Statin-Induced Muscle Necrosis in the Rat: Distribution,Development,and Fibre Selectivity

Abstract

Simvastatin and cerivastatin have been used to investigate the development of statin-induced muscle necrosis in the rat. This was similar for both statins and was treatment-duration dependent, only occurring after 10 days had elapsed even if the dose was increased, and still occurring after this time when dosing was terminated earlier as a result of morbidity. It was then widespread and affected all areas of the muscular system. However, even when myotoxicity was severe, particular individual muscles and some types of fibres within affected muscles were spared consistently. Fibre typing of spared muscles and of acutely necrotic fibres within affected muscles indicated a differential fibre sensitivity to statin-induced muscle necrosis. The fibres showed a necrotic response to statin administration that matched their oxidative/glycolytic metabolic nature: Least sensitive →I ↔ IIA ↔ IID ↔ IIB ← most sensitive. Type I and IIB fibres represent metabolic extremes of a continuum of metabolic properties through the fibre types with type I fibres most oxidative in metabolism and type IIB fibres most glycolytic. In addition, in some (nonnecrotic) glycolytic fibres from muscles showing early multifocal single fibre necrosis the only subcellular alterations present in isolation of any other changes were mitochondrial. These changes were characterised by an increased incidence of vacuolation and the formation of myelinoid vesicular bodies that accumulated in the subsarcolemmal areas. These findings suggest an important early involvement of mitochondria in selective glycolytic muscle fibre necrosis following inhibition of the enzyme HMG-CoA reductase.

Introduction

By inhibiting the key enzyme in cholesterol biosynthesis [3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase] statins are used as an effective treatment for hypercholesterolemia. Enzyme inhibition in the liver, the major site of cholesterol biosynthesis, results in a reduction of plasma cholesterol causing an increased synthesis of hepatic cell surface membrane low density lipoprotein (LDL) receptors. This induces an increased hepatic uptake of plasma LDL with a reduction in circulating levels (Alberts et al., 1980). A rare but important adverse effect in humans associated with statin therapy is myotoxicity, which ranges from mild myopathy to frank rhabdomyolysis (Farmer and Torre-Amione, 2000). Cerivastatin was recently withdrawn from the market because of a significant number of cases of fatal rhabdomyolysis (Baker and Tarnopolsky, 2001), sometimes in combination therapy with gemfibrazil. The histopathological changes in muscle have been described in patients on statins (Chucrallah et al., 1992) and there are a number of publications detailing the histopathological changes in preclinical species including rats (Smith et al., 1991; Reijneveld et al., 1996), rabbits (Nakahara et al., 1992, 1998), and dogs (Walsh et al., 1996). Investigations have been performed over the last two decades in an effort to understand the mechanism of statin-induced muscle necrosis. Attention has focused, among other factors, on the influence of lipophilicity/hydrophilicity of the statin, effects on the muscle surface membrane including electrical current conductance (particularly associated with chloride channels), changes in intracellular ubiquinone concentrations, deficiencies in selenoprotein synthesis, organic anion transporters, and muscle energy metabolism (Smith et al., 1991; Pierno et al., 1992; Sonoda et al., 1994; Pierno et al., 1995; Laaksonen et al., 1996; Nakahara et al., 1998; Moosman and Behl, 2004; Schaefer et al., 2004; Takeda et al., 2004). Although these studies provide some detail on the nature of the muscle necrosis they do not provide a complete picture, and the precise mechanism is still unknown. We wished to contribute to the understanding of statin-induced muscle necrosis by studying the precise histopathology changes in an in vivo model, which could then be use to further elucidate the mechanism. We used the rat as preliminary investigations with the rabbit (unpublished data) showed that the latter species, in our studies, had a relatively high incidence of myositis in untreated animals that might confuse interpretation of any induced changes. This paper attempts to define more clearly, with simvastatin and cerivastatin (chosen as representative hydrophilic and lipophilic statins, respectively) in the female rat, the earliest muscle changes, their distribution and temporal development, and the exact muscle fibre types primarily affected.

