One big unhappy family: transfusion alloimmunization, thrombosis, and immune modulation/inflammation
In this issue of TRANSFUSION, Yazer and colleagues1 provide evidence that febrile reactions, and thus inflammatory responses to transfused leukoreduced platelets (PLTs), predispose the recipient to red blood cell (RBC) alloimmunization. This article provides some early supportive evidence to studies in animal models and if confirmed by more extensive studies in animals and man, could revolutionize the way we think about transfusion reactions and the generation of alloimmune responses in transfusion recipients.
Yazer and colleagues report on a pilot study of 190 patients who had a febrile reaction to PLT transfusion (or rigors and chills, without fever) and received additional transfusions, either RBCs and/or PLTs, within 10 days before or after the febrile reaction. The control group was 245 PLT transfusion recipients who did not experience a febrile reaction, but were similarly transfused with additional RBCs and/or PLTs, in temporal proximity to their index PLT transfusion. The appearance of new RBC antibodies occurring beyond 10 days was determined serologically over the ensuing 5 months or more. Their underlying hypothesis is that inflammatory stimuli can potentiate the humoral immune response to antigens presented simultaneously.
They found that new RBC antibodies were significantly more common among patients in the febrile transfusion group than in the control group (8% vs. 3%, p < 0.026). The authors suggest that recipient inflammation caused by prior PLT transfusion may affect the development of RBC alloimmunization. These findings in humans support prior innovative work in mice examining the role of inflammation as a predisposing factor to RBC alloimmunization.2,3 In these studies, pretransfusion treatment of mice with poly(IC), a double-stranded RNA, which creates a “viral-like” inflammatory state, was associated with increased RBC antibody formation. It was also associated with increased RBC antigen presentation by dendritic cells, increased costimulatory molecule expression, and enhanced proliferation of CD4+ T cells compared to unstimulated mice. In contrast, a different inflammatory compound, namely, bacterial lipopolysacchride, was not associated with increased RBC alloimmunization suggesting that different inducers of inflammation cause disparate effects on the immune system.
The authors propose that the mechanism of this increased RBC antibody production in mice and humans could be explained by the induction of an inflammatory response in the recipient by infused cytokines in PLT transfusions, including CD40L, which then nudges the immune system toward a Th2 or Type-2 response. As a result there is a greater propensity to form antibodies to foreign antigens, presumably both RBC and other antigens such as HLA. This proposal is consonant with basic immunologic wisdom of long standing that successful humoral immune response to nonpathogenic, nondangerous antigens (e.g., viral protein in a vaccine) invariably requires an accompanying inflammatory response stimulated by an adjuvant.
The conventional wisdom in transfusion medicine is to minimize the clinical significance of a febrile reaction and consider such reactions as distinct from the immunomodulatory effects of transfusion. However, if the concepts proposed in this article are confirmed, febrile reactions will be understood as a clinical sign that more profound immunologic effects are actually taking place in the patient. Similarly the development of RBC antibodies is not an isolated immunologic finding, but reveals more widespread alterations in the immune system that are occurring, albeit silently. One candidate mediator for both reactions and favoring Type-2 humoral immune responses is CD40L, which primarily originates from PLTs4 and has been implicated in febrile transfusion reactions5 and transfusion-related acute lung injury (TRALI)6 and is known to stimulate B-cell activation and immunoglobulin synthesis.7,8
The Th1/Th2 hypothesis was originally proposed in 1986 by Mosmann and colleagues9 stemming from observations in mice. It was thought then to be quite simple. In the mouse two distinct subsets of helper T cells exist, Th1 and Th2, distinguished by different patterns of cytokine production after T cells are activated. Th1 and Th2 cells are important regulators of the immune response that mediate different regulatory and effector functions. Since the 1980s this hypothesis has been adapted to human immunity and has become an important paradigm for understanding the immune response in health and disease. During the decades following, the original report on the effects of Th1 and Th2 in human disease became one of the major research foci in immunology.10 Over that period, the picture has become much more complex. New components such as T regulatory cells11 and Th312 and Th1713 subsets have been described, and attempts to interpret complex immunologic diseases by Th1/Th2 model have not always succeeded. However, the model still provides a clear testable paradigm of T-cell function on which to base further research questions and develop more sophisticated models.10 Indeed, it now appears that both cytotoxic CD8+ T cells and macrophages also differentiate into Type-1 or Type-2 effector cells.14
Antigen-presenting cells (APCs) process antigens and present fragments to relatively undifferentiated naïve T-helper cells. Depending on the antigen and the cytokines secreted by the APC, these naïve cells can become committed to differentiate or polarize into Th1, Th2, or other cells.15 Th1 cells secrete interferon-γ and to a lesser extent interleukin (IL)-2 and IL-12. This drives the Th1 pathway (or cellular immune response) to fight viruses, bacteria, and other intracellular pathogens; eliminate cancer cells; and stimulate delayed-type hypersensitivity. In contrast, if the naïve T-helper cells are induced to a Th2 pathway, they will produce predominantly IL-4 and IL-13, as well as IL-5, IL-6, and IL-10. This group of cytokines drives the Th2 pathway toward humoral immunity and up regulates antibody formation to fight extracellular organisms. Type-2 dominance or deviation of immune responses is credited with contributing to the tolerance of allografts and of the fetus during pregnancy.16
Each pathway has a positive feedback loop that increases the number of Th1 or Th2 cells involved. Each cytokine set also down regulates the other pathway. Thus, a Th2 response will inhibit the Th1 pathway and deviation to a Th1 pathway will down regulate Th2 cytokine production.15 Overproduction in either pathway is associated with clinical changes and disease states. Th1 overproduction can generate organ-specific autoimmune diseases such as rheumatoid arthritis, multiple sclerosis, and Type 1 diabetes. The Th2 pathway is associated with allergy, asthma, and systemic autoimmune disease.
