Different Impact of ALS on Laminin Isoforms in Human Extraocular Muscles versus Limb Muscles

Abstract

Purpose.: To determine the impact of amyotrophic lateral sclerosis (ALS) on the extraocular muscles (EOMs) by examining the laminin isoform composition of the basement membranes (BMs) in EOMs and limb muscles from donors with ALS.

Methods.: Muscle samples collected at autopsy from ALS donors and from transgenic mice overexpressing human superoxide dismutase type I mutations (D90A or G93A), and age-matched controls were analyzed with immunohistochemistry using antibodies against laminin chain (Ln)α2, Lnα4, Lnα5, Lnβ1, Lnβ2, and Lnγ1. Neuromuscular junctions (NMJs) were identified with α-bungarotoxin.

Results.: Lnα2, the hallmark chain of skeletal muscle, and Lnβ2 were absent or partially absent from the BMs in a variable number of muscle fibers in most of the ALS EOMs. Three ALS donors showed dramatic decrease in the levels of these chains around their muscle fibers and NMJs. Changes in Lnα2 were not age related and were also present in EOMs of ALS mouse models. Lnα4 was preserved in the majority of NMJs in EOMs but absent in the majority of NMJs in limb muscle of ALS. The BMs around muscle fibers, NMJs, nerves, and blood vessels of the majority of EOMs of ALS donors had rather normal appearance and laminin composition, but heterogeneity was observed among EOM samples of individual ALS donors and between ALS donors.

Conclusions.: The present study showed distinct impact of ALS on EOMs compared with limb muscles. The EOMs maintained a normal laminin composition in their NMJs, which may be instrumental for the fact that they are not typically affected in ALS.

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease involving limb, trunk, and bulbar muscles. It is characterized by a rather selective loss of motor neurons and supporting cells in the brain and spinal cord, resulting in progressive muscle wasting, paresis, and inevitably death from paralysis of the respiratory muscles. ^1,2 The majority of ALS cases are sporadic (SALS), but approximately 10% of cases are familiar forms of ALS (FALS). Mutations in nine genes have been found to cause ALS. Most common are mutations in the free-radical scavenging protein superoxide dismutase type I (SOD1); 12%–23% of FALS patients have mutations in this gene. ALS is a biologically heterogeneous syndrome in which genetics, environment, and aging are interconnected. ² Our understanding of the pathophysiology of ALS is still fragmentary, and a variety of mechanisms have been suggested to play important roles in ALS. ^{1,3 –6}  

Although the ocular motor nuclei are considered to be spared and disturbed ocular mobility is not typical of ALS, abnormalities of the oculomotor system in ALS patients have been reported. ^{7 –13} We recently showed that the extraocular muscles (EOMs) of terminal ALS donors show signs of disease involvement, with the presence of hypertrophic and atrophic muscle fibers, increased connective tissue, as well as changes in myosin heavy chain composition. ¹⁴ These pathologic changes were present in both FALS and SALS cases, irrespective of whether they had a bulbar or spinal disease onset. ¹⁴ Remarkably, although the EOMs are not completely spared in ALS, they are by far less affected than the spinal-innervated muscles from the same donors. 

EOMs differ from other skeletal muscles in numerous ways and are classified as a separate muscle allotype with intrinsincally distinct gene repertoires from those of the limbs and trunk and the masticatory muscle allotypes. ^15,16 One of the most striking features of EOMs is their selective involvement and/or resistance to certain diseases. The EOMs remain unaffected in lethal neuromuscular diseases affecting all other muscles in the body, such as Duchenne muscular dystrophy and merosin-deficient muscular dystrophy. In contrast, they are the most vulnerable or affected muscles in other diseases, such as myasthenia gravis, thyroid-associated orbitopathy, and Miller Fisher syndrome. ^17,18  

Laminins are extracellular matrix (ECM) proteins and major components of the basement membrane (BM) surrounding each muscle fiber. At neuromuscular junctions (NMJs), laminins are present in the BM of the synaptic cleft both on the nerve ending and on the muscle cell as well as on terminal Schwann cells (tSCs). In addition, laminins are present in the BM of nerves and blood vessels. Laminin are large heterotrimeric glycoproteins composed of an α-chain, a β-chain, and a γ-chain. To date, five different α-chains, four β-chains, and three γ-chains have been identified in humans, resulting in at least 16 different laminin isoforms whose names reveal their chain combination. ^19,20 In general, laminin chain (Ln)α2 (in heterotrimers laminin-211 and -221) is the predominant laminin chain in the BM of muscle fibers. ^{21 –23} Lnα4 (laminin-411 and -421) and Lnα5 (laminin-511 and -521) are seen in the peripheral nerves. ²⁴ At NMJs, the BM of tSC mainly contains Lnα2 and Lnα4 chains. ²⁵ The expression of laminin isoforms is spatially and temporally regulated in a precise manner. ²⁶ We have previously demonstrated that the EOMs have a distinct laminin composition. ²⁷ In addition to Lnα2, Lnα4 and Lnα5 are present in the extrasynaptical BM of muscle fibers in the EOMs, whereas Lnα4 is not present in the BM of mature limb muscles. ^27,28 At NMJs of adult EOMs, there is increased expression of Lnα4 as well as maintained expression of Lnα2, Lnα5, and Lnβ2. The perineurium and the endoneurium contain Lnα4, Lnα5, Lnβ2, and Lnα2. The intramuscular blood vessels and capillaries contain Lnα4, Lnα5, and Lnβ2. ²⁷  

