Supported by a grant from the Department of Defense, Contract W81XWH-05-2-0001. This work was supported by a Defense Appropriations Bill for applications to improve the safety and availability of transfused blood for the US Armed Forces. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the views of the Department of Defense.
Evaluation of potential immune response and in vivo survival of riboflavin-ultraviolet light–treated red blood cells in baboons
Article first published online: 14 OCT 2008
© 2008 American Association of Blood Banks
Volume 49, Issue 1, pages 64–74, January 2009
How to Cite
Goodrich, R. P., Murthy, K. K., Doane, S. K., Fitzpatrick, C. N., Morrow, L. S., Arndt, P. A., Reddy, H. L., Buytaert-Hoefen, K. A. and Garratty, G. (2009), Evaluation of potential immune response and in vivo survival of riboflavin-ultraviolet light–treated red blood cells in baboons. Transfusion, 49: 64–74. doi: 10.1111/j.1537-2995.2008.01940.x
Disclosures: RG, SD, CF, KBH, and HR were employees of Navigant Biotechnologies during the course of the study. KKM is a consultant for Navigant Biotechnologies. GG is a consultant for Navigant Biotechnologies.
- Issue published online: 23 DEC 2008
- Article first published online: 14 OCT 2008
- Received for publication February 13, 2008; revision received June 30, 2008; and accepted July 20, 2008.
BACKGROUND: Pathogen reduction methods have the potential to modify blood components, resulting in immunologic reactions or compromised blood components. This study evaluated the hypothesis that there is no immune response to riboflavin-and-ultraviolet [UV]-light–treated red blood cells (RBCs), as observed by serology and by survival of RBCs in circulation.
STUDY DESIGN AND METHODS: Three baboons were in each treatment group: 1) untreated (negative control), 2) quinacrine mustard (QM)-treated (positive control), and 3) riboflavin-and-UV light–treated (test group) RBCs. In the immunization phase, autologous test or control RBCs were injected subcutaneously on Days 0, 21, 42, and 49. Plasma samples from these days were tested against test or control RBCs by flow cytometry and standard serology. On Day 56, autologous 51Cr-labeled test or control RBCs were injected. Blood samples were taken over 21 days after injection to determine RBC survival (t1/2).
RESULTS: Untreated and riboflavin-and-UV-light–treated RBCs showed no evidence of significant immunoglobulin G (IgG) binding after incubation with autologous plasma. RBC-bound IgG was detected on QM-treated RBCs after incubation with autologous plasma. This antibody was inhibited by QM, as demonstrated by a hapten inhibition study. t1/2 values for the untreated and riboflavin-and-UV-light–treated RBCs were 7.3 ± 0.8 and 7.5 ± 1.7 days, respectively; the t1/2 value for QM-treated RBCs was 2.3 ± 2.9 days.
CONCLUSION: Treatment with riboflavin and UV light did not render RBCs immunogenic. Positive controls indicated that immunization promoted an immune response. In the 51Cr-labeled RBC survival phase of the study, riboflavin-and-UV-light–treated RBCs exhibited behavior similar to negative control RBCs. Detrimental immunologic or functional side effects were not observed.
Southwest Foundation for Biological Research
The development of pathogen reduction technologies for blood components has several challenges. Not only must the technique that is employed be demonstrated to be effective against a wide variety of pathogens that may contaminate blood products, the technique must also be demonstrated to carry out its action without significant damage or alteration to the cellular or protein components of blood.1 Achieving the former goal (inactivation of pathogens) is relatively straightforward. Simultaneously doing the latter (no harm to cell or protein components) has been difficult to implement for cellular components.2
The treatment of red blood cells (RBCs) poses unique development hurdles. Photochemical methods, which have been employed for platelets (PLTs) and plasma, have limitations on the wavelengths of light that can be used, due to the optical properties of the RBCs and hemoglobin (Hb). Photosensitizers that have been tested for the treatment of RBCs are primarily those that absorb wavelengths of light longer than those absorbed by Hb.3,4 Methods for inactivation of pathogens in RBC products have focused on the use of chemical agents activated by changes in pH that occur when they are added to blood. Activation by these means generates reactive species that can damage nucleic acids in the products, the desired effect for pathogen and white blood cell (WBC) inactivation.5-8
Under normal conditions, RBCs carry oxygen. Activation of chemical processes in the presence of oxygen can lead to the generation of reactive oxygen species.9,10 These reactive species are relatively nonspecific in their action, an effect that can cause indiscriminate damage to cell membranes or protein constituents of the therapeutic elements of blood.11,12 Preventing this nonspecific, oxidative pathway for reaction of photochemical or chemical sensitizers in blood poses a concern in RBC preparations.13-17
Several adverse effects on RBC quality have been reported when products are treated with processes for pathogen inactivation. These include observations of increased hemolysis, membrane modification, methemoglobin formation, and even the formation of antibody against the agent used to treat RBCs, in human subjects.6,18 All of these effects raise considerable practical issues for routine use of these products.
