The most important obstacle in organ transplantation
and blood transfusion is the immunological response
to the transplanted tissues and cells. Red blood cells
(RBCs) are the simplest transfused allogeneic cells (1).
The membrane of RBCs architecturally contains complex structures that carry defied polymorphic epitopes, identifid serologically as blood group antigens.
The fist reason of RBC rejection is the immune reaction against the transfusion of mismatched RBCs (2).
Following Landsteiner’s discovery of ABO blood group
antigens, the blood transfusion was changed from a
highly risky surgical procedure to a mundane yet clinically crucial procedure (2). Thus, it is essential to identify the appropriate donors, with ABO and D (Rhesus)
blood typing, in all transfusions.
However, it is diffilt to identify appropriate blood
donors (3-5). The chronically transfused patients, who
suffr from thalassemia and sickle cell disease encounter the problems of alloimmunization (6, 7). Immunoprotection of RBCs can be achieved by covalent attachment of methoxy polyethylene glycol (mPEG) to the
proteins expressed on the cell membrane (PEGylation)
(8, 9). Owing to the mPEG flxibility, a large steric exclusion volume surrounding the RBCs inhibits the binding
of large molecules (e.g. antibodies) or other cells (e.g.
immune cells) to them (8). Several reactive derivatives
of mPEG, such as mPEG-succinimidyl valerate (mPEGSVA) and mPEG-succinimidyl carbonate (mPEG-SC), have
been employed to covalently attach mPEG to the surface
of RBCs (5, 8, 10-15).
Recently, Wang et al. have investigated the effct of PEGylation using mPEG-SVA on reducing the antigenic recognition of blood group antigens expressed by normal
human RBCs (5). Lee and Scott used mPEG-SC and mPEGSVA for PEGylation of polystyrene beads to evaluate the
charge camouflge and protein adsorption (8).
Despite the development of this strategy, optimization
of the reaction conditions for simultaneous camouflge
of the major and minor RBC antigens has not been thoroughly investigated. In our previous works, we have investigated the attachment of activated mPEG to RBCs in
order to optimize the reaction conditions for protection
of major antigens on RBCs (16, 17).
Ongoing work on the modifiation of RBC suggests
that PEGylation may be a practical method for decreasing the risk of transfusional alloimmunization, and
treatment of alloimmunized patients (5, 8).
In the present study, RBCs were coated with two derivatives (mPEG-SC and mPEG-SVA) in two diffrent molecular weights to investigate simultaneous camouflge of
the major and minor antigens (A, Rh (D) and Kell antigens). Then, the effct of polymer size and linker chemistry on camouflge of these antigens was assessed.
Additionally, central composite design (CCD) methods
were used to fid the optimal conditions for PEGylation
of RBCs. Finally, the immunocamouflge and immunogenic potential of unmodifid and PEGylated RBCs were
evaluated in a murine transfusion model.
3. Materials and Methods
In our research, we used several ingredients from different sources, which are listed as follows: mPEG (10 and
20 kDa) purchased from Sigma; mPEG-SVA (10 and 20
kDa) obtained from Laysan Bio, Inc. (Arab, AL, USA); N, N′-
Disuccinimidyl carbonate purchased from Aldrich (U.S.A);
sodium chloride, isopropanol, dry diethylether, 4-(dimethylamino) pyridine, cyclohexane and dioxin obtained from
Merck (Germany); packed Rh-positive RBCs (A+-Kell+) obtained from Iranian Blood Transfusion Organization; antiD and anti-A sera purchased from Cinna Gen Inc. (Iran);
and anti-Kell sera obtained from Diagast (France).
3.2. Preparation of Samples
3.2.1. mPEG-SC Derivatization
We performed derivatization of 10 and 20 kDa mPEGs
with SC using a modifid method described by Miron
and Wilchek (18). Briefl, 1 mmol of vacuum dried mPEG
(overnight at 80°C) was dissolved in 100 mL dioxan and
warmed slightly to help the dissolution and in the following step, the solution was cooled at room temperature.
Then, threefold excess of N, N′-Disuccinimidyl carbonate was slurried in 10 mL of dry acetone and added to
the above solution. Next, 6 -mmol of 4-(dimethylamino)
pyridine was dissolved in 20 mL of dry acetone, added to
the reaction mixture, and is allowed to proceed at room
temperature under dry nitrogen atmosphere for 24 h.
