Proteomic profiling of sporadic late-onset nemaline myopathy.

OBJECTIVE: To define the proteomic profile of sporadic late-onset nemaline myopathy (SLONM) and explore its pathogenesis.
METHODS: We performed mass spectrometry on laser-dissected frozen muscle samples from five patients with SLONM, three of whom with an associated monoclonal protein (MP), and four controls, to determine the proteomic profile of SLONM. Furthermore, we assessed the role of the MP by evaluating the expression of the immunoglobulin light chain variable regions (IGVL).
RESULTS: There were 294 differentially expressed proteins: 272 upregulated and 22 downregulated. Among the top 100 upregulated proteins, the most common categories were: nuclear or nucleic acid metabolism (24%), extracellular matrix and basal lamina (17%), immune response (13%), and actin dynamics (8%). Downregulated proteins consisted mostly of contractile proteins. Among upregulated proteins, there were 65 with a role related to the immune system, including eight proteins involved in major histocompatibility complex 1 (MHC1) and antigen processing, 15 in MHCII complex and phagocytosis, and 23 in B and/or T-cell function. Among nine upregulated immunoglobulin proteins, there were two IGVL genes. However, these were also detected in SLONM cases without an MP, with no evidence of clonally dominant immunoglobulin deposition. In muscle sections from SLONM patients, nemaline rods tended to accumulate in atrophic fibers with marked rarefaction of the myofibrils. Increased MHC1 reactivity was present in fibers containing nemaline rods as well as adjacent nonatrophic fibers.
CONCLUSION: Our findings suggest that aberrant immune activation is present in SLONM, but do not support a direct causal relationship between the MP and SLONM.

Introduction

Sporadic late‐onset nemaline myopathy (SLONM) is a rare acquired myopathy that affects adults, usually after the age of 40. , Muscle weakness can be rapidly progressive, and lead to respiratory failure and death if left untreated. , , About 61% of patients have an associated monoclonal protein (MP). A causal relationship between the plasma cell disorder and SLONM has not been well established, especially since MPs are also identified in 3–5% of the general population of the same age. However, the higher than expected frequency of an MP in SLONM raises suspicion for association between the two entities. , Except in rare cases, the MP in SLONM is classified as monoclonal gammopathy of uncertain significance (MGUS), and patients usually die from neuromuscular respiratory failure rather than from progression to hematologic malignancies. Fortunately, when accurately diagnosed, the majority of patients respond to treatment with intravenous immunoglobulin. , In patients with an associated MP, autologous stem cell transplant (ASCT) or plasma cell‐directed therapy may also be considered. , , , On the other hand, treatment response to corticosteroids or oral immunosuppressant’s remains poor. ,

Histologically, SLONM is characterized by accumulation of nemaline rods in muscle fibers with minimal or no inflammation. , , Aside from its response to IVIG and ASCT, it remains uncertain whether the disease is immune mediated, and little is known regarding the underlying disease mechanisms. In this study, we performed laser dissection liquid chromatography and electrospray tandem mass spectrometry (LMD‐MS) to define the proteomic composition of SLONM and shed light into its pathogenesis.

Methods

Selection of cases and controls

The study was approved by Mayo Clinic Institutional Review Board. The study was considered minimal risk; therefore, the requirement for informed consent was waived. However, records of any patient who had not provided authorization for their medical records to be used for research, as per Minnesota statute 144.335, were not reviewed. Patients with SLONM were identified from our recently published cohort. We selected five patients with SLONM who were not on immunosuppression for their myopathy at time of the biopsy, and who had remaining frozen muscle tissue. We identified four controls from our muscle laboratory database. To be included as a control, the patient had to have no clinical evidence of a myopathy and normal creatine kinase level.

Processing of muscle samples

Fibers with nemaline rods from SLONM patients and normal muscle fibers from controls were collected in separate tubes by laser pressure catapulting (LPC) using a Zeiss Palm MicroBeam scope controlled by RoboPalm software. In order to do that, 10 μm sections of frozen muscle on polyethylene naphthalate membrane slides (ThermoFisher Scientific, Waltham, MA) were stained with trichrome to visualize fibers with rod aggregation. Fibers with rods were outlined, laser microdissected, and catapulted into the cap of 0.5 mL tubes containing 0.05% Protease Max/0.005% Zwittergent 3–16/100 mmol/L Tris, pH 8.5. Nuclei were avoided where possible. Similar process was applied to sections from controls containing structurally normal fibers. An area of 0.3 mm2 was collected for each sample. Samples were trypsin‐digested in the same collection tube after protein reduction and alkylation. After acidification, each peptide digest was transferred to a mass spectrometry (MS) sample vial, dried down, and brought up in the same volume of 0.1% trifluoroacetic acid for injection on the MS.