Materials and Methods

Test Materials

Simvastatin, purity 98.4%, was obtained from Wuhan S&M Biochemie (Wuhan Hubei, China) and cerivastatin, purity 98.5%, from Cadila Healthcare (Ankleshwar, Gujarat, India). Mevalonic acid lactone was supplied by Sigma Aldrich UK. The statins were formulated for dosing as suspensions in water containing 0.5% hydroxypropyl methylcellulose and 0.1% w/v polysorbate 80. Mevalonic acid lactone was dosed as a solution in the same vehicle.

Animals and Treatments

This study was planned in accordance with the standards of animal care and ethics described in “Guidance on the Operations of the Animals (Scientific Procedures) Act 1986” issued by the U.K. Home Office and was conducted so that any clinical expression of toxicity remained within a moderate severity limit as described in guidelines agreed with the U.K. Home Office Inspector.

Female Wistar Hannover rats, substrain Crl:WI-(Glx/BRL/Han)BR, were obtained from Charles River UK. They were multiple-housed appropriate to each study and were acclimatised for at least 6 days before treatment was started. Animal rooms were illuminated in a 12-hour light/dark cycle and temperature and humidity were controlled within the limits 21 ± 2°C and 55 ± 15%RH. Pelleted RM1 (E) SQC rodent diet and drinking water were freely available. The animals used were within an age range of 6 to 8 weeks at the start of dosing.

Details of dosing and necropsy schedules are included in Table 1. With each statin, 2 separate studies were performed: A preliminary study to select doses for a second study in which the time-course of muscle changes would be defined. The formulated test materials were dosed by oral gavage. Statins were given once daily using a standard dose volume of 5 ml/kg body weight but the daily dose of mevalonic acid was given in 2 equal fractions, the first at about 1 hour before administration of the statin dose and the second at about 6 hours after administration of the statin. The standard fractional dose volume for mevalonic acid was 2.5 ml/kg body weight. The animals were observed and weighed daily and, where necessary, the degree of toxicity expressed in individual animals was contained within predefined limits of severity by withholding dosing when the body weight loss from day 1 exceeded 10%. Dosing was then restarted when the body weight gain showed recovery. Standard plasma chemistry assessments were made during the time-course studies (blood sampling in lithium heparin restricted to time of necropsy) including an evaluation of plasma creatinine kinase (CK) activity (spectrophotometric assay).

Necropsy and Histology

In the preliminary studies, heart, kidneys, liver, stomach, and a range of muscle tissues were sampled from at least 3 rats per group, including all premature decedents. The muscle tissues sampled were biceps femoris, extensor digitorum longus, gastrocnemius, semimembranosus, semitendinosus, soleus, tibialis cranialis, and vastus medialis from the left hind limb; biceps brachii, extensor carpi radialis longus, flexor carpi ulnaris, supraspinatus, triceps brachii caput laterale, and triceps brachii caput longum from the left forelimb; abdominal peritoneal, diaphragm, longissimus lumborum, masseter superficialis, panniculus carnosus (skin), tongue, and trapezius. In the time-course studies, kidneys, liver, stomach, and a selected range of muscle tissues (comprising biceps femoris, extensor digitorum longus, gastrocnemius, soleus, tibialis cranialis from the left hind limb, and longissimus lumborum) were sampled from all animals.

Tissues were fixed in buffered 10% formalin, processed to wax blocks, and then sectioned and stained with haematoxylin and eosin for examination by light microscopy. During the histopathological examination of the muscle sections, necrosis was graded subjectively: Minimal (*)-up to 10 single fibres affected in the whole section; mild (**)-up to approximately 20% of fibres in section affected; moderate (***)-up to 50% of fibres in section affected; severe (****)-more than 50% of fibres in section affected. For the time-course studies samples of biceps femoris, extensor digitorum longus, gastrocnemius, and soleus muscles were also sampled and processed for ultrastructural or histochemical assessment as described next.

Electron Microscopy

Samples for ultrastructural assessment were immersion fixed in 2.5% glutaraldehyde fixative. Glutaraldehyde-fixed samples were postfixed in 1% osmium tetroxide and processed to Araldite resin blocks. Thin, 70–90-nm resin sections were cut and stained using uranyl acetate and lead citrate. Ultrastructural morphology was examined on a Hitachi H7100 transmission electron microscope using a 75 kV accelerating voltage.