Over the past several years, a photochemical method for the treatment of blood products utilizing riboflavin and ultraviolet (UV) light has been developed.19 This process is now commercially available in Europe for the treatment of PLT concentrates. It has been evaluated extensively in preclinical studies, including studies of PLT and protein function, for the ability to inactivate pathogens (virus, bacteria, parasites) and WBCs and for the clinical performance in supporting transfusion practice for patients with thrombocytopenia.20 In addition, the product has been subjected to extensive toxicologic evaluation, including assessment of genotoxicity, systemic toxicity, and neoantigenicity in animal and in vitro models.21
The use of the same technology for the treatment of RBCs appears to be an obvious extension of this work. Martin and colleagues22 investigated the potential to achieve pathogen inactivation by utilizing a unique feature of the riboflavin molecule when it associates with nucleic acids. This effect results in the movement of the bound riboflavin compound into a spectral region that can be targeted with an appropriately selected external light source. Such a feature might have several benefits. First, it might allow the activation of the bound riboflavin in a selective manner, thus preventing activation of compound free in solution. Such selectivity avoids or minimizes reactions with dissolved oxygen, which can create reactive oxygen species. Second, it might prevent nonspecific photochemistry or chemical modification of proteins and membranes and thereby prevent the creation of foreign surfaces for immune recognition or nonimmunologic adsorption of proteins.23,24
Given these features of the photochemistry of riboflavin and nucleic acids, we hypothesized that treatment of RBCs with this process should not result in immunogenic modifications of the RBC surface. To test this, baboons were immunized with treated preparations, and plasma from the immunized baboons was tested for immunoglobulin G (IgG) binding to treated cells using flow cytometry and standard serologic methods. Nonimmunologic adsorption of proteins (e.g., IgG) onto RBCs was also monitored. The immunized baboons were used to measure survival of treated RBCs after radiolabeling and intravenous (IV) infusion.
MATERIALS AND METHODS
The in vivo portion of this study was performed at the Southwest Foundation for Biological Research (SFBR), which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International, and was performed in compliance with the “Public Health Service Policy on Humane Care and Use of Laboratory Animals,” revised September 1986. The Armed Services Committee on Animal Laboratory Safety also approved the study protocol before initiation.
As a positive control for this study, a quinacrine mustard (QM) compound was utilized. QM is similar in nature and structure to compounds that bond to cells and proteins and that thereby become immunogenic in animals and human subjects.6,25-28
Nine baboons (Papio cynocephalus) were assigned to one of three study groups (negative control, positive control, riboflavin-and-UV-light–treated) for a total of three animals per study group. Of the nine baboons, five were female and four were male. Animal weights ranged from 17 to 27 kg. Units of whole blood (75 mL) were collected from each animal before the start of the study (Day −7) and on Day 49 of the study. The units of whole blood were collected into citrate-phosphate-dextrose anticoagulant (1:10 ratio) and then shipped overnight for preparation the next day. Each unit was leukoreduced with a pediatric leukoreduction filter (Pall NEO1, Pall, East Hills, NY) and then separated by centrifugation (7277 × g for 12 min). The plasma was removed, aliquots of the plasma were stored for use by the indirect antiglobulin test (IAT) and flow cytometry assays, and resulting RBC products were then handled to produce untreated RBCs for the negative control group, QM-treated RBCs for the positive control group, and RBCs treated with riboflavin and UV light under the conditions specified for the test group. The RBCs prepared from Day −7 blood were used for the immunization phase of the study. The RBCs prepared from Day 49 blood were used, on Day 56, for the flow cytometry and serology testing, and for the in vivo radiolabeled recovery and survival.
