Data Loading...
The pathogenesis of immune thrombocytopaenic purpura Flipbook PDF
Fifthly, why do some patients respond to certain therapies and others do not? The ... pathogenesis of ITP has yet to be
99 Views
111 Downloads
FLIP PDF 271.93KB
review
The pathogenesis of immune thrombocytopaenic purpura Nichola Cooper and James Bussel Department of Pediatrics, Weill Medical College of Cornell University, New York, NY, USA
Summary Immune thrombocytopaenic purpura (ITP) is an autoimmune bleeding disease that is rarely fatal. However, in many adults treatment is unsatisfactory, with as much morbidity from the immunosuppressive effects of treatment as from bleeding. Identifying the underlying disease process should help us to identify more targeted therapies and improve not only the treatment but also the quality of life of patients with this disorder. Keywords: thrombocytopaenia, immune thrombocytopaenic purpura, pathology, T cells, cytokines. Immune thrombocytopaenic purpura (ITP) is a heterogeneous disease characterised by increased platelet destruction and thrombocytopaenia. A number of features suggest this destruction is immune-mediated and that it may involve not only the destruction of the platelet, but also inhibition of platelet release by the megakaryocyte. The exact mechanism of the immune dysfunction, however, is generally not known. For example it is unclear whether ITP is initially caused by a B-cell abnormality, a T-cell disorder, an abnormality of thrombopoiesis, or even from increased mononuclear phagocyte activation. In addition, certain patients who apparently have ITP may have an indolent form of myelodysplasia that is not yet evident on bone marrow examination. Some of the difficulties in defining the pathology of ITP arise because it a heterogeneous disease with individual patients having different causes of thrombocytopaenia and other difficulties relate to the limited nature of assays, such as the antiplatelet antibody. There are many issues connected to the pathology of ITP that need resolution or at least improved understanding. First, what initiates ITP; what is the underlying defect resulting in the accelerated platelet destruction? Secondly, what maintains the disease; why do some people recover while others have persistent thrombocytopaenia? Thirdly, what is the mechanism of the thrombocytopaenia in different patients? Fourthly, why do some patients have severe thrombocytopaenia and bleeding, even serious bleeding, while others with the same platelet
Correspondence: James Bussel, Professor of Pediatrics, Weill Medical College of Cornell University, 525 East 68th Street, P-695 New York, NY 10021, USA. E-mail: [email protected]
doi:10.1111/j.1365-2141.2006.06024.x
count are relatively asymptomatic? Fifthly, why do some patients respond to certain therapies and others do not? The answers to the latter, more clinical questions presumably reflect the physiological processes in the first three questions. Finally, does ‘ITP’ associated with other diseases, such as systemic lupus erythematosus (SLE) and hypothyroidism have the same pathology as primary (idiopathic) ITP? A better understanding of these questions will lead to improved management of ITP, including more appropriate choice of when to treat and which treatment to select.
The role of the spleen In 1916, Kaznelson, while a medical student in Vienna, prevailed upon the attending surgeon to perform a splenectomy in a patient with ITP. The splenectomy was successful in normalising the platelet count and, with other cases, first established the critical role of the spleen in ITP (Kaznelson, 1916). However, the cause of thrombocytopaenia remained unclear. Was the spleen destroying the platelets or did it secrete a suppressive substance that inhibited platelet production and/ or release into the circulation? Doan et al (1960) examined a number of spleens from patients with ITP. They demonstrated sea blue (lipid laden) histiocytes in the spleen, suggesting it was the platelet ‘destroyer’. What directed the spleen to prematurely destroy platelets, however, remained unclear.
