ARTICLE Immune Recognition of Exposed Xenoantigens on the Surface of PEGylated Bovine Red Blood Cells Sharon I. Gundersen,1 Melanie S. Kennedy,2 Andre F. Palmer3 1

Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, Indiana 2 The Ohio State University, Columbus, Ohio 3 Department of Chemical and Biomolecular Engineering, The Ohio State University, 140 West 19th Ave., Columbus, Ohio 43210; telephone: 614-292-6033; fax: 614-292-3769; e-mail: [email protected] Received 18 January 2008; revision received 18 March 2008; accepted 19 March 2008 Published online 24 March 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bit.21908

ABSTRACT: Due to potential problems that can occur during blood transfusion and increasing blood shortages, our group engineered methoxypolyethylene glycol conjugated bovine red blood cells (mPEG-bRBCs) as a potential universal oxygen therapeutic. This current work investigates the immunological properties of mPEG-bRBCs incubated with human plasma (hP) and correlates these properties to exposed Gala(1,3)Gal xenoantigens. After mPEG-bRBCs were incubated with hP, the amount of bound IgG and IgM was assessed via flow cytometry. Flow cytometry also assessed the amount of GS-IB4 bound to exposed Gala(1,3)Gal xenoantigens. The results of this study demonstrate that most hP samples strongly promote agglutination of mPEG-bRBCs regardless of the extent of mPEG surface coverage or donor blood type. IgG and IgM from hP bound strongly to mPEG-bRBCs. In general, the Gala(1,3)Gal xenoantigen remains exposed at all levels of PEG surface coverage. PEGylation did block some of the xenoantigens as the amount of exposed Gala (1,3)Gal decreased with increased mPEG surface coverage. However, this was not sufficient to prevent a strong agglutination reaction. Taken together, the results of this study indicate that the current strategy for PEGylating bRBCs is unsatisfactory for the development of immunologically silent oxygen therapeutics. Biotechnol. Bioeng. 2008;101: 337–344. ß 2008 Wiley Periodicals, Inc. KEYWORDS: bovine red blood cells; xenotransfusion; artificial blood substitutes; blood cells; PEGylation; mPEG; erythrocytes

Introduction Blood transfusion is a commonplace critical life saving procedure in modern medicine for the treatment of trauma victims and patients with a variety of diseases (Riess, 2001; Correspondence to: A.F. Palmer Contract grant sponsor: United States Public Health Service Contract grant number: HL078840; DK070862

ß 2008 Wiley Periodicals, Inc.

Scott et al., 1997; Stollings and Oyen, 2006). However, with increasing blood demand and cost per unit of blood, decreasing donor supply, and risks associated with cross matching different blood types, transfusion research has focused on developing several types of universal oxygen therapeutics ranging from hemoglobin-based oxygen carriers (HBOCs), PEGylated red blood cells (RBCs) to perfluorocarbons (Moore, 2003; Patton and Palmer, 2005; Riess, 2001; Sakai et al., 2007; Stollings and Oyen, 2006; Winslow, 2003). HBOCs possess several disadvantages, which has sparked a movement to investigate modifications of the intact RBC by conjugating methoxypolyethylene glycol (mPEG) to the surface of human RBCs (Bradley et al., 2002; Davey, 2004; Garratty, 2004; Nacharaju et al., 2005; Tan et al., 2006). Although transfusion of mPEG-hRBCs may decrease the probability of an adverse immunological reaction, it is still reliant on human blood donors, which are already in low supply (Riess, 2001; Standl, 2004; Stollings and Oyen, 2006). Our group has conjugated mPEG to the surface of bovine RBCs (bRBCs), as a plentiful source of RBCs, and has previously characterized the reaction efficiency and physical properties of bRBCs surface conjugated with 20 kDa mPEG (Gundersen and Palmer, 2007). PEGylated bRBCs are considered to be a promising alternative to hRBCs by easing the demand on human blood donors (Gundersen and Palmer, 2007). It was found that bRBCs had increased surface coverage of conjugated mPEG when reacted with higher initial concentrations of the reactive mPEG molecule. Also, PEGylated bRBCs (mPEG-bRBCs) maintained their ability to bind and release oxygen as well as deformability under shear flow at all mPEG concentrations studied. In order to assess the potential in vivo biocompatibility of mPEG-bRBCs, it is vital that mPEG-bRBCs not initiate an immediate immune reaction when transfused into the human systemic circulation.

