|Home | About | Journals | Submit | Contact Us | Français|
Transfusion therapy is currently an effective therapeutic intervention in a number of diseases, including sickle cell disease. However, its use is complicated by a high incidence of red blood cell (RBC) alloimmunization in the transfusion recipients. The identification of T regulatory cells (Tregs) among the CD4+CD25+ T cell subset as key regulators of peripheral tolerance in mice as well as humans has opened an exciting era in the prevention and treatment of autoimmune disease and for improving organ transplantation. However, their potential in inducing transfusion tolerance remains to be explored. We used red cells from mice transgenic for human glycophorin A blood group antigen as donor cells and transfused wild-type mice to induce alloantibodies, as an experimental system to study RBC alloimmunization. We found that depletion with anti-CD25 enhanced the alloantibody production, indicating that CD25 Tregs play an important role in regulation of alloantibody responses. More importantly, adoptive transfer of purified population of CD4+CD25+ but not CD4+CD25– cells from naïve mice prevented the induction of IgG and IgM alloantibody production in transfusion recipients, with a concomitant reduction in activated splenic B cells and macrophages. Similarly, adoptive transfer of purified populations of CD4+CD25+ cells from naïve mice into naïve syngeneic recipients inhibited the anti-Ig response to rat RBCs in the recipients but transfer of control CD4+CD25– cells did not. Altogether, our results demonstrate that Tregs participate in the control of trans-fusion-associated RBC alloantibody responses, opening up the possibility that Treg immunotherapy may be exploited for suppressing transfusion immunization events.
Chronic transfusion therapy is increasingly used as a secondary prevention of life-threatening events in a number of disease indications, including sickle cell anemia. Indeed, data from two National Heart, Lung, and Blood-sponsored Stroke Prevention Trials in Sickle Cell Anemia (STOP1 and 2) studies showed that transfusion therapy is effective in the primary prevention of stroke in a subset of sickle cell patients [1,2]. However, alloimmunization to minor red blood cell (RBC) antigens is frequent in chronically transfused recipients. In such transfused patients, almost 50% will have developed alloantibodies by adulthood, with a significant portion having made alloantibodies to several red cell antigens  causing complications ranging in severity from life-threatening delayed hemolytic transfusion reactions and autoimmunization to practical difficulties in obtaining matched blood [4–7]. While several factors including host genetics can influence the recipient's immune system to react to RBC alloantigens, the inflammatory status of transfusion recipients appears to be critical in determining the immunogenicity of transfused RBCs .
A strategy to reduce RBC alloimmunization has been to use phenotypically matched units prior to transfusion, although there is controversy in the implementation of this approach . Induction of immune tolerance to prevent alloimmunization from transfusion is a potential approach that remains to be explored. The identification of CD4+ CD25+ regulatory T cells (Tregs) for controlling immune tolerance has opened the possibility of developing novel immunotherapeutic strategies in suppressing pathologic immune responses in autoimmune diseases, transplantation, and graft versus host disease . Different populations of Tregs have been described, including thymically derived naturally occurring naïve Tregs which constitute about 1–2% of peripheral blood mononuclear cells or about 5–10% of the CD4+ T cells and adaptive Tregs that are induced in the periphery through exposure to antigen [9,10]. The recent study showing that adoptive transfer of Tregs does not result in attenuated responses to pathogens  underscores the potential of Treg-based tolerance induction protocols for treatment or prevention of a number of diseases.
Several transgenic lines expressing human blood group antigens on their RBCs [12,13] are currently available, thus allowing the development of mouse models of red cell alloimmunization in which RBCs from the transgenic mice can be used as donor cells and transfused into wild-type “antigen null” mouse recipients [8,14]. In addition, mouse models of xenogeneic transfusion have been used extensively to study the regulation of transfusion-associated humoral responses [15,16], although antibody regulation by Tregs in such models has not as yet been described. Play-fair and Marshall-Clarke described an interesting mouse model of RBC immunization, in which repeated injections with rat RBCs resulted in the development of anti-rat RBCs and subsequently in induction of autoantibodies against self RBCs and a disease process similar to warm type AIHA in man . Antibody responses to both rat and mouse RBCs are T helper dependent [18–20]. Interestingly, a differential regulation of antibody responses against self and nonself red cell antigens appears to exist in this mouse model. Specifically, spleen cells from “primed” mice immunized with rat RBCs transferred to naïve recipients suppress the subsequent induction of autoantibodies, but not the xenoantibody response [21,22]. In contrast, splenocytes from naïve mice appear to reduce the xenoantibody response, but not the autoantibody response upon transfer [21,22]. Using Tregs from mice previously immunized with rat RBCs, we previously showed that such primed Tregs were able to dampen the autoantibody response but not the anti-rat RBC response. However, the role of naïve Tregs in the control of antibody responses in this model was not investigated.
