started in the 1990s because of concerns regarding HIV in donor blood. At present, red blood cells substitutes in phase III clinical trials are nothing more than oxygen carriers [1–5]. For example, the most commonly used first generation blood substitutes in phase III clinical trial consist of Ringer-Lactate solution with an oxygen carrier in the form of modified haemoglobin (Hb) or perfluorochemicals.

The first attempt at modified haemoglobin was to prepare artificial red blood cells encapsulating haemoglobin and red blood cell enzymes, but this proved to be an overambitious approach at that time [6, 7]. A much simpler approach is based on polyhaemoglobin (PolyHb) which is prepared based on the use of bifunctional agents to cross-link haemoglobin intermolecularlly [6, 7] including the use of glutaraldehyde to cross-link haemoglobin into soluble PolyHb (Fig. 1) [8]. This type of cross-linking prevents the breakdown of the haemoglobin into dimers that was a problem when unmodified stroma-free haemoglobin (SF-Hb) was used [9]. Promising results using PolyHb are being gathered in phase III clinical trials involving mostly perioperative uses where oxygen carriers are needed in addition to volume replacements [10–17]. One type has been approved for routine use in South Africa [15]. Another oxygen carrier is based on the conjugation of haemoglobin molecules to polymers to form conjugated haemoglobin (Fig. 1) [7, 18–20]. Improved conjugated haemoglobins in phase II clinical trials have been reported recently [21, 22]. Still another type of oxygen carrier in phase III clinical trial is based on perfluorochemicals [23, 24]. These oxygen carriers are an intermediate step between volume replacement and red blood cells. While they work well when the conditions only require short-term oxygen carriers, there are other conditions requiring more than just oxygen carriers. Furthermore, progress in the field requires further basic fundamental knowledge. This paper concentrates on three areas: (1) modified haemoglobin, nitric oxide and vasoactivity; (2) oxygen carriers with antioxidant acitivity for those conditions with potentials for ischaemia–reperfusion injuries and (3) red blood cell substitutes that resemble more closely their biological counterpart.

Figure 1.

Different types of modified haemoglobin. Updated, modified and reprinted with permission from: Chang, TMS, Blood Substitutes. Karger Publisher, Basel, 1997, courtesy of copyright holder [10].

Vasoactivity and nitric oxide

Cross-linking each haemoglobin molecule intramolecularly can also prevent its breakdown into dimers [25, 26]. Another approach is the use of recombinant technology for Escherichia coli to produce ‘fused’ single human haemoglobin molecules that do not break down into half molecules (Fig. 1) [27, 28]. Both of these have very good oxygen releasing characteristics (high P50). However, infusion of 2 units or more in clinical trials can result in vasoconstriction [26, 28]. One theory is that unlike PolyHb and conjugated haemoglobin, the smaller single tetrameric haemoglobin molecules can cross the intercellular junction of endothelial cells lining the vessel walls and enter the interstitial space where they bind nitric oxide [10, 11]. When 2 units or more are used, this can result in the removal of sufficient amount of nitric oxide to result in vasoconstriction and other smooth muscle effects. It should be noted that PolyHb and conjugated haemoglobin usually contain a small fraction of intramolecularly cross-linked tetrameric haemoglobin. The PolyHb containing <1% of tetrameric Hb did not result in vasoactivity when very large volumes were infused in clinical trials. Polyhaemoglobin with <5% of tetramers results in slight vasoactivity when large volumes were infused. Polyhaemoglobin with higher concentration of tetramer results in vasoactivity when lower volumes were infused. As mentioned above, when only intramolecularly cross-linked tetrameric haemoglobin was used, there was marked vasoactivity when small volumes were used. This seems to support the possible role of tetrameric haemoglobin in vasoacitivity [10, 11]. Another possible contribution to vasoconstriction has also been proposed related to the difference in the flow and oxygen release characteristics of oxygen carriers in the microcirculation of oxygen carriers as compared with RBC [3]. Second-generation recombinant human haemoglobin has recently been developed in which the addition of an amino acid, tryptophan resulted in steric hindrance for the nitric oxide receptor site [29]. This new tetrameric recombinant human haemoglobin does not cause vasoconstriction when infused into animals. However, there was a change in the oxygen release characteristics (lower P50). This problem was solved by replacing the distal histidine of the recombinant human haemoglobin with glutamine resulting in normal oxygen release characteristics. This new approach is now being actively developed.

