Autoantibodies in immune thrombocytopenia affect the physiological interaction between megakaryocytes and bone marrow extracellular matrix proteins

Abstract

Immune thrombocytopenia (ITP) is characterized by its heterogeneity among patients with regard to both clinical features and response to treatment. Although there is still much to investigate, this heterogeneity could be the reflection of multiple mechanisms contributing to thrombocytopenia, affecting individual patients in variable proportions. These mechanisms include increased platelet clearance by autoantibodies targeting main platelet glycoproteins (GPs) (IIbIIIa, IbIX and IaIIa) and impaired megakaryopoiesis (McMillan et al, 2004). Recently, we reported that ITP plasma samples inhibit proplatelet (PP) formation by normal megakaryocytes (MKs) (Lev et al, 2014), demonstrating that thrombopoiesis, the key step of platelet production, is altered, a finding further confirmed by others (Iraqi et al, 2015). In agreement with these experimental data, studies on platelet kinetics show inadequate platelet production according to the degree of thrombocytopenia (Ballem et al, 1987). Interaction between these GP receptors of the megakaryocytic lineage and their corresponding extracellular matrix ligands is essential for MK development and platelet production in the bone marrow (BM). Proof of the relevance of this interaction include the abnormalities in platelet generation found in several inherited conditions: (i) patients with MYH9-related thrombocytopenia (Pecci et al, 2009) carrying mutations in myosin IIA (MYH9), downstream of collagen type I receptor, GPIaIIa; (ii) macrothrombocytopenia due to mutations that partially activate fibrinogen receptor, GPIIbIIIa (Kunishima et al, 2011); and (iii) mutations in GPIbIXV, the von Willebrand factor (VWF) receptor, leading to Bernard-Soulier syndrome (Strassel et al, 2009). Given that the main ITP autoantibodies target GPIIbIIIa, GPIbIX and GPIaIIa, we hypothesized that their binding could impair GP function, altering MK interaction with extracellular matrix proteins and thus, megakaryocytic behaviour within the BM environment. In order to test this assumption, we studied the effect of ITP plasma on normal MK adhesion, spreading, activation of downstream signalling pathways and thrombopoiesis. After approval was given by the institutional Ethics Committees and written informed consent was obtained, blood samples from 14 ITP patients (Table SI) and controls were obtained and used to prepare recalcified plasma and IgG fractions (Lev et al, 2014). Autoantibodies were detected using the PAKAUTO kit (GTI Diagnostics Inc., Waukesha, WI, USA). Normal mature MKs were obtained after 13-day culture of human cord blood-derived CD34+ cells, as described (Balduini et al, 2008), and incubated with recalcified plasma or purified IgG on type I collagen, fibrinogen and VWF-coated surfaces. Then, adhesion, spreading and activation of downstream signalling pathways were assessed. Details are provided in the Supporting Information. The percentage of MK adhesion to type I collagen in the presence of anti-GPIaIIa ITP samples (n = 2) was below the normal reference range. These values were also lower than those obtained for ITP plasmas bearing anti-GPIIbIIIa (n = 4), or negative for autoantibodies (n = 4), which behaved normally (Fig 1A). Similar results were obtained when MKs were incubated with the corresponding purified IgG fractions, indicating that autoantibodies are responsible for this inhibitory effect (Fig 1B). MK spreading followed the same pattern (Fig 1C), demonstrating that autoantibodies also affect the ability of MKs to actively attach to this substrate (Fig 1D). To test GPIaIIa intracellular signalling after collagen engagement, we evaluated myosin light chain 2 (MLC2) phosphorylation, a downstream mediator of GPIaIIa and a critical regulator of several aspects of MK physiology, including proper myosin contractility during spreading (Malara et al, 2011) and modulation of PP production (Chang et al, 2007). MKs showed decreased phospho-MLC2 levels in the presence of anti-GPIaIIa-bearing plasmas (Fig 1E). According to the classical haematopoiesis model, MKs begin their development in the osteoblast niche, a type I collagen-enriched environment. MK binding to this extracellular protein through GPIaIIa inhibits