Label‐free liquid chromatography MS

Digested samples were loaded onto an nLC–MS/MS system consisting of a Q‐Exactive Orbitrap high resolution mass spectrometer and a Dionex Ultimate 3000 RSLC nano liquid chromatograph (nLC) (ThermoFisher Scientific, Waltham, MA). Peptides were pre‐concentrated onto a 0.33‐μL EXP2 trap (2.7 μm Halo Peptide ES‐C18, Optimize Technologies, Oregon City, OR) in the Dionex column oven and washed using a 10‐μL/min flow of aqueous 0.05% TFA and 0.2% formic acid before switching the trap in‐line with the nLC column. Peptides were separated on a 35 cm long × 75 μm id PicoFrit column, self‐packed with ThermoFisher Acclaim RSLC 2.2 μm C18 stationary phase, using a 120 min reversed phase LC method with a gradient from 2% to 40% B in 100 min at 250 nL/min flow rate, followed by increasing to 85% B for 6 min before re‐equilibrating the column and trap at 2% B. Mobile phase A was 2% acetonitrile in water with 0.2% formic acid; mobile phase B was 80% acetonitrile, 10% isopropanol, and 10% water with 0.2% overall concentration of formic acid. The trap and column were heated to 40°C.

MS and MS/MS spectra were acquired in a data‐dependent acquisition mode with MS precursors acquired from m/z 340 to 1800 (60,000 resolving power (RP) FWHM at m/z 200, AGC = 3e6, max fill time = 150 msec). MS/MS spectra were collected for the top 20 precursors with charge 2–5, at a normalized collision energy = 26 (15,000 RP, AGC = 1e5, max fill time = 80 msec, isolation window = 3, window offset = 0.5, fixed first m/z of 140, precursor dynamic exclusion = 30 sec).

Bioinformatics and statistical analysis

A previously published bioinformatics pipeline was utilized to process the raw LC–MS/MS data and perform peptide intensity‐based label‐free quantification of proteins present in the samples. Raw data files were loaded into MaxQuant software (version 1.6.0.16) configured to search the MS/MS spectra against a database containing Uniprot human protein sequences (downloaded on 20 November 2020) and common contaminants (like sheep keratin, cotton proteins, etc.). Reversed protein sequences were appended to the database to estimate peptide and protein false discovery rates (FDRs). MaxQuant was instructed to use trypsin as digestion enzyme and the following posttranslational modifications when matching the MS/MS against the sequence database: carbamidomethyl cysteine (+57.021 Da), oxidation of methionine (+15.995), and deamidation of asparagine (+0.985). The software identified the peptides and proteins present in the samples at an FDR ≤1%, grouped protein identifications into groups and reported protein group intensities.

A previously published, in‐house developed R script was utilized to process the reported protein group intensities and find differentially expressed proteins between any two experimental groups (Data S1). For this, protein group intensities of each sample were log2 transformed and normalized using trimmed mean of M‐values method. For each protein group, the normalized intensities observed in any two experimental groups of samples were modeled using a Gaussian‐linked generalized linear model. An ANOVA test was utilized to detect the differentially expressed protein groups between pairs of experimental groups. Differential expression p‐values were FDR corrected using Benjamini–Hochberg–Yekutieli procedure. Protein groups with an FDR < 0.05 and an absolute log2 fold change of at least 0.5 were considered as significantly differentially expressed and saved for further analysis.

Investigation of immune system‐related proteins

We investigated the role of each of the top 100 upregulated proteins and the 22 downregulated proteins via literature search. To further investigate the potential immune basis of SLONM, we also searched the immune role of all the differentially expressed proteins of interest. These proteins were identified from immune‐related pathways detected by Broad's Gene Set Enrichment Software and Ingenuity Pathway analysis software.

Assessment of immunoglobulin clonality

In order to determine the role of the MP in SLONM and the clonality of the detected immunoglobulins in muscle, we searched the raw data files against a light chain variable region (IGVL) sequence database as previously described. IGVLs are unique for each plasma cell clone and can therefore be used as a surrogate for clonality of the deposited immunoglobulins. In brief, a database of immunoglobulin light chain variable genes was assembled and the MS/MS spectra in each sample were matched against the database using MyriMatch database search engine. Reversed sequence entries were appended to the database to estimate FDRs. MyriMatch was configured to use the same peptide/fragment mass tolerances, digestion enzyme, and posttranslational modifications as the MaxQuant‐based procedure described above. IDPicker software was utilized to filter the peptide and protein identifications at 1% FDR. Identified peptides that matched to light/heavy chain constant/variable regions were grouped by their gene family and their MS/MS counts were added together. Spectral counts of each Ig gene were normalized, as previously described. A Wilcox rank test was used to evaluate for differential expression of Ig genes between the two experimental groups, and genes with a p‐value <0.05 were considered as statistically significant. We also computed the log2 ratio of average counts observed for each Ig gene in samples from SLONM patients versus control subjects.