Negative control: no treatment
After the separation from plasma described above, AS-3 was added to each negative control RBC product to obtain a hematocrit (Hct) of 60 ± 5 percent and a volume of 25 ± 5 mL. No additional manipulations of these products were performed.
Positive control: quinacrine treatment
After the separation from plasma as described above, RBCs were washed and resuspended at a 50 percent Hct in 0.9 percent saline with 200 µmol per L QM hydrochloride. The RBCs were kept at room temperature for 24 hours with gentle mixing. The positive control RBC products were then washed once with a volume of 0.9 percent saline that resuspended the RBCs to at least a 50 percent Hct to remove unbound chemical. After the wash, the RBCs were suspended in AS-3 at a 60 percent Hct before storage.
Test product: riboflavin and UV light (Mirasol) treatment
After the separation from plasma as described above, the RBCs were washed with a solution of 0.9 percent saline and 0.2 percent dextrose at approximately two times the weight of the RBCs to remove plasma and then the RBCs were resuspended in a solution of 500 µmol per L riboflavin and 0.9 percent NaCl to obtain 3 percent Hct in 480 mL for illumination. The contents of the illumination bag were gently mixed manually. The illumination bag was placed in the illuminator and exposed to UV light with linear agitation of 120 cpm. During illumination, the temperature of the RBC suspension was in the range of 20 to 23°C. The bag was removed after delivery of the target energy (37.3 J/mLRBC, 250- to 400-nm light source) was delivered to the bag. This energy was equivalent to twice the dose required for effective pathogen reduction conditions established in separate studies (unpublished data). After illumination, the treated RBC products were centrifuged to concentrate and stored in 10 mL of AS-3 at a 60 ± 0.5 percent Hct until subsequent use.
The energy delivered to the RBCs was normalized for the volume of RBCs in the suspension. Normal treatment volume for the riboflavin-UV light–treated human RBCs is 480 mL with 30 percent Hct. The maximum volume that was approved by the Institutional Animal Care and Use Committee (IACUC) at SFBR from the baboon blood draw was 75 mL; therefore, a relatively low Hct (3%) was necessary to treat the RBC suspension at a volume of 480 mL. Use of the characterized illumination volume (480 mL) ensured that the mixing dynamics within the illumination bag would be consistent with prior experiments (unpublished data), which showed that normalization of energy delivery with respect to RBC volume (joules per milliliter of RBCs, J/mLRBC) is appropriate.
Before storage, approximately 1 mL was removed from each of the RBC preparations to measure Hct and total Hb with a cell counter (AcT Diff, Beckman Coulter, Fullerton, CA) and to measure free plasma Hb in the supernatant with a photometer (HemoCue Plasma/LowHb, HemoCue, Lake Forest, CA). Samples for the initial RBC immunization (prepared from the Day −7 blood draw) were frozen without cryoprotectant, in 2-mL cryovials, shipped frozen to the study site, and stored frozen until the time of preparation for immunization. Samples for the RBC survival portion of the study, for the flow cytometry, and for the serology testing (prepared from the Day 49 blood draw) were shipped cold and stored refrigerated, in 15-mL conical tubes, for 7 days until the time of testing or of radiolabeling and infusion.
Immunization against RBC preparations
The immunization schedule is depicted in Table 1. Animals received subcutaneous injections of the test or control articles on Days 0, 21, and 42 of the study. All preparations (frozen RBCs in AS-3) were thawed and then formulated with alum as an adjuvant to obtain 1 mL of RBCs and alum mixture for subcutaneous injection. A 1-mL dose in total was injected, utilizing four injection sites per animal, 250 µL per site. All injections were made subcutaneously to facilitate the possible formation of antibodies via dendritic cells and other antigen-presenting cells present in the lymphatic system.