Muscle Histochemistry

Muscle samples were trimmed, orientated on a cork disk, and frozen in isopentane (Fisher Scientific) precooled with liquid nitrogen. Serial cryosections of 7-μm thickness were cut from each sample for fibre typing. Sections were stained for mATPase activity following pre-incubation at high and low pH. One section was placed in an incubating solution at pH 9.4 consisting of 0.5% ATP (Sigma) in 0.1 M glycine/NaC1 buffer with 0.75 M CaC1₂ for 45 minutes at 37°C. A further section was pre-incubation in 0.1 M sodium acetate buffer with 10 mM ETDA (pH 4.1–4.3) for 10 minutes at 4°C before placing in the incubation solution noted previously. Following incubation the slides were transferred to 2% CoCl₂ for 5 minutes followed by 30 seconds in 10% ammonium sulphide solution. Sections were washed thoroughly in distilled water between each step. Sections were lightly counterstained with Carazzi’s haematoxylin before being dehydrated, cleared, and mounted in Histomount. Using this method types I, IIC, IIA, IID, and IIB fibres (see results in “fibre typing in spared muscles” for description of fibre types) could be discriminated.

Muscle Immunohistochemistry

For the simvastatin time-course study, further serial cryostat sections were stained for fast and slow myosin heavy chains using antibodies from NovoCastra (NCL-MHCf at 1:10, and NCL-MHCs at 1:20, respectively). The sections were incubated in the primary antibody for 60 minutes, then incubated in the secondary antibody (rabbit anti-mouse HRP conjugate–Dako P0260 at 1:100) for 30 minutes, before being visualised by incubation with 3,3 diaminobenzidine tetrahydrochloride (DAB)(Dako K3468) for 5 minutes. All incubations were at room temperature, and sections were washed thoroughly in tris-buffered saline between each step. Sections were counterstained with Carazzi’s haematoxylin before being dehydrated, cleared, and mounted in Histomount. For the cerivastatin time-course study, 4 μm sections of formalin-fixed, paraffin-embedded muscle were stained for fibre type, using the dual-labelling immunohistochemical technique described by Behan et al. (2002). Dewaxed sections were subjected to 2 minutes’ full pressure in a microwave pressure cooker containing 0.01 M citrate buffer at pH 6.0, and then 5 minutes’ digestion at room temperature by proteinase K (Dako S3020). Endogenous peroxidase activity was blocked by incubation in Dako’s Peroxidase Block (K401111) for 20 minutes, followed by 15 minutes in 20% normal rabbit serum. Mouse monoclonal antibody to slow myosin (Sigma M-8421) at a dilution of 1:2000 was applied for 30 minutes, followed by 30 minutes in peroxidase-conjugated rabbit anti-mouse antibody (Dako P0260) at 1:100. Vector Laboratory’s SG peroxidase substrate kit (SK4700) was then applied for 10 minutes. Following a further 15 minutes in 20% normal rabbit serum, Sigma’s alkaline phosphatase-conjugated mouse monoclonal antibody to fast myosin (A4335) was applied at a dilution of 1:50 for 60 minutes. This was visualised using Vector Red alkaline phosphatase substrate kit (Vector Labs SK5100) for 10 minutes. All incubations were at room temperature, and sections were washed thoroughly in tris-buffered saline between each step. Sections were dehydrated, cleared, and mounted in Histomount.

Results

Clinical Observations

Simvastatin and cerivastatin at high doses resulted in essentially similar clinical observations including reduced body weight gain, body weight loss, and clinical signs of hunched posture, pilo-erection, and cold extremities. Based on the severity of these changes animals in the preliminary studies were excluded from dosing for a period of time or sacrificed as detailed later. For the simvastatin time-course study, the dose level set caused 12/30 animals to express a degree of toxicity requiring suspension of dosing for 1 to 4 days to avoid unscheduled terminations. For the cerivastatin time-course study, body weight loss remained below 10% compared with day 1, and there were no adverse clinical signs, and the study was completed without any suspension of dosing.