The antiplatelet factor The first direct evidence that ITP is caused by a plasma-derived antiplatelet factor was provided by Harrington et al (1951), when they showed that infusion of plasma from patients with ITP-induced thrombocytopaenia in normal recipients. Shulman et al (1965) demonstrated that this thrombocytopaenic effect of patient plasma was dose-dependent and that a larger dose was required to create equivalent thrombocytopaenia in recipients who had been splenectomised. They showed that the thrombocytopaenic factor was in the immunoglobulin (Ig)Grich serum fraction that it could be adsorbed by platelets and that it reacted with autologous platelets. On this basis and because of transient neonatal thrombocytopaenia in infants of mothers with ITP, this factor was conjectured to be an antiplatelet antibody. Platelet-associated IgG was first quantified by Dixon and Rosse (1975). The initial studies utilised the
ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 133, 364–374
Review whole platelet as the target and were very exciting because they demonstrated a high degree of sensitivity, >85–90%, in patients with ITP (reviewed by McMillan, 1981). However, it soon became clear that this high sensitivity was accompanied by a low specificity. The specificity of the antiplatelet antibodies to individual platelet glycoproteins (GP), such as GPIIb/IIIa, was demonstrated by van Leeuwen et al (1982) using the platelet immunofluorescent test (PIFT). In 32 of 42 cases, platelet IgG eluated from patients with ITP bound to normal platelets but not to platelets from patients with Glanzmann’s thrombasthenia, who lack platelet GPIIb/IIIa. Subsequently, both platelet-associated and plasma-derived antibodies have been identified with varying degrees of specificity and sensitivity. Current studies using platelet GP-specific assays, modelled on the monoclonal antibody immobilised platelet antigen (MAIPA; Kiefel et al, 1987), maintain a high specificity, 85–90%, at the expense of a lower sensitivity, 50–70%. The coating of platelets with antibodies rarely seems to interfere with their function, as patients with ITP very infrequently suffer from major bleeding when their counts are >50 · 109/l. Rather these antibodies target them for early destruction by the mononuclear phagocyte system (MPS), primarily in the spleen but also in the liver and bone marrow. The destruction of platelets by the MPS seems to result in the presentation of additional platelet antigens to the immune system by the antigen-presenting cells with epitope spreading. Hence, patients with chronic ITP often have antiplatelet antibody with specificity directed to multiple GP; i.e. antiGPIIb/IIIa and anti-GPIb/IX and anti-GPIa/IIa (Cines & Blanchette, 2002). What causes the initial development of the antiplatelet antibodies, however, is not clear. The VH3-30 heavy chain has been found to be highly represented among plateletreactive Fab fragments from patients with ITP when compared with its prevalence in the general library (Roark et al, 2002). This VH3-30 heavy chain gene has also been implicated in the pathogenesis of diseases such as autoimmune haemolytic anaemia (AIHA), SLE, chronic lymphocytic leukaemia (CLL), common variable immunodeficiency (CVID) and human immunodeficiency virus (HIV) infection, which may explain why ITP often occurs with these diseases and is consistent with the finding that some patients with ITP have been shown to have oligoclonal B-cell populations (van der Harst et al, 1990). Quite separately, other studies have shown light chain restriction of antibodies in patients with ITP (Stockelberg et al, 1995; McMillan et al, 2001). These two studies suggest that antiplatelet autoantibodies may be clonally restricted and that antiplatelet antibodies are produced from a limited number of B-cell clones. As antigen-driven affinity selection and somatic mutation are involved, this indicates T cell-driven antibody production. The above observations, suggesting B-cell abnormalities in some patients and T-cell abnormalities in others, expose one
facet of the heterogeneity of the thrombocytopaenia occurring in patients with ITP.
Immune tolerance The immune system has a number of mechanisms to establish and maintain self-tolerance. The primary mechanism is central thymic tolerance as a result of the deletion of differentiating T cells that express antigen-specific receptors with high binding affinity for intrathymic self-antigens. Only self-reactive T cells of low affinity and T cells with receptors specific for antigens that are not represented intrathymically mature and join the peripheral T-cell pool (Kruisbeek & Amsen, 1996). However, a number of potentially pathogenic, self-reactive T cells survive and form part of the normal T-cell repertoire (Fowell & Mason, 1993; Hafler & Weiner, 1995; Sakaguchi et al, 1995). Control of these cells involves post-thymic or peripheral tolerance of which there are a number of mechanisms. For example, self-antigens may be sequestered from the circulation in tissues, such as the lens and the testes (Doherty, 1997). Secondly, under certain conditions, self-reactive T cells are deleted or rendered anergic because they do not receive the required co-stimulatory signals (Schwartz, 1996). Thirdly, tolerance to self-antigens can be maintained by T-regulatory cells, which keep self-reactive T cells in check (Le Douarin et al, 1996; Takahashi et al, 1998; Seddon & Mason, 1999). Finally, peripheral tolerance is also achieved by B-cell deletion in the bone marrow. Differentiating B cells that express surface immunoglobulin receptors with high binding affinity for membrane-bound self-antigens are progressively deleted as they mature in the bone marrow. The majority of antibodies expressed by early immature B cells are self-directed. Almost all of these cells are removed from the developing population at two checkpoints during B-cell development (Wardemann et al, 2003).
