56 A. Cristea et al. in the growing population of elderly citizens and an improved understanding of the mechanisms underlying the impaired motor function is needed for several reasons (Mahoney et al. 1999). First, the growing population of elderly persons and the consequent enormous social and economic impact of aging-related problems in general, along with motor handicap and dependency in particular, have led to a need for focused research efforts to increase the quality of life in the aged. Second, normal aging processes may have a negative influence on the progression of various neu- romuscular disorders. Aging may also play an important etiological role in some neuromuscular diseases, such as amyotrophic lateral sclerosis, muscular dystro- phies, mitochondrial myopathies and the post-polio syndrome. Third, a significant proportion of frail elderly people have to use all their muscle power even to rise from a chair, and an additional small impairment in muscle function may dramatically change their life from an independent to a dependent one (Bean et al. 2007; Frontera et al. 2008). The cause of falls and fall-related injuries in old age are com- plex and involve multiple risk factors (Nevitt et al. 1991). It is important, however, to separate the different components of the fall, i.e., what initiates the fall and the ability to restore standing balance when it is disturbed; or to safely arrest a fall that may occur when standing balance cannot be recovered. It is widely accepted that somatosensory function has a strong impact on posture and locomotion in humans. The aging-related impairments in somatosensation, vision and vestibular function will, accordingly, have a negative effect on maintaining postural balance. The diffi- culty in recovering from an impending fall which is impaired in old age, especially in old women, is not related to an impairment of the sensory process or to the motor planning that leads to the initiation of muscle contraction (Schultz et al. 1997). The sources of these aging and gender differences have been reported to lie primarily in events after depolarization of skeletal muscle, i.e., in force-generation and contrac- tile speed (Schultz et al. 1997). In support of this, the decline in mobility and lower extremity disability have been reported most influential in predicting falls in the elderly (Welle et al. 1993; Robbins et al. 1989; Mahoney et al. 1994, 1999), and the predictive value of the muscle weakness is increased further when muscle force is measured at a speed of movement resembling more functional limb velocities. It is becoming increasingly evident that the aging-related muscle wasting repre- sent a wasting condition that is different from other types of muscle atrophy, i.e., in response to denervation, microgravity, bed rest, immobilization in plaster or the cachexia associated with cancer, renal failure or chronic obstructive lung disease. This review will focus on aging-related changes at the motor unit level. The motor unit concept was first introduced by Lidell and Sherrington (1925) and in a report 5 years later Eccles and Sherrington (1930) emphasized the importance of the motor unit as being the “final functional unit” that forms the basis of all graded muscle contractions. However, the spatial organization of the muscle fibers belong- ing to a single motor unit, motor unit fibers, was not resolved until the introduction of the single muscle fiber EMG (Stålberg and Ekstedt 1973a) and glycogen- depletion techniques (Edstrom and Kugelberg 1968; Burke et al. 1971; Brandstater and Lambert 1969, 1973; Doyle 1969) almost half a century later. Visual evaluation of spatial organization and enzyme-histochemical properties of motor unit fibers 57Aging-Related Changes Motor Unit Structure and Function indicated that motor unit fibers were randomly distributed within a defined territory, motor unit territory, and that motor unit fibers expressed metabolic homogeneity (Burke and Tsairis 1973; Edstrom and Kugelberg 1968). However, with the intro- duction of quantitative enzyme-histochemical methods and computer assisted detailed analyses of spatial arrangement of motor unit fibers, it was shown that there is inhomogeneity within motor units, but these differences are significantly smaller than the differences observed between different normal motor units from young individuals (Monti et al. 2001; Ansved et al. 1991; Bodine-Fowler et al. 1990; Martin et al. 1988). Further, a systematic difference in the