Mevalonic acid lactone (MA) was well tolerated when given daily for up to 16 days in 2 equal fractions totalling 60 or 120 mg/kg. When administered with simvastatin and cerivastatin it clearly ameliorated the adverse effect of both on body weight and clinical signs.

Plasma Chemistry

Daily administration of 80 mg/kg simvastatin or 0.5 mg/kg cerivastatin in the time-course studies was associated with moderate to marked increases in plasma CK activities on or after day 12. Animals co-administered simvastatin or cerivastatin and MA showed no changes through the treatment period or at termination. Salient data are shown in Table 2.

General Pathology

In the liver, hepatocyte single cell necrosis, increased mitotic figures and cytoplasmic basophilia occurred following simvastatin administration at 80/60 mg/kg/day. With cerivastatin equivalent changes were present with a dose-related severity in animals dosed 1, 2, and 3, but not 0.5 mg/kg/day. Similar changes were not present in animals dosed with simvastatin or cerivastatin when mevalonic acid lactone was co-administered.

In the nonglandular stomachs of animals dosed 80/60 mg/kg/day simvastatin and 0.5, 1, 2, and 3 mg/kg/day cerivastatin there was a dose- and time-related epithelial hyperplasia with hyperkeratosis and gastritis. These changes were also present in simvastatin- and cerivastatin-dosed rats co-administered mevalonic acid lactone, although the changes were reduced in incidence and severity with cerivastatin.

Animals dosed 1 mg/kg/day cerivastatin for up to 9 days showed myocardial vacuolation and in 1 animal ventricular dilatation and minimal myocardial necrosis. There were no important changes in the hearts of rats dosed 80/60 mg/kg/day simvastatin.

Muscle Necrosis: Temporal Development, Distribution, and Histopathology Control Muscles

The incidence of fibre necrosis in control rats was very low. It was characterised by necrosis of a single or a few isolated individual muscle fibers generally with an associated mononuclear cell infiltration. The soleus was the most commonly affected muscle in control animals. These findings are illustrated in Table 3, which includes findings for control rats in the preliminary simvastatin study. For control rats from the time-course simvastatin study (data not shown) terminated from day 5 to 12, findings were restricted to the soleus muscle from 1 animal where a minimal necrosis was seen. In controls from the cerivastatin time-course study (data not presented) terminated from day 5 to 12 (there were no controls in the preliminary cerivastatin study), findings were restricted to the soleus muscle from 2 rats and the gastrocnemius muscle from a further animal where a minimal necrosis was seen.

Preliminary Studies

All aspects of induced muscle necrosis were remarkably similar between simvastatin and cerivastatin. Table 3 details the changes for animals terminated prematurely following dosing with simvastatin at 80 mg/kg/day (preliminary study). Muscle necrosis was not present in samples from the animal that was terminated on day 7. Up to severe necrosis was present in 3 animals sacrificed between days 13 and 16. The necrosis affected most muscles sampled. However, the soleus muscle from the hind limb, the region of the flexor carpi ulnaris muscle from the fore limb, and the tongue were totally spared. Some sparing was also seen in the masseter muscle and diaphragm. In this study, no muscle necrosis was detected in 4 animals following reduction of the simvastatin dose to 60 mg/kg/day (dosed 80 mg/kg/day simvastatin to day 7, undosed to day 14, then 60 mg/kg/day to day 42) or in 4 rats co-administered simvastatin (as detailed before) and mevalonic acid lactone (data not presented in table).

Table 4 shows findings from the preliminary cerivastatin study. At the doses of 1, 2, and 3 mg/kg/day in animals sacrificed prior to day 10, only minimal muscle fibre necrosis was sporadically detected. At day 9, this minimal change in 4 muscles of 1 rat (106) was of uncertain relationship to administration of cerivastatin, however, none of the other changes could be attributed to dosing as they were consistent with the control incidence and severity. At the lowest dose of 0.5 mg/kg/day all 3 rats (sacrificed on schedule at day 15) showed up to severe necrosis in most muscles sampled. As with simvastatin, there was no necrosis in the soleus muscle or tongue, and the flexor carpi ulnaris muscle and diaphragm were virtually spared.