|Study group||Number of animals||Treatment||Route of injection†||Dosage and volume||Frequency|
|1||3||Negative control: autologous, untreated RBCs||S/C||1 mL total, 250 µL at each of four sites‡||On Days 0, 21, and 42|
|2||3||Positive control: autologous RBCs treated with QM||S/C||1 mL total, 250 µL at each of four sites‡||On Days 0, 21, and 42|
|3||3||Test: autologous RBCs treated with riboflavin and light||S/C||1 mL total, 250 µL at each of four sites‡||On Days 0, 21, and 42|
Standardization of antisera used with baboon RBCs
All reagents were standardized for use with baboon RBCs as previously described.29 Briefly, the reagents (goat or rabbit source polyclonal antisera) were initially tested with untreated baboon RBCs (Bioreclamation, Hicksville, NY) to determine the presence of heterophile antibodies. Dilutions of antisera that did not react with untreated baboon RBCs were then tested against the same baboon RBCs that had been coated in vitro with either purified rhesus monkey IgG (Southern Biotechnology, Birmingham, AL) or purified rhesus monkey albumin (Nordic Immunological Laboratories, Tilburg, The Netherlands) using a chromic chloride coupling method.29 Purified baboon IgG or albumin was not commercially available but interspecies cross-reactivity is a normal feature of antibodies to mammalian serum proteins. The RBCs coated with either monkey IgG or monkey albumin were then tested with anti-monkey IgG or anti-monkey albumin by a low-ionic-strength saline (LISS) antiglobulin test or a fluorescein isothiocyanate (FITC)-labeled anti-monkey IgG by flow cytometry. Two different lots of an anti-monkey IgG reagent (Serotec, Oxford, UK; cross-reactive with baboon, rhesus monkey, and human RBCs) were found not to be suitable for LISS antiglobulin tests (results were never stronger than microscopically positive); a commercial anti-human IgG (Ortho Diagnostics, Raritan, NJ) gave much better results (i.e., reacted 3+ with baboon RBCs coated with rhesus monkey IgG and did not react with the same RBCs without IgG sensitization). A commercial anti-monkey albumin reagent (Nordic Immunological, Tilburg, The Netherlands) for the LISS antiglobulin test method reacted 3+ with baboon RBCs coated with rhesus monkey albumin. The commercial FITC-labeled anti-monkey reagent (Serotec Ltd, Oxford, UK) for flow cytometry was strongly fluorescent with baboon RBCs coated with rhesus monkey IgG (i.e., fluorescence was nine times stronger than results with uncoated RBCs).
The IAT utilized a low-ionic-strength environment by a test tube method. The baboon plasma samples from Days 0, 21, 42, and 49 were incubated with autologous RBCs (untreated RBCs in the untreated group, untreated and QM-treated in the QM-treated group, or untreated and Mirasol-treated in the test group) in the presence of commercial LISS (Ortho AES) for 15 minutes at 37°C. After centrifugation and examination for agglutination, the RBCs were then washed four times with isotonic phosphate-buffered saline (PBS) and tested with either polyclonal anti-human IgG (Ortho) or polyclonal anti-monkey albumin (Nordic Immunological). The latter was used to test for nonimmunologic uptake of plasma proteins.29
For flow cytometric analyses, the autologous baboon RBCs (untreated RBCs in the untreated group, untreated and QM-treated in the QM-treated group, or untreated and Mirasol-treated in the test group) were incubated in autologous plasma from Days 0, 21, 42, and 49 at 37°C for 60 minutes, washed four times with PBS, and then incubated at room temperature for 30 minutes with FITC anti-monkey IgG (Serotec Ltd). Ten thousand events were collected on a flow cytometer (FACSort, BD Biosciences, San Jose, CA) using logarithmic amplification. The median fluorescence for each sample was determined.
Owing to low levels of IgG on normal RBCs and/or nonspecific binding of FITC anti-IgG, the following formula was used to analyze the flow cytometry data:
A relative intensity value of 1 means that the results obtained after incubation of RBCs with plasma were no different than the results from RBCs without plasma added. A relative intensity value of 2 meant that the fluorescence median results obtained from RBCs after incubation with plasma were twice as high as the results from RBCs without plasma added.
Hapten inhibition study
An equal volume of baboon plasma and 1 or 5 mg per mL QM dihydrochloride (Sigma-Aldrich, St Louis, MO) diluted in PBS (or PBS as a dilution control) were mixed together and incubated at 37°C for 1 hour. One drop of 4 percent (vol/vol) QM-treated RBCs was added and the mixture was incubated at 37°C for an additional hour. The RBCs were washed and tested with anti-human IgG.