As a first step to explore the role of Tregs in control of antibody responses to incompatible RBCs, we have manipulated Treg numbers in mice prior to transfusion of allogeneic RBCs from transgenic mice expressing human glycophorin A (GPA) blood group antigen and measured alloanti-body production. In addition, we have examined the antibody responses to xenogeneic rat RBCs following adoptive transfer of naïve Tregs. Our data indicate that naïve Tregs do indeed participate in regulation of RBC alloantibody responses and that they may have potential as immunotherapeutic targets for prevention of RBC alloimmunization and induction of transfusion tolerance.
Hendrickson et al. recently demonstrated that inflammatory signals through polyinosinic polycytidylic acid (poly(I:C)), a synthetic double-stranded RNA molecule, significantly increased both the frequency and the magnitude of alloimmunization in a mouse model . We also found, using a different model of alloimmunization, namely, that of transfusing leukoreduced RBCs from mouse transgenic for human glycophorin A (huGPA)  as a donor model blood group antigen into wild-type (huGPA antigen negative) recipient mice, that alloantibody production was significantly increased when the mice were also treated with unmethylated bacterial CpG dinucleotides (CpG ODN) known to induce an inflammatory signal  (Fig. 1, P = 0.0002). Depletion of CD25-expressing cells with anti-CD25 prior to huGPA RBC/CpG ODN transfusion further enhanced alloantibody production (Fig. 1, P = 0.03), indicating that Tregs are likely to play a role in regulating the alloimmune response. To confirm that Tregs do indeed play a role in regulation of RBC alloimmunization, we adoptively transferred sorted splenic populations of CD4+CD25+ and CD4+CD25– from naïve mice prior to transfusion of huGPA RBC/CpG ODN and found that only the CD4+CD25+-transferred mice had blunted alloantibody (IgM (P = 0.02) and IgG (P = 0.048)) responses comparable to background levels (Fig. 2A,B, respectively). This downregulation of alloantibody production lasted for at least 5 weeks post-transfer (Fig. 2C).
We have previously shown that depletion of mice with anti-CD25 prior to immunization with rat RBCs which elicit T cell responses to several red cell antigenic determinants  results in increase in both the magnitude and frequency of antibody responses to the rat RBCs . To test whether adoptive transfer of naïve Tregs can suppress antibody responses against multiple incompatible red cell antigens, we performed adoptive transfer experiments in mice prior to immunization with rat RBCs. Similar to the studies described above, CD4+CD25+ adoptively transferred mice had reduced levels of anti-rat RBC xenoantibodies comparable to background levels (Fig. 3, P = 0.002), indicating that naïve Tregs participate in regulation of transfusion-associated antibody responses to multiple allogeneic red cells.
Given the well-established role of splenic macrophages in antibody-mediated red cell destruction, we examined the activation status of macrophages (Mac1+) in Treg treated and untreated mice using the activation marker CD86 (Fig. 4A). In addition, we analyzed the frequency of activated splenic B cells (B220+) using CD69 activation marker (Fig. 4B). Transfusion of huGPA RBC alone without adjuvant had little effect on the frequency of activated macrophage (Fig. 4A) and caused a slight increase in the frequency of activated B cells, although the increase did not reach significance (Fig. 4B). We found that mice injected with CpG ODN alone have increased activated macrophage (32%, SD 2%, in controls, compared with 45%, SD 7%, in CpG ODN alone; P < 0.001) and activated B cells (mean 3%, SD 0.1%, in controls, compared with 4.1%, SD 0.7%, in CpG ODN alone; P < 0.05), consistent with the known adjuvant activity of CpG ODN on antigen-presenting cells and B cells (see Fig. 4) . In the presence of the combination of red cells (huGPA RBCs) and CpG ODN, the frequency of activated macrophage and B cells more than doubled (75%, SD 7%; and 6.6%, SD 1.2%, respectively, Fig. 4). Upon pretreatment with control CD4+CD25– cells prior to transfusion of huGPA RBC plus CpG ODN, the frequencies of activated macrophage and B cells did not change when compared with those in mice with no pretreatment but injected with huGPA RBC plus CpG ODN (see Fig. 4). In contrast, pretreatment with CD4+CD25+ resulted in significant decrease in frequencies of activated macrophage (40%, SD 5%, P = 0.006) as well as activated B cells (4.4%, SD 0.6%, P = 0.046) when compared with those in mice transfused with huGPA RBC plus CpG ODN (see Fig. 4), consistent with dampened alloantibody production in CD4+CD25+-transferred mice (see Fig. 2).