metabolic proper- ties of motor unit fibers along the superficial deep axis within the motor unit terri- tory in the fast-twitch tibialis anterior muscle suggests a biological etiology of this inhomogeneity (Larsson 1992). This also challenges the concept that motoneuron properties, firing frequency and/or trophic factors, are the sole factors controlling muscle fiber properties within single motor units (Larsson 1992). One factor of importance for the inhomogeneity observed along the superficial-deep axis of the tibialis anterior muscle is the impact of baseline cell tension, i.e., tensegrity, on cell structure and function (Ingber 1991, 1993, 1997, 2002a, b) in a large muscle with a complex muscle fiber arrangement such as the rat tibialis anterior muscle. In addi- tion, aging-related changes in tensegrity may have a significant impact on motor unit structure and function. A number of different classifications of motor units into different types have been used, based on biochemical, enzyme-histochemical, contractile and fatigue properties (Burke 1981). However, it is important to emphasize that all motor unit types display considerable variation of e.g. physiological and biochemical proper- ties, representing a continuum rather than discrete units (Burke 1981; Kugelberg and Lindegren 1979). However, there is a need to group units into specific catego- ries to systemize and communicate experimental observations (Henneman et al. 1965; McPhederan et al. 1965; Wuerker et al. 1965; Burke 1981). In this short review, motor units will be classified based on contractile properties into the fast- or slow-twitch types or as types I, IIa, IIx or IIb based on motor unit fiber myosin heavy chain (MyHC) isoform expression. 2 Effects of Aging on Motor Unit Organization In human skeletal muscle, motor unit fiber organization was first resolved using the single fibre EMG technique introduced by Stålberg and Ekstedt (1973b), i.e., elec- trophysiological data supporting previous experimental animal studies using the glycogen depletion technique. The glycogen depletion technique was originally proposed by Krnjevic’ and Miledi (1958) and successfully used first by Edström and Kugelberg (1968), Burke et al. (1971), Brandstater and Lambert (1969, 1973), and Doyle and Mayer (1969). After isolating single motor units by ventral root teasing or intracellular stimulation of the motoneuron soma in the ventral horn, single motor unit contractile and physiological properties were characterized and 58 A. Cristea et al. followed by the depletion of the motor unit fibers of their glycogen by prolonged repetitive stimulation. Serial cross-sections of the muscle were subsequently stained for glycogen, and different enzyme- or immuno-histochemical methods, allowing the identification of the motor unit fibers and characterizing them accord- ing to metabolic properties (mitochondrial enzyme activities), myofibrillar ATPase type, MyHC isoform expression etc. By using the glycogen-depletion method, it was for the first time demonstrated that motor unit fibers were scattered in the muscle, resulting in considerable overlap in territories of different motor units and different motor units being highly intermingled in non-pathological muscle. In response to denervation reinnervation via collateral sprouting motor unit fibers are reorganized, resulting in groups or whole fascicles innervated by a single alpha motoneuron and the fibre type grouping observed in enzyme-histochemically stained muscle cross-sections from patients with chronic denervation and reinner- vation (Kugelberg et al. 1970). During aging, several studies have documented a loss of myelinated neurones in both peripheral nerves and ventral roots in humans as well as in different experimental animal models, although some studies in animals have reported and unaltered number of myelinated neurones during aging (for refs. see Larsson and Ansved 1995). Conflicting results may at least in part be related to methodological differences, but more importantly different muscles and rodent strains are affected differently by aging (Lionikas et al. 2003, 2005a, 2006). The aging-related loss of a-motoneurones is a slow process and the ultimate loss of the neurone is preceded by a gradual loss in motoneuron structure and function, referred to as the “sick” motoneuron (McComas et al. 1971). The decreased protein synthesis in the neuronal cell body and the decreased rate of axonal transport in old age may accordingly be a reflection