The first function of antibodies is to act as opsonins to promote phagocytosis of sensitized platelets (19). Antibodies are fixed to the target, leading to activation of Fc receptors and phagocytosis. Anti-HLA-I antibodies or auto-antibodies directed against platelets, in case of immune thrombocytopenia (characterized by an isolated thrombocytopenia), appear to be partially caused by IgG-mediated platelet destruction in the spleen (20). Moreover, internalization of platelets with high HLA-I density by macrophages has been shown to be significantly increased in contrast to platelets coming from patients with low HLA-I density at the surface. Thus, the degree of antibody opsonisation and antigen expression directly correlates with antibody-mediated internalization of platelets by macrophages, suggesting that platelets expressing fewer antigens could be used in priority to treat refractory patients (21).

Platelet activation mediated by anti-HLA-I antibodies could partially explain immune platelet clearance during refractoriness. Some antibodies directed against HLA-I molecules have been reported to induce platelet activation (22, 23). Depending on both the antibodies and the recognized epitopes, two mechanisms have been described. First, anti-HLA-I antibodies can mobilize the FcγRIIa/CD32a expressed on the cell surface of platelets, which leads to platelet activation. FcγRIIa is a low affinity receptor for IgG subclasses. It bears an intracellular immunoreceptor tyrosine-based activation motif (ITAM) that mediates cell activation upon its crosslinking by IgG-immune complexes (24). HLA-I antibodies can initiate the signaling cascade, leading to ITAM phosphorylation and consequent platelet activation (22). Using human monoclonal antibodies that recognize different epitopes of HLA-I on platelets, the same group showed that blocking the classical pathway of the complement, by using anti-C1q or anti-C5 (Eculizumab) antibodies, inhibited to some extent the level of platelet activation. Finally, the activation induced by anti-HLA-I antibodies can be fully abrogated by the inhibition of the CD32a receptor with monoclonal antibody IV.3 in combination with complement classical pathway inhibitor Eculizumab, meaning that both FcyRIIa and complement activation pathways can act in synergy (23).

Antibody-mediated platelet clearance can also be explained by deglycosylation. In immune thrombocytopenia, anti-GPIbα antibodies cross-link GPIbα subunits, initiating activation and signaling. The activation of platelets conducts to granule secretion, after which released CD62P and NEU1 desialylate surface sialic acid residues, particularly on GPIbα. The removal of the bulky terminal sialic residues facilitates assembly of the GPIb complex and also triggers activation, thus forming an activation loop (25). Once desialylated, platelets are trapped in the liver where they can be cleared by Kupffer cells, a subset of liver professional phagocytes. A collaboration between Ashwell Morell receptors and macrophage galactose lectins, both expressed by Kupffer cells, has been described to eliminate efficiently desialylated platelets (26). More recently, the C-type lectin receptor, CLEC4F, expressed specifically on Kupffer cells, has been reported to recognize desialylated glycans and to participate in the clearance of aged platelets (27). The question of potential desialylation of platelets during PTR is still open.

The isotypes involved in PTR are predominantly IgG (28). IgGs, the major immunoglobulins in serum, are divided into four subclasses: IgG1, IgG2, IgG3 and IgG4. The differences between each subclass lie in the amino acid composition of their heavy chain and have a decisive impact on the hinge region as well as in their relative abundance. Each subclass has therefore a particular profile, particularly in terms of ability to activate the complement cascade, with IgG1 and IgG3 being more efficient in complement activation than IgG2 and IgG4, and their affinity for the Fc receptors (FcR), meaning that anti-HLA-I antibody subclasses could correlate with distinct effects on platelets (29). Moreover, a study performed in baboons suggested that the subclasses could influence the location of platelet sequestration. They showed that platelets were sequestered in the spleen when the anti-platelet antibody was IgG1, whereas IgG2 led mostly to liver elimination (30). A link between anti-HLA-I antibody subclasses and allograft outcome has also been studied, suggesting a correlation between complement fixation and allograft failure (31). All of these observations indicate that it might be useful to consider the HLA-I antibody subclass when facing an immune PTR.

IgG glycans are essential for the maintenance of functional structure and can evolve depending on time, age, disease, or environmental factors. The glycosylation of antibodies affects their affinity for FcR and their ability to mobilize complement (32) thus potentially modulating their ability to activate platelets.