The immune deviation polarization process can also be affected by cytokines and other biologic modifiers produced by other cell types including CD40 ligand (CD40L, CD154), a proinflammatory mediator which is predominantly derived from PLTs and promotes a Th2 response and increases in regulatory T cells.17 In addition, the cytokine microenvironment of the patient is also important, being affected by both endogenous cytokine production in response to disease and transfusion and cytokines infused in transfused blood components. Noncytokine mediators such as chemokines, lipid mediators such as prostaglandins (e.g., PGE2), and free radicals also contribute to the signals for differentiation and influence outcome.
Placing the immunomodulatory/inflammatory effects of blood transfusion into the Th1/Th2 paradigm may help explain, in part, the apparently contradictory effects of transfusion.16 The hypothesis that the Th1/Th2 paradigm is involved in transfusion immunomodulation is supported by changes in the circulating cytokine profile in mice and in humans after transfusion.18-20 In both species allogeneic transfusion is associated with a Th2-biased immune response, consonant with clinical observations of increased antibody formation and allergic reactions. Similarly down regulation of Th1 cytokines may contribute to impaired host defenses against microorganisms, decreased macrophage antigen presentation and reduced cancer surveillance. Increased B-cell activity, which, in the current study, may be due to stimulation by transfused PLT-derived CD40L,7 could explain the increased risk of alloimmunization to RBC antigens, development of HLA antibodies, and the allergic reactions to plasma proteins after PLT transfusions, particularly those characterized by febrile reactions. Altered cellular immunity and impaired macrophage function after transfusion supports the observed increased risk of postoperative infections after transfusions and the improved survival of renal allografts.21 The exact details of how this occurs and to what extent may vary with the blood component, contaminating white blood cells (WBCs), and mediators released during storage and may also vary with the immune and cytokine status of the patient at the time of transfusion.
In the article by Yazer and coworkers, the PLT transfusions in the study arm were specifically chosen because they had caused a febrile reaction in the recipient, demonstrating an inflammatory response. All PLTs were leukoreduced, although some were leukoreduced at the time the PLTs were issued for transfusion and thus would contain both WBC-associated cytokines (IL-1, IL-6, IL-8, and tumor necrosis factor-α among others)22 as well as those derived from PLTs (IL-6, CD40L, transforming growth factor-β, PLT-derived growth factor, vascular endothelial growth factor, and RANTES).23 For reactions occurring after PLT transfusions that were leukoreduced before storage, PLT cytokines, such as the CD40L, would be primarily responsible for both the inflammatory effects and the immune deviation.4 In the study by Yazer and colleagues, it was not possible to distinguish between pre- versus poststorage leukoreduction of PLTs and potential effects on RBC alloimmunization.
This study comes in the midst of a revolution in our understanding of the relationship between thrombosis and inflammation, a subject in which the PLT plays a significant role.24 It is clear that inflammation promotes thrombosis, and vice versa, and these clinical and biologic events are also commonly seen in close temporal proximity to blood transfusions, even after PLT transfusions to thrombocytopenic hospitalized patients with cancer.25,26 Thus PLT transfusion sometimes profoundly influences innate immunity, adaptive B- and T-cell immunity, and the balance between hemostasis and thrombosis.27 Given that these functions are already dysregulated to varying extents due to a primary illness in patients requiring transfusions,25 it should be no surprise that additional untoward clinical events are associated with, and sometimes caused by, transfusions.