Biotechnology and Bioengineering, Vol. 101, No. 2, October 1, 2008

337

This work will evaluate the immediate immune reaction of mPEG-bRBCs towards human antibodies derived from human plasma (hP) through a series of in vitro experiments. Initial experiments determined the ability of hP to initiate mPEG-bRBC agglutination with serologic tube techniques according to normal blood bank antibody screening procedures (Brecher, 2005). Since two of the primary immunoglobulins that recognize foreign cells are IgG and IgM, the extent of IgG and IgM binding to mPEG-bRBCs was determined via flow cytometry. Although bRBCs do not have similar ABO blood groups as hRBCs, there are still many membrane bound protein and carbohydrate xenoantigens that could initiate an immune reaction when introduced into a patient’s blood system (Goddeeris, 1998; Johnstone et al., 2004). The primary xenoantigen recognized by the human immune system is the xenoantigen Gala(1,3)Gal found on the surface of mammalian cells, with the exception of humans and ‘‘higher primates’’ (Johnstone et al., 2004; Lin et al., 1998; Milland et al., 2006; Rayat et al., 1998). Gala(1,3)Gal is an unbranched terminal disaccharide that is bound to glycoproteins and membrane glycolipids and has been a significant obstacle in xenotransfusion studies (Milland et al., 2006; Yu et al., 2005). To determine if PEGylation of bRBCs effectively masked the Gala(1,3)Gal antigen, a fluorescein isothiocyanate (FITC) conjugated lectin derived from Griffonia simplicifolia was used to detect exposed antigen. The fourth subunit of this lectin, GS-IB4, has been found to bind specifically to the antigen on mammalian cells (Lin et al., 1998).

Materials and Methods The methods presented to determine the agglutination of mPEG-bRBCs are the same serologic tube technique methods used in standard blood banking procedures (Brecher, 2005). Since this is based on a subjective visual ranking system, the amount of IgG and IgM bound to the surface of bRBCs was quantitatively determined via flow cytometry. By monitoring the reaction between mPEGbRBCs and IgG and IgM in hP, this assay can give a reasonable estimation of the biocompatibility of mPEGbRBCs. To quantify the amount of exposed Gala(1,3)Gal xenoantigen, the amount of GS-IB4 bound to mPEG-bRBCs was also determined via flow cytometry.

bRBCs and mPEG Solutions bRBCs suspended in 0.85% saline were purchased from Quad Five (Ryegate, MT). All bRBC samples were washed three times with phosphate buffered saline (PBS) (pH 7.2) then suspended to approximately 40% hematocrit (hct) in PBS. Quad Five carefully selects and quarantines their donor animals. All animals are maintained under veterinary care and diagnostic protocols, and bled in separate facilities to

338

Biotechnology and Bioengineering, Vol. 101, No. 2, October 1, 2008

maintain quarantine conditions. The cattle are routinely screened for tuberculosis, brucellosis, blue-tongue virus, anaplasmosis, leptospirosis, IBR, P13, BVD, and bovine leukosis virus (Wiley Micks, Quad Five, fax transmittal, May 16, 2005). The 20 kDa succinimidyl ester of methoxypolyethylene glycol propionic acid (SPA-mPEG) was purchased from Nektar Therapeutics (Huntsville, AL). The reactive succinimidyl ester moiety of mPEG was designed for protein surface conjugation and was recommended by Nektar for this purpose (Harris and Kozlowski, 1995; Nektar, 2006). These linear monofunctional polymers are capped on one end by a methoxy group to prevent the generation of crosslinked products. The N-hydroxysuccinimide active ester on SPA-mPEG reacts with amino acid residues with amine groups (such as lysines) producing a stable amide link, free of hydrolysis-prone ester linkages and is unlikely to elute or fragment in vivo.

Conjugation of mPEG to bRBCs bRBCs were diluted to 12% hematocrit (hct) in PBS prior to reaction. Solutions of SPA-mPEG were prepared at varying concentrations, 0 mM (buffer control) to 6 mM, in 1 mM increments, in 50 mM K2HPO4 and 105 mM NaCl (mPEG buffer). An additional unreacted control of bRBCs mixed with PBS was also evaluated to determine if reaction procedures had any effect on the bRBCs. bRBCs and SPAmPEG solutions were mixed in equal volumes, and allowed to react for 1 h at room temperature as described previously (Armstrong et al., 1997; Bradley et al., 2002; Gundersen and Palmer, 2007; Murad et al., 1999). At the end of the reaction, modified bRBCs were washed three times in PBS and resuspended to 3 mL (
12% hct).