Immunohistochemical and immunofluorescence studies

Major histocompatibility complex class I (MHC1) was immunolocalized using monoclonal antibodies on frozen sections from the five included SLONM patients and controls. We also included five patients from our database with congenital nemaline myopathy. Consecutive trichrome‐stained sections were performed to identify the fibers with nemaline rods. Capillary endothelia were localized with Ulex europaeus agglutinin I (UEA), and the complement C5b9 membrane attack complex (MAC) with monoclonal antibodies in frozen sections from patients and controls and visualized under immunofluorescence microscopy.

Data availability

Data not provided in the article and additional information on methods and materials may be shared upon request.

RESULTS

Baseline characteristics of patients and controls

The demographics and disease characteristics of patients with SLONM are shown in Table 1. Controls had a median age of 52 (range 40–60) and included two males and two females. The muscle biopsies were obtained to rule out a myopathy for the following symptoms: myalgia (two patients), dyspnea on exertion, and fatigue.

Table 1

Baseline characteristics of patients with sporadic late‐onset nemaline myopathy.

	All patients	P1	P2	P3	P4	P5
Age at biopsy	59 (48–68)	59	46	58	63	68
Sex	4M/1F	M	M	F	M	M
Disease duration (symptom onset to biopsy date)	23 (4–138)	39	6	4	138	23
Associated monoclonal protein	3/5 (60%)	N	Y	N	Y	Y

F, female; M, male.

Proteomics

Of the 1458 total detected proteins, 294 proteins were differentially expressed in SLONM, of which 272 upregulated and 22 downregulated. Among the top 100 upregulated proteins, the most common categories were as follows: nuclear or nucleic acid metabolism (24%), extracellular matrix and basal lamina (17%), and immune response (13%) (Table 2). Downregulated proteins consisted mostly of contractile proteins. A complete list of differentially expressed proteins is shown in Table S1.

Table 2

Top 100 upregulated and all downregulated proteins.

