Wiskott-Aldrich syndrome: current research concepts.

Abstract

Wiskott-Aldrich syndrome (WAS) is a rare X-linked condition characterized by the triad of immune deficiency, eczema and thrombocytopenia ( Sullivan et al, 1994 ; Remold-O'Donnell et al, 1996 ). The disease is invariably fatal unless treated by allogeneic bone marrow transplantation ( Mullen et al, 1993 ). Progress of research into the pathogenesis of WAS has been greatly enhanced by the identification of the gene that is mutated both in WAS and in the milder condition, X-linked thrombocytopenia (XLT) ( Derry et al, 1994 ). The protein product of this gene, named Wiskott-Aldrich syndrome protein (WASP), has not been identified previously, and initial inspection of its amino acid sequence gave few clues to its probable function. Subsequent biochemical studies have indicated that WASP is a cytoplasmic protein, expressed only in haemopoietic cells, and that it plays a key role in regulating changes in cytoskeletal structure in response to external stimuli. This conclusion is supported by recent data from cell biology studies which suggest that abnormal motility and morphology of haemopoietic cells could account for many features of WAS. The present hope, therefore, is that further studies of WASP will not only lead to improved understanding of the pathogenesis of WAS and, perhaps, to development of novel therapies, but will also shed light on the biochemistry of cytoskeletal regulation in haemopoietic cells and on the role of cell migration in the regulation of inflammatory and immune responses. Most WAS patients who are symptomatic before diagnosis have clinical manifestations of thrombocytopenia and increased susceptibility to infection ( Sullivan et al, 1994 ). Although thrombocytopenia is present from birth, and is persistent unless treated, overt symptoms of immune deficiency, including eczema, appear later in the course of disease and are variable and progressive. Some individuals with XLT never exhibit symptoms of immune deficiency, even though molecular genetic studies have demonstrated that XLT and WAS result from mutations in the same gene ( Schwarz, 1996). The thrombocytopenia of WAS is believed to result primarily from accelerated platelet destruction, although thrombopoiesis may also be intrinsically abnormal ( Semple et al, 1997 ). Splenectomy typically results in an increase in platelet number and size towards normal, but this response is often incomplete ( Corash et al, 1985 ). Clinical manifestations of the immune disorder include susceptibility to pyogenic, viral and opportunistic infection, eczema, autoimmune phenomena, and increased incidence of lymphoproliferative disease ( Sullivan et al, 1994 ). Progressive decrease in T-cell number and function during childhood is associated with restricted defects in proliferative responses of WAS T cells, deficient antibody responses, particularly to polysaccharide antigens, and low or absent levels of isohaemagglutinins ( Remold-O'Donnell et al, 1996 ). Examination of T-cell lines derived from WAS patients showed that WAS T cells have abnormal morphology ( Remold-O'Donnell et al, 1996 ; Galego et al, 1997 ). When derived from normal individuals, such cells are round and have surfaces covered with many tiny projections called microvilli. Electron microscopy reveals a network of actin filaments beneath the cell membrane, extending into the microvilli. WAS cells, in contrast, were found to be irregular in shape and to have greatly reduced numbers of microvilli and a poorly delineated submembrane actin network. Taken with the observation that WAS platelets are abnormally small, these data provided the first hint that WAS might result from abnormal cytoskeletal organization in haemopoietic cells. In general, heterozygous female carriers of WAS are clinically unaffected and exhibit non-random X-inactivation in all of their haemopoietic cells, including CD34-positive progenitors ( Wengler et al, 1995 ). Thus, the X chromosome carrying the wild-type WASP gene is active in all haemopoietic cells. In contrast, X-inactivation in non-haemopoietic tissues follows a typical random pattern. It is therefore possible that in heterozygous female carriers of WAS, cells with an active mutated X chromosome fail to compete successfully with normal cells at some critical stage of haemopoiesis, perhaps even during development of the haemopoietic system in utero. The possible cellular mechanisms of this failure will be discussed later. Recently, X-linked WAS has been described in a heterozygous girl with unbalanced inactivation of the normal X chromosome in non-haemopoietic cells, as well as in haemopoietic cells. The reasons for this are unclear at present, but studies of other family members suggest that it may have arisen as a result of an independently determined genetic abnormality of X-inactivation ( Parolini et al, 1998 ). Linkage analyses showed that WAS resulted from