Neurology India Medknow Publications on behalf of the Neurological Society of India ISSN: 0028-3886 EISSN: 1998-4022 Vol. 59, Num. 4, 2011, pp. 513-520

Neurology India, Vol. 59, No. 4, July-August, 2011, pp. 513-520

Original Article

Immunohistochemical differentiation of inflammatory myopathies

Jayalakshmi B Panicker¹, Geeta Chacko¹, Anil Kumar B Patil², Mathew Alexander², Jayaprakash Muliyil³

¹ Section of Neuropathology, Department of Neurological Sciences and Pathology, Christian Medical College, Vellore, India
² Section of Neurology, Department of Neurological Sciences, Christian Medical College, Vellore, India
³ Section of Neurology, Department of Community Health, Christian Medical College, Vellore, India
Correspondence Address: Geeta Chacko, Section of Neuropathology, Department of Neurological Sciences and Pathology, Christian Medical College, Vellore , 632 004, India, geetachacko@cmcvellore.ac.in

Date of Submission: 16-Jan-2011
Date of Decision: 09-Feb-2011
Date of Acceptance: 27-Apr-2011

Code Number: ni11162

PMID: 21891925

DOI: 10.4103/0028-3886.84329

Abstract

Background: Idiopathic inflammatory myopathies are a heterogeneous group of acquired muscle disorders with considerable overlap in the histological features, making histological diagnosis difficult at times.
Aims: To determine the immunohistochemical profile of clinically suspected cases of inflammatory myopathies, using monoclonal antibodies to HLA-1 and membrane attack complex (MAC), and to correlate the clinical, serological, and electromyographic profile and the histopathological picture, with the immunohistochemical profile.
Settings and Design: This was a retrospective study analyzing the clinical and histopathological features in muscle of clinically suspected cases of inflammatory myopathy and correlating it to their HLA-1 and MAC immunostaining profiles.
Material and Methods: The study subjects included 33 cases with suspected inflammatory myopathy and 59 with non-inflammatory muscle disease, as controls. Clinical data, electromyographic findings, serological profile, and details of therapy were obtained from patient records.
Statistical Analysis: Student 'T' test, Pearson's Chi square test, and Kappa statistics were used appropriately.
Results: Although HLA-1 and MAC immunostaining did not help to differentiate the individual subtypes of inflammatory myopathy, when either HLA-1 or MAC was positive, inflammatory myopathy could be ruled in with 86.5% certainty and when both HLA-1 and MAC were negative, it could be ruled out with 95% certainty.

Conclusions: A combination of clinical presentation, serological profile, electromyographic and histopathological features, together with the immunoprofile for HLA-1 and MAC, contribute toward making a diagnosis of inflammatory myopathy.

Keywords: HLA-1, inflammatory myopathy, membrane attack complex

Introduction

Idiopathic inflammatory myopathies comprise of dermatomyositis, polymyositis, and inclusion body myositis, and pose difficulty in diagnosis on account of an overlap in clinical, histological, and immunohistochemical features. A muscle biopsy may not show inflammation, and other muscle diseases may also display inflammatory infiltrates. ^[1],[2],[3] Dermatomyositis is thought to be complement mediated ^[1],[4] while polymyositis and inclusion body myositis are considered to be major histocompatibility (MHC) antigen restricted, cytotoxicT-cell mediated, with overexpression of the MHC-1 antigen. ^[5] However, studies attempting to define a diagnostic, immunohistochemical profile for inflammatory myopathies, based on the aforementioned pathogenetic theories have shown conflicting results. ^{[6],[7],[8],[9],[10],[11],[12],[13],[14],[15],[16],[17],[18],[19],[20]} This study was undertaken to determine the validity of HLA-1 and membrane attack complex (MAC), in the diagnosis of inflammatory myopathies, and to define criteria for the diagnosis of these potentially treatable muscle diseases.

Material and Methods

In this retrospective study from October 2006 to May 2008, we examined 92 muscle biopsies, 52 from patients clinically suspected to have inflammatory myopathy and 40 from patients with suspected non-inflammatory muscle disorders. Cases in which fresh frozen samples were not available were excluded from the study. Clinical data, electromyographic findings, and serological profiles were obtained from patient records.

Fresh muscle biopsies were immediately divided into two parts. One was fixed in 10% neutral buffered formalin, processed routinely, and 4μ sections of paraffin-embedded tissue cut and stained with hematoxylin and eosin (H and E). The other part was snap frozen in isopentane cooled in liquid nitrogen at −80°C. Five micron cryostat sections were prepared and stained with H and E, modified Gomori's trichrome, and enzyme histochemical stains. A portion of the muscle was also submitted for electron microscopic examination.