Because the laminins play a central role (1) in the maintenance of muscle fiber integrity and (2) in NMJ regeneration and (3) they are present in the BM of muscle fibers, tSCs, nerves, and blood vessels, ^{21,25,26,29,30} in the present paper we used antibodies against laminin chains α2, α4, α5, and β2 to further investigate the impact of ALS on the EOMs. Particular focus was put on the NMJs as the ECM is known to play a crucial role in NMJ repair and remodeling. 

Material and Methods

Human Samples

Twenty-seven EOM muscle samples and seven limb muscle samples collected at autopsy from eight ALS donors (Table 1) were used. For comparison, eight EOM samples from six age-matched subjects (males, 54, 76, 82, 86, 87 years old; female, 72 years old) and five samples of biceps brachii from five age-matched males (58, 58, 78, 81, 83 years old) were used as controls. None of the control subjects was known to suffer from neuromuscular disease. The muscle samples were collected and studied with the approval of the Regional Ethical Review Board in Umeå, following the legal and ethical recommendations of the Swedish Transplantation Law, and adhering to the tenets of the Declaration of Helsinki. 

Table 1.

 

Characteristics of ALS Donors and Muscles Examined in Each Donor

Donor	Sex	Age at Death (y)	Symptom Duration (mo)	Diagnosis	SOD1 Genotype	Site of First Symptom	Muscle Studied
							EOM	Limb
Da*	Male	78	84	FALS	D90A/D90A	Bulbar onset	—	Vastus lateralis
D1	Male	80	31	SALS	WT/WT	Right hand	RS, OS, RL, RM	Biceps brachii
D2	Male	75	317	FALS	D90A/D90A	Left leg	RS, RL, RM	Tibialis anterior
D3	Female	64	132	FALS	D90A/D90A	Left leg	RS, OS, RL, RM	Biceps brachii
D4	Female	80	12	SALS	WT/WT	Bulbar onset	RL, RM	Vastus lateralis
D5	Male	66	13	SALS	WT/WT	Bulbar onset	OS (2), RL (4)	Biceps brachii
D7	Female	58	50	SALS	WT/WT	Bulbar onset	RS (2), OS, RL, RM (2)	Vastus lateralis
D8	Male	71	21	SALS	WT/WT	Thorax	OS, RM	—

*  The donors were numbered according to their collected order. For consistency with the previous donor numbers, ¹⁴ the newly added and earliest donor was named Da. The middle part of each EOM listed was examined. In some cases, the distal and proximal parts of the same muscle or the middle part of the contralateral muscle was also examined, and the total number of samples studied is given in parentheses. ALS, amyotrophic lateral sclerosis, EOM, extraocular muscle; FALS, familiar forms of ALS; OS, obliquus superior; RL, rectus lateralis; RM, rectus medialis; RS, rectus superior; SALS, sporadic ALS; WT, wild-type phenotype.

Mouse Samples

Transgenic mice overexpressing human SOD1 with D90A or G93A mutations were obtained and bred as previously described. ³¹ Mice were regarded as terminal stage when they were no longer able to reach for food. C57BL/6 mice were used as controls. Both SOD1^D90A (n = 2) and SOD1^G93A (n = 3) mice at terminal stage and age-matched wild-type controls (n = 5) were killed, and the extraocular muscles and hind limb muscles were rapidly excised. Experiments complied with ARVO Statement for the use of Animals in Ophthalmic and Vision Research and were approved by the Ethical Committee of the Medical Faculty, Umeå University. 

Antibodies and Immunofluorescence

The muscle samples were mounted and processed for histochemistry as previously described ¹⁸ with a large battery of antibodies (Table 2). Neuromuscular junctions were detected by labeling with rhodamine-conjugated α-bungarotoxin (Molecular Probes, Eugene, OR). The fluorescein-conjugated goat anti-mouse/rabbit secondary antibodies (Alexa 488 for green fluorescence; Molecular Probes) or donkey anti-rat secondary antibody (FITC; Jackson Immunoresearch Laboratories, West Grove, PA) were used. 

Table 2.

 

Antibodies Used for Immunocytochemistry

Antibody	Specificity	Short Name	Reference
5H2*	Laminin chain α2	Anti-Lnα2	32
168FC10†	Laminin chain α4	Anti-Lnα4	28
4C7*	Laminin chain α5	Anti-Lnα5	33
DG10†	Laminin chain β1	Anti-Lnβ1	34
C4†	Laminin chain β2	Anti-Lnβ2	35
113BC7†	Laminin chain γ1	Anti-Lnγ1	36
MCA-1982‡	Lutheran glycoprotein	Anti-Lu	27
A4.951§	MyHCI	Anti-MyHCI	37, 38
A4.74§	MyHCIIa	Anti-MyHCIIa	37
N2.261§	MyHCI+IIa+eom	Anti-MyHCI+IIa+eom	38, 39
CD56‖	Neural cell adhesion molecule	Anti-N-CAM	40
4H8–2#	Laminin chain α2**	Anti-Lnα2	41, 42
1117#	Laminin chain β2**	Anti-Lnβ2	42
1129 637#	Laminin chain α4**	Anti-Lnα4	42
585#	Laminin chain α5**	Anti-Lnα5	42

*  Novocastra Laboratories (Newcastle upon Tyne, UK).

