Physiology, structure, and susceptibility to injury of skeletal muscle in mice lacking keratin 19-based and desmin-based intermediate filaments

Richard M. Lovering, Andrea O'Neill, Joaquin M. Muriel, Benjamin L. Prosser, John Strong, Robert J. Bloch


Intermediate filaments, composed of desmin and of keratins, play important roles in linking contractile elements to each other and to the sarcolemma in striated muscle. Our previous results show that the tibialis anterior (TA) muscles of mice lacking keratin 19 (K19) lose costameres, accumulate mitochondria under the sarcolemma, and generate lower specific tension than controls. Here we compare the physiology and morphology of TA muscles of mice lacking K19 with muscles lacking desmin or both proteins [double knockout (DKO)]. K19−/− mice and DKO mice showed a threefold increase in the levels of creatine kinase (CK) in the serum. The absence of desmin caused a larger change in specific tension (−40%) than the absence of K19 (−19%) and played the predominant role in contractile function (−40%) and decreased tolerance to exercise in the DKO muscle. By contrast, the absence of both proteins was required to obtain a significantly greater loss of contractile torque after injury (−48%) compared with wild type (−39%), as well as near-complete disruption of costameres. The DKO muscle also showed a significantly greater misalignment of myofibrils than either mutant alone. In contrast, large subsarcolemmal gaps and extensive accumulation of mitochondria were only seen in K19-null TA muscles, and the absence of both K19 and desmin yielded milder phenotypes. Our results suggest that keratin filaments containing K19- and desmin-based intermediate filaments can play independent, complementary, or antagonistic roles in the physiology and morphology of fast-twitch skeletal muscle.

  • costameres
  • muscular dystrophy
  • myopathy

force generated during the contraction of skeletal muscle cells, or myofibers, is ultimately transmitted to the tendons at either end of the muscle. Sarcomeres are the basic functional unit of contraction and are arranged in series to form myofibrils, many of which are organized in parallel in a myofiber. The end-to-end arrangement of sarcomeres is the basis of traditional models of force transmission, in which contractile force is transmitted from sarcomere to sarcomere along myofibrils, ultimately reaching the myotendinous junction. Much of the force generated by contracting skeletal muscle is transmitted radially, however, between neighboring myofibrils and from the peripheral myofibrils through the plasma membrane to the extracellular matrix and neighboring cells (reviewed in Refs. 46, 52, 70). As a result, a significant amount of force is exerted on the cell membrane and the structures that link it to the underlying contractile apparatus and the extracellular matrix (5, 67). The membrane and its associated cytoskeletal and matrix components are collectively termed the sarcolemma (1), which under most physiological conditions is more than strong enough to withstand the stresses caused by contraction (5).

Several conditions can lead to the exertion of forces that exceed the strength of attachment between contractile structures and the sarcolemma, however. In some muscular dystrophies, proteins like dystrophin, which play a role in these attachments, are missing. This leads to a weakening of the attachments between the peripheral myofibrils and the membrane (12, 20), so that normal contractile forces damage the sarcolemma. Even in healthy muscle, excessive force can be generated during a maximal eccentric (lengthening) contraction. Under these conditions, the contractile strength actually exceeds the maximal isometric tetanic tension (32, 56). In some cases, the resulting damage can lead to necrotic cell death, resembling that seen in some muscular dystrophies.

The sarcolemma of fast-twitch muscle is organized into specialized domains that participate in the radial transmission of force from the contractile apparatus to the extracellular matrix and the tendon. These domains, termed “costameres” (51), are oriented transversely, over the Z disks and M bands of nearby myofibrils, and longitudinally, to form a rectilinear, membrane-bound network comprised of integral membrane proteins (such as dystroglycan, the sarcoglycans, and the Na-K-ATPase), proteins of the extracellular matrix (such as laminin), and proteins of the membrane-associated cytoskeleton (such as dystrophin, syntrophins, spectrin, ankyrin, and vinculin; Refs. 18, 31). Dystrophin and dystrophin-associated proteins (DAPs), including several of the proteins just listed, are enriched at costameres (4, 18, 22, 30, 31, 42, 50, 54, 55, 64). Here, they are thought to play an essential role in transmitting force from the contractile apparatus across the sarcolemma to the extracellular matrix (10, 12, 15, 18, 20, 53), through the ability to bind simultaneously to cytoskeletal filaments emerging from the contractile apparatus and to laminin and other proteins of the muscle basal lamina (10, 19, 20).

