Diagnostically important muscle pathology in DNAJB6 mutated LGMD1D

Limb girdle muscular dystrophy (LGMD) 1D is an autosomal dominant myopathy usually with late onset of weakness at the age of 30–60 and slow progression, although more aggressive forms can occur [1]. The disease affects mainly the lower limb muscles with a particular pattern of preferential involvement [2]. Typically the first symptom is difficulty in climbing stairs. Loss of ambulation does not occur until late senescence, except for the aggressive form of the disease [3]. Some patients may suffer from dysphagia or dysarthria. Phenotype descriptions and families from different ethnic populations have been reported [4, 5].

Myofibrillar myopathies (MFMs) are a group of myopathies that share common morphological findings but have different molecular etiologies [6, 7]. MFMs are characterized by a distinct pathological pattern of dark and hyaline cytoplasmic changes on trichrome staining corresponding to myofibrillar dissolution and accumulation of myofibrillar aggregates causing accumulated expression of several proteins, including desmin, myotilin, alphaB-crystallin and occasional amyloid material. On electron microscopy pathological alterations are marked in Z-disc structures leading to Z-line streaming and dispersion of dark Z-disc components [8]. A novel term: Protein aggregate and vacuolar myopathy, comprises a larger spectrum of myopathies unified by their pathological features, including the myofibrillar myopathies and various rimmed vacuolar myopathies with protein aggregates or inclusions [9].

We have previously reported preliminary data on muscle biopsy findings in patients with p.F93L mutation in DNAJB6 [1, 10] and Sato et al. have reported biopsy findings in five LGMD1D patients [11] with the p.F93I mutation. In our current study of a large biopsy material of LGMD1D patients we show consistent pathological changes in the routine histopathology that can be used to direct the molecular genetic confirmation of the final diagnosis. Moreover, we show that the abnormal protein accumulations follow a sequential evolution from the early stage of small central myofibrillar lesions to the later stage of widespread myofibrillar disintegration with and autophagic rimmed vacuolar pathology in atrophic degenerating fibers.

Clinical summary of the patients is presented in Table 1.

On light microscopy general myopathic or dystrophic features of variable severity depending on the stage of the disease were observed in all cases: fibrosis and fatty replacement of the muscle tissue, fiber size variation with atrophic and occasional hypertrophic fibers and central nucleation (Fig. 2a, b). There were occasional split fibers, necrotic and regenerating fibers, as well as focal endomysial mononuclear cells (Table 3). A uniform microscopic finding was the presence of distinct areas of myofibrillar abnormalities. In addition, rimmed vacuolar pathology was found in all cases. In the early stage pathology the myofibrillar lesions were characteristically small in size and often in a few fibers only, making them difficult to distinguish. On transverse sections they appeared as rounded dense aggregates on H&E staining and were often more easily identified on the modified trichrome stain as dark green areas. In these regions oxidative enzyme stainings were often weak, however without very distinct core or rubbed-out pathology. On longitudinal sections the myofibrillar lesions were wavy in shape extending over several sarcomeres in length. They occurred mainly in fibers with normal caliber and were usually not located in the periphery of the fibers. Both fiber types were equally involved. In the late stage of the disease the rimmed vacuolar pathology was usually very prominent. In a single biopsy a large number of fibers containing pleomorphic hyaline myofibrillar masses reminiscent of a classic myofibrillar myopathy were seen (Fig. 3a). These pleomorphic regions were occasionally congophilic on Congo red staining appearing bright red visualized through Texas red fluorescent filters (Fig. 3b). Increased expression of beta-amyloid precursor protein (APP) was also observed in some of these fiber regions associated with the hyaline deposits.