Conditions with potentials for ischaemia–reperfusion injury

Red blood cells contain catalase (CAT), superoxide dismutase (SOD) and other enzymes. However, the present first generation blood substitutes are only oxygen carriers with no enzyme activities [30].

Lack of oxygen supply in severe haemorrhagic shock, stroke, myocardial infarction, organ transplantation and other conditions may result in ischaemia. Ischaemia leads to alterations in metabolic reactions producing hypoxanthine and activating the enzyme xanthine oxidase. The level of hypoxanthine increases with the duration and severity of ischaemia. When the tissue is reperfused with oxygen carrying fluid, xanthine oxidase converts oxygen and hypoxanthine into superoxide. By several mechanisms, superoxide results in the formation of oxygen radicals that can cause tissue injury. Superoxide dismutase and CAT in red blood cell converts superoxide into hydrogen peroxide that is in turn converted into water and oxygen. First, generation oxygen carriers do not contain these enzymes and thus could cause increased ischaemia–reperfusion injury in certain conditions. These conditions include severe sustained haemorrhagic shock, stroke, mycarodial infarction, organ transplantation and others.

We are studying the cross-linking of trace amounts of CAT and SOD to haemoglobin (Hb) to form PolyHb–SOD–CAT (Figs 2–4) [31–34]. Compared with PolyHb, PolyHb–SOD–CAT removes significantly more oxygen radicals and peroxides and stabilizes the cross-linked haemoglobin resulting in decreased oxidative iron and haem release [31]. Cross-linking these enzymes to PolyHb is important because otherwise, free SOD and CAT are removed rapidly from the circulation with a half-time of <30 min. In the form of PolyHb–SOD–CAT these enzymes circulate with a half-time more comparable with PolyHb which is about 24 h in human. In the reperfusion of ischaemic rat intestine, PolyHb–SOD–CAT significantly reduced the increase in oxygen radicals casued by PolyHb as measured by an increase in 3,4-dihydroxybenzoate (Fig. 2) [10, 32]. We have just completed our studies on global cerebral ischaemia–reperfusion [34]. This is based on bleeding anesthetized rats to hypotensive level combined with transient occlusion of both common carotid arteries [35]. After different length of time, this was followed by the release of the occlusion of the carotid arteries and reinfusion using different types of oxygen carrying fluids [34]. In one of our studies, animals were subjected to 0, 20, 30, 40 or 60 min of ischaemia before reperfusion and effects on blood–brain barrier (BBB) were followed as Evans blue extravasation (Fig. 3) [34]. With 60 min of ischaemia, Evans blue extravasation in the groups receiving oxygenated saline, SF-Hb and PolyHb were significantly higher than that of the sham control (P < 0.01). In another study, after 60 min of ischaemia, PolyHb–SOD–CAT was used as reperfusion fluid and then followed for 6 h [34]. The effect on BBB (Evans blue extravasation) was compared with the use of other reperfusion solutions after 60 min of ischaemia. PolyHb–SOD–CAT significantly attenuated the severity of BBB disruption as compared with saline, SF-Hb, PolyHb or a solution of free haemoglobin, SOD and CAT (P < 0.01) (Fig. 3) [34]. In the same study, brain oedema was followed as changes in brain water content. The changes in brain water content of PolyHb–SOD–CAT treated animals were not significantly different from that of the sham control (Fig. 4). The increase in water contents of saline, SF-Hb, PolyHb and the solution of free haemoglobin, SOD and CAT were significantly higher than that of the sham control and PolyHb–SOD–CAT group by the 4th hour and increased thereafter with time (P < 0.01).

Figure 2.

Oxygen radicals in ischaemic intestinal reperfused with PolyHb or PolyHb–SOD–catalase. Effects on oxygen radicals as measured by effluent 3,4-dihydroxybenzoate. Reprinted with permission from: S Razack, F D'Agnillo, and TMS Chang. Artificial Cells, Blood Substitutes and Immobilization Biotechnology, 25: 181–192, 1997. Courtesy of Marcel Dekker Publisher Inc., NY.