The filamentous structures that link superficial myofibrils to costameres include microfilaments and intermediate filaments (IFs). The microfilaments are primarily composed of γ-actin (17), which associates with the sarcolemma, at least in part, through its ability to bind to dystrophin (20, 28, 62). Dystrophin has an actin-binding domain at its NH2 terminus, which suggested that its primary linkage to the contractile apparatus is mediated by actin (47). Contrary to this notion, however, is the observation that the absence of γ-actin has no effect on the coordinated organization of costameres with the underlying contractile apparatus (20). Thus structures other than microfilaments are likely to maintain the links between the myofibrils and the sarcolemma. We have found that dystrophin's actin binding domain can also associate with filaments composed of keratins 8 (K8) and 19 (K19) (65, 71), expressed in adult skeletal muscle (65, 71), suggesting an alternative connection between dystrophin and the underlying contractile apparatus. There is now considerable evidence to suggest that at least two sets of IFs, composed in part of desmin and in part of K19 and K8, help to link the contractile apparatus to costameres at the sarcolemma.

Desmin, a type III IF protein of adult striated muscle, is thought to form the majority of IFs in cardiac and skeletal myofibers and so has been extensively studied by several laboratories since its discovery by Lazarides and his colleagues in 1977 (see Ref. 29) and the generation two decades later of desmin-null mice by Milner et al. (44) and Li et al. (34). Desmin in adult striated muscle forms IFs that surround the contractile apparatus at the level of Z disks and links Z disks to overlying domains of costameres at the sarcolemma (33, 48). Consistent with this, these domains of costameres are selectively lost in fast-twitch muscles of mice lacking desmin (48). A mechanical role for desmin in force transmission is suggested by studies showing that muscles of desmin-null mice generate less force and are more susceptible to damage due to eccentric exercise (34). Studies of other forms of exercise suggest, however, that muscles lacking desmin may be less susceptible to injury (63). Despite the uncertainty about its role in protecting muscle from injury, the finding that desmin in otherwise healthy muscle is disrupted immediately after a large-strain eccentric injury (36) is consistent with its postulated role in force transmission.

Desmin plays a role in linking myofibrils to the sarcolemma (34, 35), but it is not the only IF protein to do so: keratins are also present in striated muscle (65, 66, 71) and are likely to play important roles in morphogenesis and force transmission, as they surround the myofibrils at Z disks and link peripheral myofibrils to costameres at the level of both Z disks and M bands (48). We have identified K8, a type II keratin, and K19, a type I keratin, in mature striated muscle (71). These molecules heterodimerize and then polymerize further to form IFs (14). Our evidence indicates that K8 and K19 are present at costameres, which they can stabilize even when desmin is absent (48). Ultrastructural evidence and studies of mice lacking either of these proteins as a result of homologous recombination (“knockout,” or KO mice) suggest that desmin-based and K19-based IFs are important for the organization of costameres and for the stability of the sarcolemma (7, 24, 48, 63, 66).

Our previous results and those of others have highlighted the distinctive effects on muscle of null mutations for desmin and K19, including differences in the distribution of mitochondria in fast-twitch and slow-twitch or cardiac muscle (43, 66) and differences in the stability of costameres in fast-twitch skeletal muscle (48). However, these studies were performed on mice of different strains, making direct comparisons difficult. The effects of null mutations on contractile force, ability to exercise, and susceptibility to injury have been difficult to compare for the same reason. To address the potential similarities and differences in the roles of desmin-based and K19-based IFs in fast-twitch skeletal muscle, we have bred the desmin-null mutation into the FVB background, in which we have studied K19-null muscles. We also generated double mutants lacking both desmin and K19 [double knockout (DKO)]. Our experiments comparing the structure and function of fast-twitch muscles in these mice support the hypothesis that desmin and keratin filaments play distinct roles in skeletal myofibers.



Male and female FVB mice homozygous for the K19-null mutation (68) were the kind gift of Dr. M. Bishr Omary (now of the Department of Molecular and Integrative Physiology, University of Michigan Medical School, Ann Arbor, MI). The mice were bred and genotyped from tail snips, as described previously (27, 68), with the Nucleon Genomic DNA kit (Tepnel Life Sciences, Manchester, UK). The PCR reaction used the Accuprime PFX kit (Invitrogen, Carlsbad, CA). We also examined 129SVJ mice that are null for desmin (desmin−/−; Ref. 44), the kind gift of Dr. Y. Capetanaki (now of the Bioacademy, Athens, Greece). The genotyping of desmin−/− mice was performed as above, with the following primers: forward: 5′-TGATGTCAGGAGGGCTACA-3′; reverse wild type (WT): 5′-CTCACTTGGCCTTGAGCCTCTG-3′; reverse null: 5′-TCCTCGTGCTTTACGGTATC-3′. Desmin−/− mice in the FVB background were generated by breeding the desmin−/− 129SVJ strain with FVB mice (48), selecting heterozygous offspring, and breeding these with WT FVB mice. This was repeated more than six times. FVB mice homozygous for the desmin−/− genotype were obtained by mating heterozygotes and bred as homozygotes thereafter. DKO mice, lacking desmin and K19, were obtained by breeding K19-null and desmin-null mice, both in the FVB background, and then breeding offspring that were heterozygous for both alleles. Mice null for both proteins obtained in this way were fertile and were used for subsequent breeding.