of this “sickness” (Knox et al. 1989; Komiya 1980; McMartin and O’Connor 1979). The mechanisms underlying the aging-related loss of alpha motoneurons are not known. However, it is interesting to note that non- enzymatic glycosylation (glycation), a post-translational modification regarded as one of the biochemical bases underlying the pathophysiology of aging (Brownlee 1995), has been forwarded as a potential mechanism underlying the progressive loss of motoneurons in patients with amyotrophic lateral sclerosis (Shinpo et al. 2000). In addition, glycation has recently been shown to induce impairment of myosin function, i.e., resembling aging-related changes in myosin function (Ramamurthy et al. 2001). It may therefore be speculated that an aging-related decrease in the prevention of advanced glycation endproduct formation by antioxi- dants, such as glutathione, may provide an aging-related factor of importance in the loss of large motoneurons and sarcopenia, as well as for regulation of muscle con- traction at the motor protein level (Larsson 2003). The aging-related changes in the number, structure and function of alpha motoneurones have significant effects on the spatial organization of both fast- and slow-twitch motor units. An increased innervation ratio, i.e., muscle fibers per motor unit, and expansion of the motor unit territory were observed in the old fast- and slow-twitch motor units over and above the borders of the motor units in the young animals (Larsson et al. 1991a). Kanda and Hashizume (1991) 59Aging-Related Changes Motor Unit Structure and Function reported that the motor unit function in the rat medial gastrocnemius was restored after nerve crush injury even in old age, suggesting that aged motoneurons have a preserved capacity for axonal regeneration and reinnerva- tion. In fact, aged reinnervated motor units of both fast- and slow-twitch types produced very large tensions, indicating that the ability to reinnervate and hold extra fibers was well-maintained. Further, slow-twitch motor units may have an advantage compared with fast-twitch units in innervating previously denervated muscle fibers even in old age and that the largest, most rapidly conducting motoneurons with the lowest oxidative enzyme activity, i.e., innervating fast- twitch muscle fibers, are preferentially lost during aging (see Einsiedel et al. 1992; Einsiedel and Luff 1992; Hashizume et al. 1988; Ishihara et al. 1987; Kanda and Hashizume 1989; Pettigrew and Gardiner 1987). By calculating interfiber and nearest neighbor distances between all motor unit fibers within specific motor units, by using a computer-assisted algorithm where all motor unit fibers are given their x,y-coordinates in the muscle cross-section, significant rearrangements of the spatial distribution was observed, although mechanisms may vary according to muscle type (Ansved et al. 1991; Edstrom and Larsson 1987). For instance, an increased proportion of short distances, as judged from the nearest-neighbor distances, was observed in fast-twitch motor units from the rat tibialis anterior muscle, whereas the interfiber distances did not differ between young and old animals. A trend towards an increased proportion of short distances was also observed in slow-twitch motor units from the rat soleus, according to nearest-neighbor distance analyses, and interfiber distance distribu- tions revealed a greater proportion of short and long distances in the old slow- twitch motor units. These signs of motor unit fiber rearrangements, together with the increased innervation ratio, the increased size of the motor unit territo- ries and a decrease in the total number of muscle fibers (motor units) suggest an aging-related denervation- reinnervation process. The different patterns of fast- and slow-twitch motor unit fiber rearrangements in old rat skeletal muscle indi- cate differences in the type of reinnervation (Ansved et al. 1991). Less grouping of fibers is expected after nodal sprouting than after terminal sprouting, and conceivably more readily detected by interfiber distances. Terminal sprouting, on the other hand, would be expected to cause a change in the distribution of nearest neighbor distances (Kugelberg et al. 1970). In humans, there are electrophysiological evidence of an ongoing denervation- reinnervation process during the aging process reflected by needle-EMG exami- nations showing motor unit potentials with longer duration, greater amplitude and a larger number of phases and satellite potentials in old than is found in young individuals (e.g. Buchtahl and Rosenfalck 1955; Campbell et al. 1973, Carlson et al. 1964). This is supported by more modern electromyographic techniques such as the macro-, scanning- and single fiber-EMG recordings confirming an increased motor unit size and redistribution of motor unit fibers in aging human skeletal muscle (see Cavanagh et al. 1993; Gilchrist et al. 1992; Stalberg and Antoni 1980; Stalberg and Fawcett 1982; Stalberg and Thiele 1975; Thiele and Stalberg 1975). 