Scientific Commentaries

Inclusion body myositis (IBM) is the most prevalent acquired myopathy in adults beyond the age of 50. The disease is relentlessly progressive and largely resistant to immunosuppressive therapy (reviewed by Hohlfeld, 2011). In IBM, and less frequently also in polymyositis, non-necrotic muscle fibres are focally surrounded and invaded by CD8+ cytotoxic T cells (CTL) (Fig. 1) (Engel and Arahata, 1986). The CTLs show immunohistological features of activation and form close contacts with muscle fibres. The muscle fibres that are attacked express major histocompatibility complex (MHC) class I molecules on their surface, and are therefore capable of presenting MHC class I-bound peptides to CD8+ T cells (Fig. 1). It is thus assumed that the muscle-invading CTLs recognize an unknown antigen (or antigens) on muscle fibres. This is further supported by the observation that CTLs tend to orient their cytotoxic granules towards the contacting muscle fibre, providing tell-tale evidence that an immunological synapse has formed between the CTL and its target cell (Goebels et al., 1996). Furthermore, the muscle-infiltrating CTLs are clonally expanded, and identical clones of CTLs may be detected in muscle and blood for prolonged periods of time, providing evidence for chronic antigen-driven proliferation of the CTLs (Hohlfeld, 2011). In this issue of Brain, Greenberg and co-workers report that 50% of patients with IBM harbour clonally expanded populations of CTLs that meet the haematological criteria for ‘T cell large granular lymphocytic leukaemia’ (T-LGL leukaemia) (Greenberg et al., 2016). These novel, somewhat provocative findings could have implications not only for understanding the pathogenesis, but also for improving the diagnosis and treatment of IBM. The term ‘large granular lymphocyte’ (LGL) refers to a population of white blood cells containing azurophilic cytotoxic granules. LGL comprise both CTL and natural killer (NK) cells. Normally, the number of LGLs in peripheral blood is below 0.25 10/l (Steinway et al. 2014). In T-LGL leukaemia, CD8+ CTLs are elevated and clonally expanded. Despite the term T-LGL ‘leukaemia’, the disorder really belongs to a grey zone between chronic lymphocytic proliferation and malignancy. Accordingly, the spectrum of T-LGL leukaemia ranges from relatively benign ‘clonal T-cell expansions of unknown significance’ to a more aggressive haematological disorder accompanied by anaemia, neutropenia or thrombocytopenia. Furthermore, T-LGL leukaemia may be associated with autoimmune diseases, such as rheumatoid arthritis, in particular presenting with extra-articular clinical signs characteristic of Felty’s syndrome, e.g. splenomegaly and neutropenia (Steinway et al., 2014). Greenberg et al. (2016) investigated 38 IBM cases for features of T-LGL leukaemia. In all, 22/38 (58%) of patients met the criteria of an abnormal expansion of LGL in association with an autoimmune disorder (assuming that IBM is indeed an autoimmune disorder), and 13/38 (34%) met the more stringent criteria of an absolute LGL count of4500/ml, presence of clonal T cell receptor rearrangements, and association with an autoimmune disorder (Greenberg et al., 2016). The expanded LGLs had phenotypic features of T-LGL leukaemia cells, as demonstrated by flow cytometry. Moreover, the authors looked for haematological changes that can be assessed with routinely available laboratory tests. They found that elevated absolute counts of CD8+ T cells and a reduced CD4/CD8 ratio closely correlated with the presence of T-LGL clonal expansions, and interestingly, also with the intensity of the histopathological changes in patients’ muscle. Although muscle biopsy remains the gold standard for the diagnosis of IBM, haematological markers such as those described here might assist in the diagnosis and be useful for monitoring the response to immunosuppressive treatment (see below). We do not know which of the two disorders, T-LGL leukaemia or IBM, developed first in these patients. The authors suggest a scenario whereby chronic antigenic stimulation and BRAIN 2016: 139; 1312–1324 | 1312