51Cr labeling procedure for baboon RBCs
Ten milliliters of test or control RBCs (in AS-3, pH approx. 6.4) was gently mixed with 200 µCi of sodium chromate 51Cr (pH 7.5 to 8.5; Mallinckrodt, St Louis, MO) by swirling at 5- to 10-minute intervals at room temperature for 30 minutes. A small aliquot was used for obtaining preinfusion counts of the suspension in a gamma counter (Wallac LKB 1282, Wallac, Turku, Finland) and approximately 10 mL of labeled autologous RBCs was infused IV. Before infusion a blood sample (0 hour) was collected from each animal, and after infusion, 10-mL blood samples were collected at 1 and 24 hours and on Days 7, 14, and 21 for analysis. Whole blood samples were counted with the gamma counter, followed by centrifugation to separate plasma and cell fractions. The plasma fraction from each sample was then counted. The whole blood samples were also analyzed for clinical chemistry and hematology parameters.
Evaluation of survival of 51Cr-labeled RBC preparations
The evaluation of survival of autologous negative control, positive control, and riboflavin-and-UV-light–treated RBCs labeled with 51Cr was evaluated as depicted in Table 2. The sample times (1 hr, 24 hr, 7 days, 14 days, and 21 days after infusion) were chosen to limit the stress to the animals (per the directions of the SFBR IACUC) while providing enough data to determine if Mirasol treatment led to substantial differences in RBC survival in comparison to the untreated controls.
|Study group||Number of animals||Treatment||Route of injection*||Dose and volume (mL)||Frequency|
|1||3||Negative control: autologous, untreated 51Cr-labeled RBCs||IV||10||Once on Day 56|
|2||3||Positive control: autologous 51Cr RBCs treated with QM||IV||10||Once on Day 56|
|3||3||Test article: autologous 51Cr RBCs treated with riboflavin and light||IV||10||Once on Day 56|
The 51Cr-labeled RBC survival was calculated from the values for percent recovery of the 51Cr radioactivity observed in the 1-hour sample. The calculation of recovery at each sample time (%recovery(t)) also accounted for differences in animal weights and Hct levels over the 21 days of the study:
where the sample time, t, is 1 hour, 24 hours, 7 * 24 hours, 14 * 24 hours, or 21 * 24 hours. The values for percent recovery were used to calculate the t1/2 (time at which 50% of the 51Cr-labeled RBCs have been cleared from circulation) from an exponential fit of the values for percent recovery for each baboon with respect to time.
After treatment of each of the preparations according to the procedures outlined under Materials and Methods, hemolysis levels were low. Untreated RBCs exhibited less than 1 percent hemolysis; positive control RBCs and riboflavin-and-UV-light–treated RBCs exhibited less than 2 percent hemolysis.
Results from the serologic evaluation of RBC preparations using anti-human IgG are shown in Table 3. An initial response against the QM-treated RBCs was observed at Day 0 in one of the three baboons that received QM-treated RBCs. The strength of the reactions observed in this baboon increased with increasing exposure on Days 21, 42, and 49, suggesting the continued development of an antibody response against the QM-treated RBCs. The two other baboons receiving QM-treated RBCs developed weak reactivity with QM-treated RBCs by Day 42; the reactivity was transient for one (not apparent on Day 49) but was increased on Day 49 for the other baboon. No IgG binding was observed for any of the animals receiving either riboflavin-and-UV-light–treated or untreated RBCs. When anti-albumin was exposed to riboflavin-and-UV-light–treated and untreated baboon RBCs after incubation with autologous plasma, no reactivity was observed. Thus, there was no evidence for the presence of nonimmunologic absorption of protein by the test RBCs or controls.
|Group and animal number||Day 0||Day 21||Day 42||Day 49|
|Negative, Animal 1||0||0.9||0||1.0||0||1.0||0||1.0|
|Negative, Animal 2||0||1.0||0||1.0||0||1.0||0||1.0|
|Negative, Animal 3||0||1.0||0||0.9||0||1.0||0||1.0|
|Positive, Animal 4||0||1.1||0||1.2||1+W||1.4||0||1.2|
|Positive, Animal 5||3+||2.1||3½+||2.1||3+†||2.1||3+†||2.1|
|Positive, Animal 6||0||1.2||0||1.6||½+||1.5||2+||2.4|
|Test, Animal 7||0||1.1||0||1.1||0||1.2||0||1.2|
|Test, Animal 8||0||1.1||0||1.1||0||1.1||0||1.1|
|Test, Animal 9||0||1.1||0||1.0||0||1.1||0||1.1|
The results from the serologic evaluations were confirmed by the flow cytometric analysis of the RBC preparations with plasma from Days 21, 42, and 49 (Table 3). Flow cytometry results also correlated with the serology results for QM-treated products, with the initial positive response seen with one of the baboons and increasing levels of response with increased exposure of the animals to the test article. In the case of the animals receiving untreated or riboflavin-and-UV-light–treated products, the flow cytometry values indicated no significant binding of IgG. Figure 1 presents the flow cytometric results for one set of reactive and nonreactive samples.
