Gain and Loss of Extracellular Molecules in Sporadic Inclusion Body Myositis and Polymyositis—A Proteomics‐Based Study

Abstract

Sporadic inclusion body myositis (sIBM) contains non‐necrotic myofibers that are surrounded and/or invaded by inflammatory cells. In this study we aimed to identify selective molecules that are present at this site. Myofibers of four biopsies of sIBM that were surrounded and/or invaded by inflammatory cells were microdissected, pooled and profiled by proteomic studies using mass spectrometry. Normal skeletal muscle tissue served as control. Based on the table of proteins that were detected in sIBM only, we selected nine extracellular matrix molecules and validated the results performing immunofluorescence. Seven out of nine proteins that were detected in sIBM by mass spectrometry showed different immunohistochemical results in myositis and normal controls. Of these, the small leucine‐rich repeat proteins proline arginine‐rich end leucine‐rich repeat protein (PRELP) and biglycan were deposited precisely at myofibers surrounded and/or invaded by inflammatory cells both in sIBM and polymyositis. The basement membrane (BM) molecules merosin, perlecan, nidogen‐2 and collagen IV were variably destroyed or increased at these sites. P component, which ensheathed all myofibers in normal controls, was absent from invaded myofibers. Similar to BM remodeling, the specific deposition of PRELP and biglycan may represent a mechanism to defend against immune attack. Loss of P component may affect the anchorage of the myofiber in the endomysium.

INTRODUCTION

Sporadic inclusion body myositis (sIBM) is an idiopathic inflammatory myopathy of the elderly population with a prevalence of as high as 51 patients >50 years per 1 million inhabitants (29). The disease is not amenable to immunomodulatory therapy. Muscle biopsy is required to establish the diagnosis. Histologically, the disease is characterized by immune cells invading non‐necrotic myofibers (3). In addition to the inflammatory process, degenerative changes are hypothesized to play a pathogenetic role, represented by rimmed vacuoles and amyloid deposits (28). In the present study, we aimed to assess the molecular repertoire of the typical feature of myofibers being surrounded and/or invaded by inflammatory cells.

Tandem mass spectrometry (MS) has emerged as a powerful technique to identify components of protein complexes. Because of the high complexity of biological samples, the mass spectrometer has time to select only a small subset of the peptides that are sufficiently abundant for detection of fragmentation (1). The workflow applied to separate peptides and proteins before subjecting them to MS has a great impact on the number of proteins obtained. In human skeletal muscle, a systematic study that was performed across a range of myopathic biopsies identified larger numbers of proteins using two‐dimensional peptide separation when compared with one‐dimensional peptide separation. Moreover, using sodium dodecyl sulfate (SDS) gels enhanced the yield of soluble proteins (31).

Accordingly, results of MS studies that addressed the molecular repertoire of sIBM varied. A previous investigation that complemented MS with microarray, histochemical, immunohistochemical, and immunoblot studies used high performance liquid chromatography to separate peptides. It documented loss of many fast‐twitch specific structural proteins and glycolytic enzymes in sIBM and, to a lesser extent, in polymyositis, despite relative preservation of transcript levels. Also, increased abundance of a nuclear membrane protein, immunoglobulins, and two calpain‐3 substrates were found in sIBM (32).

Different results were obtained by Li and colleagues who performed two‐dimensional polyacrylamide gel electrophoresis to separate peptides before subjecting them to proteomic analysis (24). They detected 16 proteins that were upregulated and 6 proteins that were downregulated in sIBM compared with cases of non‐IBM inflammatory myopathy. IBM‐specific proteins included proteins which have been associated with amyloidosis, detoxification, energy metabolism, and protein folding, respectively. The IBM‐downregulated proteins mainly serve as carriers for muscle contraction and other normal muscle functions. Western blot and immunohistochemistry verified the proteomic findings (24).