In this study, we have demonstrated that naïve Tregs participate in the control of RBC alloimmunization. Specifically, using RBCs from mice transgenic for GPA and transfusing them together with an adjuvant into wild-type antigen negative recipients to induce RBC alloimmunization, we found blunted alloantibody responses with concomitant reduction in activated splenic B cells and macrophages in mice treated with naïve Tregs. Similarly, in a second model using xenogenic rat RBCs to induce RBC alloimmunization, antibody responses were dampened following transfer of naïve Tregs. Thus, despite using disparate blood group antigens as antigenic stimuli with or without adjuvant in our two mouse models, Treg immunotherapy was effective in suppression of RBC alloimmunization.
Immunization with huGPA RBCs required an inflammatory signal through the use of the CpG ODN adjuvant to induce robust alloantibody production (Fig. 1), whereas administration of rat RBCs alone was sufficient to induce anti-rat RBC production, probably reflecting the highly immunogenic nature of xenogeneic RBCs. Hendrickson et al. recently found that mouse RBC alloimmunization to hen egg lysozyme is significantly increased in the presence of poly(I:C) known to induce viral-like inflammation . Using a different inflammatory signal (CpG ODN) , we also found enhanced RBC alloimmunization to human GPA blood group antigen in our mouse model. Possible mechanisms that contribute to the adjuvant activity of CpG ODN include their effect on B cells, antigen-presenting cells as well as T cells . Specifically, CpG ODN treatment results in B cell stimulation ; increased antigen uptake ; enhanced maturation and differentiation of antigen-presenting cells, which in turn result in activation of T cells ; as well as a decrease in Treg suppressive activity, which in turn enhances T effector cell function . Interestingly, we found an increase in both the activation status of macrophages in mice treated with CpG ODN alone (although it did not reach significance) as well as B cells (see Fig. 4). However, antigenic stimulation using huGPA RBCs in combination with CpG ODN resulted in significant upregulation of activation markers on both B cells and macrophages (see Fig. 4), consistent with increase in alloantibody production. Treatment with Tregs prior to immunization with huGPA RBC/CpG ODN significantly reduced the macrophage and B cell activated cell frequencies and these were comparable to the frequencies in mice injected with CpG ODN alone. We have also found that that there is a reduction in the frequency of Tregs in mice that have made an alloantibody response to huGPA RBCs alone, and that the Treg numbers are further reduced when mice are challenged with huGPA RBCs in combination with CpG ODN (our unpublished data). Although preliminary, these data indicate that the suppression of Treg activity may also contribute to the mechanism of enhancement of RBC alloimmunization by inflammatory signals. Consistent with this data, we found that 5 times more naïve Tregs were needed to suppress the alloantibody responses to huGPA RBC plus CpG ODN than for suppression of anti-rat RBC antibody responses where no adjuvant was used (see Materials and Methods).
The exact mechanism of Treg-mediated suppression of RBC alloimmunization remains to be elucidated. Naturally occurring naïve Tregs inhibit the responses of T effector cells by direct cell contact in vitro [9,10]. In contrast, adaptive Tregs that develop from naive CD4 cells in the periphery during the course of an immune response [30,31] are thought to modulate immune responses exclusively via cytokine-mediated effects and can include IL-10- and TGF-β-producing subsets (Tr1 and Th3, respectively) [32,33]. More recent data indicate that contacts with antigen-presenting cells dictate whether Tregs directly or indirectly alter the activation and differentiation of pathogenic T cells . Specifically, it appears that the type of antigen-presenting cells that the antigen is targeted to as well as the immunogenicity of antigen are among the factors thought to contribute to the cytokine secretion phenotypes of adaptive Tregs . There are also in vitro studies indicating that Tregs can suppress B cells directly without having to suppress T helper cells  which appears to be dependent on activation status of Tregs as well as that of the B cells  Based on the potential mechanism of Treg suppression in other experimental systems [38,39], it may be that by transferring Tregs in our mouse models the ratios of naïve Tregs to target cells such as T effector cells, B cells, and/or antigen-presenting cells are increased. Alternatively, the cytokine environment may be altered by the increase in the absolute numbers of Tregs in the transferred mice, which in turn affect the outcome of the immune responses to RBC immunization . Using the rat RBC immunization model in which repeated injections with rat RBCs results in breakdown of tolerance against self RBC antigens, we recently showed that autoantibody production against mouse red cell is also under the control of Tregs, albeit a subset of Tregs that had been previously exposed to autoantigens . Interestingly, these primed Tregs had little effect on suppression of alloantibody responses to rat RBC . Conversely, adoptive transfer of naïve Tregs at concentrations that suppressed anti-rat RBC responses did not have an effect on the development of autoantibodies following rat RBC immunization (our unpublished data), indicating that at least in this mouse model naïve Tregs and primed Tregs differ in their regulation of allo- and autoantibody responses against RBC antigens. Elucidating the mechanism of suppression responsible for prevention of RBC allo- and autoimmunization by naïve and primed Tregs, respectively, remains to be defined. Similarly, defining the target cells such as activated CD4+ effector cells, B cells, and/or antigen-presenting cells that are either directly or indirectly via inhibitory cytokines under the control of naïve or primed Tregs  will be critical for the future manipulation of these cells for use in cell-based therapies for prevention of red cell allo- and autoimmunization in the transfusion setting.