Figure 3.

Effects of PolyHb–SOD–CAT on blood–brain barrier in ischaemia–reperfusion compared to other solutions. Reprinted with permission from Artificial Cells, Blood Substitutes and Immobilization Biotechnology, an International Journal, 30: 25–42, 2002. Courtesy of Marcel Dekker Inc., NY.

Figure 4.

Effects of PolyHb–SOD–CAT on brain oedema in ischaemia–reperfusion compared to other oxygen carrying solutions. Reprinted with permission from: D Powanda & TMS Chang, Artificial Cells, Blood Substitutes and Immobilization Biotechnology, an International Journal, 30: 25–42, 2002. Courtesy of Marcel Dekker Inc., NY.

Another group has added antioxidant activity to intramolecularly cross-linked alpha–alpha tetrameric haemoglobin [36]. This is done by adding 2,2,6,6-tetramethyl-piperidinyl-1-oxyl (Tempo) to form polynitroxylated αα-Hb (PN-αα-Hb). The new compound PN-αα-Hb acts as an antioxidant in their in vitro and in vivo assays [36].

More complete artificial red blood cells

The above-modified haemoglobins are simpler and therefore the first red blood cell substitutes ready for clinical trials. Synthetic red blood cells formed by encapsulation of haemoglobin and red blood cell enzymes are more complete red blood cell substitutes. However, being more complete, they are also more complicated. Therefore, they take longer to develop. There are three stages in the development of artificial red blood cells.

Micro-dimension artificial red blood cells

The first study on microencapsulated haemoglobin or artificial red blood cells was reported by the author in 1957 [37] and 1964 [6] (Fig. 5). In this approach, synthetic membranes and cross-linked protein membranes are used to replace the natural red blood cells membrane [6, 10, 37]. The resulting artificial red blood cells have an oxygen dissociation curve similar to red blood cells [10, 37]. The membrane used at first was coated with a thin layer of organic liquid [37]. This retained 2–3-diphosphoglycerate inside. These artificial red blood cells do not have blood group antigens on the membrane. As a result, they do not form aggregates in the presence of blood group antibodies [7]. Red blood cell enzymes like carbonic anhydrase [6] and CAT [38] retained their activities inside the artificial RBC. Microcapsules containing haemoglobin and CAT can act as an antioxidant against the toxic effects of hydrogen peroxide in acatalasemic mice with an inborn error of metabolism in their CAT enzyme [38]. Microcapsules containing haemoglobin and CAT from a heterologous source, unlike the free enzyme, did not result in immunological reactions when given to immunized animals [39]. Artificial cells in the micron dimensions are now being explored extensively in other areas of medicine and biotechnology for cell therapy, enzyme therapy, gene therapy, drug delivery and other areas [40]. However, as far as red blood cell substitute is concerned, the single major problem at that time was rapid removal from the circulation after intravenous infusion [7]. Decreasing the diameter down to 1 micron increased the circulation time but it was still too short [7]. We found that removal of sialic acid from the red blood cell membrane resulted in the rapid removal of RBC from the circulation thus showing the importance of surface property and survival in the circulation [7]. This observation led us to prepare artificial red blood cells with modifications of surface properties including use of different synthetic polymers, cross-linked protein membranes, membrane with surface charge, polysaccharide surface as sialic acid analogues, lipid-protein and lipid-polymer [7]. Some of these improved the circulation time. However, the circulation time was still not enough for practical applications. These basic findings formed the basis for the next phases of study.

Figure 5.

Erythrocytes (RBC), micron-dimension and nano-dimension artificial RBC containing haemoglobin and enzymes. This nano-dimension combined with the use of a composite biodegradable polymeric membrane (polylactide–polyethylene glycol) have resulted in a circulation time that is double that of PolyHb (49). Updated and reprinted with permission from: Chang, TMS. Blood Substitutes. Karger Publisher, Basel, 1997, courtesy of copyright holder [10].