60 A. Cristea et al. 3 Effects of Aging on Motor Unit Types An aging-related increase in the proportion of slow-twitch (type I) fibers, according to enzyme-histochemical mATPase staining, was originally reported in the lateral portion of the human quadriceps muscle (Larsson et al. 1978, 1979). This observa- tion was originally considered controversial and caused significant scientific debate (Larsson 1983). Three decades later, the fast to slow fiber type transitions appear less controversial and fiber type transitions have been reported in human tibialis anterior, quadriceps and biceps brachi muscles (Scelsi et al. 1980; Caccia et al. 1979; Jakobsson et al. 1988; Larsson et al. 1978; Monemi et al. 1998, 1999; Tomonaga 1977). These observations at the protein level have been confirmed at the gene level, demonstrating a significant aging-related decreases in fast, but not slow, MyHC mRNA levels (Balagopal et al. 2001; Welle et al. 2000). However, aging-related fiber type transitions are muscle specific and a slow-to-fast MyHC isoform transition has been observed in human cranial nerve innervated masticatory muscles (Monemi et al. 1999). It is important to emphasize that the magnitude of these aging-related fiber type transitions is small and the physiological importance may be questioned. In the rat slow-twitch soleus muscle, a fast-to-slow fiber type transition is observed during development and maturation and adult rats have almost 100% slow fibers prior to the aging-related loss of muscle fibers (see Larsson and Ansved 1995). In fast-twitch muscles, aging-related fiber type transi- tions have frequently gone undetected when using conventional enzyme-his- tochemical myofibrillar ATPase stainings (Larsson et al. 1991a, 1993). However, the identification of a third fast myosin isoform in rat skeletal muscle according to immunocytochemistry (Schiaffino et al. 1989) and electrophoretic separation (Termin et al. 1989) named type IIx and IId, respectively, is of specific interest in this context. By using the set of monoclonal antibodies developed by Schiaffino and co-workers (Schiaffino et al. 1989), i.e., type II, IIa and IIb MyHC antibodies, it was possible to identify the IIx MyHC isoform in glycogen-depleted motor unit fibers. It was shown that the type IIx MyHC motor unit is a specific motor unit with spatial arrangement, physiological, biochemical and morphological properties separating it from the type IIa and IIb MyHC motor units (Larsson et al. 1991b). It appears as if the IIx MyHC motor unit plays an integral role in the aging- related motor unit transition in fast-twitch muscles (Larsson et al. 1991a). First, it is the dominating motor unit type in the tibialis anterior muscle in old rats, whereas in young animals it only constitutes a small proportion (Larsson et al. 1991a, b). Second, the reorganization of motor units in old age with the appearance of type IIx motor unit fibers in regions, which in the young animals are restricted to type IIb motor units (Figs. 1 and 2). Third, a large number of the IIx MyHC motor units in the old animals co-expressed IIa and IIx or IIx and IIb MyHCs. However, it is important to the emphasize that the motor unit fibers in the glycogen depleted motor unit fibers were characterized by the monoclonal bodies reactive with (a) all type II MyHCs, (b) the type IIa MyHC, and (c) the type IIb MyHC (Larsson et al. 1991a). This means that the type IIx motor unit fibers are identified as those 61Aging-Related Changes Motor Unit Structure and Function Fig. 1 Camera lucida tracings of glycogen-depleted cross-sections of fast-twitch motor unit fibres from young animals, classified according to their MyHC composition, in 21 cross-sections of tibialis anterior muscle. The superficial part of the muscle is facing the top of the figure. The type IIa, type IIx and IIb MyHC units are identified by red, green and blue filled circles, respec- tively. The horizontal bar represents 1 mm (The graph is