†  Gift from the late Ismo Virtanen (University of Helsinki, Helsinki, Finland).

‡  Serotec Ltd. (Kidlington, Oxford, UK).

§  Developmental Studies Hybridoma Bank, University of Iowa (Iowa City, IA).

‖  Becton Dickinson Immunocytochemistry System, San Jose, CA.

#  Generously provided by Madeleine Durbeej (Lund University, Lund, Sweden) and Takako Sasaki (Friedrich-Alexander University of Erlange-Nürnberg, Germany).

**  Used on mouse tissue.

The sections were examined using a microscope (Nikon Eclipse, E800; Nikon, Tokyo, Japan) equipped with a digital camera (Spot RT color; Diagnostic Instruments, Sterling Heights, MI). Computer-generated images were processed using commercially available imaging software (Photoshop; Adobe Systems, Mountain View, CA). 

Results

Control EOMs

We previously examined the distribution of laminin chains in EOMs from both young and old normal subjects. ²⁷ The patterns of distribution for the different laminin chains in the EOMs of the control subjects in the present study (Fig. 1A1–D1) were identical with those previously described, ²⁷ with the following exceptions: (1) the muscle fiber BM was mostly unstained or very weakly stained with anti-Lnα4 in old normal EOMs (Fig. 1B1); (2) the BM of a small number of EOM fibers was unstained with anti-Lnα2 in the 82-year-old male; (3) the BM of some muscle fibers was unstained with anti-Lnα5 antibody in EOMs of 82- and 86-year-old males; and (4) the endoneurium was unstained with anti-Lnα5 in the EOMs from the 86-year-old male. 

Figure 1.

 

Immunoreactivity with antibodies against Lnα2 (A), Lnα4 (B), Lnα5 (C), and Lnβ2 (D) in the BM of EOM fibers from old normal controls (A1, B1, C1, D1) and ALS donors (A2–A5, B2, B3, C2, C3, D2, D3). Note the heterogeneity in immunoreactivity with laminin chains in the BM of muscle fibers, nerves (N), and blood vessels (BV) in different ALS EOMs, and the different intensity with anti-Lnα2 in the global and orbital layers in RL from donor 2 (A5). GL, global layer; OL, orbital layer.

EOMs from ALS Donors

There was broad heterogeneity in the staining patterns obtained with the antibodies against laminin chains in different EOMs from each individual donor and between donors, irrespective of ALS type or onset. The results are summarized in Table 3. 

Table 3.

 

ALS Donors and Their Staining Patterns with Laminin Antibodies

Donor	Muscle	Lnα2			Lnα4					Lnα5					Lnβ2
		fc	p	e	fc	p	e	bv	c	fc	p	e	bv	c	fc	p	e	bv	c
1	RS	+/−/++* †	−	+	−	++	++	++	++	−	+++	−	+++	+++	−	+	−	++	−
	OS	+/−/++	−	+	−	+++	+++	++	++	−/+	+++	−	+++	+++	+/−	++	−	+++	−/+
	RL	++	−	+	−	+++	+++	+	++	−	+++	−/+	+++	+++	−	+	−	++	−
	RM	++	−	+	−	+++	+++	++	++	−/+	+++	−	+++	+++	−	+	−	++	−
2	RS	+	−	+	−	+++	+++	++	++	−/+	+++	+/−	+++	+++	−	++	−‡	++	−
	RL	+++/+*	+	+/++	−	+++	+++	++	+	−/+	+++	−†	+++	+++	−/+	+	−	++	+
	RM	++/+/−*	−	+/−	−	+++	+++	++	++	+/−	+++	+/−	+++	+++	−	+	−	++	−
3	RS	+/−/++*	−	+/−	−	+++	+++	++	+	−/+	+++	+/−	+++	+++	−/+	++	−/+	++	−
	OS	++/+	−	++	−	+++	+++	+	++	+	+++	++	+++	+++	+/−/++	+	−	+++	−
	RL	++/+/−*	−/+	+/−	−	+++	+++	++	++	−/+*	+++	−/+	+++	+++	+/−/++	++	+/−	++	−
	RM	++/+	−	+/−	−	+++	+++	++	++	−/+	+++	−/+	+++	+++	−	+	−	++	−
4	RL	+/++/−*	−	−/+	−	+++	++	++	++	−/+	+++	−†	+++	+++	−/+	+	−	++	−/+
	RM	++/+	−	+	−	+++	+++	++	++	−/+	+++	−†	+++	+++	−	+	−‡	+	−
5	OS	+/−/++	−	+/−	−	+++	+++	++	++	−/+	+++	−†	+++	+++	−	+/−	−	++	−
	OS	++/+/−*	+/−	+/−	−	+++	+++	++	++	−/+	+++	−†	+++	+++	+/++/−	+++	+/−	+++	−/+
	RL	+/++/−	−	+	−	+++	+++	++	++	−/+	+++	−	+++	+++	+/−	+	−	++	−
	RL	+/++	−	+/++	−	++	++	++	++	−	+++	−/+	+++	+++	+/−	+	−	++	−
	RL	+/++/−	−	+/−	−	+++	+++	+	++	−/+	+++	−/+	+++	+++	−/+	+/−	+/−	++	−
	RL	+/++/−	−	+/−	−	+++	+++	+	++	−/+	+++	−†	+++	+++	+/−	+	−‡	++	−
7	RS	+++/++	+	++	−	+++	+++	+	++	+	+++	−†	+++	+++	++/+	++	−	+++	−
	RS	+++/++*	+/−	+	−	+++	+++	+	++	+	+++	−/+	+++	+++	+/++	+++	−	+++	−
	RM	+++/++*	−/+	++	−	+++	+++	++	++	−/+*	+++	−†	+++	+++	++/+*	+++	+	+++	+
8	OS	++/+/−*	−	++/+	−	+++	+++	+/−	++	−/+*	+++	−†	+++	+++	+/−	++	−	+++	+/−
	RM	++/+	−	++	−	+++	+++	+/−	++	−/+*	+++	+/−	+++	+++	−†	+	−	+++	+
Compared with controls		No/↓	↓	↓/No	↓/No	No	No	↓	No	↓/No	No	↓	No	No	No/↓	↓	↓	No	↓

