Contractile proteins exist as a number of isoforms that show a developmental and tissue-specific pattern of expression. Using gene-specific cDNA probes, the expression of the sarcomeric myosin heavy chain (MHC) multi-gene family and of cardiac (foetal) α -actin was analysed in three different rat hindlimb muscles immobilised for 5 days in either the shortened or lengthened positions. For each of the MHC genes normally expressed in adult muscle (slow, DA and IIB), the effect of disuse alone (immobilisation in the shortened position) upon expression was markedly different to that of passive stretch (immobilisation in the lengthened position) in each of the three muscles. However, the same adult sarcomeric myosin heavy chain gene can be affected in a different, or even opposite, manner by either disuse or passive stretch depending on the muscle in which it is being expressed. The fast IIB MHC gene, for example, exhibits a rapid induction in the slow postural soleus muscle, in response to disuse but no such induction occurs in the faster plantaris and gastrocnemius muscles. Furthermore, the induction of this gene in the soleus was prevented by passive stretch. The MHC gene, normally only expressed in embryonic skeletal muscle, showed a similar response in all three muscles and was reinduced in adult muscle in response to passive stretch but not by disuse alone. In contrast, the isoform of α -actin which is normally only present in significant quantities in embryonic skeletal muscle and which is reduced postnatally, is not reinduced by passive stretch but is reduced still further by immobilisation in the shortened position.
Striated muscle contractile proteins generally exist as a number of closely related isoforms that are expressed in a tightly regulated developmental and tissue-specific manner. For many contractile proteins post-transcriptional processing contributes significantly to isoform diversity (Breitbart et al. 1985; Ruiz-Opazo and Nadalginard, 1987). However, this is not the case for the myosin heavy chains (MHCs) or the α -actins, which are major structural components of the thick and thin filaments, respectively. Different isoforms of both of these proteins in striated muscle are products of multigene families (Buckingham, 1985).
In some cases, a contractile protein isoform found in one adult striated muscle accumulates only transiently during the development of another muscle. This is exemplified by a nonphosphorylatable myosin light chain, which seems to be expressed at early developmental stages in some tissues, such as ventricles and skeletal muscles, but in the adult animal is only expressed in the atria (Cummins et al. 1980; Whalen and Sell, 1980). In a similar manner, the two striated muscle α -actin (cardiac and skeletal) transcripts exhibit different developmental patterns of accumulation. The cardiac-actin transcript has been shown to accumulate in large amounts in mouse and rat foetal skeletal muscle but declines postnatally and accounts for less than 5 % of actin message in the adult (Buckingham, 1985). Conversely, in cardiac muscle, the skeletal α -actin transcript is present at high levels in the foetus but at very low levels in the adult. A similar situation in the rat is observed with the ventricular α -MHC isoform which is identical to, and coded for by the same gene as, skeletal muscle slow (type 1) MHC (Lompre et al. 1984). The slow α -MHC is the most abundant isoform present in late fetal ventricular tissue, but at birth α - MHC increases and is the predominant isoform throughout perinatal and adult life. In slow postural skeletal muscles, slow (α) MHC accumulates postnatally such that by several months postpartum it is the major isoform. The order of expression of MHC isoforms during skeletal muscle development is generally considered to be embryonic → neonatal → adult. However, it appears that the slow (β) MHC can be expressed at lower levels embryonically or postnatally (Whalen, 1985). This pattern of developmentally regulated expression is not, however, without exceptions. In the extraocular muscles of the rat, the embryonic and neonatal MHCs are expressed in the adult (Wieczorek, 1985). Furthermore in other striated muscle tissues it has been observed that, in the adult, contractile proteins that are normally only expressed at significant levels during early developmental stages can be reexpressed in response to altered functional or endocrine environments. In the adult rat, the β MHC gene in the ventricle and embryonic and neonatal MHC genes in skeletal muscle are reinduced in response to hypothyroidism (lzumo et al. 1986). However, it does not appear that foetal isoforms of other contractile proteins are similarly affected by altered thyroid hormone levels (Izumo et al. 1988). It has also been observed that, in adult skeletal muscle, embryonic MHC is produced in type IIA fibres in response to denervation and paralysis (Schiaffino et al. 1988). This suggests that the expression of these isoforms is not exclusively under the regulation of a temporal developmental programme.