Further evaluation of the reactivity with the QM-treated RBCs was carried out using hapten inhibition studies. The QM-treated RBCs demonstrated a strong serologic response, which could be prevented if excess amounts of free QM hydrochloride were added to one of the reactive plasma samples before QM-treated RBCs were added (Table 4). The antiglobulin test reactivity was inhibited by both concentrations of QM, thus indicating that the antibody was directed against QM alone (not the RBC membrane). The results suggest that the binding of IgG to the QM-treated RBCs is a result of an antibody to QM present in reactive baboon plasma. Presumably QM is left on reactive QM-treated RBCs after processing.
|Reactants mixed with QM-treated RBCs||Anti-IgG|
|Plasma plus 5 mg/mL QM hydrochloride||0|
|Plasma plus 1 mg/mL QM hydrochloride||0|
|Plasma plus PBS||3+|
Table 5 displays the values measured by the gamma counter for each sample. The 1-hour samples for the negative controls have the lowest amount of 51Cr in the plasma, and the positive controls have the highest amount. Samples taken at later times displayed 51Cr values in plasma that were similar across the three different groups. A sample from one of the baboons in the positive control group (Animal 4) exhibited anomalously low counts in the 1-hour sample. The counts measured were substantially higher in the 24-hour sample from that animal and in the 1-hour samples from the other two animals in the positive control group. Therefore, the 1-hour value for gamma counts from that animal was not used to calculate percent recovery; instead, the theoretical 100 percent value was estimated (based on 51Cr radioactivity infused) with the calculation described in Table 5 and used to calculate the recovery values for that animal.
|Animal number||Negative control group||Positive control group||Test group|
|Day 7 counts|
|Day 14 counts|
|Day 21 counts|
Radiolabeled RBC recovery values for all three groups are depicted in Fig. 2 and Table 6. Samples withdrawn from negative control group baboons and test group baboons had little or no label present in the plasma fraction. Results for cell survival in circulation for RBCs in the latter two groups were nearly identical, with a mean t1/2 in circulation measured at 7.3 ± 0.8 days for untreated controls and 7.5 ± 1.7 days for the riboflavin-and-UV light–treated samples. This contrasts dramatically with the results for the QM-treated products, which exhibited a t1/2 in circulation of 2.3 ± 2.9 days on average. Cell survival for the QM-treated RBCs was severely compromised. After IV infusion, these cells were removed from circulation at an accelerated rate when compared to untreated RBCs and RBCs treated with riboflavin and UV light. As Fig. 2 shows, one of the animals in the positive control group (Animal 6) had a negative value for percent recovery of RBCs at 21 days (−0.6%). This value was not used in the exponential fit of the recovery values for that baboon. Exclusion of this value provided a conservative (larger) value for the calculated t1/2 of the 51Cr RBCs for this animal.