Gene	Protein name	Log2 fold change	Gaussian FDR
Nuclear or nucleic acid metabolism
HNRNPU	Heterogeneous nuclear ribonucleoprotein U	37.6845	0
HNRNPC	Heterogeneous nuclear ribonucleoprotein C1/2	37.4689	0
XRCC6	X‐ray repair cross‐complementing protein 6	37.3472	0
LMO7	LIM domain only protein 7	37.2114	2.10E‐117
HNRNPA3	Heterogeneous nuclear ribonucleoprotein A3	36.7089	0
H1F0	Histone H1.0	36.6937	1.92E‐61
DDX17	Probable ATP‐dependent RNA helicase DDX17	36.4066	0
MATR3	Matrin‐3	35.9364	4.32E‐90
XRCC5	X‐ray repair cross‐complementing protein 5	35.8049	0
CRIP2	Cysteine‐rich protein 2	35.4522	0
SFPQ	Short of Splicing factor, proline‐and glutamine‐rich	34.8728	1.31E‐228
IMPDH2	Inosine‐5′‐monophosphate dehydrogenase 2	34.8171	0
HNRNPR	Heterogeneous nuclear ribonucleoprotein R	34.6812	1.06E‐207
RBMX	RNA‐binding motif protein, X chromosome	34.5301	4.93E‐250
EIF3L	Eukaryotic translation initiation factor 3 subunit L	34.2818	4.61E‐157
EIF3C	Eukaryotic translation initiation factor 3 subunit C	34.0884	5.02E‐220
TNPO1	Transportin‐1	34.0577	0
TARDBP	TAR DNA‐binding protein 43	33.9986	0
HDGF	Hepatoma‐derived growth factor	32.9447	3.15E‐198
NCL	Nucleolin	32.5568	1.70E‐05
HNRNPH1	Heterogeneous nuclear ribonucleoprotein H	30.2065	0.0009048
HNRNPM	Heterogeneous nuclear ribonucleoprotein M	29.1451	0.003114204
HNRNPAB	Heterogeneous nuclear ribonucleoprotein A/B	27.9319	0.003636039
SUN2	SUN domain‐containing protein 2	27.8461	0.003623759
TMEM43	Transmembrane protein 43	27.4371	0.000481811
SMYD1	Histone‐lysine N‐methyltransferase SMYD1	−3.5626	0.00296268
Contractile proteins
MYOM3	Myomesin‐3	37.1096	4.70E‐303
MYO18B	Unconventional myosin‐XVIIIb	36.5623	0
MYL6	Myosin light polypeptide 6	29.8760	0.000987854
TNNC1	Troponin C, slow skeletal and cardiac muscles	−2.5635	0.000239163
TPM3	Tropomyosin alpha‐3 chain	−2.7234	0.000811562
MYBPC1	Myosin‐binding protein C, slow‐type	−3.4846	7.70E‐05
TPM2	Tropomyosin beta chain	−3.6542	0.006675157
TTN	Titin	−3.7723	1.25E‐05
MYOM1	Myomesin‐1	−3.9935	0.004184786
MYL1	Myosin light chain 1/3, skeletal muscle	−4.0273	0.007628146
MYL2	Myosin regulatory light chain 2	−4.0926	0.001551271
MYL5	Myosin light chain 5	−5.4719	0.003475986
MYL3	Myosin light chain 3	−5.5703	0.009064135
MYH1	Myosin‐1	−5.7091	0.001422474
MYBPC2	Myosin‐binding protein C, fast‐type	−5.8789	0.003197956
MYH13	Myosin‐13	−5.9548	1.20E‐07
MYOM2	Myomesin‐2	−6.7592	8.61E‐13
MYLPF	Myosin regulatory light chain 2, skeletal muscle isoform	−7.1407	0.001092562
MYBPC1	Myosin‐binding protein C, slow‐type	−29.8774	0.000435494
Extracellular matrix/basal lamina
POSTN	Periostin	38.1531	5.29E‐77
NID2	Nidogen	37.8429	0
PRELP	Prolargin	36.2660	4.02E‐66
MFAP5	Microfibrillar‐associated protein 5	35.8636	2.68E‐109
COL15A1	Collagen alpha‐1(XV) chain	35.1866	8.55E‐227
CILP	Cartilage intermediate layer protein 1	35.0617	5.31E‐247
TGFBI	Transforming growth factor‐beta‐induced protein Ig‐H3	34.8320	3.38E‐136
FBLN2	Fibulin‐2	34.2103	5.77E‐121
EPDR1	Mammalian ependymin‐related protein 1	34.0783	0
BGN	Biglycan	34.0641	0
FBLN5	Fibulin‐5	29.4811	0.000667082
DPT	Dermatopontin	28.9198	0.003623759
COL2A1	Collagen alpha‐1(II) chain	28.4937	0.003636039
ELN	Elastin	27.8247	0.003636039
MFAP4	Microfibril‐associated glycoprotein 4	27.6941	0.003852888
TNS1	Tensin‐1	27.4890	0.003623759
OGN	Mimecan	27.3773	0.002134546
Immune response