Submicron lipid vesicles containing haemoglobin

Haemoglobin can be microencapsulated inside small 0.2-micron diameter lipid membrane vesicles [41]. This substantially increased the circulation time, although the circulation time was still rather short. Many groups have since then carried out research to improve the preparation and the circulation time with emphasis on the modification of the surface properties. Modifications of surface properties including surface charge and the use of sialic acid analogues have further improved the circulation time. Among these many groups, two have made the most extensive progress [42–44]. Their studies show that there are no adverse changes in the histology of brain, heart, kidneys and lungs of experimental animals. They have also used this successfully in animal studies for exchange transfusion, haemorrhagic shock and other uses [42, 44]. Effects on the reticuloendothelial system have been studied by a number of groups. A recent finding is the use of lipid containing polyethylene glycol (PEG) to form liposome encapsulated haemoglobin [43]. This has resulted in a much longer circulating red cell substitutes that is double that of PolyHb. With increase in circulation time, there is also a problem with methaemoglobin formation. Many studies are being carried out to solve this problem [44]. Detailed reviews on haemoglobin lipid vesicles are available from these researchers [42, 44].

Nano-dimension biodegradable polymeric membrane artificial red blood cells

Success and progress in the testing of haemoglobin lipid vesicles stimulates research into a further generation of encapsulated haemoglobin how to improve even further the following:

  • 1Increasing stability in storage and after infusion.
  • 2Decreasing the potential effects of lipid on the reticuloendothelial systems.
  • 3Avoiding lipid peroxidation.
  • 4Solving the problem of methaemoglobin formation.
  • 5Inclusion of most of functioning red blood cell enzymes.

We are using our background in the use of biodegradable polymer like polylactide for the encapsulation of haemoglobin and other biologically active material started in 1976 [45]. Polylactides and polyglycolides are degraded in the body into water and carbon dioxide. These biodegradable polymers are in routine use in surgical sutures, drug delivery and other applications. The rate of degradation can be adjusted by changes in molecular weight and type of polymer or copolymer. It can also vary with particle size. We are now using this to prepare nano-dimension biodegradable polymer membrane haemoglobin to have mean diameter of between 80 and 200 nm (Fig. 1) [46–50]. The membrane material is made up mostly of biodegradable polymer. As polymer is stronger and porous, less membrane material is required. Polylactide is degraded in the body into lactic acid and then water and carbon dioxide. For a 500-mL suspension, the total lactic acid produced is 83 mEq [10]. This is far less than the normal resting-body lactic acid production (1000–1400 mEq day−1). This is equivalent to 1% of the capacity of the body to breakdown lactic acid (7080 mEq day−1).

Bovine haemoglobin after encapsulation has the same P50, Bohr and Hill coefficients [10]. The content of haemoglobin can match that of red blood cells [10]. One can extract the whole content of red blood cells and then nanoencapsulate this extract (Fig. 6). Furthermore, additional enzymes can be added to the solution before the nanoencapsulation process. Thus, additional SOD and CAT can also be included with the haemoglobin [10]. We have used our background in artificial cells containing multienzyme cofactor recycling systems [51] to help solve the problem of methaemoglobin formation. Applying this to haemoglobin nanocapsules has helped to solve the problem related to methaemoglobin formation. In nanocapsules, the biodegradable polymeric membranes can be made permeable to glucose and other molecules [10]. This allows us to prepare haemoglobin nanocapsules containing the methaemoglobin reductase system. External glucose can diffuse into the nanocapsules. Products of the reaction can diffuse out and therefore do not accumulate in the nanocapsules to inhibit the reaction. In vitro studies show that this can convert methaemoglobin to haemoglobin. Furthermore, reducing agents from the plasma can diffuse into the nanocapsules to reduce methaemoglobin to oxygen carrying haemoglobin. Animals have been infused with one-third the total blood volume. Most recently, we use a composite biodegradable polymeric membrane consisting of co-polymer of PEG with polylactic acid [49, 50]. After extensive research using this approach, we have now prepared nano-dimension artificial red blood cells that can retain their circulating haemoglobin level at twice the duration of PolyHb [49, 50].