Most interestingly in the adult ventricular muscle MHC, skeletal α -actin and tropomyosin genes are all re-expressed during pressure overload-induced hypertrophy where it is thought that mechanical stretch may be the stimulus for the hypertrophic response and the associated reinduction of a ‘foetal programme’ (Lompre et al. 1979; Izumo et al. 1988). This reexpression of proteins normally only expressed at an earlier stage of ventricular development appears not to be confined to contractile protein isoforms, as the atrial natriuretic factor gene is also reinduced in this tissue in response to overload (Izumo et al. 1988).
Passive stretch in skeletal muscle has been shown to produce a similar hypertrophic response caused by elevated rates of protein synthesis (Loughna et al. 1986). The aim of this study was to examine the effects of passive stretch and reduced levels of activity upon expression of MHC genes coding for ‘adult’ and ‘developmental’ isoforms as well as upon the fetal (cardiac) α - actin isoform in fast and slow adult skeletal muscles.
Materials and methods
Animals and experimental protocol
Three month-old male Sprague Dawley rats were divided into two groups of five animals. In the first group, one hindlimb was immobilised using glass fiber casting tape in a fully plantar-flexed position; in the second group, one hindlimb was immobilised in a dorsi-flexed position. After 5 days, the three muscles of the plantar group, the slow postural soleus, fast phasic plantaris and the mixed gastrocnemius, were removed from the immobilised limbs in both groups and also from the contralateral limbs that acted as controls. In the plantar-flexed limb the three muscles of the plantar group were immobilised at less than resting length whereas in the dorsi-flexed limb, these muscles were immobilised at greater than resting length.
RNA preparation and SJ nuclease mapping
Total cellular RNA was extracted from pooled soleus, gastrocnemius and plantaris muscles by a modification of the hot phenol procedure (Soeiro et al. 1966). Sl nuclease mapping was carried out as previously described with 20µg of total cellular RNA and 150 units of Sl nuclease per assay (Lompre et al. 1984). After digestion with nuclease Sl, the products were fractionated on 6 % polyacrylamide/8.3 M urea sequencing gels. The protected labelled fragments were detected by autoradiography.
Effects of diruse and passive stretch on ‘adult’ MHC gene expression
We examined the pattern of expression of three adult skeletal muscle (slow, IIA, 11B) MHC genes in three rat hindlimb muscles after 5 days immobilisation at less than or greater than resting length. It has previously been demonstrated that the slow (type I) skeletal MHC and the cardiac MHC are encoded by the same gene (Lompre et,al. 1984). To detect changes in slow MHC mRNA levels, we used a 347 nt long fragment of the cDNA clone pCMHC 5, which is specific for cardiac slow skeletal MHC (Fig. 1). This probe contains 3’ untranslated sequences and 43 nt of poly (A) tail and poly (dG) cloning linker at the 3’ end. It was found that the amount of slow MHC mRNA (Fig. 1) is high in the soleus in comparison with both the gastrocnemius and the plantaris. In the soleus, the level of expression of the slow Ml-IC gene is not greatly affected by 5 days of disuse (immobilisation in the shortened position) but expression is much reduced after 5 days of passive stretch (immobilisation in the lengthened position). In contrast, in both the plantaris and the gastrocnemius, the normally low levels of slow MHC mRNA are reduced to levels at which, using these techniques, they are no longer detectable after 5 days of disuse. Passive stretching of these muscles prevents this reduction in the plantaris and causes an accumulation of this mRNA in the gastrocnemius.