The mixture was fitered to eliminate any solid sediment and then, was poured slowly under high shear into
the diethyl ether to precipitate the derivatized mPEG.
The precipitate was then refitered and washed with diethyl ether, resuspended in isopropanol to remove any
unreacted N, N′-disuccinimidyl carbonate, fitered and
then washed with isopropanol. Next, the precipitate
was resuspended in cyclohexane, fitered and washed
with cyclohexane. Finally, the product (mPEG-SC) was
dried under a stream of dry nitrogen at room temperature for 12 h.
3.3. Characterization of mPEG-SC
A Fourier transform infrared (FTIR) spectra of mPEG-SC
and mPEG samples were recorded using an FTIR spectrophotometer (FTIR, Perkinelmer) at room temperature
over the frequency range of 450–4000 cm-1 to verify the
derivatization of mPEG-SC.
3.4. RBC Coating With Activated mPEGs
Packed Rh-positive RBCs (A+-Kell+) were suspended at
10% hematocrit in phosphate buffred saline (PBS) solution. Fresh cold solutions of the derivatized polymer
(mPEG-SC or mPEG- SVA, separately) were prepared in
PEG-buffr (50 mmol.L-1 dibasic potassium phosphate,
105 mmol.L-1 NaCl) to produce polymer concentrations
of 7–22 mg.mL-1 with diffrent molecular weights and pH,
according to the experimental design (presenting in the
following sections), and immediately added to the RBC
Derivatization reactions were carried out with gentle
mixing at room temperature for 45 min. After washing
three times with isotonic PBS (pH 7.4) at 200 × g for 10
min, the packed RBCs were prepared for assessment of
the polymer coating quality (6).
3.5. RBC Agglutination by Antisera and Cell
RBC agglutination by anti-D, A, and Kell sera was used
to assess the effctiveness of RBC PEGylation. Four hundred microliters of uncoated or PEGylated RBC suspension (6% hematocrit in an isotonic saline buffr), was
separately mixed with a solution of anti-D, A, or Kell in
PBS with a known concentration (antisera/PBS: 1/3) and
then, incubated with a gentle mixing at room temperature for 30 min.
Next, the RBCs were centrifuged at 200 ×g for 1 min
(19). One microliter of the pellet was resuspended in 1
mL of PBS. Next, by using a dye exclusion test with Trypan blue and light microscopic system (Nikon, E200),non-agglutinated viable free cells were counted on a hemocytometer (improved Neubauer ruling). Under this
condition, the viable cells remained uncolored and the
dead cells would turn blue. The higher number of viable
free cells, the greater would be the effctiveness of RBC
3.6. Scanning Electron Microscopy (SEM)
The morphology of both control and mPEG-derivatized RBCs was assessed by a scanning electron microscope (SEM 3200, China) to ensure that they are structurally appropriate for transfusion. For preparation of
the samples, the procedure explained by Kayden and
Bessis was followed (20).
3.7. Statistical Designs
The fist step identifis the most important factors in
the PEGylation reaction on human RBCs with mPEG-SVA
and mPEG-SC. Selection of these variables and their levels (polymer concentration 7–22 mg.mL-1, polymer molecular weight 10–20 kDa, and reaction pH 8.0–9.2) were
based on our preliminary experiments (not presented
here), and previous studies in the literature for RBC PEGylation (5, 8, 11, 16, 17, 19).
3.8. Central Composite Design
We employed Central Composite Design (CCD) to investigate the impact of multiple variables on the inhibition of agglutination and antibody binding (as a single
response), to estimate the contribution of individual
factors, and fially to obtain the optimal conditions for
The total number of experiments in a CCD is calculated by the formula; 2k + 2k + x0, where k and x0 are the
number of factors and central points, respectively (21).
This methodology allows the modeling of a second-order polynomial equation that defies the process. The
nonlinear model obtained with statistical Design Expert
(version 7.0.0, Stat-Ease, Inc.) is as follows:
Y = b
0+∑biXi + ∑biiX2i + ∑bijXij (1),
where Y is the predicted response, Xi shows the independent variable, b0 is an intercept, bi is the linear effect, bii is the squared effct, and bij shows the interaction effct.
