Anti-Ku + myositis: an acquired inflammatory protein-aggregate myopathy

Abstract

Myositis with anti-Ku-autoantibodies is a rare inflammatory myopathy associated with various connective tissue diseases. Histopathological studies have identified inflammatory and necrotizing aspects, but a precise morphological analysis and pathomechanistic disease model are lacking. We therefore aimed to carry out an in-depth morpho-molecular analysis to uncover possible pathomechanisms. Muscle biopsy specimens from 26 patients with anti-Ku-antibodies and unequivocal myositis were analyzed by immunohistochemistry, immunofluorescence, transcriptomics, and proteomics and compared to biopsy specimens of non-disease controls, immune-mediated necrotizing myopathy (IMNM), and inclusion body myositis (IBM). Clinical findings and laboratory parameters were evaluated retrospectively and correlated with morphological and molecular features. Patients were mainly female (92%) with a median age of 56.5 years. Isolated myositis and overlap with systemic sclerosis were reported in 31%, respectively. Isolated myositis presented with higher creatine kinase levels and cardiac involvement (83%), whereas systemic sclerosis-overlap patients often had interstitial lung disease (57%). Histopathology showed a wide spectrum from mild to pronounced myositis with diffuse sarcolemmal MHC-class I (100%) and -II (69%) immunoreactivity, myofiber necrosis (88%), endomysial inflammation (85%), thickened capillaries (84%), and vacuoles (60%). Conspicuous sarcoplasmic protein aggregates were p62, BAG3, myotilin, or immunoproteasomal beta5i-positive. Proteomic and transcriptomic analysis identified prominent up-regulation of autophagy, proteasome, and hnRNP-related cell stress. To conclude, Ku + myositis is morphologically characterized by myofiber necrosis, MHC-class I and II positivity, variable endomysial inflammation, and distinct protein aggregation varying from IBM and IMNM, and it can be placed in the spectrum of scleromyositis and overlap myositis. It features characteristic sarcoplasmic protein aggregation on an acquired basis being functionally associated with altered chaperone, proteasome, and autophagy function indicating that Ku + myositis exhibit aspects of an acquired inflammatory protein-aggregate myopathy.

Keywords: Autoantibodies; Autophagy; Connective tissue diseases; Myositis; Overlap syndrome; Systemic sclerosis.

Fig. 1

Histopathological and immunohistochemical stains of a representative Ku + myositis specimen. Histopathology of Ku + myositis showed a spectrum of mild to very severe myositis with all biopsies showing myopathic features including fiber size variation, rounded fibers and internalized myonuclei, furthermore, commonly necrotic myofibers, as well as myophagocytosis and presence of vacuoles (black arrow in b). Many biopsies showed thickened or enlarged capillaries in the endomysium (black circles, see also Fig. 4) (a, b). The inflammatory infiltrates were located in the endomysium, again diffuse and with a focal enhancement around necrotic myofibers, mainly consisting of CD68 positive cells (c) (for other inflammatory cells, please see Supplementary Fig. 5). All muscle biopsies stained with MHC-class I on the sarcolemma of the myofibers (d), and the majority of samples additionally stained with MHC-class II on the sarcolemma with patchy distribution (e). Complement deposition with C5b-9 showed decoration on the sarcolemma in the majority of biopsies (black arrows) (f). No COX negative-SDH positive fibers above age-related numbers were detected (g). Some biopsies showed subsarcolemmal dark aggregates representing areas of pathologic mitochondrial accumulation (h). Staining with non-specific esterase showed frequent myophagocytosis in most biopsies (i)

Fig. 2

Autophagy pathways are activated in Ku + myositis on both gene and protein levels. Immunohistological staining of autophagy markers p62 revealed large aggregates in the sarcoplasm of non-necrotic myofibers, with aggregates located in the center of the sarcoplasm or appearing subsarcolemmally as large caps (a1–4). The larger aggregates were also positive for BAG3, αB-Crystallin and HSP70 (b–e). With an unbiased proteomic approach we identified multiple proteins being involved in regulation of autophagy (see Supplementary Table 1). As demonstrated by the STRING graph, these proteins form a dense interacting network (f). Of note, when comparing Ku + myositis to NDC we identified upregulated proteins, which, when comparing mild versus severe Ku + myositis patients, showed even higher extent in the development of the disease, as shown for the top 10 regulated autophagy and cell stress proteins (g) (for comparison of these proteins in IBM and IMNM, see Supplementary Table 4). Of note, three of the highest expressed proteins are part of the hnRNP machinery acting as key proteins in the cellular nucleic acid metabolism. Interestingly, in the analysis of autophagy genes, we could not find differences of p62/SQSTM1 expression between Ku + myositis, NDC, and IMNM, but a significantly higher LC3 expression in Ku + samples compared to IMNM samples (h). CRYAB was significantly more expressed in Ku + myositis and IMNM compared than in NDC (i), as was HSPB8 for Ku + patients, while other chaperons (BAG3, HSPA8) were not changed compared to controls. However, a difference was seen between Ku + patients and IMNM, overall highlighting that not the quantity of gene expression but their functional implications (which seem variable) is more relevant in Ku + myositis compared to the other entities studied (j)

Fig. 3

Aspects of pathological protein aggregation in Ku + myositis. Large sarcoplasmic and subsarcolemmal aggregates stain positive with the immunoproteasome molecule LMP7/β5i (black arrow, a1), the structural protein myotilin (as well as the intermediate filament desmin; not shown) delineates aggregates of dense staining (black arrows, a2), while most of the non-affected myofibers show a fine physiological positivity, and the autophagy molecule p62 (as well as LC3; not shown) highlights sarcoplasmic aggregates with different quality of the staining pattern as well (black arrows, a3). Myotilin-positive aggregates (AF488; green) co-localize with the immunoproteasome marker LMP7/β5i (Cy3; red) (b). The autophagy marker p62-positive aggregates (AF488; green) co-localizes with the immunoproteasome marker LMP7/β5i (Cy3; red) (c). And additional stainings show co-localization of myotilin stained fibers (Cy3; red) with the chaperone αb-Crystallin (AF488; green) (d), as well as small aggregates that are p62 + (Cy3; red) and αB-Crystallin + (AF488; green) (e). Myotilin-positive aggregates (AF488; green) also co-localize with the autophagy marker p62 (Cy3; red) (f). On the gene expression level, proteosomal markers (LMP2, PSME2) are significantly increased in Ku + patients versus non-diseased controls, while there is no significant difference compared to IMNM patients (g). Yellow arrows = double staining of both markers; green arrows = single staining of respective marker

Fig. 4

Ultrastructural analysis reveals disturbed contractile elements and changes in capillary basement membranes. Ultrastructural analysis highlights focally aggregated and disrupted contractile elements which is very pronounced in (a1) and less pronounced probably in statu nascendi in (a2). Those alterations are very similar to the ones detectable in genetic protein-aggregate myopathies. The capillaries frequently showed multilayered basement membranes exceeding 4 layers (black arrows, b1), and basement membranes may also appear more homogeneously thickened (black arrows, b2); note the prominently activated rough endoplasmic reticulum (white star, b2). Pericytes (white small stars) are intermingling with thickened basement membranes (black arrows, b3).