Our study showed a reproducible relationship between dysferlin expression in skeletal muscle and in PBM by WB analysis. Analysis of 17 dysferlinopathy patients with two pathogenic mutations in the gene DYSF and 21 patients with other neuromuscular diseases confirmed normal levels of dysferlin in all the pathological controls and abnormal dysferlin (severe reduction or absence) in all dysferlinopathy patients, demonstrating a high sensitivity and specificity for diagnosis of dysferlinopathy.
IH results, on the other hand, were misleading because dysferlin expression was deficient both in patients with mutations in DYSF and in patients with other myopathies. Most patients with DYSF mutations showed absence of dysferlin in IH, but some showed residual dysferlin expression. A large proportion of pathological controls with other myopathies (13/21) showed abnormal patterns of dysferlin expression using IH but their dysferlin expression in skeletal muscle and PBM was normal according to WB.
Several authors have described different IH patterns of dysferlin expression in muscle biopsies from patients with different myopathies, but none of them compared this expression with that found in PBM. The authors in the first of these studies found IH patterns which were slightly different from those in our study. However, they did not correlate their findings with muscle WB 
. In two later papers, the authors observed mismatching results between WB and IH in muscle biopsies from different LGMD 
. In the paper by Tagawa et al
the authors found defective expression of dysferlin by IH, 19% of patients with LGMD and 50% of patients with other neuromuscular diseases, and established four different patterns of dysferlin staining (normal, negative, faint and abnormal cytoplasmic accumulation) The authors concluded that other proteins are necessary for a normal localization of dysferlin at the sarcolemma. In the paper by Lo et al the authors found abnormal dysferlin expression in a significant proportion of LGMD patients not LGMD2B (25 out of 76). They grouped these patients in four categories according to IH pattern of dysferlin expression in the muscle biopsy including: 1) deficiency, 2) reduced membrane staining, 3) reduced membrane staining with homogenously increased sarcoplasmic staining in all fibers and 4) reduced membrane staining and variable sarcoplasmic staining in some fibers. To explain their results the authors suggested that either the membrane damage is too excessive for satellite cells to repair or that the proteins needed for dysferlin function are defective in these patients. The results of both studies matched well with those found in our study (). In yet another study, the authors found a patchy pattern of sarcolemmal dysferlin in two patients with 5′ donor site mutations, but these mutations could not be related to a milder clinical phenotype 
In a more recent paper the authors described three patients with a suspicion of dysferlinopathy who displayed abnormal skeletal muscle IH (reduced or absent) and increased dysferlin protein by WB.In two of the three patients, genetic analysis showed a non-pathogenic change in one allele and no change in the second allele. The third patient displayed one pathogenic mutation in one allele and a non-pathogenic change in the second allele. WB in PBM was performed in two of the three patients and dysferlin expression was normal in both. The quantification method and the WB, however, were not shown 
. Together with our findings these results indicate that normal dysferlin expression has not yet been observed when two pathogenic mutations in the DYSF
gene are found.
Dysferlin interacts with other proteins to form a complex, such as annexin-1 and 2 
, calpain-3 
, caveolin-3 
, AHNAK 
, affixin 
, dihydropyridine receptor 
, MG53 
and tubulin 
. Disruption of proteins in the dysferlin complex can affect the subcellular localization of dysferlin in the absence of mutations in the gene DYSF
and may explain the abnormal immunohistochemical results described in the above-mentioned studies 
. For this reason it is important to quantify dysferlin expression in skeletal muscle WB as well as by IH. However, in some cases the amount of muscle biopsy is insufficient to perform WB. It may also happen that in patients at end stages of the disease the muscle biopsy may mainly consist of fat and fibrotic tissue, also leading to inaccurate quantification results of muscle proteins. To overcome these difficulties, here we propose dysferlin expression should be studied in PBM to help interpret the reductions noted in the muscle biopsy by IH. Indeed, when the clinical suspicion of dysferlinopathy is high, dysferlin in PBM can be studied even before the muscle biopsy is taken. The mean expression of dysferlin by skeletal muscle WB in our patients was 1.1±4.3% and linked with that in monocytes. It is of note that the authors of a very recent report showed that when the expression of dysferlin in skeletal muscle analysed by WB was lower than 20% it was due to mutations in the gene DYSF 
. The authors, however, did not compare dysferlin expression in PBM in the same patients, as we did in our study.
The analysis of dysferlin in PBM is very helpful to rule out a dysferlinopathy when no muscle sample is available. However, when available, the study of the muscle biopsy is very informative. In fact, the presence of characteristic inflammatory infiltrates 
and sarcolemmal and interstitial amyloid deposits 
can be helpful to support the diagnosis of dysferlinopathy and to study the pathogenesis of the disease.
The high parallelism between dysferlin expression in skeletal muscle and PBM by WB found in our study has improved the diagnostic approach in our routine practice, avoiding misleading diagnoses when the IH pattern is abnormal. In the diagnostic work up of a patient with proximal or distal weakness, we suggest to investigate the presence of dysferlin by WB in PBM in combination with the muscle biopsy prior to molecular analysis of a large gene such as DYSF.