Time-Course Studies

In the time-course simvastatin study (Table 5) for animals terminated up to day 10, again no muscle fibre necrosis was apparent. From day 12, a variable severity and distribution of muscle necrosis occurred reaching severe in the muscles of some rats. The soleus muscle was spared in all animals showing statin-induced muscle necrosis. In the time-course cerivastatin study (Table 5), 1 rat at day 10 showed a mild necrosis in the gastrocnemius muscle; however, no further important changes were seen in this or other animals at or before this time point. As in the simvastatin study, from day 12, a variable severity and distribution of muscle necrosis occurred reaching severe in the muscles of some rats. The soleus muscle was spared in all animals showing statin-induced muscle necrosis (the minimal necrosis in the soleus of 1 rat was considered spontaneous). No muscle necrosis was seen in 5 rats co-administered cerivastatin and mevalonic acid lactone for 16 days (data not presented in table).

The muscle necrosis when acute was segmental, affected individual fibres, and was characterised by cytoplasmic eosinophilia with loss of cytoplasmic structure, vacuolation, and little or no inflammatory infiltrate (Figure 1a). Subsequently the necrosis was more widespread, and there was infiltration by mononuclear and polymorphonuclear cells, oedema, and vacuolation with fragmentation and loss of cytoplasm. The basement membrane was commonly retained (Figure 1b). In more advanced lesions there was a more profound infiltration by mononuclear cells and some regeneration with myotube formation. Mineralisation of necrotic fibres was also seen.

Ultrastructure

The soleus, EDL, gastrocnemius, and biceps femoris muscles from 5 control rats in the simvastatin study, and the biceps femoris muscles from two control rats in the cerivastatin study were assessed ultrastructurally. Occasional mitochondrial vacuolation with some minor accumulation of concentric myelinoid bodies were present in all muscle and fibre types. They were not accompanied by any further notable ultrastructural changes.

Muscles from simvastatin- and cerivastatin-dosed rats (see Table 5) showing the earliest histological changes of low severity and with only minor or no frank inflammation were selected for further ultrastructural assessment. These included from simvastatin-dosed rats: EDL muscle from animals 49 and 54, gastrocnemius muscle from animals 53 and 58, and biceps femoris muscle from animals 52 and 53, and from cerivastatin-dosed rats; gastrocnemius muscle from animals 38, 43, and 58, the biceps femoris muscle from animals 39, 43, 51, 53, and 54 and the EDL muscles from animals 44 and 60.

A spectrum of changes, consistent between the 2 compounds, were detected in these samples: Changes in excess of those seen in control samples were restricted to glycolytic type II fibres showing few mitochondria and abundant sarcoplasmic glycogen deposits. Two findings that occurred in isolation of any other ultrastructural changes were increased vacuolation of mitochondria and accumulation of aberrant organelles throughout the sarcoplasm but accumulating particularly in the subsarcolemmal regions. These organelles (Figure 2) were variable in size and characterised by accumulation of concentric multilaminated membranous whorls and spheroids, often with retained mitochondrial cristae and occasional lysosome-like dark staining inclusions. Many appeared clearly to be degenerate mitochondria or derived from degenerate mitochondria. These changes were also present in fibres that showed a range of further ultrastructural alterations from simple swelling of sarcoplasmic reticulum to disintegration of contractile proteins.