B2M	Beta‐2‐microglobulin	37.1546	5.21E‐132
ILF2	Interleukin enhancer‐binding factor 2	33.9899	0
C9	Complement component C9	33.9726	0
ORM1	Alpha‐1‐acidglycoprotein1	33.9338	3.31E‐124
HLA‐A	HLA class I histocompatibility antigen, A alpha chain	33.6618	4.12E‐171
ILF3	Interleukin enhancer‐binding factor 3	32.7224	0
IGHM	Immunoglobulin heavy constant Mu	29.2927	0.003636039
IGKV3D‐20	Immunoglobulin kappa variable 3D‐20	28.9261	0.003636039
AMBP	Alpha‐1‐microglobulin/Bikunin precursor	28.7914	0.003623759
IGKV2‐29	Immunoglobulin kappa variable 2–29	28.7343	0.004076039
AHSG	Alpha‐2‐HS‐glycoprotein	28.4887	0.004105626
IGKV1‐33	Immunoglobulin kappa variable 1–33	28.3757	0.003642231
GBP1	Guanylate‐binding protein 1	27.9825	0.004076039
Actin dynamics
CCT4	T‐complex protein 1 subunit delta	36.2017	1.23E‐220
CALD1	Caldesmon	31.0103	0
MYL12A	Myosin regulatory light chain 12A	30.7955	0.000547829
ABLIM1	Actin‐binding LIM protein 1	30.5517	0.001530352
EPS8L2	Epidermal growth factor receptor kinase substrate 8‐like protein 2	29.0159	0.004255693
ACTR3	Actin‐related protein 3	28.0870	0.000859807
SEPTIN7	Septin‐7	27.7239	0.003636039
DIAPH1	Protein diaphanous homolog 1	27.5552	0.004105626
Protein homeostasis
PSAP	Prosaposin	39.3859	0
RPN2	Dolichyl‐diphosphooligosaccharide‐‐proteinglycosyltransferase subunit 2	34.1529	0
HSP90B1	Endoplasmin	28.9454	0.00101187
LMAN2	Vesicular integral‐membrane protein VIP36	28.0242	0.00374991
VPS35	Vacuolar protein sorting‐associated protein 35	27.8197	0.003814025
CES2	Cocaine esterase	27.7888	0.003075447
CMBL	Carboxymethylenebutenolidase homolog	−4.1632	6.51E‐06
Cytoskeleton
SYNC	Syncoilin	36.4357	1.55E‐290
MAPRE1	Microtubule‐associated protein RP/EB family member 1	33.8011	2.49E‐196
LAMA5	Laminin subunit alpha‐5	33.7594	0
NES	Nestin	33.6240	9.13E‐05
CNTRL	Centriolin	−28.5006	0.000638065
Ribosome
RPS4X	40S ribosomal protein S4	34.4797	1.10E‐87
RPL15	60S ribosomal protein L15	31.6865	1.61E‐87
RPS17	40S ribosomal protein S17	28.7062	0.000771255
RPL14	60S ribosomal protein L14	27.8408	0.003636039
Metabolism/mitochondria
ADH1B	All‐trans‐retinol dehydrogenase [NAD(+)]ADH1B	37.0445	0
STOML2	Stomatin‐like protein 2,mitochondrial	34.5410	0
PGD	6‐phosphogluconate dehydrogenase, decarboxylating	29.3747	0.004839224
GLUD1	Glutamate dehydrogenase 1, mitochondrial	29.1598	7.40E‐05
Cytoplasmic vesicles
ARF5	ADP‐ribosylation factor 5	37.1513	6.17E‐186
EHD2	EH domain‐containing protein2	35.8474	0
RAB11A	Ras‐related protein Rab‐11A	34.3366	2.80E‐255
DYNC1I2	Cytoplasmic dynein 1 intermediate chain 2	28.1554	0.000402775
Calcium homeostasis
S100A4	Protein S100‐A4	37.1418	6.41E‐96
S100A1	Protein S100‐A1	−3.8520	0.002102357
Sarcolemmal
SGCG	Gamma‐sarcoglycan	29.1324	0.000330293
CLIC1	Chloride intracellular channel protein 1	28.4947	0.003924595
Others
ATP1A1	Sodium/potassium‐transporting ATPase subunit alpha‐1	35.4325	0
RRAD	GTP‐binding protein RAD	35.2894	2.22E‐250
GNA13	Guanine nucleotide‐binding protein subunit alpha‐13	33.1138	1.72E‐48
FTL	Ferritin light chain	29.7657	0.000893115
NIBAN2	Protein Niban 2	29.0158	0.004255693
A2M	Alpha‐2‐macroglobulin	28.6148	0.002652335
TTR	Transthyretin	28.2200	0.003860949
A1BG	Alpha‐1B‐glycoprotein	27.6996	0.003636039
PLCD4	1‐phosphatidylinositol 4,5‐bisphosphate phosphodiesterase delta‐4	−25.7326	0.001141208
AAMDC	Mth938 domain‐containing protein	−28.3072	0.000637142

