ABSTRACT
Objective To define the diagnostic utility and clinicopathologic features correlating with anti-cytosolic 5′-nucleotidase 1A (NT5C1A) antibody positivity in idiopathic inflammatory myopathies (IIMs).
Methods 4987 patients had anti-NT5C1A status clinically tested between 2014 and 2019 in the Washington University neuromuscular clinical laboratory. Using clinicopathologic information available for 630 of these patients, we classified them as inclusion body myositis (IBM), dermatomyositis, antisynthetase syndrome, immune-mediated necrotizing myopathy (IMNM), non-specific myositis, or non-inflammatory muscle diseases.
Results Anti-NT5C1A was found in 182/287 patients with IBM (63%), 11/53 patients with dermatomyositis (21%), 7/27 patients with antisynthetase syndrome (26%), 9/76 patients with IMNM (12%), 20/84 patients with non-specific myositis (24%), and 6/103 patients with non-inflammatory muscle diseases (6%). The sensitivity and specificity of anti-NT5C1A seropositivity for the diagnosis of IBM was 63% and 85%, respectively. Anti-NT5C1A positive IBM patients had a higher frequency of finger flexion weakness (p<0.01) and their biopsies harbored more COX negative fibers (p=0.04). 18% of anti-NT5C1A negative IBM patients seroconverted to anti-NT5C1A positive on repeat testing.
Conclusions This is the largest description of patients tested by a clinical diagnostic lab for anti-NT5C1A. We confirm the sensitivity and specificity of anti-NT5C1A for IBM and identified clinicopathologic features in IBM which correlate with anti-NT5C1A status. Notably, anti-NT5C1A testing increased both the diagnostic sensitivity and specificity of IBM when combined with patient age, gender and creatine kinase (CK) level. We propose that future IBM diagnostic criteria include anti-NT5C1A testing.
INTRODUCTION
Inclusion body myositis (IBM) is an idiopathic inflammatory myopathy (IIM) that typically affects patients over the age of 50 [1]. Patients with IBM are clinically characterized by asymmetric finger flexion and knee extension weakness [1]. In 2013, anti-cytosolic 5′-nucleotidase 1A (NT5C1A) antibody was detected in the sera of patients with IBM and recognized as a potential diagnostic marker for IBM [2, 3]. Subsequently, the antibody was detected in patients with dermatomyositis, Sjögren’s syndrome, and systemic lupus erythematosus [4, 5, 6, 7, 8, 9, 10]. This suggests that anti-NT5C1A can be detected in autoimmune diseases other than IBM. The relationship between seropositivity for anti-NT5C1A and other clinicopathologic features in IBM or dermatomyositis have been discussed and some report that seropositivity for anti-NT5C1A in IBM or juvenile myositis predict a more severe phenotype [4, 10, 11, 12]. However, this relationship has not been assessed in other IIMs.
Diagnostic testing for myositis specific and myositis associated antibodies are routinely performed in patients with suspected IIM. However, their sensitivity, specificity and clinical utility vary from testing lab to testing lab. This feature makes it essential to describe testing lab specific data and its relation to clinical practice. Since 2014, the Washington University neuromuscular clinical laboratory has tested anti-NT5C1A via Western blot on 4987 patients, from >100 distinct clinical institutions, who were suspected of neuromuscular disease. 1401 patients (28%) were seropositive for anti-NT5C1A. In this study, we aimed to confirm anti-NT5C1A’s utility in diagnosing IBM, clarify the relationship between seropositivity for anti-NT5C1A and other clinicopathologic features in IIMs, and determine the usefulness of the antibody in clinical practice.
METHODS
Patients
We retrospectively identified 4987 patients, who underwent anti-NT5C1A testing at Washington University in St. Louis from 2014 to 2019.
The clinical chart, biopsy reports, and results of autoantibodies status of 1210 patients were then reviewed from Washington University School of Medicine in St. Louis, University of California, Irvine, The University of Texas Dell Medical School, University of Washington, and Integris Southwest Medical Center (Fig.1). We excluded those patients with diseases whose primary pathology is outside of skeletal muscle or whose final diagnosis was still under investigation.