Immunohistochemical localization of HLA-1 and MAC was carried out using the streptavidin-biotin peroxidase complex method. The HLA-1 antibody was a monoclonal antibody against HLA-ABC antigen (DAKO, Patts, Denmark) used at a 1:200 dilution. The MAC antibody was a monoclonal antibody to C5b-9 (DAKO, Patts, Denmark) used at a 1:50 dilution. If immunohistochemistry for dystrophins, sarcoglycans, spectrin, and merosin had been performed, these were reviewed. HLA-1 staining was not done in two cases, due to the inadequacy of the frozen sample; one case was of dermatomyositis and the other of unclassified inflammatory myopathy.

Diagnosis of inflammatory myopathy was based on the Bohan and Peter criteria ^[21],[22] and diagnosis of non-inflammatory muscle disease was based on the clinical, histological, and immunohistochemical profile. Evaluation of immunohistochemical staining was done by two observers in a blinded manner. Strong staining of the sarcoplasm or sarcolemma by HLA-1 was taken as positive. Weak staining, background staining, and endomysial capillary staining were considered as negative. Strong staining of blood vessels in the endomysium or perimysium by MAC was considered as positive. Weak staining and no staining were taken as negative.

Statistical analysis was performed to determine associations by using the Student 'T' test. Statistical significance of the associations was measured using Chi square test. The agreement between the two immunohistochemical markers, HLA-1 and MAC, was assessed using Kappa statistics.

Results

Thirty-three cases were given a final diagnosis of inflammatory myopathy using the Bohan and Peter criteria. ^[21],[22] The remaining 19 cases were included as controls. There were thus 59 controls. Amongst the 33 diagnosed as inflammatory myopathy, 23 were dermatomyositis, five were polymyositis, and five could not be definitively classified into any subtype of inflammatory myopathy. There were no cases of inclusion body myositis. The demographic, clinical, and biochemical features of the cases and controls are given in [Table - 1].

All the 23 cases of dermatomyositis showed myofibre necrosis, regenerative activity and perifascicular atrophy [Figure - 1]. An inflammatory infiltrate was seen in the endomysium in 19/23 cases and in the perimysium in four cases. Perivascular inflammation in the perimysium was seen in 15/23 cases; 5/23 cases showed vasculitis as evidenced by transmural infiltrates of inflammatory cells. All the five cases of polymyositis showed myofibre necrosis, regeneration and invasion of necrotic and non-necrotic muscle fibres by inflammatory cells [Figure - 2]. An endomysial inflammatory infiltrate was seen in all five cases and perimysial and perivascular inflammation was seen in three out of five cases. Five cases that could not be subtyped showed fibre atrophy, myofibre necrosis, inflammation, and regeneration. There was no evidence of perifascicular atrophy or invasion of non-necrotic muscle fibres by inflammatory cells. Inflammation was seen in the endomysium in three out of five cases. There was no perimysial inflammation in any of the cases and perivascular inflammation was seen in three out of five cases.

Of the 52 controls, 15 cases showed inflammatory infiltrates in the endomysium (three out of six Duchenne muscular dystrophy, one out of two limb girdle muscular dystrophy, one out of two congenital muscular dystrophy, one muscular dystrophy (unclassified), two out of eleven neurogenic atrophy, and seven out of twenty-one cases of nonspecific myopathy) and 10 cases showed inflammation in the perimysium (one of six Duchenne muscular dystrophy, one of two limb girdle muscular dystrophy, one of two congenital muscular dystrophy, one facioscapulohumeral dystrophy, two of eleven neurogenic atrophy, one of two spinal muscular atrophy, and five of twenty-one nonspecific myopathy) [Figure - 3]. Muscle biopsies from the four patients with vasculitis showed transmural inflammation involving the endomysial vessels with additional evidence of neurogenic atrophy in one case.

The HLA-1 immunostaining profile of cases and controls is illustrated in [Figure - 4] and [Figure - 5]. HLA-1 had 74.2% sensitivity and 72.9% specificity in the diagnosis of inflammatory myopathy (P=0.0001). HLA-1 had a sensitivity of 72.7% and a specificity of 66.2% in diagnosing dermatomyositis (P=0.001). The sensitivity of HLA-1 in the diagnosis of polymyositis was 60% and specificity was 57.6% and the difference was not statistically significant (P=0.649). Sixteen of fifty-nine (27.11%) controls showed immunostaining of sarcolemma by HLA-1.