The presence of the HA (fast oxidative) MHC mRNA is indicated by a fully protected 360 nt fragment of the clone pMHC40 (Izumo et al. 1986). Full protection of this probe indicates the presence of the HA MHC mRNA which was present in similar amounts in all three control muscles. Due to the high degree of sequence conservation in the coding region of this probe, a number of partially protected fragments were produced as a result of hybridisation with other MHC mRNAs (fast IIB, slow, embryonic). The expression of the II a MHC gene responds to disuse in a very similar manner to that of the slow MHC gene in all three muscles (Fig. 2) in that it has no observable effect upon expression of these genes in the soleus and it reduces mRNA levels in both the plantaris and the gastrocnemius. The effects of passive stretch upon IIA MHC gene expression also showed a similar pattern to its effects upon slow MHC gene expression in these three muscles. When compared to control levels, the amount of IIA MHC mRNA is increased in the gastrocnemius and plantaris but reduced in the soleus.
The levels of the type II B (fast glycolytic) MHC mRNA in the three plantar muscles (Fig. 3A) are shown by the presence of a 304 nt fully protected fragment of the pMHC62 probe and this transcript is present at high levels in both the control gastrocnemius and plantaris muscles but at very low levels in the soleus. The partially protected band at 202 nt represents one or more MHC mRNAs but its significance has not been characterised (Wieczorek et al. 1985). After 5 days of disuse (Fig. 3A) there is a very dramatic induction of the II b MHC gene in the soleus, whereas in both the plantaris and gastrocnemius muscles there was no observable change in the level of expression. Passive stretch causes a significant reduction in the expression of this gene in both of the fast muscles and prevents the accumulation of this transcript observed in the immobilised unstretched soleus. To determine how rapidly disuse produces a significant increase in JIB MHC mRNA levels in the soleus, a time course study was carried out over a 5-day period. Detectable increases in the amount of IIB MHC mRNA were observed after 2 days of the soleus being immobilized in the shortened position (Fig. 38).
Re-expression of the embryonic MHC gene is induced by passive stretch
Myosin isozymes in skeletal muscle follow a sequential transition from embryonic to neonatal to adult isofonn, which in the rat is nonnally completed in the first few weeks of postnatal life (Whalen et al. 1981; Mahadavi et al. 1986). We examined the possibility that the embryonic and/or neonatal MHC genes could be reinduced in skeletal muscle in response to disuse or passive stretch. The probe used for the detection of embryonic MHC mRNA was a 210 nt long fragment of the embryonic cDNA clone pMHC25 and contains 16 nt of poly (G) linker and the codons for amino acids 1805-1870 of the MHC protein. Sl nuclease mapping analysis using this probe yielded a full length protection of the probe in RNA isolated from differentiated L6E9 cells in which the embryonic MHC gene is expressed at high levels (Fig. 4). Embryonic MHC mRNA, detected by a fully protected 194 nt fragment, was observed to be present in all three muscles after immobilisation in the lengthened but not the shortened position. However, the neonatal MHC gene was not induced by either passive stretch or disuse in any of the three musdes (results not shown).
Effects of disuse and passive stretch upon foetal (cardiac) α -actin gene expression in skeletal muscle
The re-expression of the embryonic MHC gene during stretch-induced hypertrophy raises the possibility that this stimulus may produce a coordinate reinduction of contractile protein isoforms normally expressed as the major isoform only in foetal development. To examine this possibility, we examined the expression of the skeletal muscle foetal (cardiac) isoform of α -actin, the major component of the thin filament of the sarcomere and the second most abundant protein in muscle. There are six isoforms of α -actin found in vertebrates of which two are present in striated muscle, namely cardiac α -actin and skeletal α -actin. The coding; sequences of cardiac and skeletal α -actinare highly homologous but their 3’ untranslated sequences are completely divergent (Mayer et al. 1984) and therefore the presence of cardiac α -actin mRNA was detected by a 75 nt long 3’ end fragment of the pAC2 probe, which contains 20 nt of carboxy-terminal coding sequence amd 55 nt of 3’ untranslated portion of the rat cardiac,α -actin mRNA (Izumo et al. 1988). It has been observed previously that cardiac α -actin is present in very small amounts in adult skeletal muscle. However, Fig. 5 shows that there is a considerable difference in the levels of expression between controls of the three muscles examined in this study. The control soleus muscle exhibits, a considerably higher level of expression of this gene than do the other two muscles. In all three muscles, disuse caused a marked reduction in the expression of the cardiac α -actin gene. In fact, in the gastrocnemius and plantaris, levels of this mRNA become undetectable (Fig. 5). Continuous passive stretch over 5 days., although not producing a marked increase of this mRNA, prevents or decreases the reduction in cardiac α -actin mRNA levels that would occur when muscles are immobilised in the shortened position (Fig. 5).