Proteins are shown by categories in a descending order based on the log2 fold change. Downregulated proteins have negative log2fold change values. FDR, false discovery rate.

Proteins related to the immune system

Among 272 upregulated proteins, 65 unique proteins play a role related to the immune system, some with overlapping roles (Table 3). Eight proteins are involved in MHC1 and antigen processing, 15 in MHCII antigen presentation and phagocytosis, and 23 in B and/or T‐cell function. Furthermore, C9 was upregulated, in addition to 22 proteins with various other immune roles.

Table 3

Immune system‐related proteins.

Gene	Protein name	Comments
MHC1 and antigen processing
B2M	Beta‐2‐microglobulin	MHC1 complex
CALR	Calreticulin	Endoplasmic reticulum processing
CANX	Calnexin	Endoplasmic reticulum processing
HLA‐A	HLA class I histocompatibility antigen, A alpha chain	MHC1 complex
LMO7	LIM domain only protein 7	Endoplasmic reticulum processing
PDIA3	Protein disulfide‐isomerase A3	Endoplasmic reticulum processing
PSMD1	26S proteasome non‐ATPase regulatory subunit 1	Proteasome
PSMD3	26S proteasome non‐ATPase regulatory subunit 3	Proteasome
MHCII and phagocytosis
ASAH1	Acid ceramidase	Neutrophil degranulation
CALM3	Calmodulin‐3	Transmembrane signaling
COL1A1	Collagen alpha‐1(I)chain	Dendritic cell maturation
COL1A2	Collagen alpha‐2(I) chain	Dendritic cell maturation
CTSD	Cathepsin	Phagosome
DYNC1H1	Cytoplasmic dynein 1 heavy chain 1	Phagosome
DYNC1I2	Cytoplasmic dynein 1 intermediate chain 2	Phagosome
GBP1	Guanylate‐binding protein 1	Phagosome
HNRNPM	Heterogeneous nuclear ribonucleoprotein M	Macrophage function
HSPD1	60 kDa heat shock protein, mitochondrial	Macrophage function
MVP	Major vault protein	Neutrophil degranulation
PRDX5	Peroxiredoxin‐5, mitochondrial	Phagosome
PSAP	Prosaposin	Neutrophil degranulation
RAB11A	Ras‐related protein Rab‐11A	Phagosome
VAPA	Vesicle‐associated membrane protein‐associated protein A	Macrophage function
T cells
ACTR3	Actin‐related protein 3	CD28 Signaling in T‐Helper Cells
AHSG	Alpha‐2‐HS‐glycoprotein	T‐cell activation, inhibits TGF‐β
ANXA5	Annexin A5	Immune checkpoint inhibitor
CALM3	Calmodulin‐3	Transmembrane signaling
DBNL	Drebrin‐like protein	T‐cell activation, TCR related
HNRNPU	Heterogeneous nuclear ribonucleoprotein U	TCR‐dependent signaling and activation
HSP90AB1	Heat shock protein HSP90‐beta	T‐cell‐mediated antitumor responses
HSP90B1	Endoplasmin	T‐cell‐mediated antitumor responses
ILF2	Interleukin enhancer‐binding factor 2	Required for T‐cell expression of interleukin 2
ILF3	Interleukin enhancer‐binding factor 3	Required for T‐cell expression of interleukin 2
PADI2	Protein‐arginine deiminasetype‐2	Regulates Th2 and Th17 T cells
PCBP1	Poly(rC)‐binding protein 1	Immune checkpoint for T cells
PSMD1	26S proteasome non‐ATPase regulatory subunit 1	TCR signaling
PSMD3	26S proteasome non‐ATPase regulatory subunit 3	TCR signaling
RAC1	Ras‐related C3 botulinum toxin substrate 1	Transmembrane signaling in B and T cells
B cells
CALM3	Calmodulin‐3	Transmembrane signaling
CDC42	Cell division control protein 42 homolog	Essential for the Activation and Function of Mature B Cells
HNRNPC	Heterogeneous nuclear ribonucleoprotein C1/2	Follicular B cell maintenance
IGHG2	Immunoglobulin heavy constant gamma 2	Plasma cells
IGHM	Immunoglobulin heavy constant Mu	Plasma cells
IGKC	Immunoglobulin kappa constant	Plasma cells
IGKV1‐33	Immunoglobulin kappa variable 1–33	Plasma cells
IGKV2‐29	Immunoglobulin kappa variable 2–29	Plasma cells
IGKV3D‐20	Immunoglobulin kappa variable 3D‐20	Plasma cells
RAC1	Ras‐related C3 botulinum toxin substrate 1	Transmembrane signaling in B and T cells
Miscellaneous
C9	Complement component C9	Membrane attack complex
FTL	Ferritin light chain	Acute phase reactant
A1BG	Alpha‐1B‐glycoprotein	Sequence similarity to the variable regions of some immunoglobulin supergene family member proteins.
A2M	Alpha‐2‐macroglobulin	Acute phase reactant
AMBP	Alpha‐1‐microglobulin/Bikunin precursor	Acute phase reactant
APCS	Serum amyloid P‐component	Acute phase reactant
ARHGAP1	Rho GTPase‐activating protein 1	Rho signaling pathway
BGN	Biglycan	Regulates inflammation and innate immunity
DCD	Dermcidin	Antimicrobial
GNA13	Guanine nucleotide‐binding protein subunit alpha‐13	Rho signaling pathway
GSN	Gelsolin	Acute phase reactant
HIST1H4A	Histone H4	Immune system activation
HIST2H2AA3PE	Histone H2A type2‐A	Immune system activation
HIST2H2BE	Histone H2B type2‐E	Immune system activation
HNRNPK	Heterogeneous nuclear ribonucleoprotein K	Acute phase reactant. Inhibits NF‐IL6
HSPA5	Endoplasmic reticulum chaperone BiP	Promotes the production of anti‐inflammatory cytokines by interacting with phagocytic cells
ORM1	Alpha‐1‐acidglycoprotein1	Acute phase reactant, inhibits NF‐IL6
TF	Serotransferrin	Acute phase reactant
TLN1	Talin‐1	Attachment of lymphocytes to the extracellular matrix
TTR	Transthyretin	Acute phase reactant
XRCC5	X‐ray repair cross‐complementing protein 5	Innate immune response activation, cGAS pathway
XRCC6	X‐ray repair cross‐complementing protein 6	Innate immune response activation, cGAS pathway

MHC1, major histocompatibility complex 1; TCR, T‐cell receptor.