To identify specific proteins from myofibers that are surrounded and/or invaded by inflammatory cells, we enhanced the abundance of these molecules by performing microdissection prior to separating the proteins by reverse‐phase high‐performance liquid chromatography (HPLC) followed by MS profiling. We selected extracellular matrix (ECM) proteins for further study, as marked increase of connective tissue is a salient microscopic feature of sIBM (22), and these molecules received less attention in previous studies than cellular proteins 24, 32. Subsequently, we performed immunofluorescence on biopsies of sIBM and polymyositis and normal controls to determine which results were specific for sIBM.

SUBJECTS AND METHODS

Samples

MS‐based proteomic profiling was performed on muscle biopsy samples from eight patients (sIBM n = 4, normal n = 4). Patients with IBM fulfilled criteria for definite IBM (13). Immunofluorescence was performed on 10 biopsies each of sIBM and polymyositis, and six normal controls. Patients with polymyositis fulfilled criteria for definite or probable polymyositis (18). A standard battery of histological and enzyme histochemical preparations was performed on frozen sections of all biopsies that included hematoxylin‐eosin, modified trichrome, myofibrillary adenosine triphosphatase at pH 4.6, myofibrillary adenosine triphosphatase at pH 9.4, cytochrome c oxidase, and succinate dehydrogenase. Immunohistochemistry included the leukocyte common antigen CD45, macrophage marker CD68 (clone PGM1), CD8 to label cytotoxic T cells and major histocompatibility complex class 1. The age at biopsy, gender, muscle biopsied and CK levels are summarized in Table 1. Muscle specimens were mounted on a chuck, then flash frozen at −150°C in isopentane chilled by liquid nitrogen, and stored in liquid nitrogen. The study was approved by the Ethics Committee of the School of Medicine at the University of Tübingen.

Table 1.

Diagnosis, age, gender, site of biopsy and CK value.

	Diagnosis	Gender	Age at onset (years)	Age at biopsy (years)	Site of biopsy	Creatine kinase (fold upper normal limit)
1	sIBM	Male	64	72	M. flexor digitorum profundus	1.5‐fold
2	sIBM	Male	67	69	M. vastus lateralis	2‐fold
3	sIBM	Male	58	60	M. biceps brachii	5.5‐fold
4	sIBM	Male	74	75	M. vastus lateralis	3‐fold
5	sIBM	Female	52	54	M. flexor digitorum profundus	9‐fold
6	sIBM	Male	72	74	M. vastus lateralis	3.5‐fold
7	sIBM	Male	80	81	M. biceps brachii	2‐fold
8	sIBM	Male	80	83	M. vastus lateralis	1.5‐fold
9	sIBM	Male	46	49	M. deltoid	2‐fold
10	sIBM	Female	68	69	M. vastus medialis	4‐fold
11	Polymyositis	Male	70	70	M. biceps brachii	14‐fold
12	Polymyositis	Male	56	56	M. triceps brachii	2‐fold
13	Polymyositis	Female	66	65	M. vastus lateralis	10‐fold
14	Polymyositis	Female	65	67	M. vastus lateralis	2‐fold
15	Polymyositis	Male	42	43	M. deltoideus	28‐fold
16	Polymyositis	Male	49	49	M. vastus lateralis	7‐fold
17	Polymyositis	Female	76	77	M. tibialis anterior	9‐fold
18	Polymyositis	Female	53	59	M. biceps femoris	2‐fold
19	Polymyositis	Female	56	56	M. vastus lateralis	3‐fold
20	Polymyositis	Male	68	76	M. vastus lateralis	3‐fold
21	Normal control	Female		50	M. vastus lateralis	Normal
22	Normal control	Female		70	M. vastus lateralis	Normal
23	Normal control	Male		55	M. vastus lateralis	Normal
24	Normal control	Male		58	M. vastus lateralis	Normal
25	Normal control	Male		55	M. vastus lateralis	Normal
26	Normal control	Male		59	M. vastus lateralis	Normal