To obtain the optimal conditions for camouflging
Kell antigen on the surface of the human RBCs by mPEGSVA, a CCD for three variables at fie levels was designed.
The variables, their corresponding levels and the designed experiments are presented in Table 1.
Based on the obtained results, and previous studies
in the literature for RBC PEGylation, pH of the reaction
is not signifiantly effctive on antigen coating. Thus,
other experiments were carried out at constant pH and
another CCD for two variables at fie levels was designed
to investigate the effct of mPEG-SVA on camouflging
of Rh (D) and A antigens, (Table 2).
All of the responses were presented as the average
number of free cells per 1 mL suspension, obtained by
cell counting method. Furthermore, a CCD with two
variables at fie levels was designed to fid the optimal
conditions for camouflging Kell, Rh (D) and A antigens
with mPEG-SC,. Table 3 presents the variables, their corresponding values and the designed experiments.
3.9. In Vivo RBC Immunogenicity
For in vivo studies, two genetically distinct mouse
strains (C57Bl/6 and BALB/c) were selected to investigate
the alloimmunization (5). The selected mice were adult
male aged 2.5–3.0 months (25–30 g in weight). The RBCs
of C57Bl/6 mice as the donors were collected via cardiac
puncture. Mouse RBCs were washed three times in PBS
(pH 7.4) and resuspended in a 10% hematocrit.
The RBC suspension was derivatized by mPEG-SVA at
the concentration of 10 mg.mL-1 for 30 min at 4°C in PBS
(pH 8.6) as previously described. After derivatization,
the cells were washed three times in PBS and resuspended in a 30% hematocrit before transfusion.
The mice were divided into three groups; a positive
control group which are transfused with unmodifid
RBCs; the negative control group without transfusion;
and the PEGylated group which are transfused with PEGylated RBCs.
Recipient mice (BALB/c) were transfused through the
tail vein with 300 μL of non-PEGylated or PEGylated
RBCs. After 24 hours, two mice in each group (the negative control, the positive control and the PEGylated
group) were sacrified, and their serum was collected
for biochemical analysis. The levels of alanine amino
transferase (ALT), aspartate amino transferase (AST), lactate dehydrogenase (LDH), and total bilirubin (BLT) in
the mouse were measured by an auto analyzer (Selectra
XL, The Netherlands).
4.1. Characterization of mPEG-SC
Figure 1 shows the FTIR spectra of mPEG and mPEG-SC.
The FTIR spectra of mPEG-SC (curve b and c), in comparison with the FTIR spectra of mPEG (curve a), has new
peaks appearing at 1744.16 and 1743.96 cm‾ˡ, corresponding to the succinimide group in mPEG-SC of 10 and 20
kDa molecular weight, respectively, which approves the
successful polymer derivatization.
4.2. Central Composite Design
We used Central Composite Design (CCD) to fid the
suitable variables on camouflging the human RBCs.
Tables 1 –3 presents the results of CCD experiments consisted of the predicted and experimental data after PEGylation with mPEG-SVA and mPEG-SC for evaluation of
the effcts of independent variables on masking the minor and major antigens on human RBCs. All the responses are reported as the number of viable free cells per 1
mL suspension. An increase in the number of free cells
for PEGylated versus control (uncoated) RBCs, indicates
that PEG-RBCs have been protected against agglutination
To calculate the response, Y, we used a mathematical
relationship in the form of a second-order polynomial
equation. Table 4 shows the coeffients of this model
(Equation 1) for camouflging of Kell, Rh(D), and A antigens on human RBCs using mPEG-SVA and mPEG-SC.
These polynomial equations show the quantitative effect of the process variables and their interactions on the
response Y. The values of coeffients of X1, X2, and X3 are
related to the effct of these variables on the response.
A positive value in Table 4 represents a favorable effct,
while a negative value indicates an adverse effct.
In this case, X1 and X2 have the main effcts on the response. The values of X were substituted in the equation to
obtain the theoretical values of Y. The analysis of variance
indicated that X
1 is more signifiant than X2 and then X3.
Model F values presented in Table 4 imply that the
models are signifiant. The small diffrence between the
predicted R-squared and adjusted R-squared of the models indicates that they are in good agreement with each
other (Table 5).