Mice (males and females) were studied at 3 mo of age. They were anesthetized with isoflurane (2% with an oxygen flow rate of 0.5 l/min) and euthanized by perfusion fixation with 2% paraformaldehyde in buffered saline. Activity of creatine kinase (CK) in serum was assayed as described previously (66, 69). The Institutional Animal Care Committee of the University of Maryland School of Medicine approved all our protocols.

Immunofluorescent labeling.

Tibialis anterior (TA) muscles from perfusion-fixed mice were dissected, snap frozen, cryosectioned, and stained by indirect immunofluorescent methods (48) with chicken antibodies to βI-spectrin, to label costameres (see, e.g., Ref. 71). Rabbit antibodies were to dystrophin and desmin (used at 1:100; Lab Vision, Fremont, CA). Myofibers were also labeled with 5 μg/ml propidium iodide (Sigma-Aldrich, St. Louis, MO) to stain nuclei. Samples were mounted and examined under confocal optics (48, 58) and scored for sarcolemmal organization (57) as normal (score = 3), partially organized [costameres with domains remaining only over Z disks (“Z domains”); score = 2], or extensively disorganized (absence of costameres over large sarcolemmal areas; score = 1) (58). The scores were based on examination of >100 fibers per muscle, from ≥6 muscles from each strain.

Some mice were injected intraperitoneally with Evans blue dye (EBD, Sigma-Aldrich) in buffered saline (1 mg EBD·0.1 ml PBS−1·10 g body mass−1) (38) before muscles were fixed and processed as above. Myofibers with dye-labeled sarcoplasm were quantified in cross sections under fluorescence optics (38, 66). At least 400 fibers in 10 optical fields were assessed in 3 mice; results are expressed as percentage of labeled fibers.

Ultrastructure and morphometry.

Perfusion-fixed TA muscles were removed and incubated overnight in 2% glutaraldehyde, 2 mg/ml tannic acid, 0.2 M cacodylate, pH 7.2, postfixed in 1% osmium tetroxide in 0.5 M acetate buffer, en bloc stained with 1% uranyl acetate, dehydrated, embedded in Araldite-Embed 812 resin (EM Sciences), sectioned for electron microscopy at ∼90 nm, and viewed with a Phillips 201 microscope. Images were taken on Kodak 4489 film and digitally scanned at 600 dpi. Distances between the sarcolemma and the nearest myofibrils and between neighboring myofibrils in the interior of the myofiber were measured with Scion Image (Scion, Frederick, MD) from negative images of longitudinal sections of all myofibers examined. To estimate the amount of mitochondria, micrographs of cross sections were examined blindly from at least three muscles from each genotype. The fraction of cytoplasmic surface of the sarcolemmal membrane covered by mitochondria (at levels of <10%, 10–50%, or >50%) was recorded (see Fig. 2G).

The relative sizes of myofibers were determined from frozen cross sections labeled with anti-βI-spectrin. Images were obtained under confocal optics and imported into CorelDraw (Corel, Ottawa, ON, Canada). The minimal Feret's diameter was measured for at least 100 muscle fibers from each mouse with a digital caliper (8). The number of centrally nucleated fibers (CNFs) was obtained by counting cells from frozen cross sections that were stained with hematoxylin and eosin (H & E) or that had been double labeled with propidium iodide and antibodies to dystrophin.

Muscle force measurements.

Contractile function of isolated TA muscle was measured as described previously (66). We used six FVB mice of each strain [WT, K19−/−, desmin−/−, and desmin−/−/K19−/− (DKO)] to compare contractile characteristics. We attached the TA muscles of each to a load cell and applied single twitches (rectangular pulse, 1 ms) at different muscle lengths to determine the optimal length (resting length, L0, measured with calipers as the distance between the tibial tuberosity and the myotendinous junction). With muscles set at L0, we gradually increased the stimulation frequency to establish a force-frequency relationship. A maximally fused tetanic contraction was obtained at ∼100 Hz (300-ms train duration of 1-ms pulses at a constant current of 5 mA) for all strains studied. We used 150% of the maximal stimulation intensity to induce maximal activation of contraction, P0. With muscles set at L0, fatigue was induced through tetanic contractions (100 Hz for 100 ms) delivered every 2 s for 5 min. Maximal tetanic tension was measured during continuous stimulation and expressed as a percentage of P0 to provide an index of fatigue. The cross-sectional area for each muscle was determined by dividing the mass of the muscle by the product of its optimum fiber length (Lf) and density (16). Lf was determined by multiplying L0 of the TA by 0.6, the ratio of Lf to L0 (9). The density of mammalian skeletal muscle is 1.06 g/cm3 (42a). Muscle output was expressed as specific force (kN/m2), determined by dividing the tension (P0) by the muscle cross-sectional area.

Muscle injury.