|Animal number||Percent recovery (based on 1-hour sample)†||Survival (t1/2), days|
|24 hr||7 days||14 days||21 days|
|Negative control group|
|Mean ± 95% CI||66.1 ± 5.6||52.3 ± 7.5||33.8 ± 2.2||16.9 ± 3.0||7.3 ± 0.9|
|Positive control group|
|Mean ± 95% CI||59.6 ± 30.4||25.2 ± 30.6||6.4 ± 7.6||1.8 ± 2.8||2.3 ± 3.2|
|Mean ± 95% CI||79.1 ± 5.0||46.8 ± 10.5||32.1 ± 8.6||16.6 ± 5.9||7.5 ± 1.9|
The goal for all pathogen reduction methods is the successful inactivation of pathogens in blood products and subsequent prevention of disease transmission. This is accomplished primarily by damage to the nucleic acid replicating potential of the pathogen. However, the achievement of this goal alone in the context of a treatment for blood products is inadequate. Such treatments must be accomplished with an equal consideration of potentially negative effects on cell or protein quality. Processes that target nucleic acid destruction may nonetheless have additional chemical effects on cell membranes and cell proteins or on proteins in solution. The potential for several agents employed in pathogen reduction technologies to modify cell membranes and proteins has been demonstrated in prior studies.2,18
The effects of these alterations to blood cells and proteins have not yet been fully understood or evaluated in human subjects. Petz and Garratty29,30 point out two major problems that have been encountered with RBCs. The first is associated with nonimmunologic uptake of protein onto RBCs after some pathogen inactivation procedures.10,12 This RBC-bound protein (e.g., IgG) leads to positive antiglobulin tests after in vitro experiments; possibly the same could occur in vivo, leading to decreased RBC survival, but this has not yet been proven. The same problems have been noted with therapeutic drugs.29 The second problem is due to some of the chemical, used in the pathogen inactivation process, remaining on the treated RBCs. If the patient has, or develops, an antibody to epitopes on this chemical, then it will react with the treated RBCs in vitro (e.g., crossmatching tests) and in vivo, possibly leading to decreased RBC survival. An example of this, an antibody being developed to epitopes on a chemical in the pathogen inactivation field, has been described;30-32 the QM used as a positive control in our study is also illustrative of this type of reaction. Such an antibody can be shown to be an antibody to the chemical alone (not requiring RBC membrane proteins to react) by the hapten inhibition test. Such interactions are typical for several drug-induced immune hemolytic anemias (e.g., those caused by penicillin antibodies).23,29 In contrast, it has been suggested that some antibodies may be directed against neoantigens (i.e., a new antigen created by the interaction of chemical and RBC membrane). This would be suspected if the antibody was not inhibited by the chemical alone. Although a popular hypothesis, there does not seem to be evidence yet for such antibodies to be the cause of problems associated with pathogen inactivation procedures.
The extent to which modifications of the cell surface may lead to immunologic response in subjects receiving pathogen-reduced components is of significant importance. The potential for these modifications to adversely affect component behavior is why the riboflavin and UV light treatment method was evaluated in the baboon model. Reports of riboflavin association with proteins and possible covalent modification of proteins have been reported previously in the literature.33,34 These studies were often carried out in nonphysiologic conditions that are not reflective of those used or even possible for blood storage.35 Nevertheless, these reports warranted additional investigation.
Prior studies with 14C-labeled riboflavin, using this treatment process and variations of it including excessive light dose failed to show measurable binding of the material or its photoproducts, before or after photolysis, to cell membranes or proteins.21 Additional studies in an animal model to evaluate the potential of neoantigen formation against photochemically altered proteins also yielded negative results.21 Subsequent human studies of riboflavin-and-UV-light–treated PLTs in plasma were conducted in patients with thrombocytopenia, receiving multiple infusions of product over prolonged periods of time in the course of their therapy.36 Each subject in the study was evaluated independently for the possible formation of antibodies against treated PLTs. None were observed. Although weak and potentially natural associations of riboflavin with blood components and proteins cannot be ruled out, the likelihood of covalent modification of membranes or proteins by this process appear to be extremely unlikely. In addition, the fact that riboflavin and its photoproducts are naturally occurring agents in animals suggests that antibody generation against these compounds would also be highly unlikely.
The results observed in this study with the QM-treated products suggest that an existing antibody to the acridine may be present in some primates. In this study, binding of IgG to QM-treated cells occurred on Day 0 in one of three baboons as measured by both serology and flow cytometry. The strength of that binding increased with increasing exposure, but was clearly evident in that one baboon at Day 0 upon initial exposure. This is consistent with several historical literature reports for this class of agents.25-27 The baboons selected for the study were not expected to have had exposure to QM or to similar compounds and thus were not screened before assignment. A search of the experimental and medical history of the animals did not find any history of QM exposure.
