Could muscle deformity in children with spastic cerebral palsy be related to an impairment of muscle growth and altered adaptation?


Dr Martin Gough at One Small Step Gait Analysis Laboratory, Guy’s Hospital, Guy’s and St Thomas’ NHS Foundation Trust, Great Maze Pond, London SE1 9RT, UK. E-mail:


Skeletal muscle deformity is common in children with spastic cerebral palsy (CP), but the underlying mechanisms are unclear. This review explores some possible factors which may influence the development of muscle deformity in CP. Normal muscle function and growth appear to depend on the interaction of neuronal, endocrinal, nutritional, and mechanical factors, and also on the development of an appropriate balance between muscle protein synthesis and degradation, and between the development of contractile and non-contractile components. In this context, the changes seen in muscle in children with CP are reviewed and discussed. It is suggested that the development of muscle deformity in children with CP may be related to a multifactorial impairment of muscle growth, on which adaptation of the extracellular matrix due to altered loading may be imposed.

What this paper adds

  •  Skeletal muscle growth appears to be influenced by neuronal, endocrinal, nutritional, and mechanical factors.
  •  Skeletal muscle growth appears to involve a balance between the development of contractile and non-contractile components.
  •  Muscle deformity in children with spastic CP may represent an impairment of muscle growth.

Why do children with spastic cerebral palsy (CP) develop muscle deformity? We generally associate the development of deformity with the presence of spasticity and a consequent resistance of muscle to passive stretch, but the aetiology may be more complex. Children with CP show an early reduction in gastrocnemius muscle belly size,1 an alteration in gene expression within the flexor carpi ulnaris,2 an increase in sarcomere length, and an alteration in the extracellular matrix of semitendinosus.3 It is likely that the development of muscle deformity in spastic CP is related to a number of factors, but without detailed evidence are we able to speculate as to how it may occur? In this review, the factors influencing protein turnover in skeletal muscle are discussed, followed by an outline of the development of muscle innervation and of muscle fibre growth. Muscle changes in spastic CP are then discussed, following which some possible mechanisms which lead to the development of muscle deformity in CP are suggested.

Neuronal innervation, mechanical demands, paracrine and endocrine factors, and nutrition influence skeletal muscle protein synthesis and degradation, which in turn determine muscle fibre size and metabolism.4–6 The contractile proteins in the sarcomere, myosin and actin, are supported by larger proteins such as titin and nebulin,7 which connect the actin and myosin, via the Z-disc and costameres, to the basal lamina and hence, via the endomysium, perimysium, and epimysium, to the tendon. Muscle fibres thus exist in a three-dimensional connective tissue framework which provides support, allows transmission of force, and appears to monitor muscle activity levels, particularly at the level of the Z-disc.8,9 The balance between the development of muscle fibres and of the extracellular matrix in muscle appears to be determined by the interaction of neuronal, mechanical, and growth factors. Myostatin, for example, inhibits muscle fibre growth but appears to promote both the development of the connective tissue framework and muscle cell repair,10 whereas insulin-like growth factor 1, which is released in a paracrine fashion by the muscle when active, promotes protein synthesis within the muscle fibre through pathways involving mammalian target of rapamycin (mTOR).11,12 Protein synthesis may also be stimulated directly by ingested essential amino acids, particularly leucine, acting through mTOR.13 This may be inhibited by AMP-activated protein kinase during periods of starvation or stress.14 The mitochondrial content of the sarcomere, and its utilization of energy substrates such as glucose or fat, is influenced by pathways within the cell which appear sensitive to neuronal activation such as the calcineurin–nuclear factor of activated T cells system.15,16 These pathways also appear to be important in determining the type of myosin heavy chain that is expressed, which will influence the contraction speed of the muscle.16,17 Disuse of muscle, such as that seen during cast immobilization, will lead to a reduction in protein synthesis and a shift towards protein degradation, whereby muscle fibres are tagged for proteolysis by ubiquitin through the action of myostatin.5 Muscle function and muscle metabolism are linked: cast immobilization of the human quadriceps results not only in a reduction in protein synthesis within 48 hours,18 but also in a reduction of the normal protein synthesis response of muscle to ingested amino acids.19 Muscle protein synthesis is suppressed during active contraction, but there appears to be a subsequent increase in protein synthesis.20 Stretch without contraction does not appear to lead to protein synthesis in the human gastrocnemius.21 Muscle cells in culture show a prolonged inhibition of protein synthesis after mechanical stretch without concomitant activation,22 and repeated passive stretch appears to promote muscle atrophy in the rat soleus.23 Muscle growth and function are influenced particularly by neuronal activation: following neurotomy24,25 or denervation26 in a neonatal mouse model of brachial plexus palsy, impaired growth of the involved muscles with replacement of muscle fibres by fat and fibrous tissue was noted. The importance of muscle innervation is also seen in adults with chronic spinal cord injury, who show a reduction in muscle volume and muscle cross-sectional area, an increase in muscle fat content, and a reduction in muscle oxidative capacity, which can be reversed to an extent by electrical stimulation.27