Laser‐capture microdissection

Muscle biopsies were cut into 8‐µm sections and placed on a membrane‐coated slide (MembraneSlide 1.0 PEN, Carl Zeiss MicroImagingGmbH, Munich, Germany; no. 415101‐4401‐000) that was pretreated with ultraviolet (UV) light for 30'. Sections were stained with Mayer's hemalaun (Merck, Darmstadt, Germany; no. 1.09249.2500; diluted 1:1 with distilled H₂O) treated with Complete Protease Inhibitor Cocktail Tablets (Roche Diagnostics, Mannheim, Germany). Subsequently sections were dehydrated in ethanol and cleared in xylene. A P.A.L.M. Microbeam laser capture microdissection system (P.A.L.M. Microlaser Technologies GmbH, Bernried, Germany) and AdhesiveCaps 500 clear (Carl Zeiss MicroImaging GmbH; no. 415101–4400‐255) were used for laser capture microdissection and collection of myofibers that were surrounded and/or invaded by inflammatory cells. Two hundred fifty areas of each sIBM biopsy were collected. After manual software‐aided markup of all affected myofibers in a given section, automated laser capture of these regions was performed. When catapulting control biopsy specimens, myofibers and surrounding endomysial space were selected.

Protein extraction, digestion and MS

Proteins from the tissue fragments were extracted by adding NuPAGE LDS sample buffer (Invitrogen, Darmstadt, Germany) and incubated for 1 h at room temperature. After extraction, the samples were centrifuged for 15 minutes at 13 000 rpm and supernatants from each group (sIBM, control) were pooled and sample buffer was added to a final volume of 100 µL. The samples were then subjected to a short gel run (∼1 cm in a 12% SDS gel) as a cleaning step and to ensure that equal amounts of protein were used for the measurements. Therefore each sample was loaded in equal amounts on three lanes (33 µL per lane). In order to estimate protein amounts lysates of HuH7 cells with known protein concentration were loaded in four different amounts (1, 3, 5 and 7 µg). The gel was stained using the Novex® Colloidal Blue Staining Kit (Invitrogen) according to the manufacturer's instructions. The stained gel slices were excised and the containing proteins were digested in gel using trypsin as described elsewhere (5) and prepared for liquid chromatography‐mass spectroscopy (LC‐MS) using C18 StageTips (21).

The resulting peptide mixtures were analysed using a Proxeon Easy‐LC system (Proxeon Biosystems, Dreieich, Germany) coupled to a LTQ‐Orbitrap‐XL (ThermoFisher, Dreieich, Germany) equipped with a nanoelectrospray ion source (Proxeon Biosystems). Chromatographic separation of the peptides was performed using a 15‐cm fused silica emitter of 75‐µm inner diameter in‐house packed with reversed‐phase ReproSil‐Pur C18‐AQ 3 µm resin (Dr Maisch GmbH, Ammerbuch‐Entringen, Germany). The peptide mixture was loaded onto the nano‐HPLC column using the IntelliFlow option with a maximum back pressure of 280 bar and subsequently eluted with a segmented gradient of 5%–80% of solvent B (80% acetonitrile (ACN) in 0.5% acetic acid) with a constant flow of 200 nL/min over 228 minutes. Full scan MS spectra were acquired in a mass range from m/z 300 to 2000 with a resolution of 60 000 in the Orbitrap mass analyzer (after accumulation to a target value of 10⁶ charges in the linear ion trap) using the lock mass option for internal calibration (30). The 10 most intense ions were sequentially isolated for CID fragmentation in the linear ion trap with normalized collision energy of 35% at a target value of 5000. Fragment ions were recorded in the linear ion trap. Up to 500 precursor ion masses selected for MS/MS were dynamically excluded for 90 s.

MS data processing

Acquired MS spectra were processed using the MaxQuant software (version 1.0.14.3) (9). The resulting peak lists were searched against a decoy protein database consisting of the human IPI database (v3.64) and 262 commonly observed contaminants using the Mascot search engine (v2.2, Matrix Science, Boston, MA, USA). Initial mass tolerances for database search were set to 7 ppm for precursor masses measured in the Orbitrap and 0.5 Da for fragment ions recorded in the linear ion trap and a maximum of two missed cleavages was allowed. Carbamidomethylation of cysteins was used as a fixed modification; oxidation of methionine and N‐terminal acetylation was defined as variable modifications. The maximum false discovery rates (FDRs) were set to 1% for both peptide and protein levels.