Riboflavin-and-UV-light–treated RBCs and the negative control RBCs exhibited similar recovery and survival. This indicates not only that the baboons were not immunized against the RBCs (as observed in the serology and flow cytometry data) but also that the in vivo RBC quality is similar for the two groups. The average t1/2 values of the riboflavin-and-UV-light–treated and of the untreated RBCs were comparable, at slightly more than 7 days. A 13-day t1/2 was previously reported for infused baboon RBCs.37-39 The difference between reported values and those observed here is due in part to the method used to calculate t1/2 in this study. The 1-hour values for gamma count were used in this study as the basis for the calculation of recovery, rather than the total radioactivity infused (with the exception of Animal 4, in the positive control group). In this study, all baboon RBCs were prepared from 24-hour-old baboon whole blood that was shipped from the study site to the laboratory for processing. The 13-day t1/2 values were reported for studies where the RBCs were freshly drawn, labeled with 51Cr, and reinfused. The differences in the reported values may also be due to the storage of the RBCs in conical centrifuge tubes for this study, rather than in blood bags, during the 7 days between preparation of RBCs in AS-3 and labeling with 51Cr. The lack of di(2-ethylhexyl)phthalate in the centrifuge tubes was expected to have an effect on overall RBC quality across all groups. The difference in t1/2 between the previously reported study and this study is most likely in response to storage conditions before transfusion. The differences in survival between the riboflavin-and-UV-light–treated and the untreated versus the QM-treated RBCs correlated with the responses observed in the serology and flow cytometry data.
Survival of the QM-treated RBCs in circulation over time was severely compromised. The decreased survival of RBCs in the positive control group is undoubtedly affected by the immunization of the recipient baboons. The positive control RBCs experienced an extended (24-hr) incubation at room temperature in the QM solution. That incubation was also expected to be somewhat detrimental to RBC quality, although the hemolysis levels observed during removal of QM solution and suspension in AS-3 were small (less than 2%). The likelihood that an immune response is the cause of the lower survival in this group, rather than altered function, is given support by the t1/2 calculated for Animal 4 (in the positive control group). That animal had the least response to immunization, as observed by serology and flow cytometry, and had the greatest t1/2 of all of the positive control group animals. The t1/2 for Animal 4 was determined from recovery values that were based on the theoretical 100 percent 51Cr value. With this relatively conservative estimate, Animal 4 exhibited higher recovery values at 24 hours and 7 days and a correspondingly longer t1/2. However, recovery values for all positive control group animals are substantially lower than those for test and negative control group animals (Table 6, Fig. 2) by Day 7 (144 hr), and the difference became more pronounced on Days 14 and 21.
The results from this study are consistent with prior observations21 and suggest that photochemical treatment of RBCs with a riboflavin-and-UV-light–based method is not likely to produce unwanted chemical modifications to RBCs or proteins that would result in antibody formation or nonimmunologic binding of protein, leading to poor survival of RBCs. Such effects, which were demonstrated in this study with a positive control arm for antibody production, led to the binding of IgG to the modified components and subsequent adverse outcomes relative to cell survival in vivo. This work suggests that the riboflavin and UV light treatment technique may be utilized without detrimental immunologic side effects as a consequence of photochemical or chemical alterations induced in cells or proteins.
- 2The design and development of selective, photoactivated drugs for sterilization of blood products. Drugs of the Future 1997;22:159-71., .
- 20The antiviral and antibacterial properties of riboflavin and light: applications to blood safety and transfusion medicine. In: SilvaE, EdwardsAM, editors. Flavins: photochemistry and photobiology. Cambridge: Royal Society of Chemistry; 2006. p. 83-113., , , , , , , , , , , , , .
- 29Immune hemolytic anemias. 2nd ed. Philadelphia (PA): Churchill Livingstone; 2004., .
- 30Immunogenic issues with the use of pathogen-inactivated RBCs. In: WebertKE, CsertiCM, HannonJ, LinY, PavenskiK, PendergrastJM, BlajchmanMA, editors. Proceedings of a consensus conference: pathogen inactivation —making decisions about new technologies. Transfus Med Rev 2008;22:16-7..
- 32Antibody formulation to S-303-treated RBCs in setting of chronic RBC transfusion (abstract). Blood 2004;104:112a., , , , , , , .
- 34Characterization of human riboflavin-binding immunoglobulins. In: EdemondsonDE, McCormickDB, editors. Flavins and flavoproteins. Berlin: Walter de Gruyer & Co.; 1987. p. 445-8., , .
- 36Safety and performance of MIRASOL® PRT treated platelet transfusion products. Lakewood (CO): Navigant Biotechnologies; 2006. Available from: http://clinicaltrials.gov/ct/show/NCT00263809?order=1
- 37Volume of RBCs, 24- and 48-hour posttransfusion survivals, and the lifespan of (51)Cr and biotin-X-N-hydroxysuccinimide (NHS)-labeled autologous baboon RBCs: effect of the anticoagulant and blood pH on (51)Cr and biotin-X-NHS elution in vivo. Transfusion 2002;42:343-8., , , .