Injury induced by large-strain lengthening contractions was performed as described previously (66), with the animal anesthetized with isoflurane and placed in a supine position. The hindlimb was stabilized, and the foot was secured onto a plate, the axis of which was attached to a stepper motor (model T8904, NMB Technologies, Chatsworth, CA) and a potentiometer, to measure range of motion of the ankle. A custom program (Labview version 4.1, National Instruments, Austin, TX) was used to synchronize contractile activation and the onset of ankle rotation. The foot was moved into plantarflexion through a 90° arc of motion (−10° to 80°, with the foot orthogonal to the tibia considered as 0°) at an angular velocity of 900°/s, beginning 200 ms after tetanic stimulation of the TA. The lengthening contraction was repeated 15 times, which resulted in a reproducible injury, which we quantified as loss of torque (38, 60, 66).

Treadmill stress tests.

The treadmill stress tests were adapted from Haubold et al. (24). Before running, mice were acclimated to the treadmill (Columbus Instruments 1055M-Exer 6M, Columbus, OH) by placing them on the stationary treadmill belt for 10 min for 3 consecutive days, followed by gentle running on the belt at 2 m/min for 10 min and then at 5 m/min for 5 min on the following 3 consecutive days. For testing, mice ran on the treadmill inclined 7° at an initial speed of 10 m/min. The speed was then increased by 1.5 m/min every 2 min until the mouse was unable to avoid the shock grid.


Desmin, K19, and costameres.

To compare the roles of desmin and K19 in maintaining costameres, we examined longitudinal cryosections of TA muscles from control, K19−/−, and desmin−/− FVB mice and from FVB mice lacking both K19 and desmin (DKO). The organization of β-spectrin at costameres was more disrupted in muscle lacking desmin (53% fibers with extensive disruption) than in muscle lacking K19 (22%), consistent with previous findings (48, 66). In mice lacking both of these IFs, however, the extent of disruption (78%) was considerably higher (Fig. 1). The differences between each of the individual null strains, and between these mice and DKO mice, were all significant (P < 0.001, χ2-test). These results suggest that both desmin and K19 contribute to the organization of costameres in murine TA muscles.

Fig. 1.

Costameres at the sarcolemma of desmin −/−, K19−/−, and DKO muscles. Top: frozen longitudinal cryosections of tibialis anterior (TA) muscles from wild-type (WT), desmin−/− (Des−/−), keratin 19 (K19)−/−, and double knockout (DKO) muscles [muscles lacking both intermediate filaments (IFs)] were immunofluorescently labeled with pairs of antibodies to a membrane skeletal protein at the sarcolemma (β-spectrin, red) and a protein of the contractile apparatus (α-actinin, green). Regions labeled by both antibodies are shown in yellow in the overlay image. Scale bars, 5 μm. Bottom: costameric structures were scored as 1 (“disrupted”), 2 (“partially disrupted”), or 3 (“normal”), as described in materials and methods. The results show that the normally rectilinear pattern of costameres is disrupted in all muscles from mutant mice.

Ultrastructure and morphology.

Myonuclei in healthy skeletal muscle are restricted to the periphery of the cell, near the sarcolemma; CNFs are an accepted marker of ongoing degeneration and regeneration in adult skeletal muscle (21, 40). The incidence of CNFs (Table 1) was higher in muscles from mice lacking desmin and from DKO mice (4.9% and 4.8%, respectively), compared with control mice (1.2%; P < 0.001). Myofiber size was evaluated by minimal Feret's diameter (8). The differences in fiber size and muscle mass (Table 1) were more variable in muscles from mice lacking desmin compared with control and K19−/− samples, consistent with our counts of CNFs and the idea that muscle fibers in these muscles are undergoing degeneration and regeneration at a low but measurable rate. The morphology seen with the light microscope were indistinguishable in desmin−/− FVB and 129SVJ mice (Table 1), indicating that it is the lack of desmin, and not the strain, that accounts for the changes associated with the null mutation.

View this table:
Table 1.

Properties of skeletal muscle lacking intermediate filament proteins

We also studied TA muscles by thin-section electron microscopy (Fig. 2). As reported by others, the lateral alignment of myofibrils in striated muscles is disturbed in the desmin-null mouse (44, 63). We compared the alignment of myofibrils in TA muscles of K19−/−, desmin−/−, and DKO mice by measuring the lateral displacement of Z disks of one myofibril to those of adjacent myofibrils (Fig. 2F). In healthy WT muscles, the Z disks were in register (Fig. 2A), as indicated by the low mean value (51.8 ± 4.7 nm) for the stagger between them. This was also true for the K19−/− TA muscles (61.0 ± 6.4 nm) (Fig. 2B), which were statistically indistinguishable from WT (P > 0.05). The alignment of Z disks was notably disrupted in the desmin−/− TA muscles (150.0 ± 11.5 nm) (Fig. 2C), in agreement with previous reports (44, 63). In the DKO muscles, however, the displacement of Z disks was significantly greater than in the desmin−/− mouse (368.0 ± 20.0 nm; P < 0.001) (Fig. 2D). Thus desmin and K19 each contribute to the lateral alignment of myofibrils in murine TA muscles, with desmin playing the predominant, but not the only, role.