thrombopoiesis (Sabri et al, 2004), precluding premature platelet release within this BM compartment. In previous studies, we observed the absence of this physiological inhibition in the presence of anti-GPIaIIa ITP autoantibodies (Lev et al, 2014). Our current results indicate that this lack of inhibition relies on the fact that anti-GPIaIIa autoantibodies block MK-type I collagen interaction and interfere with MLC2 phosphorylation, which is a key downstream effector mediating GPIaIIa-induced negative regulation of thrombopoiesis. Driven by a chemoattractant SDF-1 (also termed CXCL12) gradient, MKs migrate to the vascular niche, a microenvironment permissive for thrombopoiesis, where fibrinogen and VWF binding to GPIIbIIIa and GPIbIXV, respectively, enhances platelet production (Balduini et al, 2008). ITP plasma samples bearing autoantibodies against GPIIbIIIa and GPIbIXV, and the purified IgG fractions, significantly reduced MK adhesion to their corresponding ligands (Fig 1F, G, K, L). MK spreading was also compromised in the presence of these samples (Fig 1H, M). Plasma from ITP patients with no detectable autoantibodies allowed normal MK adhesion and spreading on these matrices (Fig 1I, N). Downstream signalling after fibrinogen-GPIIbIIIa interaction, evaluated by phosphorylation of β3 tyrosine residues (Tyr773 and Tyr785), was reduced in MKs incubated with anti-GPIIbIIIa autoantibodies (Fig 1J). Similar results were obtained when phosphorylation of c-Src (also termed CSK), a kinase that plays a major role in the GPIbIX pathway, was investigated after VWF-GPIbIX interaction in the presence of the ITP plasma bearing anti-GPIbIX autoantibodies (Fig 1O). These results suggest that MK interaction with fibrinogen and VWF in the vascular niche could be impaired in ITP patients with anti-GPIIbIIIa and anti-GPIbIX autoantibodies, leading to functional abnormalities in mature MKs and interfering with PP production and platelet release. Indeed, PP formation evaluated in MKs cultured on surfaces coated with fibrinogen and VWF were reduced in the presence of anti-GPIIbIIIa and anti-GPIbIX ITP plasmas, respectively (Fig 2A, D). The slight decrease in PP formation observed in the presence of patient samples with no detectable autoantibodies could be due to other factors, including the presence of autoantibodies different from those tested in this study. Considering the heterogeneous nature of ITP, the study of additional patients harbouring autoantibodies against these three main targets will be useful to determine the relevance of our findings in a larger ITP cohort. Overall, our results indicate that ITP autoantibodies induce abnormal MK interaction with different components of the bone marrow extracellular microenvironment that affect MK physiological functions and thus contribute to defective platelet production in ITP. These novel findings unveil additional mechanisms by which ITP autoantibodies hinder MK biology, highlighting their multiple effects in platelets and their precursors. Authors are grateful to Dr Felisa Molinas for her scientific guidance and constant support, to Marta Pierdominici for referral of patient samples and to the Departamento de Obstetricia, Hospital Materno-Infantil Dr Gianantonio for collection of cord blood. M Grodzielski received a short-term fellowship granted by the European Molecular Biology Organization to conduct part of this research project in the Department of Molecular Medicine, University of Pavia and Biotechnology Research Laboratories, IRCCS San Matteo Foundation, Pavia, Italy. This research was supported by grants from National Scientific and Technical Research Council (CONICET) (PIP 11220120100489) and Cariplo Foundation (2010-0807 and 2013-0717). All authors state that they have no conflict of interest. M. Grodzielski designed and performed experiments, analysed data, wrote the manuscript and prepared figures. C.A. Di Buduo and P.M. Soprano performed experiments. N.P. Goette and P.R. Lev performed experiments and analysed data. P.G. Heller discussed the results and wrote the manuscript. A. Balduini designed experiments, analysed data and supervised the project. R.F. Marta designed and performed experiments, analysed data, wrote the manuscript and supervised the project. All authors edited the manuscript. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.