		Muscle	Neurogenic fiber type grouping	Fiber type predominance	Fiber size variation	Fibrosis and fatty degeneration	Increase of internal nuclei	Fiber splitting	Necrosis	Lymphocyte infiltrates	Fiber regeneration	Rimmed vacuoles	Cytoplasmic myofibrillar inclusions	Core-like lesions	Rubbed-out fibers	Moth-eaten	Lobulated fibers	Ragged red fibers	COX– fibers
FF1
	III-6	VL	–	+		++	+	+	–	–	–	+	+	+	–	++	+	–	–
	III-8	S	(+)	–	++	++	++	++		–	–	+	++	++	–	–	–	–	–
	III-12	VL	–	+	(+)	+	–	–	–	–	+	(+)	+	–	–	–		–	–
		GLU	–	+	(+)	+	–	+	–	+	+	+	++	+	–	–		–	–
		VL	–		(+)	+	+	–	+	+	+	+	++	–	–	+	–	–	–
	IV-7	GM	–	–	+	+	+	+	+	+	–	+	+	+++	–	–	–	–	–
		GM	–	–	++	++	++	++	–	+	+	++	++	+	–	+	–	–	–
	IV-8	GM	–	–	+	–	+	+	+	+	+	+	+	+	–	+	+	–	–
	IV-9	VL	–	+	+	–	+	+	+	+	+	+	+	–	–	+	–	–	–
		VL	+	–	++	+	+	–	–	–	–	+	+	–	–	–	(+)	–	(+)
	IV-10	VL	–	–	–	–	–	–	–	–	–	+	(+)	+	–	–	–	–	–
		VL	–	+	+	–	+	–	–	–	–	+	+	+	–	–	–	–	–
FF2
	III-4	VL	–		+	–	+	–	–	–	–	+	+	–	–	–	–	–	(+)
	III-5	VL	–		+	–	+	–	–	–	–	+	++	–	–	–	–	–	(+)
	III-7	VL	–		++	++	+++	+	+	+	+	+	+	+	–	+	–	–	–
FF3
	IV-1	VL	+	–	++	+	+++	+	+	+	–	+	++	+	–	+	–	–	–
		GL	–	–	++	+	++	+	++	+	–	+	++	++	–	+	+	–	(+)
FF4
	III-2	VL	–	+	++	++	++	+	+	–	–	+	+	+	–	+	–	–	–
	IV-1	VL	–	–	(+)	–	+	–	+	–	+	+	+	++	–	+	–	–	–
FF5
	IV-1	VL	–	+	++	++	++	+	+	+	–	++	+	+	–	–	+++	–	–
FF6
	IV-1	VL	+	–	++	++	+	–	–	–	–	+	+	–	–	+	+	–	(+)

Internal nuclei amount assessed as: –, absent; (+), existent in < 1 %; +, existent up to 5 %; ++, existent in 6–20 %; +++, existent in over 20 % of the fibers. Fibrosis and fatty degeneration: –, no fibrosis, fatty degeneration; +, mildly increased; ++, moderately increased

On immunohistochemistry moderate or strong expression of several proteins including myotilin, desmin, alphaB-crystallin, proteasome and SMI310 were shown in the myofibrillar lesions with aggregates (Fig. 3a, c, d) (Table 4), and dystrophin showed ectopic expression of mild or moderate intensity in the same regions (Fig. 3e). The extent of the aggregate pathology was variable depending on the stage of the disease. The protein accumulations also contained all tested CASA complex proteins such as BAG3, heat shock protein (HSP) B8, and DNAJB6 itself (Fig. 3f, g, h). The rimmed vacuoles showed material with reactivity for several markers of defect degradation and autophagic processing such as ubiquitin, valosin-containing protein (VCP), TDP-43, p62 and SMI-31 (Table 4) (Fig. 4a, b, c, d, e). These rimmed vacuolar regions of local degeneration were filled with autophagosomes as shown by LC3 reactivity, whereas components of mature lysosomes expressing LAMP2 were less abundant or absent in the rimmed vacuolar spaces (Fig. 4f, g). Antibodies applied for the sarcolemmal proteins showed normal findings in the sarcolemma (Table 2).