Immunofluorescence

Cryostat sections of muscle cut at 8 µm were fixed in acetone for 5′. The antibodies used, including sources, species, dilutions and positive controls, are listed in Table 2. The following f(ab')₂ fragment secondary antibodies were used: Cy3‐conjugated and fluorescein isothiocyanate (FITC)‐conjugated donkey anti‐rabbit antibody (Dianova, Hamburg, Germany; 1:400); Cy3‐conjugated donkey anti‐goat antibody (Dianova, 1:200); goat anti‐mouse antibody conjugated with Alexa Fluor 488 (Mobitec, Göttingen, Germany: 1:400); and an FITC‐conjugated goat anti‐rat antibody (Dianova; 1:200). Observations were made with the Olympus BX60 microscope equipped for epifluorescence. The images were scanned using an F‐View digital camera fully integrated into the analySIS image analytical software [Olympus Soft Imaging Solutions (formerly Soft Imaging Systems), Münster, Germany].

Table 2.

Antibodies that were used in this study.

Protein name	Host	Dilution	Source	External normal control
PRELP	Rabbit	1:400	Ref. 14	Cartilage
Biglycan	Rabbit	1:500	Ref. 12	Cartilage
P component	Rabbit	1:3000	Dako	Cerebral amyloid angiopathy
Vitronectin	Mouse (clone 342603)	1:100	R&D Systems	Kidney
Alpha‐dystroglycan	Mouse (clone VIA4‐1)	1:10	Novocastra
Merosin	Mouse (clone Mer3/22B2)	1:100	Novocastra
Collagen IV	Mouse (clone CIV 22)	1:100	Dako
Perlecan	Rat (clone A7L6)	1:500	Millipore
Nidogen‐2	Goat	1:500	R&D systems

RESULTS

Tandem MS yielded a total of 1048 proteins and peptides. Of these, 337 were found exclusively in sIBM, and 37, exclusively in normal controls, with 674 proteins occurring in both categories (Table 3). For further study, we selected the ECM protein category. We performed immunofluorescence subject to the availability of antibodies. The remaining proteins that were found exclusively in sIBM are listed in Table 4.

Table 3.

Subcategories of proteins obtained by tandem mass spectrometry. Abbreviation: sIBM = sporadic inclusion body myositis.

Protein subcategory	All	Overlap sIBM/normal controls	sIBM only	Normal controls only
Cytoskeletal	99	63	33	3
Cytoplasmic	271	173	89	9
Nuclear	184	123	54	7
Mitochondrial	160	109	46	5
Extracellular matrix	66	36	26	4
Membranous	193	118	69	6
Unassigned	75	52	20	3
Total	1048	674	337	37

Table 4.

Proteins that were exclusively found in sporadic inclusion body myositis (sIBM) in addition to the ones listed in Table 2. Abbreviation: SDR = short‐chain dehydrogenases/reductases.

Ig kappa chain C region

Ig kappa chain V‐III region

Light chain Fab

Endonuclease domain‐containing 1 protein

Isoform 1 of Tolloid‐like protein

Fibulin‐2

Complement component C4B

Dermatopontin

Dehydrogenase/reductase SDR family member 7C

Prolactin‐inducible protein

Reticulon‐3

Haptoglobin

Vitamin D‐binding protein

1,4‐beta‐N‐acetylmuramidase C

Zymogen granule protein 16 homolog

Apolipoprotein O

Neutrophil defensin‐1

Biglycan and the extracellular matrix molecule proline arginine‐rich end leucine‐rich repeat protein (PRELP) belong to the small leucine‐rich repeat protein (SLRP) family of connective tissue proteins (35). While PRELP was absent from normal muscle (Figure 1A), biglycan was found at blood vessels and the ECM surrounding them (Figure 1B). However, we failed to detect this molecule in the endomysial space of normal controls, as reported previously (37). Deposits of PRELP were found exclusively at myofibers that were surrounded and/or invaded by inflammatory cells (Figure 2A), in the same way biglycan was deposited at these sites (Figure 2C). These findings were identical for sIBM and polymyositis.