It is thought that PLTs promote immunity and inflammation mainly by means of the CD40/CD40L pathway. In addition, in animals, blockade of this pathway is a key element in preventing allosensitization.28 Engagement of CD40, one of the receptors for CD40L, induces synthesis of many proinflammatory mediators including cytokines, chemokines, and cyclooxygenase-2, which produces PGE2, the main mediator of fever.4,5,27 The concentration of CD40L varies enormously between individual PLT units, whether collected as whole blood–derived or apheresis PLTs. High levels of CD40L in PLT transfusions are likely clinically relevant and have been linked to adverse reactions including febrile reactions5 and TRALI.6 Immumodulation by PLT transfusions and their supernatants is also linked to impaired outcomes in human acute leukemia,29 liver transplantation,30 cardiac surgery,31 newborn necrotizing enterocolitis,32 and animal model solid tumors.33
Until recently, recognized predisposing factors for RBC alloimmunization apart from the genetic disparities between donor and recipient have included the intrinsic immunogenicity of the antigenic stimulus, ill-defined genetic factors governing the immune response, the number and frequency of RBC transfusions, the patient's underlying illness, immunosuppressive treatment regimens, and possibly the age of the patient. However, it is becoming evident that, as with other transfusion reactions such as TRALI, transfusion complications may depend on both the properties of the transfusion and the immunologic milieu of the patient into whom the blood products are transfused. A two-or-more-hit phenomenon, as proposed for TRALI and multiorgan failure, may be involved.34,35
Immunologic changes mediated by the patient's own cytokines and chemokines released as a result of infection, inflammation (e.g., sickle cell disease), trauma, surgery, anesthesia, stress, and drugs may mean that the infusion of a bolus of additional cytokines from blood products tips the patient over a cytokine threshold for a reaction to occur. In other patients, there may be insufficient endogenous cytokines circulating such that exogenous transfused cytokines do not reach the necessary triggering level. As proposed by Yazer and coworkers, levels of cytokines from a prior PLT transfusion, sufficiently high to cause a febrile reaction, may be yet another predisposing factor to reaching the level necessary to generate a humoral immune response or other immunomodulatory effects.
We found that leukoreduction of RBC transfusions was associated with a RBC alloimmunization prevalence that decreased from 8.2% to 2.8% in a carefully studied cohort of patients with acute myeloid leukemia, along with similar findings in an analysis of all alloimmunized transfused patients.36 We hypothesized that removal of allogeneic WBCs reduced the Th2 stimulus of the RBC transfusions. In these clinical situations the WBC stimulus is applied in the same transfusion rather than separated in time as in the report by Yazer and colleagues. Confirming data in an animal have been reported in preliminary form.37 Similarly, in a small study Carr and colleagues38 reported in 1990 that HLA alloimmunization occurred earlier and more frequently when ABO-incompatible PLTs were transfused compared to ABO-identical ones. Nearly 69% of patients receiving ABO-incompatible PLTs became refractory to PLT transfusion compared with 14% in those receiving only ABO identical PLTs. Refractoriness was associated with the formation not only of high-titer ABH isoagglutinins, but also of anti-HLA– and PLT-specific antibodies, which appeared simultaneously with the increase in ABH isoagglutinin titers. Patients receiving ABO-mismatched PLTs also had a higher incidence of broad specificity HLA antibodies than the ABO identical group.38 Transfusion of ABO-incompatible PLTs and soluble ABO antigens may act as a stimulant for HLA sensitization via Th2 polarization. In addition ABO immune complexes may form quickly between the infused antibody or antigen and the patient's naturally occurring ABO antibody and soluble ABO antigen.39 Immune complexes have also been shown to cause deviation of the immune system toward Th2 by inhibiting IL-12 production via IL-10, a finding that may be relevant to the common practice of transfusing ABO-nonidentical PLTs.40
The ability of one immunologic stimulus to modify a seemingly unrelated immune function is not without precedent even beyond the Th1/Th2 paradigm. It is known that use of anti-inflammatory drugs impairs in vitro B-cell activation and reduces in vivo humoral immune responses to viral antigen, which may be the flip side of increased alloimmunization when inflammation is present.41,42 Anti-PLT agents are being proposed as anti-inflammatory agents, a dual function that has been known for decades for aspirin, the prototypic drug with both anti-inflammatory and antithrombotic activity.43 It has been reported that hydatidiform molar pregnancies among patients with an ABO-nonidentical husband develop both HLA antibodies and aggressive trophoblastic tumors more commonly than patients with molar pregnancies and ABO-identical husbands. Thus continued trophoblastic proliferation after evacuation of a mole is favored by ABO disparity between the mother and the molar tissue. This could speculatively be due to impaired immune surveillance due to formation of ABO immune complexes. It is also compatible with up regulation of Th2 and down regulation of Th1 function as contributing causes of impaired antitumor immunity.44 In renal transplantation across ABO barriers, HLA antibody–mediated rejection episodes are associated with increases in ABO isoagglutinin titers, suggesting generalized Th2 activation.45
Yazer and colleagues are now planning a prospective study that will help support their early but very provocative findings. This exciting and innovative line of inquiry may lead to further unraveling of the immune modulation that occurs with transfusion and the role of inflammation in alloimmunization.
CONFLICT OF INTEREST
The authors claim no conflict of interest.