IBM was classified by European Neuromuscular Center (ENMC) criteria [13]. 37 patients with definite or probable clinical criteria of IBM; yet a muscle biopsy had not been performed were also included. ENMC diagnostic criteria was used for the diagnosis of dermatomyositis and immune-mediated necrotizing myopathy (IMNM) [14,15]. Anti-synthetase syndrome was classified by the following, the presence of anti-aminoacyl tRNA synthetase antibody and one or more of the following clinical features (Raynaud’s phenomenon, arthritis, interstitial lung disease, fever (not attributable to another cause), mechanic’s hands or biopsy finding of immune myopathies with perimysial pathology) [16, 17, 18]. In this way, the cohort included patients with IBM (n=287), dermatomyositis (n=53), antisynthetase syndrome (n=27), and IMNM (n=76). Patients that did not meet this criteria and had a muscle biopsy with endomysial or perimysial inflammation, MHC-class 1 up-regulation on muscle fibers or complement C5b-9 complex deposits on endomysial capillaries were categorized as non-specific myositis (n=84).
Patients with a non-inflammatory, hereditary or presumed hereditary muscle disorder were also included. The diagnoses of these patients were VCP-related multi-system proteinopathy, limb-girdle muscular dystrophy, Becker muscular dystrophy, myotonic dystrophy, centronuclear myopathy, myofibrillar myopathy, mitochondrial myopathy, rhabdomyolysis, and McArdle’s disease (n=103).
Serological testing for autoantibodies
Anti-NT5C1A testing was clinically performed by the Washington University Neuromuscular Laboratory via Western blotting as previously described [11]. In some cases, anti-Jo-1 antibody, anti-HMGCR antibody, anti-SRP antibody, and anti-MDA5 antibody were also performed by Washington University Neuromuscular Laboratory [19]. Serums were tested and interpreted blindly without knowledge of the diagnosis. All other antibody data was extracted from the clinical chart. Panels of serum muscle specific antibodies were evaluated by clinical laboratories in Oklahoma or California.
Myopathologic analysis
The percentages of cytochrome oxidase (COX) negative and succinate dehydrogenase (SDH) positive muscle fibers were determined by photographing a random field with a 10x objective and counting 200 fibers. A cutoff value of 1.0% was used for qualitative analysis of COX negative fibers as with our previous report [20].
Definition of clinical features and assessment of responses to immunotherapy
Dermatologic features were defined by clinical description (i.e., Heliotrope rash, Gottron’s sign, V-sign, shawl sign, calcinosis cutis, and mechanic’s hands), skin biopsy data, and/or expert opinion of a dermatologist. Inflammatory arthritis was defined by imaging features and/or expert opinion of a rheumatologist. Interstitial lung disease was defined by imaging features, pulmonary function tests, biopsy data, and/or expert opinion of a pulmonologist.
Responses to treatments were determined for patients followed by more than 1 year as: effective, muscle weakness had diminished and the level of activities of daily living had improved to the same as premorbid level; no worsening, some symptoms had improved but there were still several symptoms or abnormality on the blood test; worsening, muscle weakness was progressive and there was no improvement with any immunosuppressive treatments.
Statistical analyses
Fisher exact test compared categorical variables and t-test compared continuous variables. We considered p values less than 0.05 as statistically significant. To control for multiple comparisons, we applied Bonferroni correction to provide an adjusted threshold for significance in Table 3 and Table 4. Multiple logistic regression models were used to determine the diagnostic utility of anti-NT5C1A for IBM versus IIMs.
Protocol approvals and patient consents
All studies were approved by the Human Studies Committee at Washington University in St. Louis.
Data availability
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
RESULTS
Among 527 patients with clinical or pathologically defined myositis, 229 patients (43%) were seropositive for anti-NT5C1A of which 182 patients (79%) were classified as IBM (Table1). Among all 287 IBM patients, 63% were seropositive for anti-NT5C1A. The sensitivity of anti-NT5C1A seropositivity for the diagnosis of IBM was 63%. Anti-NT5C1A seropositivity was 85 % specific for IBM among all patients with muscle diseases, and 80% specific for IBM among IIMs (Table 1). To understand how anti-NT5C1A testing changes existing sensitivity and specificity using minimal clinical criteria (age, gender and creatine kinase (CK) value) within our 527 patient IIM cohort, we performed multiple logistic regression model analysis (Fig. 2). The addition of anti-NT5C1A status increased the AUC from 0.76 to 0.84 demonstrating an improvement in diagnostic accuracy (Fig. 2 compared curve 1 (age, gender, CK) to curve 2 (anti-NT5C1A, age, gender and CK)). Logistic regression model with the addition of a biopsy with or without endomysial inflammation further increased the AUC to 0.92 (Fig. 2 curve 3 (age, gender, CK, anti-NT5C1A status, and +/- endomysial inflammation)).