This study demonstrates that the expression levels of members of the sarcomeric MHC multi-gene family are highly susceptible to altered activity levels and mechanical stimuli such as stretch. Genes coding for the ‘adult’ isoforms (type I, type IIA, type JIB) demonstrated a highly tissue specific-response to immobilisation in both the lengthened and shortened position. This is analagous to the similar tissue-specific effects upon gene expression demonstrated in this multigene family in response to altered thyroid hormone levels. This tissuespecific response is most dramatically demonstrated by the IIB MHC gene in this study. This gene shows a rapid induction in the soleus immobilised in the shortened position. Butler-Browne and Whalen (1984) suggest that muscle fibres can be induced to accumulate slow (type I) myosin at any point during development by the action of the nerve but that the embryonic → neonatal → adult fast pathway is nerve independent. While it is open to question as to whether slow myosin at the foetal or newborn stage is nerve dependent, there is considerable evidence that slow myosin accumulation in the adult animal is a nerve-driven process (Jolesz and Sreter, 1981; Narusawa et al. 1987). The results presented here suggest that, at least in adult slow postural muscles such as the soleus, where slow MHC is the most predominant isoform, the normal activity pattern actively inhibits expression of the IIB MHC gene. This hypothesis is supported by the observation that workoverload inhibits IIB MHC gene expression in both fast and slow muscles (Morgan and Loughna, 1989). Furthermore, the pattern of innervation experienced by such muscles may not be the only, or even the major factor, inhibiting the expression of the IIB MHC gene. The fact that the rapid accumulation of IIB MHC transcript in the slow postural soleus during inactivity is prevented by passively stretching the muscle suggests that stretch or tension during the ‘normal’ activity of this muscle could play a part in determining its phenotype by retarding IIB MHC expression. These data do not preclude, however, possible differences in neural activity between the immobilised shortened and the immobilised lengthened soleus from contributing to these differences in IIB MHC mRNA accumulation. We have recently, however, demonstrated that passive stretch also prevents the rapid accumulation of IIB mRNA that occurs in the soleus following denervation (Loughna and Morgan, 1990). The possibility of tension playing an important role in regulating IIB MHC gene expression is also suggested by the observation that, although immobilisation in the shortenened position did not affect IIB mRNA levels in either of the fast twitch muscles, passive stretch did cause a reduced expression of this gene in both. The IIB MHC mRNA is not detectable even in fast rat muscles until 6-7 days postpartum (Russell et al. 1988). During early postnatal growth the rapid elongation rate of the long bones of the limbs stretches the associated musculature causing an increase in sarcomere number. Stretch may therefore play a role during early development in retarding the expression of the type IIB MHC gene.
Tissue-specific responses to immobilisation in shortened and lengthened positions were also observed in type I and type IIA MHC gene expression. Immobilisation in the shortened position over a 5day period had little effect upon expression of these genes in the slow postural soleus but caused an induction of both in the two fast twitch muscles. In contrast stretch caused a reduction in expression of both genes in the soleus though not in the plantaris and gastrocnemius muscles. Alterations in MHC mRNA levels in response to disuse and passive stretch, in the plantaris and gastrocnemius muscles, were similar except for the effect of stretch upon slow MHC mRNA levels in these two muscles (Fig. 1). The increase in slow mRNA levels in the gastrocnemius but not the plantaris in response to stretch may be a result of differences in the proportions of the fibre types between these muscles. Whereas the smaller plantaris muscle has a fairly uniform distribution of mainly fast fibres, Armstrong (1984) has divided, based upon fibre type composition, the gastrocnemius into white, red and mixed regions each of which may respond differently to passive stretch. These tissue-specific responses in expression of ‘adult’ sarcomeric MHC genes measured at the whole muscle do not allow conclusions to be drawn with regard to the responses of discrete fibre populations within a muscle or even whether similar responses occur in the same fibre type within different muscles. In situ hybridisation analysis would be needed to examine this gene expression at this level.