Fig. 2.

Ultrastructural studies of TA muscles lacking desmin, K19, or both proteins. Micrographs show representative transmission electron microscopic images of longitudinal sections of TA muscles from WT (A), K19−/− (B), desmin−/− (C), and DKO (D) mice. Scale bar in D applies to A–D (0.5 μm). E: compared with WT muscles, all 3 mutants showed a significant increase in the distance between the outermost myofibrils and the sarcolemma (Z line to membrane distance), with the largest in K19−/− muscles. The absence of desmin in the DKO muscle reduces this distance to a value identical to that seen in desmin−/− muscle. F: sarcomere alignment was determined by the horizontal distance from 1 Z line to the Z line in an adjacent myofibril (Z line to Z line distance). The sarcomeres were not significantly displaced in K19−/− compared with WT muscles, but the displacement increased significantly in the desmin−/− and was greatest in DKO muscles. *P < 0.001. G: quantitation of mitochondria under sarcolemma. The fraction of cytoplasmic surface of the sarcolemmal membrane covered by mitochondria (at levels of <10%, 10–50%, or >50%) was recorded. The large increase in subsarcolemmal mitochondria in the K19-null mouse and the smaller increase in the desmin-null mouse are absent in the DKO mouse.

We previously reported (66) that the distance between the Z disks of the most superficial myofibrils and the sarcolemma is significantly greater in K19−/− muscles than in WT muscles (WT = 0.15 ± 0.02 μm; K19−/− = 0.99 ± 0.10 μm). The absence of desmin also leads to an increase in the size of this gap compared with controls (0.39 ± 0.04 μm; P < 0.01), but it is considerably smaller than that seen in the K19−/− TA muscle (P < 0.001) (Fig. 2E). The gap between the Z disk of the superficial myofibrils and the sarcolemma was also present in the DKO muscle (0.46 ± 0.04 μm; P < 0.01), but it was statistically indistinguishable from the value obtained with desmin−/− muscle and significantly smaller than that seen when only K19 was missing. Thus the absence of desmin partially rectifies the much larger subsarcolemmal gap created when K19 alone is missing.

The marked increase in the sarcolemma-to-myofibrillar gap in the K19−/− TA muscle is accompanied by a large increase in the subsarcolemmal accumulation of mitochondria (66). We measured a smaller increase in subsarcolemmal mitochondria in the TA muscles of desmin−/− muscle (Fig. 2G), consistent with the smaller gaps observed. This suggests that at least some of the large increase in the size of the gap in the K19−/− TA muscle was due in part to the accumulation of mitochondria under the sarcolemma, and that this redistribution of mitochondria depends at least in part on the presence of desmin. This is consistent with earlier reports that mitochondria accumulate under the sarcolemma of cardiac and slow-twitch skeletal muscles of desmin−/− mice (44). Remarkably, however, the DKO muscles showed the same low level of mitochondrial accumulation under the sarcolemma as WT muscles, suggesting that keratin- and desmin-based IFs play active, and in some respects antagonistic, roles in the distribution of mitochondria in fast-twitch skeletal muscle.

Whole muscle contractility.

We isolated TA muscles and measured both the contractile force in situ and the specific force of isolated muscles (66). For in situ measurements, we released the TA muscle distally and attached its distal tendon to a load cell (see materials and methods). Resting length (the length at which the muscle generates maximal tension, L0) did not differ significantly among muscles lacking K19, desmin, or both proteins (Table 1). Measurements of specific tension (force) (Fig. 3) showed a decrease in K19−/− TA muscles (21 ± 2.4 g/mm3) compared with controls (25 ± 4.1 g/mm3; P < 0.05) and an even greater decrease in the desmin-null and DKO TA muscles (15 ± 1.9 and 15 ± 2.3 g/mm3, respectively). Compared with controls, these differences in desmin-null and DKO values were significant (P < 0.001). They were also significantly different from K19−/− muscles (P = 0.015), but they were indistinguishable from each other. Desmin, therefore, has a greater effect on specific force that predominates in the DKO.

Fig. 3.

Specific tension is reduced in mice lacking desmin, K19, or both IF proteins. Maximal tetanic tension of TA muscles was measured in situ in WT mice and in mice lacking desmin, K19, or both IF proteins. Data are reported as specific tension (force normalized to cross-sectional area). The absence of desmin in the desmin-null and DKO muscles yields identical changes in specific force. *Significantly different from WT (P < 0.001); †significantly different from WT and K19−/−.

Susceptibility to injury.