When does muscle innervation develop? Embryonic muscle fibres become innervated, soon after they are formed, by motor neurons which grow out from the anterior horn of the spinal cord and which are activated by the developing intrinsic networks in the spinal cord.28 Unlike the situation in mature muscle, each skeletal muscle fibre is initially innervated by a number of alpha motor neurons.29 In a similar manner, sensory fibres from muscle are initially greater in number than in mature muscle and make extensive connections with intrinsic neurons in the cord and with alpha motor neurons.30 The axons of the corticospinal tract reach the spinal cord in humans between 17 and 29 weeks following conception and subsequently innervate the cord between 31 and 35 weeks following conception.30 The ingrowth of the corticospinal tract appears to promote the development of cord networks with selective loss of some neuronal connections and facilitation and growth of other connections. This in turn appears to lead to the development and enhancement of the motor and sensory connections between the spinal cord and skeletal muscle. Competition between individual motor neurons as a result of the increased innervation appears to lead to the development of mononeuronal innervation, with a single larger motor neuron innervating each muscle fibre. There is also a reduction in the number of sensory fibres from muscle and in the number of connections they make with the spinal cord.30–33

There is limited evidence on the early development of muscle fibre types in humans, but studies in mice suggest a sequential expression of myosin heavy chain isoforms following birth with downregulation of developmental (embryonic and perinatal) isoforms and upregulation and stabilization of adult isoforms.34 Maturation of the rat soleus, with the associated development of a predominantly slow phenotype, appears to be associated with the activation pattern imposed by the motor neuron. Eken et al.33 studied motor unit activity in rat soleus muscles and found that the mean duration of muscle activity increased from 3.4 seconds 7 days postnatally to 62 seconds in adults, with most of this increase occurring after the age of 3 weeks. They suggested that postnatal development of tonic firing in soleus motor units depends on the concurrent appearance of monoamine-dependent plateau potentials in motor neurons due to interaction between the developing corticospinal tract and intrinsic cord networks. The neuromuscular junction also undergoes development, with a gradual change in the rat from immature, low-threshold calcium channels associated with the acetylcholine receptor to faster, higher threshold channels in the mature muscle.35 These immature channels may be re-expressed after denervation.36

Antenatal and perinatal muscle fibre growth is considerable. The sartorius muscle fibres in humans appear to double in diameter between midgestation and term,37 and an accelerated growth in overall muscle fibre diameter in humans has been noted between 35 weeks’ gestation and term.38 This early marked growth of muscle fibres is dependent on an appropriate supply of amino acids, which in pigs may be most marked in late gestation and in the perinatal period39 when muscle may be particularly sensitive to growth and nutritional factors. Muscle protein synthesis in neonatal pigs is associated with an increased sensitivity to growth hormones, insulin, and amino acids, which decreases with age.40 In rats, undernutrition before weaning has been shown to cause permanent stunting of muscle growth, whereas the effect of undernutrition subsequent to weaning can generally be reversed.41 A relative inhibition of skeletal muscle growth in the fetus in a sheep model of late placental insufficiency/intrauterine growth retardation has been noted, suggesting that skeletal muscle growth may be preferentially downregulated if early nutrition is compromised.42