*  Indicates staining variability among orbital (higher staining intensity) and global (lower staining intensity) layers. Staining intensity: +++, strong; ++, moderate; +, weak; −, negative; +/−, the majority are weakly labeled; −/+, the majority are negative.

†  +/−/++ indicates the staining intensity in order of fiber abundance: the largest number of fibers showed weak staining (+), followed by fibers that were unstained (−) and the lowest number of fibers showing moderate staining (++).

‡  Residual stainings in endoneurium.

 ALS, amyotrophic lateral sclerosis; bv, blood vessels; c, capillaries; e, endoneurium; fc, fiber contour; No, no change in immunoreactivity; p, perineurium; OS, obliquus superior; RL, rectus lateralis; RM, rectus medialis; RS, rectus superior; ↓, decreased immunoreactivity.

Anti-Lnα2.

The BM of muscle fibers generally showed decreased immunoreactivity to anti-Lnα2 to different extents in all muscle sections from all donors, with the exception of donor 7, whose fiber contours showed being as strong and even as much immunoreactivity as the controls (Fig. 1A2; Table 2). The BM of muscle fibers was usually moderately to weakly stained, and lack of staining was also observed either entirely or partially around the fiber surface in each cross section (Fig. 1A3). The BM of most hypertrophic fibers was unstained, whereas the BM of adjacent normal size fibers was moderately to strongly stained (Fig. 1A4). In general, the BM of muscle fibers in the orbital layer displayed higher immunoreactivity than those in the global layer in 10 ALS samples (Fig. 1A5; Table 3), whereas there was no difference between layers in the controls, as previously described. ²⁷ Two rectus superior (RS) from donors 1 and 3 and one rectus lateralis (RL) from donor 4 showed exceptionally weak staining and unstained fiber BMs in the global layer. The BM of these fibers was also unstained or weakly stained with anti-Lnβ2 (see below). 

Immunoreactivity with anti-Lnα2 was not detectable in the vast majority of NMJs of ALS EOMs (Table 4; Fig. 2A3–A4), whereas the NMJs of old controls were labeled with this antibody, as previously described ²⁷ (Fig. 2A1–A2). The positive synaptic sites were generally found in the orbital layer and showed only thin labeling confined to the postsynaptic regions. However, four out of eight positive NMJs in oblique superior (OS) in donor 5 were also labeled presynaptically as normally seen in controls (not shown). In general, NMJs of hypertrophic fibers were not labeled by anti-Lnα2 or anti-Lnβ2, but they were strongly labeled by anti-Lnα4 and anti-Lnα5 (Fig. 2E). Most of these hypertrophic fibers labeled neither slow nor fast myosin heavy chains and were therefore classified as EOM fibers. ³⁹ To determine whether the EOM fibers lacking Lnα2 in their BMs were denervated, we investigated their reactivity with an antibody against N-CAM. EOM fibers lacking Lnα2 were in general not labeled with anti-N-CAM, and fibers labeled with anti-N-CAM antibody usually showed anti-Lnα2 labeling in their BMs (not shown). Anti-Lnα2 staining was generally not present in the perineurium and was weak in the endoneurium in the intramuscular nerves of the majority of ALS specimens whereas it was present in the peri- and endoneurium of normal EOMs (Fig. 1A). 

Table 4.