There were nine upregulated immunoglobulin proteins in muscles of SLONM patients: five heavy chain constant region proteins (IGHG1‐02, IGHG1‐03, IGHG2‐02, IGHG3‐03, and IGHG4‐02), two heavy chain variable region proteins (IGHV3‐30 and IGHV3‐74), and two light chain variable region proteins (IGKV3‐11 and IGKV3‐20). Detailed results are shown in Table S2. As mentioned in the methods section, IGVLs are unique to each plasma cell clone and they serve as surrogates of immunoglobulin clonality. Log2 fold ratio for SLONM versus controls was 1.64 IGKV3‐20 and 1.07 for IGKV3‐11. Upregulation of IGKV3‐20 was seen in all samples from SLONM patients, with or without an MP.

Immunohistochemical studies

Given the upregulation of proteins of the MHC1 complex in our cases, we aimed to confirm this finding by immunohistochemistry and to evaluate if it is helpful in differentiating SLONM from congenital nemaline myopathy. Therefore, we evaluated our cases and five patients with congenital nemaline myopathy for MHC1 expression. The latter included: two 31‐ and 19‐year‐old females with ACTA1 myopathy and a 1 year‐old female, a 3‐year‐old male, and a 2‐month‐old male with unknown genes. In SLONM, nemaline rods tend to accumulate in atrophic fibers with marked rarefaction of the myofibrils. Whereas in congenital nemaline myopathies, nemaline rods occur in nonatrophic fibers and are located subsarcolemmally or scattered throughout the sarcoplasm (Fig. 1). In muscle specimens form patients with SLONM, increased MHC1 reactivity was present in fibers containing nemaline rods, as well as adjacent nonatrophic fibers without structural abnormalities or rod accumulation (Fig. 1). This pattern was seen in four out of five SLONM patients and the fifth patient had scattered muscle fibers with increased MHC1 reactivity. In contrast, fibers containing nemaline rods from patients with congenital nemaline myopathy did not display increased MHC1 reactivity (Fig. 1). There was no evidence of a microangiopathy on UEA staining, and only rare fibers had sarcolemmal complement deposit on MAC staining (data not shown). Of note, none of the SLONM patients had inflammatory collections on muscle biopsy.

Figure 1

Muscle biopsy findings. Trichrome‐stained frozen sections (A, C, E, and F) and MHC1‐reacted sections (B and D) from patients with SLONM (A, B, and E) and patients with congenital nemaline myopathy (C, D, and F). In SLONM patients, increased MHC1 reactivity was present in fibers containing nemaline rods as well as adjacent nonatrophic fibers (B), whereas fibers containing nemaline rods from patients with congenital nemaline myopathy did not display increased MHC1 reactivity (D). In SLONM, nemaline rods tend to accumulate in atrophic fibers (A and E) with marked rarefaction of the myofibrils. Whereas in congenital nemaline myopathies, nemaline rods occur in nonatrophic fibers and occur subsarcolemmally (F) or scattered throughout the sarcoplasm (C). Scale bar: 50 μm. MHC1, major histocompatibility complex 1; SLONM, sporadic late‐onset nemaline myopathy. [Colour figure can be viewed at wileyonlinelibrary.com]

Discussion

In this study, we provided a descriptive overview of the proteomic profile of SLONM and we highlighted the immune‐mediated pathways. Our results suggest that despite the lack of inflammatory collections on muscle biopsy, several immune‐related proteins in muscle samples obtained from patients with SLONM are upregulated, including those related to MHC‐1 antigen processing and antibody‐mediated damage. Furthermore, there was no evidence of direct deposition of MP in muscle tissue.

MHC1 upregulation allows a nucleated cell to present antigens recognized by T‐lymphocytes. MHC1 and several related proteins were upregulated in SLONM, including proteins involved in MHC1 processing in the endoplasmic reticulum such as β2‐microglobulin, calnexin, calreticulin, and PDIA3; and antigen processing by proteasomal enzymes such as LMO7, PSMD1, and PSMD3. The MHC1/antigen complex is then presented at the cell membrane (sarcolemmal MHC1 reactivity on immunohistochemistry). The MHC1/antigen complex interacts with the corresponding CD8 T‐cell via two main receptors: the CD8 receptor or and the T‐cell receptor (TCR). Among the CD8 T‐cell‐related upregulated proteins, several of which are involved in TCR activation and signaling such as HNRNPU, DBNL, PSMD3, and PSMD1. , We also identified several proteins involved in B‐cell function and signaling, most prominent among which were several light and heavy chain components as discussed below.