To understand the differences between anti-NT5C1A seropositive and seronegative patients independent of IIM subtype, we compared these two groups (Table 2). There was no significant difference as to the ratio of males to females, age at onset or antinuclear antibody seropositivity. In contrast, anti-NT5C1A positive patients had a lower serum CK (868±1339 IU/L), higher frequency of endomysial inflammation (91%), and presence of MHC-class1 expression (88%), compared with seronegative patients (p<0.01). There was no difference in the presence of perimysial inflammation between the two groups. Notably, complaints of dysphagia were significantly enriched in anti-NT5C1A patients (p=0.03).
To see if these differences were due to the overrepresentation of IBM patients in the anti-NT5C1A positive group, we performed a similar analysis that removed the 287 patients with IBM and thus looked at only non-IBM IIMs. Similar to the previous group, anti-NT5C1A positive non-IBM IIM patients had a lower serum CK (1912±2809 IU/L; p=0.01) and a higher frequency of endomysial inflammation (61%) suggesting that these two features are not affected by the increased number of IBM patients in the anti-NT5C1A positive group. As expected, independent of anti-NT5C1A status, patients with IBM were significantly older (p<0.01) and male 60% as compared to 41% in the non-IBM cohort (p<0.01). In addition, IBM patients showed lower serum CK (581±632 IU/L), higher frequency of endomysial inflammation (93%), MHC-class1 expression (87%) and dysphagia (50%).
Anti-NT5C1A positive IBM patients had a higher frequency of finger flexion weakness (p<0.01), but no significant differences were seen with regard to age at onset, maximum creatine kinase level, dysphagia, complications of other autoimmune diseases/sarcoidosis/malignancy, or co-existing autoantibodies (Table 3). Anti-NT5C1A positive IBM patients also had significantly more COX negative fibers (p=0.04). There were no other significant differences in muscle biopsy features.
We also compared clinicopathologic features and responsiveness to immunosuppressive treatments in non-IBM IIM patients who underwent anti-NT5C1A testing. Among patients with dermatomyositis (n=53), antisynthetase syndrome (n=27), and IMNM (n=76), 11 (21%), 7 (26%), and 9 (12%) patients were seropositive for anti-NT5C1A, respectively (Table 4). Dermatomyositis patients seropositive for anti-NT5C1A had a higher frequency of calcinosis cutis (p<0.01). There were no significant differences between anti-NT5C1A positive patients and negative patients with antisynthetase syndrome or IMNM. Of note, anti-NT5C1A positive dermatomyositis or antisynthetase syndrome patients did not show a higher association with interstitial lung disease. Also, anti-NT5C1A status did not correlate with treatment response in patients seropositive for anti-HMGCR antibody or anti-SRP antibody (p=0.61, p=0.53). Interestingly, among patients with dermatomyositis, antisynthetase syndrome, and IMNM that had an associated malignancy, all were seronegative for anti-NT5C1A. Among the 103 non-inflammatory muscle disease, VCP-related multi-system proteinopathy showed higher rate of anti-NT5C1A seropositivity compared with that in other non-inflammatory muscle diseases (p=0.03) (Table 1).
There were 48 patients with IBM whose anti-NT5C1A status had been retested at an interval of greater than one year. Anti-NT5C1A status did not change in 43/48 (90%) of patients. Among 22 patients who were negative for anti-NT5C1A, four patients (18%) converted to positive. There was no significant difference between these four patients and those whose antibody status remained negative regarding age at onset (p=0.6), time of onset to biopsy (p=0.63), or serum CK level (p=0.33).
DISCUSSION
In this study, we explored the clinical utility of anti-NT5C1A antibody testing in a large number of patients with IIMs. Among patients with muscle diseases, the seropositivity of anti-NT5C1A was 63% sensitive and 85% specific for the diagnosis of IBM. This result compares with the results of previous studies, which suggested the range of sensitivity was 33% to 80% and that of specificity was 87% to 100% [2, 3, 5, 6, 7, 8, 21]. The Washington University Neuromuscular Clinical Laboratory clinically performs anti-NT5C1A antibody testing via western blot. The sensitivity of anti-NT5C1A for IBM with western blot was 61% in a previous study and 63% in this study [7]. Among studies using more than 10 samples from patients with IBM, the sensitivities of western blotting were consistently higher than that of ELISA assay unless a combination of ELISA assays were used to detect IgM, IgA, and IgG anti-NT5C1A antibodies [21].