 

Percentage of NMJs Showing Positive Staining with Antilaminin Chain Antibodies

	Donor	Muscle	Lna2	Lna4	Lna5	Lnb2
EOM	1	OS	13.6% (22)	81% (21)	53.6% (28)	34.6% (26)
	2	RL	15.3% (98)	46.4% (61)	57.9% (57)	7.6% (79)
	3	RS	28.2% (131)	93.8% (81)	81.7% (60)	4.7% (85)
		RL	16.7% (6)	100% (7)	80% (10)	0% (8)
	4	RL	5.8% (86)	80% (30)	38.5% (78)	9.3% (43)
	5	OS	38.1% (21)	95.5% (22)	73.3% (30)	46.1% (13)
		RS	46.2% (13)	100% (3)	88.9% (9)	87.5% (18)
	7	RS	75% (8)	75% (8)	71.4% (7)	20% (5)
		RM	25.7% (35)	80% (10)	68.8% (16)	65.4% (26)
	8	RM	25% (68)	84.9% (86)	72.3% (112)	5.6% (36)
		OS	18.7% (16)	85.7% (21)	63.6% (22)	12.3% (14)
	Mean ± SD		28.0 ± 19.3% (504)	83.9 ± 15.1% (350)	68.2 ± 14.2% (429)	26.7 ± 28.6% (353)
Limb	1	Biceps	40% (10)	0% (11)	0% (5)	44.4% (9)
	3	Biceps	70.6% (17)	43.5% (23)	65% (20)	50% (12)
		Biceps	75% (4)	28.6% (7)	100% (5)	0% (8)
	8	Vastus	50% (2)	0% (3)	—	0% (8)
	Mean ± SD		58.9 ± 16.7% (33)	18.0 ± 21.7% (44)	55.0 ± 50.7% (30)	23.6 ± 27.3% (37)

 Total number of NMJs analyzed in each muscle samples is given in parentheses.

 EOM, extraocular membrane; NMJ, neuromuscular junction; OS, obliquus superior; RL, rectus lateralis; RM, rectus medialis; RS, rectus superior; SD, standard deviation.

Figure 2.

 

Immunoreactivity with antibodies against Lnα2 (A, E1), Lnα4 (B, E2, F), Lnα5 (C), and Lnβ2 (D, E3) at NMJs from EOMs of old normal controls (top panel) and ALS donors (three lower panels). Note the presence of Lnα4 but absence of Lnα2 and Lnβ2 in the NMJs of ALS donors. Arrows indicate positive staining; arrowheads indicate absence of staining. (F1–F3) Types of labeling with anti-Lnα4 found in NMJs of ALS donors. (F4) Anti-Lnα4 immunostaining in the BM of terminal Schwann cell.

Anti-Lnα4.

Immunostaining with anti-Lnα4 was not detected in the extrasynaptic BM of muscle fibers from ALS EOMs (Fig. 1B2–B3). In contrast, strong labeling with anti-Lnα4 was found in the vast majority of NMJs in all ALS EOMs (Fig. 2B, 2F), with the exception of donor 2/RL, where weakly labeled or unlabeled NMJs accounted for more than half of the NMJs examined (Fig. 2B3–B4; Table 4). Strong staining with anti-Lnα4 was found frequently in NMJs where immunostaining with anti-Lnα2 was not observed, in donor 3/RS+RL and donor 4/RL (Table 4; Fig. 2E). The localization of Lnα4 labeling varied in different NMJs and between donors, in contrast to the constant staining pattern seen with anti-Lnα5 antibody at NMJs (see below). Thus, we examined five additional EOM specimens with 213 new NMJs, in addition to the 350 NMJs presented in Table 3, to fully understand the distribution of laminin α4 chain at NMJs in ALS. Three distinct staining patterns with anti-Lnα4 were identified after double labeling with α-bungarotoxin in cross sections. Lnα4 was present (1) at both pre- and postsynaptic sites in the majority of NMJs (Fig. 2F1), as also typically seen in controls, (2) only at presynaptic sites (Fig. 2F2), and the least common (3) only at postsynaptic sites (Fig. 2F3). It was often seen that the BMs of terminal Schwann cells and axons in the presynaptic region were strongly stained with anti-Lnα4 in the first case (Fig. 2F4). Staining with anti-Lnα4 was strong in both peri- and endoneurium of intramuscular nerve bundles (Fig. 1B2). However, less stained peri- or endoneurium plus disrupted BM of peri- and endoneurium were occasionally encountered (Fig. 1B3). The blood vessel staining was weak or moderately, whereas the capillaries were, in general, moderately stained by anti-Lnα4 (Fig. 1B2). 

Anti-Lnα5.

In general, anti-Lnα5 did not label the BM of the majority of the fibers of ALS EOMs except for four muscle specimens from three donors (D3/OS, two of D7/RS, and D8/RM) in which the BM of muscle fibers was weakly stained (Fig. 1C2–C3). Staining of adjacent muscle sections with an antibody against Lutheran protein, a transmembrane receptor for Lnα5, showed an identical distribution pattern for both Lnα5 and its receptor; that is, Lutheran protein was present only in areas containing Lnα5 (Fig. 3). 

Figure 3.