Agglutination of mPEG-bRBCs With HP HP was acquired from donor blood samples collected in K2 EDTA 5.4 mg/3 mL tubes (BD, Franklin Lakes, NJ) at The Ohio State University Hospital’s Transfusion Service (Columbus, OH). Each of the blood samples was centrifuged for 5 min at maximum speed on a bench top centrifuge, and the plasma supernatant was pipetted to another tube for agglutination experiments, while pelleted RBCs were discarded. Forty different plasma samples were used for agglutination experiments. The following serological crossmatching techniques were preformed under the supervision of a blood bank specialist. In duplicate runs, one drop of mPEG-bRBCs or the bRBC control, at approximately 4% hct, was added to two drops of hP within a glass test tube and centrifuged for 25 s at maximum speed on a bench top centrifuge. The tubes were gently shaken in order to grade the amount of agglutination (or cell aggregation) for the immediate spin reaction (IS grade). The extent of agglutination, or agglutination reaction grade, was based on a relative scale

with 4þ corresponding to the strongest agglutination reaction, which yielded cells as one large aggregate in the test tube after gentle shaking. The weakest reaction received a grade of 1þ, in which case the cells were mostly resuspended after mixing with very small, but still visible, aggregates. Lastly, a grade of 0 agglutination was assigned for no reaction and the cells were observed to be completely resuspended within the solution. After obtaining the IS grade, the tubes were incubated for an hour at 378C in a heating block, then centrifuged again for 25 s at maximum speed on a bench top centrifuge. The tubes were gently shaken to record the 378C incubation agglutination grade (37 grade). Next, the cells were washed four times in an IEC Centra W automatic cell washer (Thermo Scientific, Waltham, MA) with 25 s centrifugation spins between washes. Two drops of anti-IgG murine monoclonal gamma-clone antibody (Immucor, Inc., Norcross, GA) were mixed with each tube of cells after washing, in order to detect bound human IgG. Cells were washed and centrifuged for 25 s. The final agglutination grade with antihuman IgG present (AHG grade) was then recorded. If the AHG (anti-human globulin) reading was negative (AHG grade of 0), one drop of Checkcell (IgG-coated pooled RBCs served as an antiglobulin control) (Immucor, Inc.) was added to the tube. This was spun and agitated again, and resuspended in order to assess the extent of agglutination thus ensuring that the experiment was performed properly and that the integrity of anti-IgG was not compromised. For hP titer experiments, hP was diluted in doubling dilutions from 1:1 to 1:512 (1:2, 1:4, 1:8, and so on). The diluted hP was used for agglutination tests as detailed above. The titer was recorded as the dilution that yielded an IS and AHG grade of 1þ or more.

Similarly to the agglutination tests, one drop of mPEGbRBCs or bRBC control, at 4% hct, was added to two drops of the diluted pHP in a glass test tube, in duplicate. The mixture of cells and pHP was centrifuged for 25 s. For the negative control, one drop of mPEG-bRBCs or control bRBCs was added to two drops isotonic saline and was carried through the same preparation steps. The tubes were gently shaken in order to resuspend the cells and incubated for 1 h at 378C in a heating block. After incubation, the cells were washed two times with 0.5 mL of PBE (ethylenediamine tetraacetic acid [EDTA] in PBS) and resuspended with 200 mL of PBE. The 200 mL suspension was separated into two 100 mL aliquots and transferred into separate flow cytometer tubes. Five microliters of goat-anti-human IgG (H&L) (Antibodies Incorporated) was added to one of the 100 mL aliquots to label the IgG antibodies. The other 100 mL aliquot was mixed with 5 mL of goat-anti-human IgM (Antibodies Incorporated) labeled IgM antibodies. Following a 30 min incubation period in the dark at room temperature, cells were washed two times with 1 mL PBE. After washing, cells were resuspended in 1 mL of PBE and loaded into the flow cytometer’s carousel. A Cytomics FC 500 Series Flow Cytometry System (Beckman Coulter, Fullerton, CA) with the Beckman Coulter’s MXP software was used to conduct fluorescent activated analysis to determine the relative fluorescence of the cells. The measured fluorescence values are reported by the software as the mean channel fluorescence of the injected cells. The results of the flow cytometry experiments are presented in terms of mean channel shift, which is the fluorescence of the cells incubated with hP subtracted from the fluorescence of the negative control.