We used a model that we previously described for mice (60, 66) to measure isometric torque before and after injury induced by large-strain lengthening contractions (Fig. 4). We found a significant loss of torque in control mice (39 ± 4%), K19−/− mice (40 ± 3%), and desmin−/− mice (35 ± 8%) after injury caused by lengthening contractions, but these differences were not statistically significant (P > 0.05). By contrast, the DKO mice showed a greater loss of torque (48 ± 5%) compared with control mice (P = 0.004). The loss of torque in the DKO mouse was also significantly different compared with each of the individual mutants (P < 0.05). These data parallel the findings of histopathology after contraction-induced injury, as all muscles showed a significant increase in the number of fibers with membrane damage, as evidenced by intracellular EBD, but only the DKO TA muscles showed a greater increase compared with controls (Fig. 5). We conclude that the absence of both K19-based and desmin-based IFs is required to increase susceptibility to large-strain injury.

Fig. 4.

Mice lacking both desmin and K19 are more susceptible to injury. Injury was induced by 15 large-strain lengthening (“eccentric”) contractions through a 90° arc of motion. Maximal torque was measured before and after injury in WT mice and in mice lacking desmin, K19, or both IFs. The individual mutants had little effect on susceptibility to high-strain injury, whereas the DKO mice were more susceptible. *P < 0.005.

Fig. 5.

Membrane stability in mutant muscle fibers before and after injury. Evans blue dye (EBD) was used to evaluate sarcolemmal integrity within the TA muscles of mice that were uninjured or mice that were injured by high-strain lengthening contractions. A and B: cryosections of TA muscles from EBD-injected DKO mice showing EBD and DAPI before (A) and after (B) injury. Sarcolemmal damage is indicated by the presence of myoplasmic EBD. As shown in C, without injury (black bars) the number of EBD-positive (EBD+) fibers was not significantly different from WT in any of the mutant muscles. All animals showed a significant increase (*) in the number of positively labeled fibers after injury (gray bars), but this increase was significantly higher in the DKO muscles (†) compared with all other groups (P < 0.05; n = 3 for all groups).

Muscle fatigue.

We tested the rate of muscle fatigue in situ (66) but observed no significant differences between groups at the end of the 5-min fatigue protocol (P = 0.76). When earlier time points in the fatigue protocol were evaluated, both desmin−/− and DKO muscles showed a significant increase in fatigue compared with K19−/− and WT muscles (Fig. 6; P < 0.001).

Fig. 6.

Muscles lacking desmin fatigue more rapidly. With muscles set at resting length (L0), fatigue was induced through tetanic contractions delivered every 2 s for 5 min. Isolated TA muscles were used to measure the rate of fatigue (0.5 Hz, 200-ms train, × 5 min). A: maximal tetanic tension was measured each minute during continuous stimulation and expressed as % of maximal activation of contraction (P0), to provide an index of fatigue. There was no difference among the groups after the 5-min protocol, but the desmin−/− and DKO muscles had a faster initial rate of decline early during the fatigue test, specifically at the 1 min mark. *P < 0.001. B: representative trace recordings from a WT mouse and a DKO mouse.

We also tested the endurance of the mice on a treadmill. Treadmill speed was gradually increased until the mice, running uphill at an angle of 7°, could no longer avoid the shock grid. WT and K19−/− mice were able to run for ∼15 min (ceased running at 20 and 22 m/min, respectively), but desmin−/− and DKO mice were able to run only for ∼3 min (ceased running at 11 and 13 m/min, respectively). The results suggest that the desmin−/− and DKO mice have significantly lower endurance than the WT or K19−/− mice. The desmin−/− and DKO mice were not significantly different from each other (P > 0.05), but they were both significantly different from the WT and K19−/− mice (P < 0.01), consistent with our measurements of fatigue. Thus desmin, but not K19, plays an important role in endurance and fatigue.


IFs resist mechanical stress (13) and are likely to play important mechanical roles in muscle (4, 11). We have been studying the IFs in skeletal muscle, to learn how they contribute to the subcellular architecture required for contraction and the pathways from the contractile apparatus to the extracellular matrix required for lateral force transmission. Our previous results indicated that K19 played a significant role in mature skeletal muscle (71), but because studies of desmin, the major IF protein of striated muscle, had been performed in a different strain of mouse, we could not reliably compare them. We therefore bred the desmin−/− phenotype into the same FVB strain in which we studied the phenotype of K19−/− muscle and subsequently created double mutants, lacking both proteins, by breeding the individual KO mice. We compared the effects of these mutations on the organization of costameres, muscle morphology and physiology, and the susceptibility to contraction-induced damage. Our results confirm that the elimination of both K19 and desmin in fast-twitch skeletal muscle disrupts the organization of muscle fibers and compromises contractile force and protection from injury. They further suggest that these two sets of IFs play distinct, complementary, or, in some cases, antagonistic roles in muscle structure and function.

Muscles lacking K19 and desmin.