We generally think of muscle fibre growth in terms of length, but muscle fibres also grow in diameter. Oertel et al.43 studied post-mortem sections of the human vastus lateralis and deltoid and noted an increase in mean fibre diameter from 10 to 12μm shortly after birth to 40 to 60μm between the ages of 15 and 20 years. Lexell et al.44 noted that the mean fibre diameter of the human vastus lateralis in post-mortem specimens increases more than twofold between the ages of 5 and 20 years, and found that this was closely associated with the similar increase noted in muscle cross-sectional area. Binzoni et al.45 noted a continued increase in the pennation angle of the human medial gastrocnemius during growth, which they felt reflected muscle fibre hypertrophy. Longitudinal growth of the mouse gastrocnemius muscle belly occurs predominantly through muscle fibre hypertrophy rather than through increased fibre length,46 but the situation in humans appears more complex. Benard et al.47 used ultrasound to study the growth of the medial gastrocnemius in children aged between 5 and 12 years. They found that longitudinal growth of muscle fibres does occur, but because of the pennate nature of the gastrocnemius, longitudinal muscle fibre growth accounts for only 20% of the longitudinal growth of the gastrocnemius medialis belly: the remaining 80% was related to the increase in diameter of the muscle fibres. The contribution of muscle fibre growth in length and in diameter to longitudinal growth and increased volume of the muscle belly will vary depending on the morphology of the muscle. It is also important to consider the growth of the non-contractile components of skeletal muscle. Development of the contractile component of muscle is accompanied by development of the connective tissue framework and by the organized arrangement of fibres within the muscle: it seems that myostatin is effectively suppressed during perinatal and early postnatal muscle growth in rats, possibly to enhance development of the contractile component of muscle.48 Muscle fibre development and growth is thus likely to reflect the interplay of a number of factors including neuronal, nutritional, and hormonal factors and the initial and subsequent pattern of muscle use. The mouse models of neonatal brachial plexus palsy noted above24–26 suggest that neuronal activation of muscle may be particularly important in muscle fibre growth in the young animal.

What do we know about muscle changes in spastic CP? Barrett and Lichtwark49 systematically reviewed studies on muscle morphology and structure in children with CP and noted marked fibre size variability, conflicting reports on muscle length changes, and variable fibre type predominance. They noted that the most consistent change evident in the muscles of children with CP was reduced size. As noted above, this may occur earlier than expected: a reduction in volume and in physiological cross-sectional area has been noted in the medial gastrocnemius in children with CP between 2 and 5 years of age.1 This occurred without significant changes in muscle fibre length but was associated with a reduction in passive dorsiflexion, thus suggesting a reduction in muscle fibre diameter leading to reduced longitudinal growth of the muscle belly. Reduced muscle fibre diameter has been noted in the semitendinosus muscle of children with CP.3 The muscles of children with CP have reduced strength because of impaired muscle activation and cocontraction of antagonist muscles,50 and they may have prolonged activation and relaxation times.51 They have also been noted to have an increased collagen52 and fat content.53

These alterations in morphology and function are associated with changes in muscle structure and function at the cellular level: the recent studies by Smith et al.2,3 show particular promise in developing our understanding of the changes seen in the muscle of children with CP. They assessed the transcriptional profile of the flexor carpi ulnaris in six children with CP and compared it with that of two typically-developing children.2 They noted upregulation of insulin-like growth factor 1 and myostatin and decreased oxidative metabolic gene transcription, despite a paradoxical increase in the slow fibre pathway genes. They also noted downregulation of fatty acid metabolism and transport. Embryonic myosin heavy chain expression was upregulated, as was expression of dystrophin and extracellular matrix collagen components. Smith et al.3 subsequently assessed biopsies from the semitendinosus and gracilis muscles in 17 individuals with CP and 14 typically-developing comparison children. They found a markedly reduced muscle fibre diameter and a reduced muscle cross-sectional area. Sarcomere lengths of the individuals with CP were increased at rest compared with those of the comparison individuals: this was more marked with increasing deformity and with a greater limitation of function. The involved muscles were also stiffer owing to increased collagen content.

Henderson et al.54 proposed that the osteopenia seen in children with CP does not represent a loss of bone but instead may represent a failure of normal growth of bone. In a similar manner, the development of deformity may represent an impairment of growth and altered development of skeletal muscle. The variability in muscle morphology noted by Barrett and Lichtwark49 and the different fibre length and pennation angle changes seen in adjacent muscles by Moreau et al.55 suggest that the process is influenced by individual muscle morphology. The variable levels of myosin heavy chain proteins expressed in the same muscle (flexor carpi ulnaris) noted by Ponten et al.,56 with a shift towards a slower phenotype seen in children with mild CP and towards a faster phenotype in children with more severe CP, suggest that the level of neurological involvement is also important. Is it possible to speculate on how this may occur?