GS-IB4 Flow Cytometry of mPEG-bRBCs Flow Cytometric Analysis of IgG and IgM Bound to mPEG-bRBCs Fluorescent-activated analysis was used to determine which of the two most likely antibodies, IgG or IgM, present within hP were the cause of the observed cell agglutination. To ensure maximum IgG and IgM binding, thus providing simple flow cytometry detection, plasma from 20 different human blood samples was collected and combined to provide one pooled plasma source (pHP). The pHP was diluted 1:2, 1:5, 1:10, and 1:20 to ensure no large aggregates of bRBCs formed in the sample, since large aggregates could arrest flow within the tubing of the flow cytometer. Goat-anti-human IgG (H&L) (Antibodies Incorporated, Davis, CA) was used to label IgG antibodies at a working dilution of 1:40, while goat-anti-human IgM (Antibodies Incorporated) were used to label IgM antibodies at a working dilution of 1:25. A negative antibody control was employed that consisted of bRBCs, from SPA-mPEG modified and unmodified (PBS and 0 mM controls) samples, suspended in isotonic saline instead of pHP.

FITC conjugated GS-IB4 lectin (referred to from now on as simply GS-IB4) was used to determine if Gala(1,3)Gal is sufficiently masked from preexisting human antibodies present in hP. GS-IB4 was purchased from Sigma (Cat #L2895, Sigma–Aldrich, St. Louis, MO) as a dry powder, and was dissolved in PBS, pH 6.8, supplemented with 0.5 mM CaCl2 (PBS-Ca) to yield a 1 mg/mL stock solution. Working dilutions of 1:10 and 1:25 were prepared from the stock solution and utilized for flow cytometry experiments. Similarly to previously described studies, mPEG-bRBCs were resuspended at a concentration of 106 cells/200 mL in PBS-Ca (Milland et al., 2006; Yu et al., 2005). Ten microliters of the working dilution of GS-IB4 was added to the cell suspension for every 200 mL of suspension, and then incubated for 1 h in the dark at room temperature. After the incubation period was over, cells were washed two times with PBS-Ca and loaded into the flow cytometer tube carousel. Beckman Coulter’s MXP software recorded the fluorescence as the mean channel fluorescence events of each bRBC passing through the detector. Again, the mean

Gundersen et al.: Immune Recognition of mPEG-bRBCs Biotechnology and Bioengineering

339

channel shift is presented in the results, and was calculated by subtracting the fluorescence of washed bRBCs incubated with saline (negative control) from the fluorescence of mPEG-bRBCs and controls incubated with the working dilution of GS-IB4. Additional negative controls using human RBCs incubated with GS-IB4 were included to ensure there was no non-specific binding of GS-IB4.

Statistical Analysis All data presented is an average of duplicate runs unless otherwise stated. The corresponding error bars represent the standard error over all measurement runs.

Results HP Agglutination and Titer The agglutination grade of mPEG-bRBCs incubated with hP derived from blood samples varied randomly. A selection of agglutination grades from six different plasma samples is shown in Figure 1, where the different colors in the bar graph designate each plasma donor. Each of the first three plasma samples were tested against bRBCs reacted with 1, 2, or 3 mM mPEG, respectively. The last three plasma samples shown in the figure were incubated with bRBCs reacted with 4, 5, or 6 mM SPA-mPEG, respectively. The values shown are averages of duplicate samples for each mPEG-bRBC, or control, as indicated. Plasma #1 (black) yielded high IS and 37 agglutination grade with control bRBCs and cells reacted with 1 and 2 mM SPA-mPEG. The AHG agglutination grade is slightly less for Plasma #3 (pink) compared to Plasma #1. Cell agglutination decreased slightly for bRBCs reacted with 3 mM mPEG. Plasma #2 (blue) yielded low agglutination for

Figure 1.

all bRBCs reacted with mPEG. All three plasmas tested against bRBCs reacted at higher mPEG concentrations, Plasma #4, 5, and 6 (red, green, pink, respectively) yielded agglutination grades mostly between 1þ and 3þ. Although there were no large aggregates (grade of 4þ), these results indicate a strong immediate immune reaction against mPEG-bRBCs that varies between individuals. Figure 2 shows the overall agglutination grade averaged over all forty plasma samples. It is evident that, on average, there is a fairly high cell agglutination grade that is independent of PEG surface coverage. Across all samples, the amount of agglutination with hP was independent of ABO blood type, and varied between different plasma samples with no apparent observable trend. Some hP samples caused no aggregation (grades of zero for IS, 37, and AHG), while others of the same blood type elicited strong aggregation (grades of 4þ). Given the scatter in the data, titer experiments on serial dilutions of plasma with saline determined if there was a strongly binding antibody that could be influencing the results. The titer results are shown in Table I, which lists the titers for two of the plasma donor samples as the inverse of the highest dilution eliciting an agglutination grade of 1þ. The titer results obtained are low, with the highest one being a eightfold plasma dilution with Plasma #1. The agglutination of bRBCs with Plasma #2 dropped off quickly as the highest titer obtained is a twofold dilution.