Oxygen carriers in radiation therapy for tumours

One group has carried out phase 1b clinical trials to evaluate the safety of PEG-conjugated haemoglobin as an adjuvant to radiation therapy in human cancer patients [20]. This is based on their earlier experimental observation that PEG-bovine haemoglobin oxygenates hypoxic tumour tissue and dramatically increases its sensitivity to radiation therapy in laboratory models and veterinary animals. Other groups are also testing different types of oxygen carriers for the same purpose.

Melanoma is a fatal skin cancer. Meadow's group has shown that lowering of tyrosine would inhibit the growth of melanoma in culture and in animals. However, the use of tyrosine-restricted diet in patients has resulted in nausea, vomiting and weight loss. We have cross-linked tyrosinase to PolyHb resulting in a novel solution that acts as an oxygen carrier to supply oxygen for radiation therapy and at the same time lower systemic tyrosine [51]. This preparation inhibits the growth of melanoma cells in cell culture [51]. When combined with daily oral administration of microencapsulated tyrosinase, we can maintain the low tyrosine level while the rats continue to gain weight and grow at the same rate as the control group [51].

Discussion and conclusions

As early as 1957, this author prepared encapsulated haemoglobin to form artificial red blood cells [37]. In 1964, the same author showed the possibility of cross-linking haemoglobin to form PolyHb [6–8]. In 1968, another group showed that intramolecularly cross-linked tetrameric haemoglobin no longer has the renal toxicity of free haemoglobin [25]. Thus, the basic ideas of modified haemoglobin, although very crude, were all there in the 1960s. Unfortunately, except for some military interest, there was little academic, industrial or public interest in these ideas until the crisis of HIV in donor blood. Unfortunately, a product cannot be ready for clinical use without years of research and development followed by time consuming clinical trials. Thus, despite more than 10 years of intense efforts there is still no blood substitute ready for routine use in the Western world. Even when the first-generation oxygen carriers become available for routine use, they are only good for some clinical conditions. Much more still needs to be done to develop new generations of blood substitutes that can be used in other conditions. It is true that effective screening tests for HIV and hepatitis C have markedly decreased the risk of infection from donor blood to negligible numbers. However, can we be sure that there will not be another unknown agent in the future? If so, it will take time to develop suitable screening tests and even more time to develop the new generations of blood substitutes needed. Futhermore, with an increasing ageing population there is an increasing demand on the short supply of donor blood. In addition, increasingly more sophisticatic and extensive surgical procedures and more frequent international conflicts also place an unrealistic demand on the short supply of donor blood. The first generation oxygen carriers have already demonstrated the feasibility of red blood cell substitutes as follows. These oxygen carriers can be ultrafiltered and pasteurized to remove microorganisms including HIV and other viruses. Because these oxygen carriers do not have blood group antigens, cross-matching and typing are not required before use. This saves time and facilities and allows on-the-spot transfusion especially in emergency situation. Furthermore, these blood substitutes can be stored for more than 1 year even at room temperature for one type. These first generation oxygen carriers already have a number of important applications. However, there is an urgent need to develop the new generations of blood substitutes to broaden the potential areas of applications. However, even if we start to do this seriously and immediately, it will still take much time for these to be ready.

Conflict of interest statement

T. M. S. Chang is Director of the Artificial Cells and Organ Research Center, and Director of the MSS-FRSQ Research Group on Blood Substitutes in Transfusion Medicine.


The author acknowledges, being the principle investigator, research supports from the Canadian Institute of Health Research, the Virage Centre of Excellence Award from the Quebec Ministry of Science and Education and the Bayer/Canadian Blood Agency/HemaQuebec/Canadian Institutes of Health Research Partnership Fund. The recent award of the MSSS-FRSQ Research Group on Blood Substitutes in Transfusion Medicine from the Quebec Ministry of Health with the author as director is also gratefully acknowledged.

Received 30 January 2003; revision received 17 February 2003; accepted 26 February 2003.

Thomas Ming Swi Chang, Faculty of Medicine, McGill University, 3655 Drummond Street, Room 1006, Montreal, Quebec, Canada H3G 1Y6 (e-mail: artcell.med@mcgill.ca; website: http://www.artcell.mcgill.ca).