Although the K19−/− mutation produced a distinct phenotype in muscle, many of the changes that occur in the DKO muscle are indistinguishable from those seen in the desmin−/− mutant alone. These include specific tension generated during tetanic contraction, the extent of degeneration and regeneration, as measured by counts of CNFs and variability of fiber diameters, the distance between the most superficial myofibrils and the sarcolemma, and fatigability. By contrast, the K19−/− mutation determined only one of the phenotypes seen in the DKO, the levels of CK in the serum. Our results therefore indicate that desmin, the major IF protein of striated muscle, plays a much more extensive role than K19.

Not surprisingly, several phenotypic changes were significantly greater in the DKO muscle than in either individual mutant. Compared with WT muscle, the disruption of costameres occurs in this order: K19−/− < desmin−/− < DKO (Fig. 1). DKO mice also showed the greatest susceptibility to injury (Fig. 4) and sarcolemmal permeability after injury (Fig. 5) and the greatest lateral misalignment of myofibrils (Fig. 2F). It therefore seems likely that these properties are determined by both desmin and K19.

Remarkably, two phenotypes were less severe in the DKO mouse than in the K19−/− mouse, the distance between the sarcolemma and the nearest myofibril and the extent to which mitochondria accumulated under the sarcolemma. The ability of the desmin−/− mutation to partially rectify or fully reverse the abnormal phenotypes seen in K19−/− muscle is not consistent with a simple model for IFs in which they play passive, biomechanical roles in skeletal muscle.

Rather, our findings suggest a complex mix of roles for desmin-based and K19-based IFs in skeletal muscle. Our results show that, although desmin plays a more predominant role in skeletal muscle than K19, the different IFs formed by these proteins have distinct functions. They have complementary actions affecting some aspects of muscle morphology and physiology, but for others, such as the size of the gap between the outermost myofibrils and the sarcolemma and the number of mitochondria in this gap, their actions appear to be antagonistic. The latter may be explained by differential effects of the null mutations on gene expression or signaling or by distinct and potentially antagonistic interactions of keratin and desmin IFs with motor proteins, such as dynein and kinesin (see, e.g., Refs. 23, 25, 26, 37).

Role of desmin.

Our studies, which examined desmin−/− TA muscles of FVB mice, revealed no significant differences in comparisons with desmin−/− 129SVJ muscles in contractility (data not shown) or in other properties (Table 1). Our results are consistent with a role for desmin in aligning the myofibrils with each other and with linking peripheral myofibrils to costameres at the sarcolemma. Our data confirm earlier morphological studies, which showed that the absence of desmin resulted in misalignment of neighboring myofibrils (2, 34), a larger gap between the sarcolemma and the nearest myofibrils (43), and a significant loss of costameres (48). We also confirm earlier findings that skeletal muscle lacking desmin generates lower specific force than normal muscle (63) and shows less capacity for running uphill on a treadmill (24), but our experiments suggest that desmin−/− muscle is no more susceptible than WT muscle to injury from large-strain lengthening contractions, contrary to a previous report (63).

Although our results and those of Sam et al. (63) suggest that desmin−/− muscle responds to different kinds of lengthening contractions in distinct ways, our results differ, likely for methodological reasons. We examined the function of the entire group of ankle dorsiflexor muscles, in which the TA muscle is predominant, and made specific force measurements of the TA muscle. Sam et al. focused on the extensor digitorum longus (EDL) muscle. Our protocols for eccentric injury also differed: we utilized a few large-strain lengthening contractions, whereas Sam et al. used many small-strain lengthening contractions. WT muscle responds differently to these two procedures (41), so it is not surprising that desmin−/− muscles would, too.

As we modeled our treadmill experiments on those of Haubold et al. (24), the fact that we obtained similar results was to be expected. Like Haubold et al. (24), we find a decrease in running endurance in desmin−/− mice. Here we also show that this impairment is similar in mice lacking both desmin and K19 (Fig. 7). Because IFs can affect respiratory function (43) and both these IFs are present in cardiac as well as skeletal muscle, the cause of impaired endurance is not clear. As none of the TA muscles showed altered fatigability after 5 min of intermittent tetanic stimulation, the mechanisms remain to be elucidated. Other processes, such as respiratory dysfunction or cardiomyopathy, may be involved, but isolated TA muscles from the desmin−/− and DKO mice did indeed show a faster rate of fatigue (Fig. 6), suggesting that changes in skeletal muscle contribute to their poor treadmill performance.

Fig. 7.

Mice lacking desmin fatigue more rapidly in treadmill running. Mice were run on a treadmill starting at 10 m/min. The speed was increased by 1.5 m/min every 2 min, until the mouse was exhausted. The desmin−/− and DKO mice fatigued much more rapidly than the WT or K19−/− mice. *P < 0.01.

Costameres and the sarcolemma.