Autoimmunity Despite these complex mechanisms to achieve and maintain tolerance, a number of self-reactive B and T cells survive. Usually these cells are not given the complex set of costimulatory signals required to proliferate and become activated, i.e. via CD40–CD40 ligand interactions or CD28–CD80/ 86; they therefore remain quiescent. Tolerance can be overcome in a variety of ways. Cross-reaction of antigens, such as those found on bacteria or viruses, may stimulate selfreacting B cells via molecular mimicry. Helicobacter pylori and varicella zoster virus may be examples of this. A number of antigens are able to provide T-independent signals to B cells. Polyclonal activation of B cells by stimuli as diverse as Epstein– Barr virus (EBV), malaria and graft-versus-host disease (GVHD) may result in autoantibodies being formed. Interruption of the checkpoints at which the immature B cells producing autoreactive antibodies are removed from the circulation will also result in an increase in the numbers of
ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 133, 364–374
365
Review autoantibodies (Wardemann et al, 2003). Finally, breakdown of the suppressor cell and also of the anti-idiotype regulatory networks can result in autoimmune reactions. This latter explanation is thought to be the one linking autoimmunity and immunodeficiency and therefore may be the one most pertinent to ITP. An additional potential factor in the autoimmunity of ITP is that platelets can present themselves to the immune system by membrane major histocompatibility complex (MHC) class I molecules, enabling destruction by cell-mediated mechanisms. When activated, platelets increase their expression of CD40L, and potentially other immune molecules, which may contribute to their immunological recognition and therefore to the development of autoimmunity to platelets. Human leucocyte antigen (HLA) types have only been clearly linked to the development or clinical course of ITP in genetically homogeneous populations, such as the Japanese (Nomura et al, 1998). Nonetheless, immune recognition as a result of HLA may also be critical in development of autoimmunity in heterogeneous populations.
Th1 and Th2 responses CD4+ helper T (Th) cells are thought to be the most important regulatory component of the immune system. They regulate the response to infection and control the immune system to prevent autoimmunity. Th cells secrete cytokines and modulate the cellular response to antigens. Two main ‘types’ of T-helper cells (Th1 and Th2) have been described. While an oversimplification, they nonetheless are useful in considering patterns of the immune responses. Generally, Th1 cytokines include interferon (IFN)-c, tumour necrosis factor (TNF)-b and interleukin (IL)-2. Th1 cells are
involved in cell-mediated inflammatory reactions and delayed hypersensitivity reactions. Several of the Th1 cytokines activate cytotoxic, inflammatory and delayed hypersensitivity reactions. In contrast, Th2 cells produce IL-4, IL-5, IL-6, IL-9, IL10 and IL-13. These cells encourage the production of antibodies and are associated with regulation of strong antibody and allergic responses.
T-cell abnormalities in patients with ITP A number of T-cell abnormalities have been demonstrated in patients with ITP (summarised in Fig 1) and it is likely that there are three main mechanisms by which T cells could be involved in the thrombocytopaenia in patients with ITP. First, a number of studies suggest a Th1 bias, compared with Th2, in adults with chronic ITP. For example, increased numbers of HLA-DR+ T cells, increased soluble IL-2 receptors, and a cytokine profile suggesting the activation of precursor helper T and type 1 helper T cells have been described (Semple et al, 1996, reviewed in Andersson & Wadenvik, 2004). Reduced levels of IL-10 have also been described in patients with active disease when compared with those in remission or healthy controls (Andersson et al, 2002) but conversely, raised IL-10 levels have been described in children with chronic ITP (Semple et al, 1996; Mouzaki et al, 2002). Further evidence of Th1 involvement in the pathology of ITP is illustrated by an increase in the Th1 cytokines, IL-2 and IFN-c, in patients with ITP when compared with controls (Panitsas et al, 2004). Interestingly, this increase was more marked in patients in remission than in those with active disease. The same study also found suppression of expression of Th2 cytokines, IL-4 and IL-5, in patients with active disease relative to patients in remission and suppression of IL-10 expression, following
Fig 1. Summary of the variety of T-cell abnormalities found in adults with ITP. PBMC, peripheral blood mononuclear cells; MLR, mixed lymphocyte reaction; ITP, immune thrombocytopaenic purpura.