mPEG-bRBC Bound IgG and IgM As described in the methods section, the amount of IgG or IgM bound to mPEG-bRBCs was expressed as the mean channel shift of the fluorescence. The mean channel shift for mPEG-bRBC labeled with anti-IgG was randomly scattered, as shown in Figure 3. The numbers given above the bars are

Average of duplicate hP agglutination grades from six random plasma donors at all concentrations of SPA-mPEG. Each plasma source is represented by its own color and the solid bars represent IS agglutination grades, the hollow bars represent 37 agglutination grades and the bars with diagonal lines represent AHG agglutination grades. The controls refer to the average agglutination grade of washed bRBCs reacted with 0 mM mPEG and the PBS reaction controls. [Color figure can be seen in the online version of this article, available at www.interscience.wiley.com.]

340

Biotechnology and Bioengineering, Vol. 101, No. 2, October 1, 2008

Figure 2. Overall agglutination grade averaged over all forty plasma donors. The solid bars represent IS agglutination grade, the bars with waved lines represent 37 agglutination grades and the bars with diagonal hashes represent AHG agglutination grades averaged over all 40 hP sources.

Figure 3. Average fluorescence from IgG bound to the surface of mPEG-bRBCs as represented by the mean channel shift generated from the anti-IgG label. Cells were reacted at the specified concentration of SPA-mPEG. PBS represents the control group of cells diluted with PBS instead of being reacted with mPEG, and washed bRBCs represent fresh cells washed in PBS. The numbers above the cells indicate the IS agglutination grades of bovine RBCs with pHP.

the IS grades for that dilution of pHP with the specified mPEG-bRBCs. bRBCs reacted with 6 mM SPA-mPEG had average agglutination grades of 2.5þ, 3þ, and 0.5þ at pHP dilutions of 1:5, 1:10, and 1:20 respectively. The average mean channel shift was 1.2 at the 1:5 pHP dilution, and 1 for 1:20 dilution. Cells reacted with 3 mM SPA-mPEG displayed low average agglutination grades, but higher mean channel fluorescence. Even though the average agglutination decreased with pHP dilution, the flow cytometer results show that IgG still bound to mPEG-bRBCs for all pHP dilutions. The mean channel shift for IgM binding is shown in Figure 4 with the serologic agglutination grades notated above the bars. For both reaction samples, 3 and 6 mM SPA-mPEG, and controls, the mean channel shift decreased in correlation with the increased pHP dilutions. Also, the average agglutination grades decreased with increasing pHP dilutions. bRBCs reacted with 6 mM SPA-mPEG exhibited a mean channel shift of
2 at a pHP dilution of 1:5 and the mean channel shift decreased to 0.6 at the highest pHP dilution of 1:20. However, contrary to what would be expected, PEGylation appears to have increased the amount of IgM binding. At the 1:20 pHP dilution, the average mean channel shift is 0.39 for the PBS reaction control and the 6 mM modified bRBCs exhibited an average mean channel shift of 0.6.

Gala(1,3)Gal Exposure on the Surface of mPEG-bRBCs

Table I. of hP.

Figure 5 shows the mean channel shift from GS-IB4 bound to the surface of mPEG-bRBCs. Frame A of the figure gives the mean channel shift for a 1:10 working dilution of GS-IB4, while frame B is for a 1:25 working dilution of GS-IB4. The mean channel shift of human RBCs incubated with GS-IB4 controls were omitted from the figures because it was consistently less than one, thus there was no nonspecific binding of GS-IB4 to human RBCs. PEGylated bRBCs exhibited decreasing GS-IB4 fluorescence with increasing mPEG surface coverage. However, strong GS-IB4 binding indicates that the Gala(1,3)Gal antigen is still exposed. The mean channel fluorescence of 6 mM modified bRBCs was about half of that compared to

Titer results of mPEG-bRBCs incubated with serial dilutions

Plasma #1

Plasma #2

bRBC sample

IS

AHG

IS

AHG

3 mM 1 mM 0 mM PBS Washed bRBCs

8 4 8 8 8

8 4 8 8 8

2 2 2 2 2

2 2 2 2 2

Figure 4. Average fluorescence from IgM bound to the surface of mPEG-bRBC as represented by the mean channel shift generated from the anti-IgM label. PBS represents the control group of cells diluted with PBS instead of being reacted with mPEG, and washed bRBCs represents fresh cells washed in PBS. The numbers above the cells indicate the IS agglutination grades of bovine RBCs with pHP.