In conclusion, in this study we have shown that naturally occurring Tregs participate in regulation of transfusion-mediated RBC alloantibody responses, opening up the possibility that Treg immunotherapy may be applied for induction of transfusion tolerance.
Mice were housed in a specific pathogen-free barrier facility with restricted access, and all procedures were performed as approved by the Institutional Animal Care and Use Committee of the New York Blood Center.
Female (8–10-week-old) C57/BL6 mice were transfused intravenously with 50 μl Ficoll-treated packed RBCs from transgenic glycophorin A (GPA) mice  (equivalent to approximately 2 × 108 cells or 1–2 packed RBC units) without or with 50 μg unmethylated CpG (cytosine-phosphodiester bond-guanine)-containing oligonucleotides (CpG-ODN, sequence 1826 TCCATGACGTTCCTGACGTT , Coley Pharmaceutical Group) on a weekly basis for 5 weeks. Staining with anti-CD45 leukocyte marker (Pharmingen) routinely demonstrated that packed red cells had undetectable levels of leukocyte contamination (data not shown). Levels of alloantibodies in blood samples obtained by retro-orbital sinus bleeding were measured by incubating diluted plasma with GPA transgenic RBCs and analysis by flow cytometry. Similar antibody reactivity was obtained when using human GPA positive (M+N+) RBCs (Table I). We did not detect any autoantibodies directed against the murine RBCs in mice injected with CpG ODN alone or in combination with huGPA RBCs (data not shown). For Treg depletion studies, 500 μg of anti-CD25 (clone 7D4, BD Pharmingen, San Diego, CA) was given 8 hr prior to RBC transfusions. For adoptive transfer experiments, 106 purified sorted splenic CD4+CD25– and CD4+CD25+ cells (>95% purity, data not shown) were injected through the tail vein, and after 1 day, the mice were transfused intravenously with allogeneic RBCs plus CpG adjuvant. Lower than 106 CD4+CD25+ cell dose numbers were not effective in blunting RBC alloimmune response (data not shown). For phenotypic analysis, splenocytes from mice were stained with FITC-anti-Mac1 (CD11b), and PE-anti-CD86 as well as PerCP-anti-B220, and PE-anti-CD69 (all from Pharmingen) and analyzed on a FACS Canto with Diva analysis software.
Female (8–10-week-old) C57/Bl6 mice were immunized on a weekly basis for 5 weeks with 2 × 108 Ficoll-treated packed rat RBCs and levels of anti-rat RBC antibodies were detected by flow cytometric analysis as previously described . Adoptive transfer experiments were preformed using 5 × 105 purified sorted CD4+CD25+ and CD4+CD25– cells (>95% purity, data not shown) from splenocytes of syngeneic naïve mice as described .
The significance of differences between groups of mice was calculated using a single factor ANOVA test.
We thank Xiaoying Zheng (NYBC) for technical assistance with some of the adoptive transfer experiments. We are grateful to Gregory Halverson (NYBC) for sharing his unpublished data on CpG ODN immunization regimen, Dr. Petra Hoffmann (University Hospital Regensburg, Germany) for helpful discussions, and Dr. Robert Reiss (NYBC) for critical reading of the manuscript.
Contract grant sponsor: NIH; Contract grant number: R01 HL69102.