Costameres are disrupted in skeletal muscles from desmin−/− 129SVJ mice (48). Here we show that they are similarly disrupted in desmin−/− FVB muscle, indeed, more so than in K19−/− muscles, but not as much as when desmin and K19 are both absent (Fig. 1). Our earlier studies indicated that keratins are present at all costameric structures, whereas desmin is present only at those costameric structures overlying Z disks of nearby myofibrils. Our results suggest that most costameric structures are maintained by desmin-based and K19-based IFs acting together, with both sets of filaments stabilizing the structures associated with Z disks and IFs containing K19 stabilizing the structures overlying M bands. The fact that costameric elements are still detectable in a small number of myofibers in DKO muscle suggests that other structures linking costameres to the contractile apparatus are present in muscle. These structures may be composed of γ-actin (3), or other filamentous proteins, including plectin (59) and perhaps other IF proteins (our unpublished results).

The fact that K19-based and desmin-based IFs both organize costameres and link them to the nearby contractile machinery may explain, at least in part, the reduced specific tension observed in muscles lacking desmin or both proteins. Sam et al. (63) suggested that connections between myofibrils are more compliant when desmin is missing, thus lowering the efficiency of force transmission. Although this is likely, other mechanisms cannot be excluded, including more compliant links between the myofibrils and the sarcolemma, as well as more general changes in the myoplasm, such as a redistribution of mitochondria (44, 66) or alterations in signaling or Ca2+ stores. Keratin filaments can anchor signaling molecules (14); as the predominant IF protein in striated muscle, desmin filaments may play a similar role, perhaps through synemin, which copolymerizes with desmin (45) and is an A-kinase anchoring protein (AKAP) (61). Changes in Ca2+ homeostasis, which have been implicated as a factor underlying muscular dystrophies (72), remain to be examined in muscles lacking IF proteins.

Haubold et al. (24) found that mice lacking desmin did not demonstrate elevated serum CK activity (similar to our findings in Fig. 5 and Table 1), even after voluntary running. Furthermore, the serum CK activity after eccentric exercise (forced downhill treadmill running) was not different from that in control mice. The authors therefore suggested that sarcolemmal damage does not play a major role in the impairment of function in muscles lacking desmin. A possible explanation for their results and ours is that desmin is more important for transmission of force and K19 is more important for membrane stability. This notion is consistent with our data on specific force and plasma CK activity, which show that contractility is reduced in K19−/− TA muscles but not as much as in desmin−/− TAs (Fig. 3) and that plasma CK activity is increased threefold in K19−/− mice compared with WT and desmin−/− mice (Table 1). Furthermore, plasma CK levels in DKO mice are not significantly greater than those in K19 −/− mice. After injury, there is a significant increase in the amount of membrane damage when both IFs are missing (Fig. 5), suggesting distinct but partly overlapping roles for desmin and K19 in stabilizing the sarcolemma. Although we have chosen to interpret our CK measurements in terms of the changes caused by the IF null mutations to skeletal muscle, we recognize that plasma CK can also derive from the heart, which we have not yet examined in our mice.


Our results show that fast-twitch muscle of mice lacking the IF proteins, desmin and K19, show altered morphology and compromised contractile force and susceptibility to injury. The absence of both IF proteins has consequences that can be equal to, greater than, or less than the effects of the absence of either protein alone. This diverse set of results suggests that keratin and desmin filaments do not simply serve as alternative structures that stabilize the myoplasm and transmit contractile force. Future experiments will be needed to examine the possible changes in the expression and organization of other structural proteins, including other IF proteins, that are altered in the absence of desmin or K19.


This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grants to R. M. Lovering and R. J. Bloch (K01-AR-053235 and 1R01-AR-059179 to R. M. Lovering; RO1-AR-55928 to R. J. Bloch), by grants from the Muscular Dystrophy Association (no. 4278 to R. M. Lovering; no. 3771 to R. J. Bloch), and by a predoctoral stipend to B. L. Prosser from T32-AR-07592 (Dr. M. Schneider, principal investigator).


No conflicts of interest, financial or otherwise, are declared by the author(s).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 7.
  7. 8.
  8. 9.
  9. 10.
  10. 11.
  11. 12.
  12. 13.
  13. 14.
  14. 15.
  15. 16.
  16. 17.
  17. 18.
  18. 19.
  19. 20.
  20. 21.
  21. 22.
  22. 23.
  23. 24.
  24. 25.
  25. 26.
  26. 27.
  27. 28.
  28. 29.
  29. 30.
  30. 31.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 40.
  39. 41.
  40. 42.
  41. 42a.
  42. 43.
  43. 44.
  44. 45.
  45. 46.
  46. 47.
  47. 48.
  48. 50.
  49. 51.
  50. 52.
  51. 53.
  52. 54.
  53. 55.
  54. 56.
  55. 57.
  56. 58.
  57. 59.
  58. 60.
  59. 61.
  60. 62.
  61. 63.
  62. 64.
  63. 65.
  64. 66.
  65. 67.
  66. 68.
  67. 69.
  68. 70.
  69. 71.
  70. 72.
View Abstract