366
ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 133, 364–374
Review mitogenic stimulation in patients with active disease. Overall, they describe significantly increased Th1/Th2 ratios in patients with both active and quiescent disease when compared with controls. These findings may be related to ongoing immune activation as part of autoimmunity. The activity of regulatory T cells and the potential for in vivo T-cell exhaustion because of prolonged in vivo activation has not been well studied in ITP. A second method of potential T-cell involvement is the release of cytokines that interfere with megakaryocyte maturation and/or platelet release. Transforming growth factor (TGF)-b1 level has been inversely correlated with disease activity (Yoshimura et al, 2000; Andersson et al, 2002). The role of TGF-b1 in ITP is thought to be as a potent inhibitor of megakaryocyte maturation. Two studies have shown increased granulocyte-macrophage colonystimulating factor (GM-CSF) levels, and one increased macrophage (M) CSF levels, suggesting that monocytemacrophage activation is associated with ITP (Abboud et al, 1996). Circulating cytokines may also alter the response of HLA class II presentation, and/or influence the interaction between B and T lymphocytes causing pre-existing B cells to proliferate and produce high-affinity autoantibodies (reviewed by Chanock, 2003). Finally, there is evidence to suggests a direct cytotoxic effect of T cells, as illustrated in Fig 2. By DNA microarray screening, Olsson et al (2003) found increased expression of several cytotoxic genes, such as granzyme A, granzyme B and perforin, as well as increased expression of genes involved in the Th1 cell response, such as INFc and IL-2 receptor-b in a small number of patients with ITP when compared with controls. As apparent compensation for this increased cytotoxicity, they also found increased expression of the killer cell immunoglobulin-like receptor (KIR) family on CD3+ T cells in patients with ITP in remission when compared with controls and to those with active ITP. KIRs downregulate cytotoxic T lymphocytes (CTL) and natural killer cell (NK) responses by binding to MHC class 1 molecules, preventing lysis of target cells. These findings suggest that CTLs may be involved in ITP. In a direct assay similar to that measuring NK cell activity by using radiolabelled K562 targets, these investigators assessed platelet destruction in vitro by T cells. They found that six of eight patients with active ITP showed platelet lysis by T cells whereas none of the patients in remission did. The effector cells were found to be CD3+CD8+ T cells (Olsson et al, 2003).
T Cell: activated CD8+
Platelet
Lysis
KiR Fig 2. Hypothetical effect of T cells on platelets in immune thrombocytopaenic purpura (ITP).
This expanded role of cytotoxic T cells may explain why not all of the patients originally described by Harrington et al (1951) had a fall in their platelet count following the infusion of plasma from patients with ITP, and may also explain a percentage of patients without measurable antiplatelet antibodies, and again points to the heterogeneity of this disease.
Complement The role of complement (C) and complement receptors in the pathogenesis of ITP has yet to be defined. Several studies have demonstrated increased platelet-associated C3 and C4 on ITP platelets but these are thought to be secondary in importance to platelet IgG and/or the result of antiplatelet IgM (Hed, 1998). Furthermore, there is an association of (especially) C4 deficiency and ITP that has not been well studied and is of additional interest because the C4 genes are in the midst of the HLA region on chromosome 6.