Figures 2 and 3 present the main interaction effct on
the number of free RBCs, as an indicator of the extent of
PEGylation. As shown in these Figures,polymer concentration and molecular weight (size) were important in
preventing the recognition of A, Rh(D), and Kell antigens
on RBCs. In addition, it shows that higher concentration
and molecular weight of these polymers have effctively
increased the number of free RBCs (P <.05). Higher polymer concentrations even at low molecular weights are
more effctive and yield the maximum camouflge of
minor and major antigens on the RBCs.
The red zones with the maximum number of free cells
in Figures 2 and 3 shows that the combination of 20 and
10 kDa polymers with a high percentage of mPEG20 is signifiantly more effctive than 20 kDa polymers alone.
In addition, Figure 2 A indicates no signifiant diffrence between pH values of 8.0 and 9.2 after derivatization. This fiding shows that pH value of the reaction
medium is not signifiantly effctive in preventing the
antigenic recognition of Kell antigens on RBCs. Thus, 8.6
were selected as the optimal pH value.
Based on these Figures and Models, Tables 6 and 7 present the optimal conditions of PEGylation (pH of the reaction medium, polymer concentration, and percentage of
mPEG20 in the polymer mixture) obtained by using free
cell counting. The validation experiments were carried
out under optimized conditions, showed a good agreement with the statistically predicted values, and confimed the model’s authenticity.
As shown in Figures 2, 3 also Tables 6, 7, signifiant diffrences are demonstrated between the linker chemistries
with respect to effctive camouflge of blood antigens.
Therefore, mPEG-SVA is more effient than mPEG-SC at all
the grafting concentrations and polymer sizes. Furthermore, these results showed that Rh (D) and Kell antigens
are effiently camouflged by mPEG. In comparison to A
or Rh (D) antigens, Kell antigen as a minor antigen, is effectively more camouflged, as shown by decreasing in
4.3. Scanning Electron Microscopy
Figures 4 and 5 present the morphology of control (uncoated) and PEG-coated human RBCs (at 17.21 and 14.50
mg.mL-1 of mPEG-SVA) under optimal conditions (percentage of mPEG20 = 91.2% and pH 8.6) . Selection of these concentrations was based on the optimal condition and our
pervious results (16, 17).
These Figures display that at polymer concentrations
higher than 14.5 mg.mL-1, the human RBCs will deform
by increasing RBC echinocytosis. It showed that polymer
concentration of mg.mL-1 is a useful level for PEGylation of
RBCs by 91.2% mPEG20-SVA in the polymer mixture. Echinocytes are not deformable and hence tend to get trapped in
the microcirculation (22). Thus, the optimal concentration
of 17.21 mg.mL-1 (obtained by cell counting method), cannot be recommended as an optimal polymer concentration due to the loss of discocytic morphology.
These results are similar to our previous fiding for
PEGylation of RBCs by mPEG, activated by cyanuric
chloride and succinimidyl carbonate (16, 17). Morever, the
cell counting method was used to investigate this lower
polymer concentration (14.50 mg.mL-1) on camouflging
of Kell, Rh (D), and A antigens. In this regard, the numbers
of obtained viable free cells were 3.2, 2.8, and 2.4 × 108
per 1 mL suspension for Kell, Rh (D), and A antigens,
respectively. Thus, there was no signifiant diffrence
between these results and the results given in Table 6 for
polymer concentration of 17.21 mg.mL-1.
The morphology of PEGylated murine RBCs (at 14.5 and
10.0 mg.mL-1 of mPEG-SVA) under the optimal conditions
(percentage of mPEG20 = 91.2% and pH 8.6) are shown in
4.4. In Vivo RBC Immunogenicity
We evaluated the effct of PEGylation on in vivo RBC
survival. In this regard, the PEGylated RBCs and unmodifid RBCs (positive control) obtained from donor mice
(C57Bl/6) at the optimal conditions according to the SEM
results, were transplanted intra the tail of the recipient
Twenty-four hours after transplantation, serum bichemical parameters of mice were measured. Figure 7 shows the
changes in the levels of serum bichemical parameters
(ALT, AST, LDH and BLT) of the recipients and negative control mice after transplantation. The serum parameters of
the modifid RBCs group remained at the normal levels.
This happened while the increased values of these parameters were detected in positive control mice that were signifiantly diffrent from negative control.