The role of the MP detected in 60% of patients with SLONM is not well understood. The upregulation of nine immunoglobulin proteins is suggestive of underlying immune activation and antibody‐mediated tissue damage. Here, we posit that the circulating MP, detected in patients' sera, may be a simple indicator of underlying autoimmunity rather than direct involvement of the monoclonal process in muscle damage. Indeed, inflammatory disorders are a risk factor for developing an MGUS. In these cases, the MP is not known to play a causal role in the development of inflammation. This is supported in this study by the absence of a clonally dominant MP at the proteomic level. This contrasts with diseases with direct deposition of clonal immunoglobulins such as AL amyloidosis, where much higher amounts of clonally deposited IGVs are typically detected. , We also reported increased expression of IGV3‐20 in SLONM patients compared to controls. This finding is more likely to represent a marker of nonspecific immune activation for several reasons. First, both SLONM patients with and without MP had higher IGKV3‐20 than controls. Second, IGKV3 is a variable region gene family that is only noted in approximately 6–7% of patients with dysproteinemias, but in about 25% of polyclonal plasma cells, making it more likely to represent polyclonal deposition in the SLONM cases. The reason 2 SLONM‐MP cases appeared to have higher levels of several immunoglobulin heavy and light chain variable regions compared to the remaining SLONM cases may suggest that the presence of an MP is a surrogate of heightened immune‐mediated muscle damage, which could explain the disease's more aggressive clinical course in some. This study also underlines the heterogeneity across SLONM‐MP cases with some having immunoglobulin deposition levels comparable to that of SLONM without MP. Therefore, we continue to suggest considering a trial of IVIG first in all patients with SLONM irrespective of the presence of an MP. ,

Finally, while the protein abundance levels of cytokines and interleukins were below the current proteome depth of coverage for the LPC and proteomics methods described here, we showed upregulation of proteins related to IL2 (ILF2, ILF3), IL6 (ORM1, HNRNPK, HNRNPM), interferons (HSP90B1, GBP1), TGF‐β (AHSG is a TGF‐ β antagonist), and IL10 (HSPA5) signaling. , , , These findings raise suspicion for a Th1‐mediated immune response, as seen in other immune‐mediated myopathies. , , It is noteworthy that some proteins although not classically classified as immune proteins play a role in the immune system. Heat shock proteins are implicated in both pro‐inflammatory and anti‐inflammatory responses. Likewise, certain extracellular histones are involved in activating the immune system. , Lastly, C9, a component of the MAC, was upregulated. However, there was no significant complement deposit on muscle fibers on MAC staining. It is possible that the amount of deposited complement escaped detection by the used immunofluorescent method.

In addition to the immune‐related proteins, which were the focus of our study, we also identified significant changes in several proteins implicated in muscle function. As expected, downregulated muscle proteins consisted mostly of structural contractile and sarcomeric proteins. In contrast, a wide array of proteins involved in DNA transcription, RNA translation, protein synthesis and homeostasis, cell adhesion, actin dynamics, and cell metabolism were upregulated. These proteins overlap with those detected in other inherited myopathies with cytoplasmic aggregates such as myofibrillar myopathies, or acquired myopathies such as inclusion body myositis, and rare reports from congenital nemaline myopathy mouse models. , , , , , Therefore, these proteins probably reflect converging pathomechanisms indicating muscle fiber injury, unlike immune‐related proteins which are rarely encountered in inherited myopathies. , , , , Regarding the most common category of nuclear proteins, this is in keeping with the findings on muscle biopsy, where fibers with rods often display internalized nuclei with prominent nucleoli, reflecting high (compensatory) protein synthesis. However, there may be some contribution from sampling bias, as it may be easier to avoid nuclei in normal fibers given their clear subsarcolemmal location. Conversely, in fibers with rods, nuclei may sometimes be difficult to dissect from surrounding sarcoplasmic material.

Our study was limited by the small sample size and the scarce available literature on SLONM pathogenesis. Limitations also included those inherent to the proteomics platform and the depth of coverage of the utilized equipment and methods. Furthermore, the high dynamic range in muscle (over‐abundance of muscle structural proteins in the samples) makes it more difficult to see less abundant proteins. The omics platforms provide immense amount of data that may be difficult to interpret and to determine their clinical relevance. Future studies should include integration with other omics' platform to better understand SLONM pathogenesis. It would also be helpful to compare the findings in SLONM to other myopathies, namely congenital nemaline myopathy and idiopathic inflammatory myopathies. Moreover, comparing the IGV of the circulating clonal plasma cells, which would require PCR of clonal bone marrow plasma cells, to the detected IGVs in tissue would be of interest.

Conflict of Interest

Elie Naddaf, Surendra Dasari, Duygu Selcen, M. Cristine Charlesworth, Kenneth L Johnson, Michelle L. Mauermann, and Taxiarchis Kourelis report no conflict of interest.

Table S1. Complete list of differentially expressed proteins in sporadic late‐onset nemaline myopathy.

Table S2. Immunoglobulin heavy and light chain proteins.

Click here for additional data file.

Data S1. R script file.

Click here for additional data file.