Interpretation of HLA-1 staining was difficult in several cases that showed non-specific background staining despite adequate protein blocking. The immunostaining profile of cases and controls using MAC is illustrated in [Figure - 6], [Figure - 7] and [Figure - 8]. MAC had 90.9% sensitivity and 64.4% specificity in diagnosing inflammatory myopathy (P=0.000). MAC had a sensitivity of 95.7% and a specificity of 58% in diagnosing dermatomyositis (P=0.0001) and a sensitivity of 80% and specificity of 46% in diagnosing polymyositis (P=0.376). Twenty-one of fifty-nine (35.59%) controls showed positivity for MAC, of which 18 showed inflammatory cell infiltrates in the muscle with the exception of one case of non-specific myopathy and two cases of neurogenic atrophy, which had no evidence of inflammation.

The agreement analysis showed a percent of agreement between HLA-1 and MAC of 0.73. The Kappa value was 0.47, indicating an intermediate agreement. This implies that HLA-1 and MAC are markers that are independent of each other.

We therefore analyzed the usefulness of testing for HLA- 1 and MAC in series and in parallel. Of inflammatory myopathies, 93.9% were positive for either HLA-1 or MAC and 46.6% of controls were either HLA-1 or MAC positive. Thus, when HLA-1 and MAC were tested in parallel, the sensitivity was 97.5% and specificity was 45.3% (P=0.000). The likelihood ratio of a negative test was 18. Keeping the pretest probability as 50%, it was found that the post test probability was 0.95. Thus, if both HLA-1 and MAC were negative, inflammatory myopathy could be ruled out, with 95% certainty.

Of the inflammatory myopathies, 71% were positive for both HLA-1 and MAC and 16.9% of controls were positive for both HLA-1 and MAC. Thus, when HLA-1 and MAC were tested serially, the sensitivity was 66% and specificity was 89% (P=0.000). The likelihood ratio of a positive test was found to be 6.4. Keeping the pre-test probability as 50%, the post-test probability was found to be 86.5%. Thus, if both HLA-1 and MAC were positive, there was an 86.5% chance that the patient had inflammatory myopathy.

A scoring system to diagnose inflammatory myopathy was designed incorporating both the Bohan and Peter criteria ^[21],[22] and the immunoprofile. The individual parameters were scored as shown in the [Table - 2].

The final score for cases of inflammatory myopathy was 17 and above, except for one case of dermatomyositis, which had a score of 11. The latter case was the only patient on treatment prior to histological diagnosis. The mean score of the inflammatory myopathy group was 21.75 (±3.81, 11-29, median 21). The mean score of the control group was 8.47 (±7.93, 0 - 25, median 8). Four cases of nonspecific myopathy had scores higher than 15, which were clinically suspected cases of dermatomyositis, inclusion body myositis, limb girdle muscular dystrophy, and metabolic myopathy. All four patients presented with symmetrical proximal muscle weakness and had elevated creatinine phosphokinase levels. The electromyography of the first three patients showed features suggestive of inflammatory myopathy and the fourth patient had a normal electromyogram. However, the muscle biopsies of all these four patients failed to show any characteristic light microscopic features of inflammatory myopathies. Keeping a cut off score of 15, the sensitivity was 96.97% and the specificity was 91.53% (P=0.000) in diagnosing inflammatory myopathy.

Discussion

Bohan and Peter's criteria remain the cornerstone for the diagnosis of idiopathic inflammatory myopathies. ^{[1],[6],[23],[24],[25],[26]} The presence of considerable overlap in the clinical and histopathological features with other muscle disorders has resulted in a lot of research aimed at determining reliable markers for the diagnosis of this group of muscle diseases, but with controversial results. ^{[7],[10],[11],[12],[13],[14],[15],[16],[17],[18],[19],[20],[25],[27],[28],[29],[30],[31],[32],[33],[34],[35],[36],[37],[38],[39],[40]} A significant and fundamental problem in most clinical studies is the classification criteria for inflammatory myopathies. We found Bohan and Peter's ^[21],[22] criteria for the diagnosis of inflammatory myopathy useful, as they provided a guideline for recognizing cases of inflammatory myopathy, although, no single criterion could be taken in isolation to make a diagnosis. However, this classification system may be too inclusive, as it allows some patients with muscle dystrophy to be classified as polymyositis. Furthermore, inclusion body myositis, an entity not known when these criteria had been proposed, have been misclassified as polymyositis. Most workers in the field, therefore, believe that there should be a collaborative effort of specialists in Neurology, Rheumatology, Pediatrics, Dermatology, Pathology, and Epidemiology, to revise the classification criteria for inflammatory myopathies. Nonetheless, in the absence of such a consensus the Bohan and Peter's criteria are still widely used.