 

Immunoreactivity with Lnα5 and Lutheran protein in the EOMs of ALS donors. Note the synchronized presence (A, B) and absence (C, D) of Lnα5 and its receptor Lutheran protein in the areas depicted.

Anti-Lnα5 staining was frequently observed overlapping α-bungarotoxin labeling at the NMJs. On average, 68.2% of the synaptic sites were strongly to weakly stained by anti-Lnα5 antibody, and the staining was confined to the postsynaptic sites (Fig. 2C3–C4). The endoneurium was generally unstained by anti-Lnα5 antibody, but weak staining could occasionally be observed. Blood vessels and capillaries were strongly stained by the anti-Lnα5 antibody, as they were in controls. 

Anti-Lnβ2.

The staining pattern found with anti-Lnβ2 was similar to that seen with anti-Lnα2. The extrasynaptic BM was either weakly stained or unstained (Fig. 1D2–D3) except for the D7 donor in which moderate staining was occasionally present. Immunoreactivity with anti-Lnβ2 was not detectable in the majority of NMJs (73.3%; Table 4; Fig. 2D). The positive NMJs were stained only on the postsynaptic regions. Lnβ2 immunoreactivity was generally decreased in the nerves of ALS EOMs, but high-intensity staining was occasionally seen in the perineurium of some muscles (Fig. 1D2). Intramuscular blood vessels were generally moderately to strongly stained with this antibody, whereas the capillaries were weakly stained or unstained (Fig. 1D2, D3). 

EOMs from Mouse ALS Models

To investigate whether the decreased immunoreactivity for the different laminin chains in the BM of human ALS EOMs was a true trait of the disease and not related to the fact that it was autopsy material, we conducted a similar study on the EOMs from two different transgenic ALS mouse models and their wild-type controls. As described above for the human specimens, the BM of EOMs from 130-day-old SOD1^G93A and 465-day-old SOD1^D90A mice showed declined and disrupted immunostaining with both anti-Lnα2 and -Lnβ2 and were completely devoid of staining for Lnα2 or Lnβ2 chains in some fiber BMs in the global layer (Fig. 4B, 4D, 4F). In contrast, the BM of the muscle fibers in the EOMs of wild-type mice showed strong and smooth staining patterns in both orbital and global layers, irrespective of age (Fig. 4A, 4C, 4E). Lnα4 was weakly expressed in the fiber BM in the orbital layer and very weakly expressed, if at all, in the fiber BM in the global layer of EOMs from 415-day-old wild-type mice (Fig. 4G). In contrast, Lnα4 was completely undetectable in the BM of both EOM layers in the 465-day-old SOD1^D90A mice (Fig. 4H). 

Figure 4.

 

Lnα2, Lnα4, and Lnβ2 chain immunoreactivities in terminal stage SOD1 transgenic (SOD1^G93A and SOD1^D90A) and age-matched wild-type (WT) mice. Notice that the BMs of the EOMs of transgenic mice (B, D, F, H) showed disrupted and/or absent immunostaining with antibodies against Lnα2 (B, F) and Lnβ2 (D), and loss of immunoreactivity with anti-Lnα4 in the orbital area (H).

Human Limb Muscles

The BM of the muscle fibers in old and younger adult control limb muscles showed similar Lnα2 and Lnβ2 immunoreactivity, but the fiber contours in old limb muscles were not always as smooth as those of younger adults (Fig. 5A2). The BM of muscle fibers was unstained by both anti-Lnα4 and anti-Lnα5 (Fig. 5B2, 5C2), although weakly stained fiber contours with the anti-Lnα5 antibody were occasionally observed in four out of five old controls (Fig. 5C1). In control limb muscles sections, there were very low numbers of NMJs compared with NMJs in the EOMs due to the lower density in limb muscles. Only two out of 16 NMJs (12.5%) were labeled with the mAb against Lnα2, whereas the rest were either unstained or very weakly stained (Fig. 6). In contrast, the majority of NMJs were labeled with antibodies against Lnα4 (71.4%), Lnα5 (82.1%), and Lnβ2 chains (63.0%). In intramuscular nerve bundles, the BM of perineurium was stained strongly by antibodies against Lnα4 and Lnα5, weakly by anti-Lnβ2, and unstained by anti-Lnα2, whereas the BM of endoneurium was strongly labeled by anti-Lnα4, moderately to weakly by anti-Lnα2, but unlabeled by antibodies against Lnα5 and Lnβ2. Blood vessels and capillaries were strongly labeled by anti-Lnα5, moderately labeled by anti-Lnα4, weakly by anti-Lnβ2, and unlabeled by anti-Lnα2. One control subject (male, 83 years old) differed from the other normal subjects and lacked immunoreactivity to mAbs against Lnα2 and Lnα5 around the BM of most muscle fibers (not shown). 

Figure 5.

 

Immunoreactivity with antibodies against Lnα2 (A, E), Lnα4 (B), Lnα5 (C, G), and Lnβ2 (D, F) in the BM of limb muscle fibers from adult and old controls and ALS donors. Note heterogeneity in immunoreactivity with Lnα2 (A3, A4) and Lnβ2 (D3, D4) in the fiber BM of ALS limb.

Figure 6.