Alloimmunization against non-ABO/Rh (D) blood group
antigens is a common result of blood transfusion. The immunocamouflging of RBCs using grafting of mPEG on
the RBC membrane is a unique approach in preventing
this alloimmunization (12, 23).
As previously demonstrated, covalent attachment of PEG
to RBC membrane greatly reduces antigenicity and immunogenicity of RBCs by the inhibition of the antigen detection via antibodies and also by lowering antibody formation in response to allogeneic and xenogeneic transfusion
in murine models (5, 17, 23).
This study was performed to optimize the reaction conditions for covalent attachment of mPEG activated with
SVA and SC to the RBCs. The choice of linker chemistry is
critical for effient coupling in aqueous solutions. SVA
and SC were used to react with primary amines (i.e. lysine)
located on the cell membrane proteins.
mPEG-SVA forms an amide linkage while mPEG-SC employs a carbamate bond to the amine groups (8). The
choice of factors and their levels for optimization were
based on the range given in the literature. Both mPEG-SVA
and mPEG-SC were found to be suffiently reactive to produce extensively coated proteins under moderate conditions, showing the highest reactivity around pH 9.3 (24).
Other researchers utilized an elevated pH (8.0–9.2) for
grafting an activated PEG to RBC (5-8, 25, 26). Fisher in his
review displayed that at polymer concentrations lower
than 1 mM, 5 kDa activated mPEG was not very effctive
for coating the antigens (27). Higher concentrations up
to 50 mg.mL-1 had been also employed, but some abnormalities in the morphology of RBCs were observed (25).
The results presented in Figures 2, 3 show that free cell
number would go up by increasing the PEG concentration and length due to the blood group antigen masking and the corresponding decrease of agglutination
between RBCs. Increasing the percentage of mPEG20 in
the mixture of mPEG10 and mPEG
20 caused an initial increase in the free cells with a maximum point at 91.2% of
The immunocamouflging of blood group antigens is
directly proportional to the concentration and length
of the grafted polymer. This covering was proportional
to the hydrodynamic thickness of the polymer layer and
was best achieved by long-chain polymers (8). The Flory
radii (RF: root mean square of end to end length of the
polymer chain; radius of gyration) of grafted polymers
can be estimated as R
F = aN(3/5) (a = 3.5 Å, N = the number
of monomers) (8). So, the calculated Rf (in nm) for 10- and
20-kDa polymers are 9.2 and 13.8 nm, respectively. Longer
polymers would result in improved efficy of cell surface charges and antigen masking (Figures 2 B–3C).
The steric exclusion effct of the attached polymer
chains primarily inhibits protein adsorption (or antigenantibody recognition) (8). This effct is maximized when
higher density of polymer chains with small separation
between the chains are grafted to the cell surface. However, it is diffilt to achieve the high density grafting with
long polymers that have a large gyration radius. In contrast, the long polymers sterically prevent binding onto
the surface, but this inhibition is not against the short
polymers (8). Therefore, higher density can be achieved
by using long and short polymers together, as shown in
Figures 2 B– 3 C.
As mentioned earlier, pH > 7.0 is required for reaction
of the activated mPEG with primary amine residues. Also,
it is well known that a non-physiological pH can be very
damaging to the cells. Thus, a mild pH of 8.6 was found as
the optimal condition. The optimal condition, at which
the number of free cells is maximal (as presented in Tables 6 and 7) was determined by counting the free cells. It
can be observed that the optimal conditions are in agreement with those deduced from Figures 2 and 3.
Linker chemistry is an effctive factor in the efficy of
immunocamouflging by the polymer (8, 12). mPEG-SVA
more effctively camouflged the blood group antigens,
as noted by the reduction in antibody adsorption and agglutination formation (Figures 2 and 3 and Tables 6 and
7). This diffrence is relative to the extended hydrolysis
half-life of mPEG-SVA (33.6 min) in comparison to mPEGSC (20.4 min) in aqueous solution (8).