Figure 1.

Normal controls. A. Proline arginine‐rich end leucine‐rich repeat protein (PRELP) was absent from normal muscle. B. Biglycan was present at perimysial blood vessels and the connective tissue surrounding them. C. P component shows linear and dotty staining in‐between myofibers. Bar, 100 µm.

Figure 2.

Proline arginine‐rich end leucine‐rich repeat protein (PRELP) (A) and biglycan (C) were present at myofibers that were invaded and/or surrounded by cells. Asterisks in (C) and (D) denote perimysial connective tissue. (A,B) Sporadic inclusion body myositis (sIBM). (C,D) Polymyositis. Bar, 100 µm. Abbreviation: DAPI = 4′,6‐diamidino‐2‐phenylindole.

The presence of PRELP and biglycan coincided with several alterations of the BM ensheathing the invaded myofiber (3, 4, 5, 6). Redundant BM loops were found (Figure 3) along with destroyed BM (4, 5) and thickened BM material (Figure 6). The results for nidogen‐2 and collagen IV were identical to those found with merosin and perlecan (not shown).

Figure 3.

Deposits of proline arginine‐rich end leucine‐rich repeat protein (PRELP) are associated with remodeled basement membrane (BM). Note redundant BM loop in inset of (B). Sporadic IBM. Bars, 100 µm (A) and 10 µm (inset of B).

Figure 4.

Deposits of proline arginine‐rich end leucine‐rich repeat protein (PRELP) (A,C) are associated with cellular infiltrate (B,C) and with a small basement membrane (BM) split (D, arrow in inset). Sporadic inclusion body myositis (sIBM). Bars, 10 µm. Abbreviation: DAPI = 4′,6‐diamidino‐2‐phenylindole.

Figure 5.

Deposits of biglycan (A,C) are associated with cellular infiltrate (B,C) and loss of basement membrane (BM) (D, arrow). Polymyositis. Bar, 10 µm. Abbreviation: DAPI = 4′,6‐diamidino‐2‐phenylindole.

Figure 6.

Deposits of proline arginine‐rich end leucine‐rich repeat protein (PRELP) (A,C) are associated with cellular infiltrate (B,C) and with thickened basement membrane (BM) material (D, arrows) and redundant BM material (D, arrowhead). Polymyositis. Bar, 10 µm. Abbreviation: DAPI = 4′,6‐diamidino‐2‐phenylindole.

P component, which has not been studied in normal muscle so far, showed linear and dotted deposits that were localized in‐between myofibers (Figure 1C). Originally identified as a normal plasma protein, P component was subsequently found to be a universal constituent of amyloid deposits, and, finally, a normal glycoprotein of basement membrane (BM) and of elastic fiber microfibrils 6, 11, 23. Myofibers being surrounded and/or invaded by inflammatory cells lacked P component (Figure 7). For the remaining myofibers, P component expression was inconsistent. Absence of this protein (Figure 7A) occurred alongside increased deposits (Figure 7B) within the same biopsy.

Figure 7.

P component staining of cases of sporadic inclusion body myositis (sIBM) (A,C,E) and polymyositis (B,D,F). (A,C,E) P component was entirely absent, except for a few coarse fibrils. (B,D,F) The protein was variably normal or increased, however, it was absent from the site of cell invasion (arrow). Bar, 10 µm. Abbreviation: DAPI = 4′,6‐diamidino‐2‐phenylindole.

There was no differential anti‐alpha‐dystroglycan labeling in normal controls and in myositis (not shown). Vitronectin was absent from normal controls and myositis (not shown).

DISCUSSION

The main results of this study are the documentation for the first time of the deposition of SLRP PRELP and biglycan at myofibers surrounded and/or invaded by inflammatory cells. In addition, we observed the remodeling of the myofiber's BM; and the loss of P component at these sites.

Both PRELP and biglycan belong to the SLRP class of proteins (35). Neither protein is present at the normal sarcolemma (Figure 1A,B). All SLRPs are thought to be involved in protein–protein interaction. They bind to fibrillar collagen via their leucine rich repeat region. Moreover, they interact with cell surface‐bound heparan sulphate proteoglycans (HSPGs), and with BM‐bound HSPGs, including agrin and perlecan 17, 25, 27, 33, 35.