 

Immunoreactivity with antibodies against Lnα2 (A), Lnα4 (B), Lnα5 (C), and Lnβ2 (D) at NMJs from old normal controls and ALS donors. Arrows indicate positive staining; arrowheads indicate lack of staining.

In ALS limb muscles, anti-Lnα2 labeling was mostly undetectable in the fiber BMs in donor A (Fig. 5A3), whereas in other donors, both smooth and disrupted staining patterns were seen around the muscle fibers in different areas of the muscle sections (Fig. 5A4). In three out of seven ALS limb muscle samples, strong and thick staining was frequently observed in the interstitial space between muscle fibers or in proximity to muscle fiber contours, the latter often with a spotted pattern (not shown). Lnα4 and Lnα5 chain immunostaining was not detected in all donors, yet the novel expression of Lnα4 was clearly observed in the BMs of areas with very small muscle fibers (Fig. 5B3), indicating that these muscle fibers were regenerating. The immaturity of these very small muscle fibers was confirmed by strong labeling observed with mAb against MyHCemb (not shown). The BM of these small fibers was also stained with anti-Lnα2 and -Lnβ2 but not with anti-Lnα5 (Fig. 5E-G). The fiber contours in all donors were weakly labeled with mAb against Lnβ2 chain (Fig. 5D3) except donor A, in which the fiber BMs were mostly unstained (Fig. 5D4). In ALS donors, 58.9% (33 total) and 55% (30 total) of NMJs were labeled with mAbs against Lnα2 and Lnα5 chains, respectively. A dramatic decrease in the population of NMJs labeled with mAbs against Lnα4 (44 total, 18%) and Lnβ2 (37 total, 23.6%) was observed. The perineurium was strongly stained by anti-Lnα4 and -Lnα5, weakly stained by anti-Lnβ2, but unstained by anti-Lnα2, whereas the endoneurium was strongly stained by anti-Lnα4 only (Fig 5). The blood vessels and capillaries were stained with anti-Lnα4, -Lnα5, and -Lnβ2 strongly to weakly and unlabeled with anti-Lnα2, as in controls. 

Discussion

The present study is the first to provide detailed information on the distribution of laminin chains α2, α4, α5, and β2 in the BM of muscle fibers, and their NMJs, and in the intramuscular nerve bundles as well as in blood vessels and capillaries, of human extraocular and limb muscles with ALS. Because the laminins are major components of all these BMs, the present study provides a thorough survey of the impact of ALS in these two muscle types. We found that the majority of EOMs of ALS donors showed a rather normal appearance and composition of BMs around muscle fibers, NMJs, nerves, and blood vessels. However, they were not totally spared, and a number of nonnegligible pathologic alterations were encountered: (1) Lnα2 and Lnβ2 were absent or partially absent from the BMs in a variable number of muscle fibers in most of the ALS EOMs; (2) three EOMs from three donors showed a dramatic decrease in the level of Lnα2 and Lnβ2 in the BM of their muscle fibers and NMJs; (3) Lnα4 was present in the majority of NMJs in EOMs but absent in the majority of NMJs in limb muscle of ALS donors; (4) decreased levels of Lnα2 and Lnβ2 and disrupted BMs with Lnα4 were noted in the peri- and endoneurium of the intramuscular nerves in ALS donors. 

A major concern when evaluating the laminin compositions of ALS muscles was to distinguish between the impact of the disease itself and the impact of aging, as all but one ALS donors in our study were over 60 years old. For the following reasons, aging alone seems unlikely to explain the present findings in ALS muscles. First, different patterns of immunoreactivity with anti-laminin chain antibodies were observed in EOMs of old controls and ALS donors. The laminin chain compositions of the different BMs in old control EOMs were similar to those of younger adults except for the absence of Lnα4 around the muscle fibers and, in some cases, also lack of Lnα2 and Lnα5 in the BM of some muscle fibers. Second, the alterations in laminin chain distribution were observed to some extent in the majority of donors irrespective of their age, which span between 58 and 80 years old. However, the most remarkable decreases in Lnα2 and Lnβ2 were found in three donors (donor 1/RS, donor 3/RS, and donor 4/RL) out of 24 EOMs. Third, there were no differences in laminin composition between wild-type control mice of different ages. In particular, the levels of Lnα2 and Lnβ2 were similar in both 133- and 415-day-old control mice. 

Laminins are the major noncollagenous components of the extracellular matrix where they are expressed in spatially and temporally regulated patterns and play crucial roles for tissue differentiation, maintenance, and regeneration. In muscle tissue, laminin 211 is the typical isoform present in the BM of muscle fibers, where it is essential to provide muscle fiber integrity, and in the BM of the Schwann cells of the peripheral nerves, where it is essential to maintain Schwann cell-axon interactions and to promote neurite outgrowth in case of nerve injury. ^{21,43 –45} In addition to laminin 211, normal EOMs also have laminins 411, 511, 421, and 521 in the BM of their muscle fibers, a feature that we have previously proposed to be instrumental in the selective sparing of the EOMs in merosin-deficient congenital muscle dystrophy. ^27,46 Here we showed that in normal aging there was a decrease or absence of Lnα4 in the EOMs, without any signs of BM disruption. Therefore, we proposed that Lnα4 is not indispensable for the maintenance of muscle fiber integrity in the human EOMs. The absence of Lnα4 in the ALS specimens can be attributed to the aging process, although we cannot exclude that ALS per se may speed up the elimination of this laminin chains in the BM of muscle fibers in the EOMs. 