Muscle biopsy remains pivotal in the diagnosis of inflammatory myopathy, but can also result in misdiagnosis, due either to inflammatory infiltrates being present in non-inflammatory muscle diseases such as muscular dystrophies ^[2],[3],[10] or, due to inflammatory myopathies being devoid of inflammation. ^[2],[3],[29] This study shows endomysial inflammation in 14/59 (23.73%) controls and perivascular inflammation in 11/59 (18.64%) controls. Dalakas ^[39] stated that histiocytic infiltrates predominate in non-inflammatory myopathies and that polymyositis can be distinguished by the presence of primary inflammation. However, in the present series even non-inflammatory myopathies show lymphocytes and plasma cells in the inflammatory infiltrates indicating that the nature of the inflammatory infiltrates is not discriminatory. Perimysial and perivascular inflammatory infiltrates are considered to be more frequent than endomysial inflammation in dermatomyositis, ^[1],[7] however, we found perimysial inflammation in only 65% and endomysial inflammation in 82% of cases with dermatomyositis. Although endomysial inflammation was seen in all five cases of polymyositis, three also showed perimysial inflammation. Thus, the distribution of the inflammatory infiltrates did not help to differentiate between dermatomyositis and polymyositis.

In patients with an acquired subacute myopathy, Dalakas et al., ^[24],[39] suggested that a probable diagnosis of polymyositis could be made if there was increased serum creatinine phosphokinase, primary inflammation in the muscle, and widespread expression of HLA-1 (MHC-1) antigen. We found that about 25% of controls showed positive sarcolemmal staining for HLA-1; the sensitivity and specificity of HLA-1 in the diagnosis of inflammatory myopathy being only 74.2% and 72.9%, respectively, similar to that reported by van der Pas et al., ^[19] and significantly lower than that reported by Jain et al.^[15] In addition HLA-1 was not useful in discriminating between dermatomyositis and polymyositis. Others ^{[16],[19],[27],[30]} reported HLA-1 expression by muscle fibres in non-inflammatory myopathies and normal muscle. Overexpression of HLA-1 in muscular dystrophies, as seen in our study, is probably secondary to an immune-mediated process. ^[36]

We found a sensitivity of 95.7% for skeletal muscle microvasculature MAC immunopositivity in the diagnosis of dermatomyositis - a finding that is well known. ^{[8],[18],[24]} Positive staining for MAC was seen in the endomysial capillaries, perimysial vessels, and necrotic fibers in these cases. On the other hand our specificity was only 58%, with high false positive rates, as similar staining patterns were seen in polymyositis, Duchenne muscular dystrophy, vasculitis, non-specific myopathy, neurogenic atrophy, and spinal muscular atrophy. This is in sharp contrast to the findings of Kissel et al., ^[17] who found deposits of MAC, exclusively in dermatomyositis. Similarly, Brunn et al., ^[10] argue that MAC expression distinguishes between polymyositis, inclusion body myositis, dermatomyositis and dysferlinopathies. The conclusions of the latter study need to be viewed with caution, as they had only one case each of dermatomyositis, polymyositis, and inclusion body myositis and three cases of dysferlinopathies. An agreement analysis was done for the cases subjected to immunostaining with HLA-1 and MAC. The per cent agreement between HLA--1 and MAC was 0.73. The Kappa value was 0.47, indicating an intermediate agreement between the two methods. This implied that HLA-1 and MAC were independent of each other, marking for different aspects of the underlying pathology. When HLA-1 and MAC were tested in series, inflammatory myopathy could be ruled in with 86.5% certainty, and when tested in parallel inflammatory, myopathy could be ruled out with 95% certainty.

In the scoring system adopted by us, a cut-off score of 15 had a sensitivity of 96.97% and specificity of 91.53% (P=0.0001) in diagnosing inflammatory myopathy. It is likely that the four controls, three of whom were clinically suspected to have inflammatory myopathy and had a score of >15, were in fact cases of inflammatory myopathy with no inflammation in the biopsy - the high score being due to clinical and electromyographic features and immunopositivity for MAC and HLA-1. The absence of inflammation does not exclude a diagnosis of inflammatory myopathy given that endomysial complement deposition is the earliest change in this entity. ^[29] All biopsies in our study were performed before treatment was instituted except in one case of dermatomyositis with a score of 11, in whom the inflammatory response was attenuated, probably contributing to the low score. The drawback of our study was the lack of a long-term outcome and response to therapy that would help to not only check and confirm the diagnosis of inflammatory myopathy, but also aid in further refining the scoring system devised by us.

In conclusion, although HLA-1 and MAC immunosta-ining may be used as part of the diagnostic workup for inflammatory myopathies, they do not aid in the subtyping of this group of diseases. Therefore, other specific markers have to be explored, which will aid in the diagnosis of this potentially treatable group of muscle diseases. For the present, based on our study, we have formulated the algorithm shown in [Figure - 9], to diagnose inflammatory myopathies.

Acknowledgment

The authors wish to acknowledge the technical help received from Ms Janet Paul and Ms Sunitha James.

References

Copyright 2011 - Neurology India

The following images related to this document are available:

Photo images