As shown in this study, PEGylation resulted in a decrease in antibody recognition of A, Rh (D), and Kell antigens. Meanwhile, the data noted that Kell antigen was
signifiantly camouflged, and also Rh (D) antigen was
masked more effctively in comparison to A antigen. This
is because there is a large diffrence between the densities of Kell, Rh (D) and A antigens on the RBC (5). Therefore, as shown in Tables 6 and 7, immunocamouflging
of these antigens is diffrent. Most importantly, antigen
masking was attenuated by increasing the number of antigens on the cell surface. The results of SEM showed that
polymer concentration of 14.5 mg.mL-1 is a useful level for
PEGylation of RBCs by 91.2% mPEG20-SVA in the polymer
mixture. Echinocytes RBCs are not deformable and hence
tend to get trapped in the microcirculation (22). Thus, the
optimal concentration of 17.21 mg.mL-1, obtained by cell
counting method, cannot be recommended due to the
loss of discocytic morphology. These results are similar
to our previous fiding for PEGylation of RBCs by mPEG,
activated by cyanuric chloride and succinimidyl carbonate (16, 17).
Another interesting observation was derived from investigation of murine PEGylated RBCs SEM results. These
fidings showed that the optimal polymer concentration
(14.5 mg.mL-1) for PEGylation of human RBCs is unsuitable
for murine RBCs. The density of antigens on the murine
RBCs is lower than human ones, so at this concentration,
the echinocytosis increases. Thus, mPEG-SVA-coated RBCs
with concentration of 10 mg.mL-1 have been employed for
the mouse model.
The SEM results indicated that the better maintenance
of mouse RBCs discocytic morphology was achieved at
this polymer concentration. While in vitro analyses indicated the efficy and potential clinical value of immunocamouflged RBCs, in vivo efficy and stability are both
crucial for evaluating potential utility. Importantly, immunocamouflging of RBCs did not have adverse effct
on in vivo survival of a murine transfusion model.
Figure 7 shows the biochemistry parameters of positive control, negative control, and polymer-modifid (10
mg.mL-1 mPEG-SVA) RBCs. As noted, no signifiant diffrences were observed between the serum biochemistry
parameters of negative control group and PEG-modifid
RBCs group and both populations fell in the normal murine biochemistry patterns.
Blood biochemistry assays revealed that ALT, AST, LDH,
and BLT levels in positive control mice signifiantly have
increased, while these parameters did not signifiantly increase in the serum of PEGylated group. Elevation of these
parameters in the blood serum depends on the donor RBCs
lysing by host immune system after recognition of antigens on the donor RBCs. Our results indicated that coated
donor RBCs by mPEG-SVA would result in reducing the immune recognition and clearance by the recipient mice.
Based on these results, we concluded that PEGylation of
RBCs is an appropriate approach for their immunocamouflging. Covalent attachment of activated mPEGs to
the RBC membrane provides a unique strategy in inhibiting the immunological recognition of allogeneic cells.
The aim of the present study was to optimize the covering of minor and major blood group antigens, simultaneously.
The variables of PEGylation reaction were optimized
using factorial and CCD methods. The optimal pH of the
reaction medium was 8.6. Also, the other optimal conditions determined for mPEG-SVA and mPEG-SC using cell
counting method were as follows: mPEG20 in the polymer mixture, 91.2 and 85.36%, and polymer concentration,
17.21 and 19.80 mg.mL-1, respectively. These conditions are
similar, but mPEG-SVA is a more effctive reagent for RBC
coating. It was also concluded that it is not suffient selecting the optimal polymer concentration only by evaluating the extent of PEGylation as the only criterion.
However, according to the SEM results, the maximum
polymer concentration of 14.5 mg.mL-1, as the best condition, was suggested for mPEG-SVA modifid human RBCs
due to the observation of enhancing the rate of echinocytosis by increasing the polymer concentration. By considering the in vitro results, it is obvious that membrane
PEGylation camouflges the blood group antigens. This
effct is observed dramatically and signifiantly for nonABO/Rh (D) antigens.
Finally, these results clearly recommend that PEGylation of non-ABO/Rh(D) matched RBCs signifiantly
reduces the risk of alloimmunization in patients requiring chronic blood transfusions. Therefore, application of
this technology may be clinically useful in chronically
transfused persons with diseases such as sickle-cell anemia and thalassemia because of the increased risk of alloimmunization. Our fidings (regarding in vivo murine)
with allogeneic donor cells suggest that the actual risk of
alloimmunization would be effctively reduced.
The authors wish to thank Iranian Blood Transfusion Organization for providing packed Rh-Kell-A positive RBCs.
All authors participated equally in the present study.
The study was self-funded.
There was no conflct of intere