Gundersen et al.: Immune Recognition of mPEG-bRBCs Biotechnology and Bioengineering

341

Figure 5.

Average fluorescence from GS-IB4 bound to the exposed Gala(1,3)Gal xenoantigens on the surface of mPEG-bRBCs as represented by the mean channel shift. The mean channel shift was generated from working dilutions of (A) 1/10 and (B) 1/25 GS-IB4 bound to bRBCs reacted with various concentrations of SPA-mPEG. Also included are two controls consisting of: 0 mM SPA-mPEG and the PBS unreacted control.

bRBCs reacted with 1 mM SPA-mPEG. bRBCs reacted with 1 mM SPA-mPEG exhibited a mean channel shift of 49 with the 1:10 working dilution of GS-IB4, and a mean channel shift of 27 for 6 mM modified cells. The 1:25 dilution of GS-IB4 yielded a mean channel shift of 29.5 for 1 mM modified bRBCs, and 12 for 6 mM modified bRBCs.

Discussion For any oxygen therapeutic to be physiologically accepted, it must be immunologically silent. The in vitro experiments preformed in this work determined the potential immediate immune reaction against mPEG-bRBCs. The first set of experiments evaluated mPEG-bRBCs against hP similarly to standard blood bank cross-matching procedures. Since the reaction with hP seemed to be nonspecific, we performed flow cytometry analysis against IgG and IgM to determine which antibody could be initiating the immune reaction. We assumed that IgM caused the agglutination reflected in the IS grade, and IgG caused the agglutination for the AHG grade. Subsequently, flow cytometry with mPEG-bRBCs

342

Biotechnology and Bioengineering, Vol. 101, No. 2, October 1, 2008

incubated with a labeled GS-IB4 lectin against the Gala(1,3)Gal epitope determined if the primary mammalian xenoantigen was exposed after PEGylation with SPA-mPEG. The results from the serologic experiments found PEGylated bRBCs to be immunologically reactive against many different plasma samples regardless of mPEG surface coverage. No discernable trend was observed in terms of which plasma samples caused strong agglutination of mPEG-bRBCs and which were unreactive. The titer data was used to determine if there was an antibody in some of the plasma population that bound very strongly to mPEGbRBCs. However, the hP agglutination titer grade dropped off rather quickly, suggesting there is not a high titer antibody in any of the plasmas and suggesting insufficient coverage of the bRBCs antigens in general by mPEG. For a more quantitative analysis of which antibodies bind to mPEG-bRBCs, the amount of IgG and IgM from pHP bound to the surface of mPEG-bRBCs was determined via flow cytometry analysis. IgG was found to bind strongly to mPEG-bRBCs at all dilutions, while IgMs binding decreased as the plasma was diluted. Since IgG, with AHG antibody, is primarily responsible for the agglutination of cells, this would explain the strong average agglutination observed with the serologic reactions with pHP. The results from the serologic and flow cytometry experiments with pHP were discouraging and confusing, since there was still strong aggregation and IgG and IgM binding independent of mPEG coverage and blood type. The increased mean channel shift of IgM binding, seen in Figure 4, could possibly be due to the mPEG stabilizing the IgM on the bRBCs membrane once it has a chance to interact with an antigen. This would also agree with the increased average agglutination shown in Figure 2. While the investigation of this possibility and mechanism would be very valuable and intriguing, it is beyond the scope of the current study. To determine a possible explanation of these results, the last set of experiments attempted to determine if the Gala(1,3)Gal antigen was still exposed after PEGylation. The Gala(1,3)Gal antigen is a primary xenoantigen found on most mammalian cells, and can initiate a strong immune reaction with a naturally occurring antibody in the human immune system via IgM (Lin et al., 1998; Sandrin et al., 1993). The decrease in the mean channel shift after incubation with GS-IB4 shows that an increase in initial mPEG concentration decreases the amount of exposed Gala(1,3)Gal antigen. This is in agreement with the results presented in our previous paper that shows that SPA-mPEG covalently binds to the surface of bRBCs in a dose-dependent manner. Therefore, the SPA-mPEG works as intended by camouflaging Gala(1,3)Gal antigens. The extent of which is dependent on the extent of PEGylation coverage. However, these antigens are not sufficiently covered as GS-IB4 can still bind and PEGylation did not completely prevent immune system reaction as shown by IgG and IgM binding and agglutination with NHS.