modified from Larsson et al. 1991a, b) 62 A. Cristea et al. Fig. 2 Camera lucida tracings of glycogen-depleted cross-sections of fast-twitch motor unit fibres from young animals, classified according to their MyHC composition, in 16 cross-sections of tibialis anterior muscle. The superficial part of the muscle is facing the top of the figure. Motor units includ- ing type IIa, type IIx and IIb MyHC muscle fibers are identified by red, green and blue filled circles, respectively. Motor units containing more than one type of muscle fiber are identified by two filled circles. The horizontal bar represents 1 mm (The graph is modified from Larsson et al. 1991a, b) 63Aging-Related Changes Motor Unit Structure and Function reactive with the general fast MyHC antibody and unreactive with the specific type IIa or IIb MyHC antibodies. Thus, the “hybrid” motor units co-expressing type IIx with another MyHCs, included both muscle fibers expressing the type IIx MyHC and fibers expressing one of the other fast MyHC isoforms and not type IIx fibers co-expressing another fast MyHC isoform. Thus, the MyHC expression in these units are controlled by other mechanisms than the motoneuron properties, although an inhomogeneity of motoneuron properties within different branches of the neuron as a consequence of an aging-related impairment of motoneuron function cannot be completely ruled out. This type of non-uniform MyHC isoform expression among motor unit fibers was never observed in the young animals (Larsson et al. 1991a) (Fig. 1). A further fast-to-slow transformation process resulting in an increased number of type IIa and type I fibers has been reported in old age, but this transfor- mation process appears to be confined to a very old age (Bass et al. 1975; Boreham et al. 1988; Caccia et al. 1979; Ishihara et al. 1987; Kanda and Hashizume 1989; Kovanen and Suominen 1987; Pettigrew and Gardiner 1987). It is therefore sug- gested that the increased number of the IIx MyHC units in fast-twitch muscles in old age reflects an aging-related motor unit transition from type IIb- to IIx, possibly preceding a transformation to types IIa and I, following the sequence IIb ®IIxb ®IIx ®IIxa ®IIa ®I (Gorza 1990; Larsson et al. 1991a). 4 Physiological Properties of Motor Units The muscle wasting associated with old age and the associated decline of maximum contractile force is the most common type of muscle atrophy and impaired muscle function observed in humans, but the rate varies between different individuals, as well as between different muscles in the same individual (Larsson 1982). A more sedentary lifestyle in old age contributes to the impaired muscle function, but can- not account for all the differences in the aging-related loss of muscle force and mass. The underlying mechanisms are complex and involve genetic factors (Lionikas et al. 2005a, b, 2006). Muscle force is proportional to the cross-sectional area and the force generating capacity (maximum force normalized to cross-sectional area, specific force) of the activated muscle tissue, and there is reason to believe that both area and specific force decrease in old age. The loss of alpha motoneu- rons, the incomplete reinnervation of previously denervated muscle fibers, the fol- lowing muscle fiber loss and fiber atrophy play an important role in the aging-related decline in muscle force documented in rodents (e.g. Ansved and Larsson 1989; Arabadjis et al. 1990; Edstrom and Larsson 1987; Larsson and Edstrom 1986). An aging-related total muscle fiber loss has also been reported in humans, based on autopsy sections and measurements of fiber numbers in selected regions of the muscle, followed by extrapolation to the whole-muscle cross-sections (Lexell et al. 1983). The aging-related decrease in cross-sectional fiber area appears to affect muscle fibers of the fast-twitch type preferentially in humans (Larsson et al. 1978, 1979; Scelsi et al. 1980; Tomonaga 1977). In addition to the loss of muscle fiber 64 A. Cristea et al. number and size, an altered spatial organization of motor unit fibers may affect the lateral transmission of forces and overall muscle force (see Monti et al. 2001). Divergent results have been reported with regard to aging-related changes in spe- cific force based on measurements at whole muscle or motor unit levels (see Larsson 2003; Larsson and Ansved 1995). This is, at least in part, due to a number of different factors that obscure measurements of aging-related differences in spe- cific force