Analysis of sera from alloimmunized patients showed that Fc glycosylation of anti-HLA-I antibodies is highly variable. Using recombinant glycoengineered anti-HLA monoclonal antibodies with variable Fc glycosylation, Van Osch et al. showed that galactosylation of human anti-platelet IgG enhances complement activation by promoting hexamerization of hIgG1 which constitutes an optimal platform for C1q to bind (33). Furthermore, Kapur et al. showed that anti-HLA-I IgG, collected from alloimmunized patients in PTR or from patients with neonatal thrombocytopenia, are more galactosylated than total serum IgG1 levels, but they do not present different levels of sialylation and fucosylation. Modulation of galactosylation can affect IgG clearance in vivo and predisposes them for further modification by the addition of sialic acid. By contrast, anti-human platelet antigens antibodies are less fucosylated than total serum IgG1, leading to enhanced phagocytosis of IgG-binded platelets and increased disease severity (34). Accordingly, HLA-specific afucosylated IgG1 have been proposed to be a potential predictor of antibody pathogenicity (35), illustrating the importance of sugars in antibody functions.

As analyzed by Wang et al., in a 204-patient cohort, alloantibodies were directed towards the most prevalent HLA-I A and B molecules in that population (36). WIM8E5, an anti-HLA-I monoclonal antibody that recognizes an epitope on all HLA-A antigens, except for HLA-A3 and HLA-A32 (and reduced binding to HLA-A2), has been suggested to fix complement depending on the number of epitopes available for it; however, it remains to be determined if the identity of these epitopes matters, or just the number of binding sites on HLA-I available on platelets for a given antibody (23).

It has also been proposed that a higher diversity in the antibody specificity correlates with a higher level of HLA-I antibodies as detected by screening. It is likely that the greater the range of specificities, the higher the probability that an antibody will recognize HLA-I antigens on transfused platelets (37). This scenario becomes particularly relevant in multi-transfused (i.e. oncohematologic) patients where multiple exposures could potentially boost titers of specific anti-HLA-I antibodies, as well as elicit antibody responses to new alloantigens (38).

It is possible that the presence of anti-idiotypic antibodies directed against the variable regions of anti-HLA-I antibodies can regulate the latter’s abundance (39). It remains to be studied if the anti-idiotypic antibodies recognize preferentially some V regions in anti-HLA-I with determined specificities for particular HLA antigens.

Platelet refractoriness could be overcome by transfusing platelets devoid of any HLA-I molecules. The first attempt to remove, or at least decrease, HLA antigens platelet expression was based on chemical modification. Acid treatment of human cells has been reported to eliminate the antigenicity of class I MHC molecules without significant cell death (40). Based on this discovery, several methods for preparing HLA-I depleted platelets have been reported. Acid treatment has been proven to reduce HLA-I molecules expression (41) and the subsequent HLA-antibody-mediated phagocytosis (42). Interestingly, proteomic analyses showed that the changes of platelet proteins after treatment with citric acid were functionally safe (43). However, despite the description of the absence of negative effects on proteins involved in coagulation and haemostasis, these results still need to be confirmed by functional tests. This approach is therefore attractive because it could theoretically be suitable for mass production at a reasonable cost. However, no method has yet been approved for clinical use and research is still ongoing in this field.

A promising but challenging alternative to the chemical treatment could be the use of HLA-I-deficient platelets generated in vitro. Different methods of in vitro platelet production have been reported, which differ in the source of the stem cells: induced pluripotent stem cells (iPSCs) or hematopoietic progenitors (CD34⁺ cells). Each of these offers advantages and disadvantages for the development of a transfusion product (44). To be expressed on the surface of cells, HLA-I molecules must form a heterodimer between a heavy chain and a light chain, the β2-microglobulin. The strategy used so far to suppress the expression of HLA-I molecules relies on the deletion of the β2-microglobulin gene, consequently leading to the absence of HLA-I molecules on the cell surface. HLA-I deficient platelets produced from iPSCs have been reported to successfully circulate in an alloimmune PTR model of mice reconstituted with human NK cells (45). More recently, results from the first clinical trial of ex vivo-generated platelets show that it is a safe product, however, it remains restricted to autologous and therefore individualized ex vivo platelet production (46). Thus, optimization is necessary before cultured platelets can become a plausible alternative to the HLA-compatible blood product.