The majority of ALS EOMs showed a decrease in Lnα5 without or with only a mild decrease in Lnα2 and Lnβ2, in the BMs of muscle fibers. However, three out of 24 ALS EOMs showed a marked decreased level of Lnα2 and Lnβ2 in addition to the loss of Lnα4 and Lnα5. In other words, the BM of these three ALS EOMs lacked α chains, and Lnβ1 and Lnγ1 were the only two remaining laminin chains, which are the ubiquitous chains in most muscles. ⁴⁷ Even though Lnα2 was no longer present around some EOM fibers in ALS, most of these fibers appeared well preserved and generally innervated since (1) they did not contain N-CAM, which is a marker for fiber denervation, (2) they were labeled with α-bungarotoxin, and (3) their NMJs contained Lnα4 and Lnα5 as normal controls. The fact that Lnα2 chain was also absent or decreased in the BM of EOM muscle fibers in the two mouse models of ALS suggests that downregulation of this laminin chain is truly related to ALS. Both transgenic mouse models overexpress mutated human SOD1, but the D90A mice survive for more than 400 days whereas the G93A mice have a very aggressive form of ALS and die at approximately 130∼150 days. ^2,48 Moreover, the Lnα2 chain is the only chain common to the BM of muscle fibers, their NMJs, and the nerves. Thereby, ALS appears to have impact on both the nerves and the muscle fibers. 

The present findings on the BM of muscle fibers correlated well with our previous report on altered myosin heavy chain composition and cellular architecture of these EOMs in ALS. ¹⁴ We have previously observed areas of fatty replacement in donor 1/RS; grouped hypertrophic fibers, scattered atrophic fibers, and areas filled with nerve trunks in donor 3/RS; and presence of fibrosis and central nuclei in donor 4/RL. In addition, different EOMs from the same donor showed different levels of involvement in the disease. In donor 4, for example, the RL showed extensive replacement of muscle fibers by connective tissue and central nuclei, whereas the RM from the same eye was less affected. ¹⁴ Accordingly, the RM from donor 4 showed less alterations in laminin chain composition than the RL. Thus, changes in BM laminin composition generally correlated to altered morphology and signs of disease involvement. There was no obvious difference between the donors with a SOD1 mutation and those without, although our series comprise only two cases of FALS (D2, D3) where both EOM and limb samples were available. 

Lnα2, Lnα4, Lnα5, and Lnβ2 are normally present in the synaptic BM. ^27,49,50 Lnα2 and Lnα4 are present in the BM of both muscle fibers (primary cleft and junction folds) and of Schwann cells at NMJs, whereas Lnα5 and Lnβ2 are present only in the BM of muscle fibers (primary cleft and junction fold) at NMJs. The laminin chains are secreted by the respective cells underlying the BM. The present study showed a decrease of Lnα2 and Lnβ2 and maintenance of Lnα4 and Lnα5 at NMJs in ALS EOMs. To determine whether the reduction of Lnα2 and Lnβ2 at NMJs was a consequence of denervation, we correlated the present findings with preliminary data on the distribution of neurofilament and synaptophysin proteins that are markers of nerve branches and presynaptic vesicles at NMJs, respectively (Liu and Pedrosa-Domellöf in preparation). Because the vast majority of the NMJs was strongly labeled by anti-neurofilament and anti-synaptophysin antibodies, the majority of the NMJs lacking Lnα2 and Lnβ2 chains are most likely physically innervated. These results indicate that muscle fiber–dependent changes in the BM composition occur before denervation and therefore indirectly support the so-called dying back principle. ^{51 –53} Data from SOD1^G93A transgenic mice showed that peripheral denervation occurs before the animals become clinically weak or motor neuron loss is detected centrally. ⁵²  

The most striking difference between EOM and limb in ALS donors was the decrease of Lnα4 in the NMJs of limb muscle, whereas this isoform was not affected in the NMJs of the EOM in terminal stages of ALS. Differences in synaptic composition may be of importance for prolonged survival of NMJs and thereby influence the impact of ALS on the EOMs. 

Footnotes

 Supported by the Swedish Research Council (K2010-62x20399-04-02), Åke-Wibergs Stiftelsen, the Swedish Brain Research Foundation, KMA, the Swedish Medical Society, the Swedish Association for the Neurologically Disabled (NHR), the Kempe Foundation, the Hållstens Research Foundation, and the Björklund Foundation for ALS Research.

Footnotes

 Disclosure: J.-X. Liu, None; T. Brännström, None; P.M. Andersen, None; F. Pedrosa-Domellöf, None

The authors thank the patients and their families for their participation in this project; Madeleine Durbeej and Takako Sasaki for kindly providing antibodies; and Margaretha Enerstedt and Ulla-Stina Spetz for their excellent technical assistance.