at the muscle and motor unit levels, such as differences in intramuscular fiber orientation, differences in the mechanical leverage provided by the bony anatomy of the joint, the elasticity of the muscle and the muscle tendons, patterns of motor unit recruitment, and activation of antagonist muscles. Single fiber preparations, on the other hand, allow investigation of the function of myofilament proteins in a cell with an intact filament lattice, but without the confounding effects related to intercellular connective tissue or protein heterogeneity between cells of multicellular preparations. Studies on aging-related changes in specific tension at the single fiber level, using the skinned fiber preparation, have reported an aging-related decline in spe- cific tension varying between 9% and 47% (Frontera et al. 2000; Larsson et al. 1997; Li 1996; Lowe et al. 2001; Thompson and Brown 1999). A decrease in spe- cific force in skinned fibers, where excitation-contraction coupling has been by passed and contractile proteins are activated directly with calcium, could be due to a decrease in the number of cross-bridges in the driving stroke per muscle fiber volume, or to a decrease in the force developed by each cross-bridge, or to a com- bination of both mechanisms. In very old rodents, there is ultrastructural evidence of myofibrillar loss and also of an increase in intermyofibrillar spaces, which is expected to influence the specific tension (Ansved and Edstrom 1991), a preferen- tial loss of myosin has been reported in rodents (Thompson et al. 2006) and similar observations have also been observed in human skeletal muscle (Cristea and Larsson, unpublished observations, Fig. 3), indicating that an altered myofibrillar protein stochiometry may contribute to the decreased specific force in old age in rodents as well as in humans. Lowe and co-workers (Lowe et al. 2001), using elec- tron paramagnetic resonance spectroscopy analyses, have shown aging-related changes in the function of the myosin head in old animals, i.e., a decreased number of myosin heads in the strong-binding structural state. We have recently observed aging-related changes in 3D myonuclear organization with regional accumulation of myonuclei resulting in an increased variablility of myonuclei domain size, but with no or only minimal changes in average myonuclei domain size (Cristea et al. 2010) (Fig. 4). It is hypothesized that the increased variability in myonuclei domain size may have a negative effect on contractile protein synthesis and transport. This may also result in increased post-translational modifications of contractile proteins. It has been known for many years that skeletal muscle generates free radicals and muscle derived reactive oxygen species and nitric oxide derivatives influence regu- lation of muscle contraction and induce posttranslational modifications, and that aging exaggerates these effects (Reid and Durham 2002). Myofibrillar proteins, such as myosin, actin and tropomyosin, are highly sensitive to free radical-mediated oxi- dative stress (Barreiro and Hussain 2010). An aging-related increase in carbonylated 65Aging-Related Changes Motor Unit Structure and Function 2,50 2,00 1,50 1,00 0,50 0,00 YM OM SOM YW OW SOW Fig. 3 Myosin:actin ratio determined on 12% SDS-PAGE gels in different age and gender groups. Subjects are divided into young (Y, 24–35 years, n=12), old (O, 65–83, years, n=12) and very old (SO, 89–96 years, n= 9) men (M) and women (W). According to two-way ANOVA there were significant age (p<0.01) and gender (p<0.001) effects, i.e., lower myosin:actin ratios among women and lower ratios in the old and very old individuals. Values are means ± SE Fig. 4 Confocal microscopy images of myonuclei in muscle fibers expressing the type I MyHC isoform. (a) Young women, age 35 years. (b) Old women, age 78 years. (c) Very old women, age 96 years. The horizontal bar denotes 100 mm . proposed by Krnjevic’ and Miledi (1958) and successfully used first by Edström and Kugelberg (1968), Burke et al. (1971), Brandstater and Lambert (1969, 1973), and Doyle and Mayer (1969). After. the elasticity of the muscle and the muscle tendons, patterns of motor unit recruitment, and activation of antagonist muscles. Single fiber preparations, on the other hand, allow investigation. et al. 1992; Einsiedel and Luff 1992; Hashizume et al. 1988; Ishihara et al. 1987; Kanda and Hashizume 1989; Pettigrew and Gardiner 1987). By calculating interfiber and nearest neighbor distances