Ultrastructural assessment of the histologically unaffected soleus from animals showing severe necrosis in other muscles (simvastatin-dosed animals 46, 54, 55, and 57; Table 5) was also undertaken and compared with the control samples. There were no statin-induced ultrastructural changes detectable in these muscles. Muscle samples (gastrocnemius and biceps femoris) from 1 rat dosed continuously with simvastatin to day 12 (animal 45) and 2 rats (43 and 44) dosed continuously to day 16 but showing no simvastatin induced fibre necrosis in any muscle were also examined ultrastructurally. Only the gastrocnemius muscle from animal 44 showed a slight increase in mitochondrial myelinoid inclusions relative to the controls. Similarly, a muscle (biceps femoris) from each of 2 rats (40 and 41) dosed continuously with simvastatin and sacrificed on day 10 (showing no necrosis histologically) showed a slight but notable increase in mitochondrial myelinoid bodies, whereas the gastrocnemius muscle from these rats and both muscles from a further rat (39) showed no changes relative to the controls. Muscles from simvastatin-(biceps femoris and gastrocnemius from animals 37, 38, and 59) and cerivastatin-dosed (biceps femoris and gastrocnemius from animals 31, 36, and 42) rats sacrificed on day 8 only showed ultrastructural changes consistent with control samples.

Fibre Typing in Spared Muscles

The muscle fibre type of mammals is determined by the particular myosin heavy chain (MHC) isoform expressed. The limb muscles of the adult rat express 1 slow and 3 fast MHC isoforms (Schiafino and Reggiani, 1994). Most fibers normally express only 1 isoform and are referred to as “pure,” although fibres are present that contain more than 1 (Staron and Pette, 1993). The pure and mixed fibre types constitute a continuum from the slowest twitch type I fibres to the fastest twitch type IIB (Staron et al., 1999): I ↔ IC ↔ IIC ↔ IIA↔ IIAD ↔ IID ↔ IIDB ↔ IIB (pure fibres shown in bold). The myosin ATPase differential staining procedure following pre-incubation at alkaline and acid pH could discriminate fibre types I, IIC, IIA, IID, and IIB. As noted before, a number of muscles were totally or relatively spared from both simvastatin- and cerivastatin-induced necrosis. In all our studies the soleus muscle was insensitive to statin-induced necrosis. Qualitative histochemical analysis of muscles from age-matched undosed rats (3 samples/muscle) indicated that the soleus muscle consisted predominantly of type I fibres, but also with a considerably smaller proportion of type IIA and a few IIC fibres. There were no type IID or IIB fibres detected. This is consistent with findings from other authors (Armstrong and Phelps, 1984; Kanbara et al., 1997; Staron et al., 1999; Soukup et al., 2002). Other muscles showing the least sensitivity to statin-induced necrosis were the muscle in the region of the flexor carpi ulnaris and the tongue. The muscles sampled from the region of the flexor carpi ulnaris also consisted of a substantial proportion of type I and IID fibres with a small proportion of type IIA and occasional IIC (Figures 3a and 3b). Again no type IIB fibres were present. The tongue was shown to consist predominantly of type IID fibres with only occasional type IIA and no IIB. Other muscles relatively insensitive to statin-induced necrosis were the masseter and diaphragm. The region of the superficial masseter muscle sampled for these studies was shown to consist predominantly of type IID with a lesser proportion of type IIA fibres (Figures 3c and 3d) No type I or IIB fibres were detected. This is to some degree consistent with findings from Kanbara et al. (1997) although they did detect a significant proportion of type IIB fibres. This discrepancy may be a result of a different sampling region. The diaphragm consisted of a substantial proportion of type I fibres with a further mixture of type IIA and IID, and no type IIB and this is also consistent with previous findings (Kanbara et al., 1997).