Children born preterm may have impaired muscle growth.57,58 An initial selective impairment of muscle growth due to nutritional and endocrine factors, as has been noted in sheep models,42 could be exacerbated in children with CP by the reduction in corticospinal input to the developing spinal cord networks in CP, which could impair the subsequent development of muscle innervation. Reduced corticospinal input to the developing spinal cord may lead to persistent polyneuronal innervation of muscle, impaired development of motor neurons, and impaired development of the neuromuscular junction.59 It may also lead to a reduced ability to refine afferent fibre terminations, which together with reduced development of inhibitory interneurons30 may explain the clinical finding of spasticity. Impaired early development of muscle fibres, in size and number, may reduce the fibres available for subsequent hypertrophy, as has been noted in the rat.41 Impairment of neuronal input to muscle in the early postnatal period, as a result of diminished corticospinal tract and cord development, may impair the development and growth of the slow muscle fibre phenotype, which may limit the ability of muscle to perform sustained contraction that is necessary for the development of posture and ambulation. The metabolic capacity of muscle will also be influenced by an altered innervation pattern: reduced activation of pathways such as the calcineurin–nuclear factor of activated T cells system, because of reduced muscle innervation, may result in reduced oxidative capacity and a reduction in the ability to use fat as a source of energy as has been noted in muscles in children with CP.2 Neurotomy or denervation in the mouse model24–26 has been shown to lead to reduced longitudinal growth of the muscle belly, which is associated with loss of passive elbow extension and with replacement of muscle by fat and fibrous tissue. Muscle overactivity does not appear to be a feature of spasticity, in which reduced muscle activation has been reported,60 but prolonged muscle contraction may occur in children with severe dystonic CP. This may result in suppression of protein synthesis20 and hence impair muscle growth.

Subsequent adaptive changes in the muscle of children with CP may reflect a combination of altered development and altered loading. Innervation appears to be important in developing the integrity of the costameres, and hence the sarcolemma, through its influence on agrin formation:61 the upregulation of dystrophin, which has been described in CP,2 may reflect a compensation for reduced agrin. Loading, particularly eccentric loading, may lead to enhancement of the connective tissue network of muscle and possible suppression of muscle fibre growth through the action of myostatin.10 Expression of the genes responsible for collagen turnover in the tendons of children with CP appears altered and suggests an adaptation to higher levels of stress.62 Muscle injury due to altered loading may lead to fibrosis and the replacement of muscle fibres by fat.63 The increased intramuscular fat noted in children with CP53 may thus reflect a combination of altered innervation, altered metabolism, and a response to injury. It is possible that overuse may be the main cause of fibrosis and fatty change in the muscles of ambulant children, while reduced innervation may be more important in non-ambulant children. A shift in balance towards development of the connective tissue framework of muscle and away from muscle fibre growth may explain the increased sarcomere lengths noted in the semitendinosus muscle in children with CP.3 A muscle fibre with sarcomeres under stretch at rest may function best at a shorter length: this may explain the contribution of the hamstring muscles to increased knee flexion. Although sarcomere lengths in the gastrocnemius have not, as yet, been measured in children with CP, increased sarcomere length at rest in these muscles could contribute to the development of an equinus gait.

Intervention may also affect muscle growth and adaptation in children with CP. Denervation with botulinum toxin may reduce muscle fibre growth and promote fibrosis or fatty change, as has been noted in animal models,25,64 and may also adversely affect the metabolic capacity of the muscle. Immobilization or prolonged stretch of muscle may suppress muscle protein synthesis,22 promote muscle fibre atrophy,23 and lead to an altered metabolic capacity and reduced sensitivity of the muscle to insulin and amino acids18,19 and impaired glucose tolerance in the limb.65 Growing muscle in animals appears particularly susceptible to the effects of denervation66 or immobilization.67 Strengthening has been shown to improve gastrocnemius muscle volume in children with CP,68 but its effect on the oxidative metabolic capacity of the muscle may be at least as significant in terms of muscle function. Surgery to lengthen a tendon may allow sarcomeres to work at an optimal length without compromising joint position, and by reducing the abnormal load on sarcomeres may shift the balance away from the development of the extracellular matrix towards that of muscle fibres. Recurrence of deformity is likely, however, if the underlying muscle growth continues to be impaired.

The concept of skeletal muscle deformity arising in children with spastic CP because of a combination of impairment of muscle growth and subsequent altered muscle adaptation is speculative but appears to fit the limited available evidence. It is a simplification of what is a complex interactive process but offers a starting point that attempts to integrate our understanding of normal muscle growth and development with our limited knowledge of the changes seen in children with CP. Recent developments in our understanding of normal muscle function and growth offer considerable potential for clinicians working with children with CP: close interaction between clinicians and scientists may be the most effective way of developing an understanding of and possible new therapeutic pathways for this important clinical problem.