mutation of a single gene mapping to Xp11.22, and in 1994 this gene was isolated by positional cloning ( Derry et al, 1994 ). The predicted protein product of the gene comprised 502 amino acids and, since the amino acid sequence bore no obvious similarity to any known protein, it was named Wiskott-Aldrich syndrome protein (WASP). Western blotting and immunocytochemical analyses subsequently showed that WASP is a cytoplasmic protein with a molecular weight of approximately 65 000, and that it is expressed in all haemopoietic lineages, including CD34-positive progenitors from bone marrow ( Parolini et al, 1997 ; Zhu et al, 1997 ). Our recent work shows that WASP is expressed at the earliest detectable stages of embryonic haemopoiesis, in clusters of CD34+ cells on the ventral wall of the dorsal aorta in 36-day-old human embryos (C. Marshall, unpublished observations). The first real clues to the function of WASP came from the discovery that it binds to a number of other proteins. Derry et al (1994 ) had noted that the predicted amino acid sequence of WASP contained several proline-rich stretches reminiscent of binding sites for SH3 domains (Fig 1). SH3 domains are found in a large number of proteins involved in intracellular signal transduction, and are involved in mediating assembly of multi-protein complexes in the cell ( Cohen et al, 1995 ). As expected, WASP was found to bind to a variety of SH3-containing proteins in vitro, and such binding was found to involve the proline-rich regions of WASP ( Cory et al, 1996 ; Finan et al, 1996 ). However, many of these in vitro interactions cannot be detected in vivo, even when WASP and its binding partner are artificially expressed at high levels in cultured fibroblasts or epithelial cells. Only three SH3-containing proteins, Fyn ( Banin et al, 1996 ), Nck ( Rivero-Lezcano et al, 1995 ) and Grb2 ( Miki et al, 1997 ), are bound to WASP naturally in haemopoietic cells. Fig 1. Schema of the Wiskott-Aldrich syndrome protein (WASP), showing the protein's domains and the protein–protein interactions in which they are known to participate in vivo. N, amino terminus; C, carboxyl terminus; PH, pleckstrin homology domain; G, GTPase-binding domain, PRR, proline-rich region; WH2, WASP homology 2 domain, including regions of homology to verprolin and cofilin. In other proteins, including N-WASP, PH domains mediate binding to membrane phospholipids. However, there is no evidence that the PH domain of WASP serves this function. WIP binds to the amino-terminal one-third of WASP, but the precise location of the binding site has not been reported. Nck and Grb2 are adaptor proteins that lack intrinsic enzyme activity but are responsible for mediating assembly of complexes of intracellular signalling proteins. The significance of their interaction with WASP is unclear, but it is possible that they couple WASP to cell surface membrane receptors. Fyn is a member of the c-Src family of cytoplasmic protein-tyrosine kinases ( Brickell, 1996). Other c-Src family members bind to WASP in vitro, but not in vivo ( Banin et al, 1996 ). Fyn is involved in transducing signals from T-cell receptor (TCR) and B-cell receptor (BCR) complexes ( Chan et al, 1994 ; Pleiman et al, 1994 ), and is also thought to mediate signalling in other haemopoietic lineages, including myeloid cells. For example, Fyn is physically associated with the LFA1/CR3/uPA-R complex, which plays a key role in regulating migration of myeloid cells ( Bohuslav et al, 1995 ). Interestingly, Fyn is involved in signal transduction pathways that regulate cytoskeletal organization in fibroblasts ( Thomas et al, 1995 ), and other c-Src family proteins are also important regulators of the cytoskeleton. For example, c-Src is required for formation of ruffled membrane borders in osteoclasts ( Boyce et al, 1992 ). These data suggested that the binding of WASP to Fyn might represent a step in transduction of signals that regulate cytoskeletal organization in haemopoietic cells ( Banin et al, 1996 ). The biochemical consequences of WASP–Fyn binding are unclear, but our preliminary results suggest that the kinase activity of Fyn is stimulated by binding to WASP, and that WASP may be a substrate for Fyn (S. Banin, unpublished observations). More direct evidence that WASP has a role in regulating cytoskeletal architecture came from the finding that WASP binds to the cytoplasmic protein Cdc42 ( Aspenström et al, 1996 ; Kolluri et al, 1996 ; Symons et al, 1996 ). Cdc42 belongs to the Rho family of small GTPases and regulates formation of distinct actin-filament containing structures, known as filopodia, in fibroblasts and macrophages ( Allen et al, 1997 ). The GTPase-binding domain shown in Fig 1 mediates binding of Cdc42 to WASP. Finally, experiments in which WASP was expressed artificially in fibroblasts demonstrated that WASP could stimulate formation of actin filaments. This effect of WASP was inhibited by