Megakaryocytes and platelet production It has always been assumed that there is compensatory but inadequate increased platelet production in patients with ITP. The two initial studies of platelet survival suggested that platelet production was increased but only by one- to threefold and not in all patients (Harker, 1970; Branehog et al, 1974). A number of follow-up studies using chromium-labelled allogeneic platelets confirmed ‘rapid platelet turnover’; several even showed a platelet survival time as short as several minutes (Mueller-Eckhardt, 1988). Calculations based on the apparent high rate of turnover suggested a substantially increased platelet production. However, in the early 1980s, the survival time of circulating autologous platelets in several studies using 111-indium showed that platelet survival was longer than expected (near normal). Therefore platelet turnover, and by inference platelet production, was found either to be decreased or at best normal in approximately two-thirds of patients with ITP (Heyns et al, 1982, 1986; Stoll et al, 1985; Ballem et al, 1987; Gernsheimer et al, 1989). This finding is presumed to be due to a direct effect of antibody on megakaryocyte maturation or platelet release. An antibody effect on megakaryocytes is consistent with the fact that megakaryocytes are known to express GPIIb–IIIa and GPIb–IX on their surfaces (Vainchenker et al, 1982) that most ITP antiplatelet autoantibodies react with one or both of these GP complexes (McMillan et al, 1987; Kiefel et al, 1992) and also that platelet antibody has been demonstrated on megakaryocytes similar to the way that antibodies have been shown to be bound to platelets. However, direct megakaryocyte or intramedullary platelet destruction is unlikely because of the adequate numbers of megakaryocytes and the absence of lipidladen macrophages in the bone marrow. Two recent studies examined the effects of ITP plasma on control megakaryocytopoiesis in vitro. Chang et al (2003) assessed the effect of plasma from patients with childhood ITP
ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 133, 364–374
367
Review (44 acute and nine chronic) on thrombopoietin (TPO)induced production of megakaryocytes from cord blood cells in liquid culture in vitro. They described suppression of in vitro production of megakaryocytes from cord blood cells by plasma from ITP patients with detectable antiplatelet antibodies. Plasma from control subjects or patients with ITP without detectable antibodies did not have the same effect (Chang et al, 2003). McMillan et al (2004) similarly studied the effect of plasma from adult patients with chronic ITP on in vitro megakaryocytopoiesis. CD34+ cells from healthy donors were cultured in medium-containing pegylated recombinant human megakaryocyte growth and development factor (PEG-rHuMGDF, a form of TPO) and 10% plasma from either ITP patients or healthy subjects. The cultures containing plasma from 12 of 18 ITP patients showed a significant decrease (26– 95%) in megakaryocyte production when compared with control cultures. ITP plasma also inhibited megakaryocyte maturation, resulting in fewer 4N, 8N and 16N cells. A study of 205 patients with chronic ITP found only one case of anti-TPO antibody, demonstrating that this was not a factor in these findings (Aledort et al, 2004). Finally, a number of studies have shown contradictory results, describing either increased numbers of megakaryocyte colony-forming units (CFU-M; Bellucci et al, 1991; Houwerzijl et al, 2004), or decreased numbers of CFU-M in patients with ITP (Abgrall et al, 1993). While antibodies appear to mediate this effect on platelet production, other mechanisms, i.e. T cell-mediated inhibition of platelet production, are possible and largely unexplored and could have an effect by altering the cytokine milieu of the bone marrow. In addition, as previously commented upon, rare cases of myelodysplasia, characterised as having poor platelet production, may resemble ‘refractory’ ITP especially because they may also have a component of autoimmunity. Recent studies with a thrombopoietic agent in ITP have suggested that there is a dose-dependent increase in the platelet count following a single or multiple injections (Bussel et al, 2003; Kuter et al, 2004; Newland et al, 2004). This is in contrast to the lack of effect of rHuIL-11 (Bussel et al, 2001). In vitro studies in mice have explored the role of TPO and discovered that, although it stimulates proliferation of precursor cells (CFU-M), other agents, such as fibroblast growth factor (FGF) and stromal cell-derived factor (SDF1), may be required to facilitate platelet release into the circulation (Avecilla et al, 2004). This may in part explain the residual platelet level (15% of normal) seen in TPO or TPO receptor (TPO-R) knockout mice. Humans with mutations of the TPO-R, who have amegakaryocytic thrombocytopaenia; however, have platelet counts that are typically considerably