Based on the data suggesting the involvement of complement in platelet HLA-I antibodies-related activation, a pilot study has been designed to assess the efficiency of eculizumab, a monoclonal antibody that inhibits C5 complement component, in platelet transfusion refractoriness (47). In this preliminary study, administration of eculizumab allows some patients to overcome PTR despite their transfusion with HLA-incompatible platelet concentrates. A clinical trial with a larger cohort should provide definitive conclusions concerning this strategy. However, it is important to keep in mind the high cost of such a treatment, which, even if it proves to be effective, will be restricted to real transfusion deadlocks but will not allow the handling of all PTR.

Drugs against underlying pathologies common in PTR patients may help in their management when transfusion with HLA-compatible products is not feasible. Rituximab, a monoclonal antibody directed against CD20 expressed on B cells, has been reported to increment platelet count in refractory patients with aplastic anemia or myelodysplatic syndrome (48, 49). Daratumumab, an anti-CD38 monoclonal antibody common in myeloma treatment, has also proven to increase platelet transfusion efficiency in refractory patients (50). However, at least for rituximab, two response profiles were identified: an early and transient response, or a late and continuous response (48). Interestingly, Bortezomib, a specific inhibitor of 26S proteasome, induced a drastic decrease in the amount of anti-HLA-I antibodies in the serum of a patient treated for multiple myeloma, resulting in clinical response to transfusion from random donors (51). In addition, several studies describe an impact of these drugs on platelet count which means that these considerations should be taken into account in the design of future pharmacological strategies to manage refractoriness to platelet transfusions (52, 53). These different approaches seem promising but should be validated with a larger cohort of patients.

One strategy used in antibody-related disease lies on the modulation of antibodies themselves. Intravenous immunoglobulin (IvIg) therapy, defined as the use of a combination of antibodies obtained from healthy human donors, or plasma exchange (PE) have been reported as a successful treatment for PTR patients (54, 55). The clinical benefit of IvIg treatment could be due to blockage of FcγR expressed by the reticuloendothelial system, where opsonized platelets would not be cleared anymore by these cells, and also due to modulation of antibodies’ effects, i.e. interference with the complement or regulation of different FcR (56). Samuelsson et al. proposed that the IvIg anti-inflammatory activity is mediated by upregulation of the inhibitory Fc receptor, FcγRIIb, as IvIg has no protective effects on a mouse model of immune thrombocytopenia lacking this receptor (57). Another interesting role for saturation of FcRn could be the increased elimination of circulating pathogenic antibodies. FcRn are MHC-class I like molecules expressed by leukocytes and endothelial cells, which control the half-life of IgG and albumin. FcRn can form a ternary complex on an IgG Fc scaffold under acidic condition, which protects IgGs from endothelial catabolism. However, Crow et al. have suggested that the mechanism of IvIg action is FcRn-independent, as IvIg improved platelet count in a murine model of ITP in FcRn-deficient mice comparable to WT mice (58).

Furthermore, evidence of its benefit in immune-mediated PTR is still lacking and a large clinical study should be conducted to draw firm conclusions about its potential efficiency in this particular context.

IgG-degrading enzyme from S. pyogenes (IdeS) is a cysteine protease that cleaves the heavy chain of IgG, generating one F(ab’)2 fragment and two Fc fragments (59). Since IdeS can selectively and rapidly neutralize the Fc-mediated effector function of human IgG, it is a promising new therapy for treating IgG-driven disorders. Recently, an elegant study describing the use of IdeS in a platelet immune disorder has been reported. A recombinant protein with the N-terminus of IdeS and the C-terminus of a single-chain variable fragment from a CD32a antibody was generated. This protein can bind specifically to all cells expressing CD32a, including platelets, and it can neutralize anti-platelet antibodies. Interestingly, platelets with surface-bound IgG-degrading enzymes are protected from clearance in murine models of immune thrombocytopenia (60). This technology could represent a very promising strategy to avoid or alleviate PTR in the future.