co-expression of WASP with a dominant negative Cdc42 mutant, which inhibits the activity of endogenous wild-type Cdc42. This suggests that WASP regulates organization of the actin cytoskeleton as a downstream effector of Cdc42 ( Symons et al, 1996 ). Deletion of the carboxyl-terminus of WASP, including part of the WH2 domain (Fig 1), abolished its ability to stimulate formation of actin filaments, suggesting that WASP may interact with the cytoskeleton via the WH2 domain ( Symons et al, 1996 ). In support of this idea, the WH2 domain contains regions of homology to verprolin, which is involved in regulating the actin cytoskeleton in yeast, and cofilin, which is an actin-binding protein with actin-depolymerizing activity ( Miki et al, 1996 ). Other recent work, however, suggests that WASP may interact with the cytoskeleton via another protein, WIP, which binds to the amino-terminal one-third of WASP ( Ramesh et al, 1997 ). Taken together, these data suggest that WASP is involved in regulating the architecture of the actin cytoskeleton in haemopoietic cells, in response to external signals (Fig 2). This fits with earlier notions, based on observations of cell morphology, that abnormalities of the cytoskeleton are involved in the pathogenesis of WAS. Fig 2. Signal transduction pathways in which WASP might participate. BCR, B-cell receptor; TCR, T-cell receptor. If WASP has an important role in regulating cytoskeletal architecture, it is surprising that its expression is restricted to haemopoietic cells. However, recent work has identified another protein, N-WASP, which has a very similar structure to WASP and is expressed in brain and, possibly, at low levels in other tissues ( Miki et al, 1996 ). When artificially expressed in epithelial cells, N-WASP (but not WASP itself) acts downstream of Cdc42 to stimulate actin de-polymerization and to generate long microspikes from the cell surface in response to epidermal growth factor ( Miki et al, 1996 , 1998). Actin de-polymerization activity required the presence of the cofilin homology region ( Miki et al, 1996 ). In vitro work showed that Cdc42 stimulates N-WASP to create free barbed ends on actin molecules, from which actin polymerization can take place, possibly through the activity of the actin-binding protein profilin ( Miki et al, 1998 ). If neurons are the site of N-WASP expression in brain, which has not yet been established, it is attractive to believe that it plays a role in regulating cytoskeletal changes in neurite outgrowth or vesicle trafficking. If the behaviour of N-WASP expressed artificially in epithelial cells is a guide to the behaviour of endogenous WASP in haemopoietic cells, it may be that WASP also responds to external signals by creating free barbed ends from which actin polymerization can occur, promoting formation of filopodia. WASP gene sequences have now been determined for a large number of WAS patients ( Schwarz, 1996; Zhu et al, 1997 ). Many different types of mutation have been found, with few occurring in more than one family. Most mutations are single base changes that introduce missense or nonsense sequences into the WASP coding region. These are found throughout the coding region, but are concentrated towards the 5′ end, in the region encoding the PH domain (Fig 1). Single codon deletions or insertions make up the second largest class of mutation, with a few cases resulting from intronic mutations. Recent studies have examined the effects of WASP gene mutations on WASP biosynthesis. Western blotting analyses of peripheral blood monunuclear cells (L. McCarthy-Morrogh, unpublished observations) and B-lymphoblastoid cell lines ( Remold-O'Donnell et al, 1997 ; Zhu et al, 1997 ) have shown that many mutations, including many missense mutations, result in complete absence, or greatly reduced levels, of WASP. Extensive efforts have been made to link the nature of mutations seen in individual patients with the severity of their disease. It has been reported that missense mutations in the PH domain tend to result in reduced levels of WASP and mild clinical disease, whereas missense mutations affecting the C-terminus, as well as nonsense and intron mutations, tend to result in absence of WASP and severe disease ( Zhu et al, 1997 ; Remold-O'Donnell et al, 1997 ). However, this correlation is far from absolute. Indeed, there are a number of individuals with severe WAS who have the same mutation as individuals with XLT ( Remold-O'Donnell et al, 1996 ). This suggests that there are other important genetic and/or environmental factors governing the biological effects of WASP gene mutations. For example, there could be genetic variation in intracellular factors that control the stability of WASP and/or in extracellular factors, such as cytokines, extracellular matrix components or other cell types, that affect the behaviour of cells containing a mutated WASP gene. Alternatively, development of immune defects