As anti-Gala(1,3)Gal antibodies are naturally occurring and acquired through various environmental factors, similar to ABO antibodies, these would have a comparable clinical importance as the ABO system. Thus, when using the same cross-matching techniques used for ABO matching, there would be some variation in reactivity between individuals as was found in the reported results (Figs. 1 and 2). The GS-IB4 experiments show that Gala(1,3)Gal is still exposed at all degrees of mPEG surface coverage. In an individual with high levels of anti-Gala(1,3)Gal, there would be a strong serological reactivity, and vise versa for an individual with low levels of anti-Gala(1,3)Gal. This individual variability could account for the varying agglutination found from the serological results. The mean channel fluorescence of the IgG and IgM binding assays is much lower than that of GS-IB4 binding assays. As explained by Lin et al. (1998) the preferential binding sites of IgM and GS-IB4 on the Gala(1,3)Gal antigen could be different thereby causing varying fluorescence levels. Also, the structural differences and different initial concentrations of the antibody versus the lectin could contribute to the ability of it to bind to the epitope. Since human IgM and GS-IB4 will interact differently with the Gala(1,3)Gal antigen, the GS-IB4 assay cannot be used to directly explain IgM interactions with the exposed antigens on the mPEG-bRBC surface, but concludes that Gala(1,3)Gal remains exposed (depending on the degree of mPEG surface coverage) and is available for IgM to bind.

Conclusions PEGylation of bRBCs with SPA-mPEG was not able to increase the bRBC’s biocompatibility by preventing agglutination with hP, or IgG and IgM binding. SPA-mPEG also appeared to fail in completely camouflaging the Gala(1,3)Gal antigen, as shown by the high mean channel shift of GS-IB4. Although bRBC modification with increasing SPA-mPEG did show some promise in camouflaging the Gala(1,3)Gal xenoantigens, mPEG-bRBCs are still immunoreactive with an immediate immune reaction when mixed with hP and had strong IgG and IgM binding. With these results in mind, continuing studies with SPA-mPEG conjugated to bRBCs are not recommended, as SPA-mPEG was inadequate in camouflaging the surface antigens of bRBCs. The Gala(1,3)Gal antigen remains the primary obstacle to xenotransfusion and will have to be effectively covered or removed from bRBCs for transfusion into humans. While there are other possible strategies for immunocamouflaging bRBCs with mPEGs (branched PEGs and replacement of the methyl end of PEG with a larger unreactive moiety), it is unlikely that mPEGs will sufficiently cover the entire bRBCs as they are only reactive with membrane bound surface proteins and not lipids. A more reliable method to create a universal oxygen therapeutic from bRBCs would be to prevent Gala(1,3)Gal synthesis during erythrocyte

development. Small interfering RNAs (siRNAs) could be used to effectively knock out the Gala(1,3)Gal synthesis protein (a1,3galactosyltransferase) and prevent the initial formation of the Gala(1,3)Gal antigen. This research was supported by United States Public Health Service grants HL078840 and DK070862 to A.F.P. We would like to thank Dr. Gerard Lozanski at The Ohio State University’s Transfusion Service for his assistance and advice with these studies and the Transfusion Service staff for all of their help.