PRELP stands out among the SLRPs on account of its N‐terminal that contains clustered arginine residues that bind HSPG and chondroitin sulphate proteoglycan 4, 34. PRELP was first identified in cartilage of various species, including humans (17). Normal mouse and human skeletal muscle contained very low levels of PRELP mRNA 14, 15 or were negative (36).

A recent proteomics study detected PRELP in 2/5 cases of polymyositis and 6/8 cases of sIBM; the histological and immunohistochemical patterns of these inflammatory myopathies were not reported (32). Furthermore, that study found the protein to be present in only one of three normal biopsies (32). A previous proteomics study also detected PRELP in normal muscle (19). This is at variance with our results that were negative for PRELP protein in normal controls. It corroborates the need to perform complementary investigations to validate the results obtained from proteomics studies.

Similar to our experience, Parker and colleagues identified biglycan in most cases of idiopathic inflammatory myopathies but not in normal controls (32). Biglycan possesses N‐linked oligosaccharide chains within its central leucine‐rich repeats [reviewed in (33)]. Furthermore, biglycan is a strong trigger of proinflammatory signalling within the innate immune system [reviewed in (20)].

While BM proteins appeared to be upregulated in sIBM according to the tandem MS results (Table 3), by immunofluorescence they were variably destroyed or increased at sites of myofibers surrounded and/or invaded by inflammatory cells (3, 4, 5, 6). This finding confirms a recent study that demonstrated that new BM material containing the developmentally regulated α4 laminin isoform is synthesized at these sites, and that both BM destruction and new synthesis occur along the myofiber segment surrounded and/or invaded by inflammatory cells (10). Additionally, the present study shows that ECM proteins are deposited at the sites of BM remodeling. Along with the newly synthesized BM material, ECM deposition may serve as a barrier to protect myofibers against immune attack. The cellular source(s) of the ECM proteins remain(s) to be established.

A normal plasma protein and universal constituent of amyloid deposits, P component is also a normal glycoprotein of elastic fiber microfibrils and of BM 6, 11. P component has not been previously demonstrated in skeletal muscle. However, it is deposited in the ECM of several normal tissues. Immunohistochemical labeling of P component revealed two staining patterns: elastic fiber microfibrils surrounding blood vessels, and linear staining of eccrine sweat glands in the skin and in renal glomeruli 2, 6, 7, 11.

Normal skeletal muscle showed dotted and linear sarcolemmal deposits of P component (Figure 7). Similarly, fibrillin, which, like P component, is a constituent of elastic fiber microfibrils, was demonstrated by immunohistochemistry to ensheathe all extrafusal muscle fibers of bovine extraocular muscles (26). As fibrillin and elastin have been implicated in lateral force transmission of myofibers (26), P component may represent another participant, although the precise function of P component in normal muscle remains to be established. Of note, myofibers surrounded and/or invaded by inflammatory cells lacked P component (Figure 7). In the remaining myofibers, it was inconsistently increased or absent (Figure 7). Absence of P component in sIBM biopsies has previously been reported (28).

Taken together, there is both gain and loss of extracellular proteins at myofibers surrounded and/or invaded by inflammatory cells in sIBM and polymyositis. While P component is lost, PRELP and biglycan are added to the tissue. BM material is variously destroyed or synthesized. The binding partners and function of the SLRPs remain to be established.

A recent study gave a detailed description of a clinicopathological entity termed “PM/IBM” that is characterized by the clinical course of sIBM and the histopathological pattern of polymyositis (8). Hence, the two disease entities overlap to a greater extent than was previously recognized. Along these lines, the results of our study may suggest that overlap also occurs at the molecular level: a shared molecule is disease‐causing in sIBM but not in polymyositis because, for example, this molecule is inactivated in polymyositis by a mechanism that remains to be established. It is hoped that future animal models of sIBM are available in order to be able to conduct experimental studies to determine the role of these molecules in the pathogenetic mechanism.