could depend on the nature of each individual's exposure to extrinsic factors such as infection. Thus, WAS, like many so-called single gene defects, is in fact a multifactoral disease to which a number of genetic and environmental factors may contribute, but in which WASP gene mutations play the major part. Abnormal regulation of cytoskeletal architecture could disrupt the structural integrity of platelets and diminish the ability of lymphocytes to polarize immune receptors, adhesion receptors, and chemokine receptors when activated, resulting in abnormal immune responses. For example, polarization of T cells towards antigen-presenting cells requires Cdc42 ( Stowers et al, 1995 ) and might therefore be defective in WAS patients. However, the contribution of individual cell types to the immune deficiency of WAS has not been determined, and in our own recent work we have focused on the effects of WASP mutations on the morphology and motility of myeloid cells. Early steps in cell migration include polarization of the actin cytoskeleton to achieve spatial asymmetry, and the extension of filopodia and lamellipodia at the leading edge of the cell ( Lauffenberger & Horwitz, 1996). Since filopodial extension in macrophages requires Cdc42 ( Allen et al, 1997 ), and therefore possibly WASP, we compared the motile characteristics of macrophages from WAS patients with those from matched normal controls, using a direct-viewing Dunn chemotaxis chamber ( Zicha et al, 1997 ). We found that WAS macrophages migrated normally on glass, but, unlike normal macrophages, failed to chemotax in a gradient of CSF-1. In contrast, neutrophils from WAS patients were indistinguishable from normal neutrophils in their ability both to migrate and to chemotax ( Zicha et al, 1998 ). Interestingly, locomotion of neutrophils is known to differ from that of macrophages, the former being more amoeboid, and the latter more fibroblastic ( Condeelis, 1993), and neutrophil chemotaxis may therefore be regulated by a Cdc42/WASP-independent signal transduction pathway. Our more recent work has also demonstrated a defect in dendritic cells from WAS patients which could explain many of the features of the immune deficiency seen in these individuals. Dendritic cells are bone-marrow-derived antigen-presenting cells required for the initiation of immune responses ( Steinman, 1991). Immature dendritic cells are present as phenotypically heterogenous subsets in most non-lymphoid tissues, and are characterized by their potent antigen capture and processing function. Following encounter with antigen, these sentinel dendritic cells migrate to T-cell-rich areas of secondary lymphoid organs. During this time they lose the ability to capture antigen, but become potent stimulators of T-cell immunity ( Ibrahim et al, 1995 ). Cells with the immunophenotype, morphology and antigen-presenting activity of dendritic cells can be derived from human CD14-positive peripheral blood precursors cultured in the presence of IL-4 and GM-CSF ( Sallusto & Lanzavecchia, 1994). We therefore examined the characteristics of peripheral-blood-derived dendritic cells (PBDC) generated from WAS patients. We found that WAS PBDC had the same surface immunophenotype as normal PBDC, but differed from normal PBDC in failing to extend dendritic processes and failing to migrate when plated on fibronectin (unpublished observations). This suggests that the major immunological deficits of WAS could be explained by failure of dendritic cells to polarize and traffic normally in vivo. Maturation of dendritic cells from sentinel cells to messenger cells in situ, rather than during migration to draining lymphoid tissue, could result in dysregulated local activation of T cells and/or breakdown of dendritic cell-mediated peripheral tolerance. In either case, the presence peripherally of dendritic cells with the phenotype of mature antigen-presenting cells, but without ability to relocate, could account for the eczema and autoimmune phenomena of WAS. Moreover, failure of dendritic cells to migrate to draining lymphoid tissue would result in diminished physiological immune responses. Defects in cell motility could also explain other aspects of WAS. For example, non-random X-inactivation in heterozygous female WAS carriers could reflect a defect in migration of mutant haemopoietic stem cells into and/or out of site of haemopoiesis during embryonic development. Formation of intimate contacts between haemopoietic stem cells and stromal cells in haemopoietic microenvironments in embryonic and postnatal life might also require WASP-mediated cytokeletal changes, and it is possible that these could be disrupted in WAS patients. Although not abolishing the ability of mutant haemopoietic stem cells to generate a fully populated haemopoietic system in affected males, such defects might render them incapable of competing effectively with wild-type haemopoietic stem cells in heterozygous