		Muscle	Myotilin	α-BC	Desmin	DYS-2	SMI-31	TDP-43	p62	Ubiquitin	LAMP-2	VCP	FHL1	Membr prot	LC3	HSPB8	DNAJB6	Ubiquilin2	Cathepsin B	Filamin C	BAG3	α-actinin	Teletonin	Actin	SERCA	α-tubulin	SMI310	Proteasome	β-amyloid	APP	WIP1
FF1
	III-6	VL			+				–	–																	+		–	–	–
	III-8	S	++	++	+	++		++	++			+
	III-12	VL
		GLU			+	+
		VL	+	+	+	+				–		–		–
	IV-7	GM	++	++	+	++				++
		GM	++	++	++	++	++	++	++	++	–	++	–		++			++	–	–	++	+		–			++	+	–	–	–
	IV-8	GM	++	++	++	+		++	++	++	–	–	–		++			–	–	–	++	+		–		–	++	+	–	–	–
	IV-9	VL	–	+	+	+								–													–	+	–	–	–
		VL	+	+	+	–		++	++			–									++	+	+	–	–	–
	IV-10	VL	++	++	+	+	–		+	+		–		–
		VL	++	++	+	+	–	+	+	++	–	–			+				–	–		–	–	+	+	–	+		–	–	–
FF2
	III-4	VL			–		–
	III-5	VL			–		–			–
	III-7	VL			+	+	+			+
FF3
	IV-1	VL				–				++				–
		GL	++	++	++	+				++				–
FF4
	III-2	VL	++	++	++	++	++	++	++	++	–	++	–		++	+	+	++	+	–	++							+	–	+
	IV-1	VL			+	++				++
FF5
	IV-1	VL	++	++	+	+	++	++	++	–	–	–	–			+	+				+						++		–	–	–
FF6
	IV-1	VL	+	–	+	–	+/–	+/–	+	+	–	+	–		+		+	+/–	–

–, normal, no immunoreactivity; +, present immunoreactivity; ++, prominent immunoreactivity

On electron microscopy the myofibrillar lesions showed severe structural sarcomere disarrangement associated with prominent excess and dispersion of disintegrated Z-disk material (Fig. 5). Other frequent findings included autophagic regions with vesicles, myeloid figures and debris material, and fibers with multiple small vacuoles appearing to arise from the sarcoplasmic reticulum. A rare finding was the presence of occasional tubulofilamentous intranuclear or cytoplasmic inclusions in two cases.

The muscle pathology in our study of DNAJB6 mutated LGMD1D shows distinct findings that should provide clues for a correct diagnostic approach even in a sporadic patient without known family history. On light microscopy the most common finding is the presence of early stage myofibrillar disintegration with small myofibrillar aggregates and rimmed vacuolated fibers. On immunohistochemistry many proteins are found in the aggregates, and the rimmed vacuoles are reactive for several proteins associated with autophagy. The light microscopy findings are in general similar to MFMs. However, the myofibrillar aggregates are commonly distributed in scattered fibers and are smaller than in a typical MFM. Only in one biopsy large and pleomorphic aggregates in groups of fibers unevenly distributed across the fascicles more characteristic of a typical MFM were seen. In addition, rubbed-out fibers commonly found in MFMs such as desminopathies and alphaB-crystallinopathies are not a common feature in LGMD1D.

Secondary inflammatory changes that often may occur in genetic myopathies were observed in approximately half of the DNAJB6 mutant biopsies. However, the endomysial inflammatory cells were few in number without larger infiltrates and no invasion in non-necrotic fibers was observed. In addition, there was no significant expression of MHC HLA Class I, thus clearly preventing any confusion with sporadic inclusion body myositis (s-IBM) despite the rimmed vacuolation. In addition, no significant numbers of cytochrome oxidase-negative muscle fibers or ragged red fibers were present. Clinically, the involvement of hamstrings was always more severe than the quadriceps and without the finger flexor weakness as with s-IBM.