References Armstrong JK, Meiselman HJ, Fisher TC. 1997. Covalent binding of poly(ethylene glycol) (PEG) to the surface of red blood cells inhibits aggregation and reduces low shear blood viscosity. Am J Hematol 56: 26–28. Bradley AJ, Murad KL, Regan KL, Scott MD. 2002. Biophysical consequences of link chemistry and polymer size on stealth erythrocytes: Size does matter. Biochim Biophys Acta 1561:147–158. Brecher ME, editor. 2005. Technical manual of the American Association of Blood Banks. 15th edition. Bethesda, MD: American Association of Blood Banks. Davey RJ. 2004. Overcoming blood shortages: Red blood cell and platelet substitutes and membrane modifications. Current Hematol Rep 3(2):92–96. Garratty G. 2004. Progress in modulating the RBC membrane to produce transfusable universal/stealth donor RBCs. Transfus Med Rev 18(4): 245–256. Goddeeris BM. 1998. Immunology of cattle. In: Pastoret P-P, editor. Handbook of vertebrate immunology. San Diego: Academic Press. Gundersen SI, Palmer AF. 2007. Conjugation of methoxypolyethylene glycol to the surface of bovine red blood cells. Biotechnol Bioeng 96(6):1199–1210. Harris JM, Kozlowski A. Shearwater Polymers, Inc, assignee. 1995 9/30/ 1997. Poly(ethylene glycol) and Related Polymers Monosubstituted With Propionic or Butanoic Acids and Functional Derivatives Thereof for Biotechnical Applications patent 5672662. Johnstone JE, MacLaren LA, Doucet J, McAlister VC. 2004. In vitro studies regarding the feasibility of bovine erythrocyte xenotransfusion. Xenotransplantation 11(1):11–17. Lin SS, Parker W, Everett ML, Platt JL. 1998. Differential recognition by proteins of alpha-galactosyl residues on endothelial cell surfaces. Glycobiology 8(5):433–443. Milland J, Christiansen D, Lazarus BD, Taylor SG, Xing PX, Sandrin MS. 2006. The molecular basis for Gal alpha(1,3)Gal expression in animals with a deletion of the alpha 1,3galactosyltransferase gene. J Immunol 176(4):2448–2454. Moore EE. 2003. Blood substitutes: The future is now. J Am Coll Surg 196(1):1–17. Murad KL, Mahany KL, Brugnara C, Kuypers FA, Eaton JW, Scott MD. 1999. Structural and functional consequeces of antigenic modulation of red blood cells with methoxypoly(ethylene glycol). Blood 93(6):2121– 2127. Nacharaju P, Boctor FN, Manjula BN, Acharya SA. 2005. Surface decoration of red blood cells with maleimidophenyl-polyethylene glycol facilitated by thiolation with iminothiolane: An approach to mask A, B, and D antigens to generate universal red blood cells. Transfusion 45(3):374– 383. Nektar. 2006. Nektar Therapeutics Catalog 2006. Huntsville, AL: Nektar Therapeutics Catalog. Patton JN, Palmer AF. 2005. Engineering temperature-sensitive hydrogel nanoparticles entrapping hemoglobin as a novel type of oxygen carrier. Biomacromolecules 6(4):2204–2212.

Gundersen et al.: Immune Recognition of mPEG-bRBCs Biotechnology and Bioengineering

343

Rayat GR, Rajotte RV, Elliott JF, Korbutt GS. 1998. Expression of Gal alpha(1,3)Gal on neonatal porcine islet beta-cells and susceptibility to human antibody/complement lysis. Diabetes 47(9):1406–1411. Riess JG. 2001. Oxygen Carriers (‘‘Blood Substitutes’’)—Raison d’Etre, Chemistry, and Some Physiology. Chem Rev 101(9):2797–2919. Sakai H, Tsuchida E, Horinouchi H, Kobayashi K. 2007. One-year observation of Wistar rats after intravenous infusion of hemoglobin-vesicles (artificial oxygen carriers). Artif Cells Blood Substit Biotechnol 35(1): 81–91. Sandrin MS, Vaughan HA, Dabkowski PL, McKenzie IFC. 1993. Anti-Pig IgM antibodies in human serum react predominantly with gal(alpha-13)gal epitopes. Proc Natl Acad Sci USA 90(23):11391–11395. Scott MG, Kucik DF, Goodnough LT, Monk TG. 1997. Blood substitutes: Evolution and future applications. Clin Chem 43(9):1724–1731.

344

Biotechnology and Bioengineering, Vol. 101, No. 2, October 1, 2008

Standl T. 2004. Artificial oxygen carriers: Hemoglobin-based oxygen carriers—Current status 2004. Transfus Med Hemother 31(4):262– 268. Stollings JL, Oyen LJ. 2006. Oxygen therapeutics: Oxygen delivery without blood. Pharmacotherapy 26(10):1453–1464. Tan YX, Qiu Y, Xu H, Ji SP, Li SB, Gong F, Zhang Y. 2006. Decreased immunorejection in unmatched blood transfusions by attachment of methoxypolyethylene glycol on human red blood cells and the effect on D antigen. Transfusion 46(12):2122–2127. Winslow RM. 2003. Current status of blood substitute research: Towards a new paradigm. J Intern Med 253:508–517. Yu LY, Miao HS, Guo LH. 2005. Effect of RNA interference on Gal alpha 1,3 Gal expression in PIEC cells. DNA Cell Biol 24(4):235– 243.