females. The effects of WASP gene mutations on platelet function remain unclear, and roles in both platelet activation and platelet production have been suggested. Platelet activation by external stimuli such as thrombin results in extensive shape changes and the appearance of lamellar and filopodial protrusions, and WASP may contribute to regulation of these events in normal cells. Recent evidence suggests that the content of filamentous actin is reduced in WAS platelets, and that there are intrinsic defects of activation and structure ( Semple et al, 1997 ). Interestingly, mice deficient in the actin filament severing and capping protein, gelsolin, exhibit abnormal platelet shape chances which result in prolonged bleeding times, and also display delayed neutrophil migration in vivo ( Witke et al, 1995 ). More recent experiments suggest that gelsolin has a role in the initiation of filopodial retraction ( Lu et al, 1997 ), acting downstream of the Rac GTPase ( Arcaro, 1998). Inhibition of WASP expression using antisense oligonucleotides has been shown to affect the structure of the actin cytoskeleton in megakaryocytes ( Miki et al, 1997 ), suggesting that platelet production could be compromised by WASP gene mutations. Platelets are generated from megakaryocytes by fragmentation of the types of beaded filopodial extensions from the cell surface, and it has been suggested that an abnormality in this process could result in production of the abnormal platelet found in WAS patients ( Miki et al, 1997 ). Although considerable advances have been made in the 4 years since the WASP gene was isolated, many questions remain unanswered. A set of key questions surrounds the biochemical consequences of the interactions between WASP and its binding partners, Cdc42, Fyn, Nck, Grb2 and WIP. Are these complexes pre-existing, or assembled in response to the binding of external factors to receptors? Does binding of WASP to these partners alter their activities towards other proteins, and is the activity of WASP itself altered as a result of binding, in the same way that Cdc42-binding appears to affect the actin de-polymerization activity of N-WASP? What is the nature of the interaction between WASP and the actin cytoskeleton, and how is this altered by external signals? A second set of questions concerns the identity of the cell types whose abnormal behaviour makes the key contribution to the symptoms of immune deficiency in WAS. As discussed above, early studies focused on T cells and B cells, without any firm conclusion, whereas our recent work has focused on the possible role of dendritic cells. There is limited scope for using patient material to address these questions, and so considerable investment has to be placed on the hope that targeted disruption of the WASP gene in mice will generate animals with WAS-like symptoms. If it does, then it will be possible to examine the contribution of different cells to the pathogenesis of WAS by generating strains of mice in which the WASP gene is inactivated in specific haemopoietic cell lineages. Although these questions are of importance to students of WAS, they will also engage the attention of those interested in the regulation of cytoskeletal architecture, those examining mechanisms by which immune responses are regulated and those studying other immune disorders. WASP is therefore likely to continue to promote interactions between scientists, as well as between signalling proteins. Finally, what hope does this research hold out for WAS patients? Already, identification of the WASP gene and development of sensitive assays for WASP protein in peripheral blood mononuclear cells has placed on a firmer footing the accurate diagnosis of WAS and XLT. Since allogeneic bone marrow transplantation is already effective in treating WAS patients, WAS is also a candidate for gene therapy based on the delivery of a wild-type WASP gene to autologous haemopoietic stem cells. This would solve the problem of graft-versus-host disease associated with existing allogeneic bone marrow transplantation strategies, which may be particularly severe in these patients. Another hint for those developing novel therapeutic approaches for WAS, and in particular for the immune defects, may lie in the observation that symptoms vary greatly between individuals. If it proves to be the case that cells expressing mutant WASP can be rescued by other factors in patients with mild WAS, it may be possible to use drugs that mimic those factors to ameliorate the symptoms of severe disease. Understanding the complex interactions between WAS cells and their environment represents a challenge for the future, but holds out the hope that, though fallen, WAS cells may be redeemable. The authors are grateful to the Leukaemia Research Fund and the Primary Immune Deficiency Association for supporting their work on Wiskott-Aldrich syndrome. A.J.T. is a Wellcome Clinical Scientist Fellow.