The ultrastructural findings in myofibrillar myopathies may show some differences in the genetically different myopathies [15]. Compared to myotilinopathy the changes in LGMD1D show less frequently tubulofilamentous bundles and no basement membrane thickening, described in myotilinopathy [15]. In zaspopathy filamentous accumulations in bundles are frequent and intrasarcoplasmic rods common, but were not observed in LGMD1D. Desminopathies characteristically show cytoplasmic electrondense reticular granulofilamentous accumulations, in addition to areas with sarcomere disorganization and Z-disk alterations. Areas of amyloid-like material often adjacent to granulofilamentous inclusions or in vacuolated fibers can also be found in desminopathy, as well as other myofibrillar myopathies [8, 16]. Reticular desmin aggregates or amyloid deposits were not found ultrastructurally in the studied LGMD1D biopsies. The large complexes of electrodense granulofilamentous accumulations and sandwich formations in desminopathy and αB-crystallinopathy and filamentous bundles and extensive Z-disk alterations in myotilinopathy and zaspopathy were not observed ultrastucturally in LGMD1D.

The autophagic degenerative component of the pathology in LGMD1D with the basic features of rimmed vacuoles on light microscopy and disorganized myofibrillar structures, Z-line dispersion, and dilatation of sarcoplasmic reticulum into vacuolar formations on ultrastructure are not particularly different compared to the similar pathology of the other myofibrillar myopathies or s-IBM. However, eosinophilic inclusions in the rimmed vacuoles that occasionally may be observed in s-IBM were rare in the LGMD1D samples.

LGMD1D is caused by mutations in DNAJB6 gene [10]. DNAJB6 belongs to the evolutionarily conserved DNAJ / HSP40 family of proteins, which regulate molecular chaperone activity by stimulating ATPase activity [17]. DNAJB6 is expressed in many tissues [18] with its highest expression in brain. Its association to skeletal muscle disease was shown in 2012 [10]. Our previous studies suggested DNAJB6 involvement in the CASA pathway, a major mechanism in protein re-cycling and turnover for the maintenance of Z-disc sarcomeric integrity. We have in our previous work found that DNAJB6 interacts with several members of CASA complex: BAG3, HSPA8, STUB1 and HSPB8 [10]. In this study we also demonstrate in patient muscle biopsies the involvement of these proteins, thus confirming the results shown with the F93I mutation [11]. The dominant toxic effect of the mutant F93L decreases the anti-aggregational effect of DNAJB6 and makes the whole chaperonal complex in which DNAJB6 is involved less effective, leading to protein aggregations and secondary autophagic abnormalities with rimmed vacuolar pathology. Because of the defective DNAJB6 function Z-disc components are disturbed and accumulate first (myotilin, desmin, αB-crystallin, BAG3) followed by defective autophagy which can be seen as staining of rimmed vacuolar markers LC3, VCP, TDP-43, SMI-31 and p62 increasing with disease duration. These findings, when observed in the clinical diagnostic praxis, in particular the early changes of smaller myofibrillar aggregates in normal sized fibers should direct the further diagnostic efforts towards DNAJB6 and LGMD1D disease.

Exceptionally, LGMD1B has been reported to cause myofibrillar changes similar to LGMD1D, however without rimmed vacuolar pathology [19].

Histopathology of LGMD1D is different from the other more prevalent autosomal dominant limb-girdle muscular dystrophies; markedly different from the usually nonspecific pathology in LGMD1B and LGMD1C and also distinct from the more closely related LGMD1A myotilinopathy.

Authors take full responsibility of the data, the analyses and the interpretation. All patients have given their informed consent prior to their inclusion in the study. All investigations have been performed in accordance of with Declaration of Helsinki and its amendments. This study has been approved by the ethical committee of Tampere University Hospital. The authors declare that they have no conflicts of interest.