Whether NT5C1A antibody positivity can be used to clinically support a diagnosis of IBM has been debated. We demonstrate in this study that ROC curve analysis using a linear regression model that included anti-NT5C1A status improved diagnostic accuracy. This analysis intentionally did not include biopsy data which may not be performed on some patients (Fig. 2 Curve 2). When ROC curve analysis also included the minimal biopsy feature of endomysial inflammation the sensitivity and specificity further improved (Fig. 2 Curve 3). These data support that the diagnosis of IBM does not rely on a single piece of clinical data and that anti-NT5C1A is an additional diagnostic tool. This is further highlighted by the fact that our data finds that when an IBM patient has finger flexion weakness and/or knee extension weakness, the presence of vacuoles is detected in 158 patients (65%) and focal invasion is detected in 128 patients (68%). Thus, the sensitivity of anti-NT5C1A alone (63%) is comparable to biopsy findings known to be characteristic for IBM, albeit the specificity of a biopsy with rimmed vacuoles is likely superior to anti-NT5C1A antibody positivity.
A previous report found that anti-NT5C1A antibody positive patients had a lower frequency of rimmed vacuoles [7]. Whether this is due to clinicians more readily ordering anti-NT5C1A antibody testing on patients without biopsy features regarded as specific for IBM (i.e. rimmed vacuoles) is not clear. This is consistent with our finding that 37 patients in our cohort with a clinical diagnosis of IBM had anti-NT5C1A testing without muscle biopsy and suggests that in clinical practice, clinicians are foregoing biopsy for less invasive testing such as serum autoantibodies.
Interestingly, the seropositivity for anti-NT5C1A in VCP-related multi-system proteinopathy was higher than that in other non-inflammatory muscle diseases. Some patients with clinico-pathologically defined IBM have been found to have pathogenic VCP mutations [22, 23]. Whether this finding is related to the presence of rimmed vacuoles seen in VCP-related multi-system proteinopathy patients remains to be established.
In IBM patients, anti-NT5C1A seropositivity has been previously associated with an increase in biopsies containing COX negative fibers [12]. Our study similarly found that anti-NT5C1A positive IBM patients had biopsies with an increased percentage of COX negative fibers. Notably anti-NT5C1A positivity also strongly correlated with the presence of endomysial inflammation even in non-IBM associated IIMs.
Previously, seropositivity for anti-NT5C1A has been reported to correlate with a more severe phenotype in IBM, dysphagia and facial weakness [11, 12]. Other reports have failed to find a difference between anti-NT5C1A positive and negative patients [24]. Although these three studies did not show the relevance between seropositivity and the distal upper limb/grip weakness, in our study, seropositive IBM patients showed more frequent involvement of finger flexion weakness. The different results may be derived from the difference of method to detect the antibody, inclusion criteria, or the number of patients evaluated.
This is the first assessment of anti-NT5C1A status correlated with the clinicopathologic features of antisynthetase syndrome and IMNM. Although there were patients who were seropositive for anti-NT5C1A, their clinicopathologic features were not different from those in the seronegative patients. Specifically pertaining to malignancy (a feature important for patient management and prognosis), seropositive patients showed no clear association in dermatomyositis, antisynthetase syndrome, and IMNM.
In this study, several IBM patients had repeat anti-NT5C1A testing. When a patient was negative for anti-NT5C1A, 18% seroconverted to positive in the repeat test for anti-NT5C1A, while 96% of patients who were seropositive for the antibody remained positive in the follow up test. In clinical practice, this information will be helpful when considering if repeat testing may be helpful in the situation when the first test was negative for anti-NT5C1A.
This study has limitations. First, this is a retrospective study of clinical records and some features which were not described in the clinical charts may not have been adequately captured. Second, more numbers of patients with dermatomyositis, antisynthetase syndrome, and IMNM are needed in order to discuss a more precise statistical significance. Nevertheless, this study demonstrates the clinical utility of anti-NT5C1A in idiopathic inflammatory myopathies. Specifically, anti-NT5C1A testing increases the diagnostic sensitivity and specificity for IBM when combined with limited clinical information. We suggest that anti-NT5C1A testing be included as part of future diagnostic algorithms distinguishing IBM from other IIMs.
Data Availability
The data that support the findings of this study are available from the corresponding author, upon reasonable request.
Author Contributions
C.I., A.R.F., N.A.G., R.C.B., T.M., A.P., and C.C.W. contributed to the conception and design of the study. C.I., A.R.F., N.A.G., S.R., J.C., Y.H., L.H.W., J.C.K., B.A.B., T.M., and C.C.W. contributed to data acquisition. C.I., S.R., M.W., and C.C.W. contributed to analysis of data. C.I. and C.C.W. contributed to drafting the manuscript. C.I., A.R.F., N.A.G., R.C.B., T.M., A.P., and C.C.W. contributed to revising the manuscript for intellectual content.
Potential conflicts of interest
Nothing to report.
Footnotes
Study funding: This research was supported by CureIBM, Catalyze a Cure and NIH grants K24AR073317 to CCW and K08AR075894 to ARF.