In contrast to the tissue-specific responses of the ‘adult’ MHC genes, in this study, the embryonic MHC gene exhibited a similar pattern of expression in all three muscles. Embryonic MHC mRNA was found to be expressed in these muscles only in response to passive stretch. One possible explanation of this observation is that passive stretch caused damage to muscle fibres leading to their degeneration and subsequent regeneration. However, the degree of stretch employed in this study was within physiological limits and under such circumstances produces little or no observable damage or degeneration of muscle fibres (Holly et al. 1980; Williams et al. 1986). The accumulation of embryonic MHC mRNAs in all three muscles suggests that either it is re-expressed in all fibre types or that it is reexpressed in type IIA fibres alone, as this is the only fibre type present in significant proportions in all three muscles. Furthermore this fibre type has demonstrated an ability to re-express embryonic MHC in adult muscle as a response to denervation (Shiaffino et al. 1988). Interestingly we observed no re-induction of this isoform in response to disuse in this study. However, as Schiaffino et al. (1988) estimate that total neonatal and embryonic MHCs only account for about 1 % of total MHCs in denervated muscle similar changes might have gone undetected in this study. The induction of the embryonic MHC gene in response to passive stretch thus appears to be dissimilar to its re-expression in hypothyroid animals which only occurs in slow muscles (Izumo et al. 1988). Stretch is known to induce hypertrophy in skeletal muscle by stimulating protein synthesis and it is probable that this is the mechanical stimulus that produces the hypertrophic response in ventricular muscle during pressure overload. Such overload-induced ventricular hypertrophy is associated with the re-induction of foetal genes, including those coding for contractile proteins α -MHC, skeletal α -actinand α -tropomyosin, as well as the proto-oncogenes c-myc and c-fos and other proteins such as Atrial naturietic factor (Schwartz et al. 1986; lzumo et al. 1988). The results obtained in this study suggest that stretch may induce a similar re-induction of foetal isoforms in skeletal muscle. Although we did not observe an induction of cardiac CY-actin in response to 5 days of passive stretch, the reduction normally occurring with disuse was totally or partially prevented. A different time course of expression of this gene or a difference in mRNA stability compared to MHC mRNAs could explain why an increase in cardiac CY-actin was not observed after 5 days of passive stretch. Recently Schiaffino etal. (1989) have demonstrated such an abscence of temporal coordination in the accumulation of skeletal CY-actin and α -MHC mRNAs in the heart in response to haemodynamic overload. Further analysis at different time points and with other contractile proteins will be necessary to confirm the possibility that stretch can trigger a re-expression of a ‘developmental programme’ similar to that produced in ventricular muscle in response to haemodynamic overload. However, the observed differences in levels of expression of the cardiac α -actin gene between the slow soleus and the two fast muscles does suggest that there might be some functional control of this gene in skeletal muscle and that it is not purely developmentally regulated. This is supported by the reduction in expression of this gene in all three muscles in response to disuse. The relative amounts of cardiac α -actinmRNA present in the controls of the three muscles studied suggest that this transcript may be mainly or even solely present in the slow muscle fibres. We are currently investigating this question using in situ hybridisation.
The molecular mechanisms by which stretch may modulate levels of mRNAs within muscle fibres are unknown. However, it seems probable that stretch may produce such changes by its effects on cell membrane conformation. The presence of single-stretch-activated ion channels have been detected both in skeletal muscle (Guharay and Sachs, 1984) and endothelial cells (Lansman etal. 1987) and it has been suggested that calcium ions entering the cell through such channels may serve a second messenger function.
This work was supported by grants from the National Institutes of Health and the American Heart Association to B.N.G., and by grants from the British Heart Foundation to P.T.L. and Wellcome trust to G.G. S.I. was supported by a Merck/American College of Cardiology Fellowship award and B.N.G. is an Investigator of the Howard Hughes Medical Institute.