Fibre Typing in Muscles Showing Low-Grade Necrosis

Muscles from simvastatin and cerivastatin time-course study rats (Table 5), which showed early histological changes of low severity and with little or no apparent inflammation were also selected for further histochemical assessment including type I and II fibre typing by immunohistochemistry and myosin ATPase histochemistry. Tissues assessed from the simvastatin-dosed rats were; EDL muscle from animals 49 and 54, gastrocnemius muscle from animals 53 and 58 and biceps femoris muscle from animals 52, 53, and 58 and muscles from cerivastatin-dosed animals were; gastrocnemius muscle from animals 38, 43, and 58, the biceps femoris muscle from animals 39, 43, 51, 53, and 54, and the EDL muscles from animals 44. In addition to immunohistochemistry and myosin ATPase stability, fibre diameter was used as an indicator of fibre type for early necrotic fibres. In fast muscles fibre size is closely related to their metabolic characteristics; the lower the oxidative activity the greater the diameter of the fibre, with the largest fibres being type IIB (Wachstein and Meisel, 1955; Fuentes et al., 1998; Staron et al., 1999). Consistent findings were made for both simvastatin and cerivastatin. Immunohistochemistry for type I and type II fibres clearly showed that in muscles containing mixtures of these fibres, when early necrosis was present then type I fibres were spared. Even when a substantial proportion of the type II fibres were necrotic the type I fibres retained their normal histological appearance (Figure 4). This was consistent for all the muscles selected for assessment that contained type I fibres (EDL, red areas of gastrocnemius, and red areas of biceps femoris). In some muscles showing acute changes and containing type IIB fibres among other fibre types, often virtually all fibres affected showed myosin ATPase staining characteristics and fibre diameter consistent with type IIB with all other fibre types spared (Figure 5). In addition, areas of muscles containing substantial proportions of types I, IIA, and IID fibres, the type I fibres were consistently spared. Fibres with staining and cross-sectional diameter characteristics consistent with type IIA were also spared relative to fibres (necrotic) that retained staining and cross-sectional diameter characteristics of type IID (Figure 6).

Discussion

In an effort to develop a reproducible and robust in vivo model of statin-induced muscle necrosis, we have assessed the incidence, severity, and development of necrosis in a wide range of female rat muscles following the administration of simvastatin and cerivastatin. In our studies, statin-induced muscle necrosis in the rat was a feature of maximum tolerated dose (MTD). This is illustrated by our preliminary simvastatin study where necrosis occurred at the MTD of 80 mg/kg/day by day 12, but dosing at 60 mg/kg/day to day 43 did not result in any necrosis. Using an MTD we achieved a high incidence of muscle necrosis with a limited number of animals. It is notable that any early dose-limiting morbidity seen at the MTD was not a result of muscle necrosis as this did not occur until at least day 10. The preliminary studies with both simvastatin and cerivastatin showed that at toxic doses muscle necrosis in the rat is widespread and commonly severe, although there is some variability in severity between muscles and sparing of particular muscles. The histological development of the changes is characterised by early segmental necrosis, inflammatory (mainly mononuclear) cell infiltration with destruction and resorption of necrotic cytoplasm generally with retention of basement membranes, and if dosing is discontinued at an early stage some regeneration and mineralisation. These histological features are entirely consistent with findings of other investigators using a variety of statins including lovastatin (Waclawik et al., 1993), pravastatin, simvastatin (Smith et al., 1991), and cerivastatin (Schaefer et al., 2004).

Other investigators using a range of statins have also found that muscle necrosis is delayed and does not occur until after approximately 10 days’ dosing (Reijneveld et al., 1996; Schaefer et al., 2004). Our studies have illustrated that animals dosed with simvastatin (80 mg/kg/day) or cerivastatin (0.5 mg/kg/day) do not develop muscle necrosis prior to dosing for 9 days. With simvastain there was no evidence of any induced necrosis by study day 12 (11 days dosing), whereas with cerivastatin only 1/5 rats at day 10 showed 1 muscle with a mild necrosis considered a result of dosing. Muscle necrosis after this time became widespread and severe. This time dependency for statin-induced muscle necrosis is further illustrated by our preliminary cerivastatin study where doses of 0.5, 1, 2, and 3 mg/kg/day were used. At the 3 higher dose levels, animals were terminated between days 5 and 9 as a result of morbidity. Only minimal necrosis, consistent in nature and severity to that seen in control rats was present although at day 9, a slightly higher incidence was perhaps present in 1 rat. At day 15 following dosing of 0.5 mg/kg/day widespread severe necrosis was present in all 3 rats assessed.

This study shows that although a particularly high dose level is clearly required for the induction of muscle necrosis, even higher dose levels do not appear to significantly modify the duration required to induce a histological lesion. Our preliminary simvastatin study resulted in unacceptable levels of morbidity by day 7 when dosing was terminated. Severe muscle necrosis was seen in 3 rats that were removed from treatment, retained undosed, and then sacrificed as a result of a return of morbidity after day 13. It can be assumed from the results of the simvastatin time-course study that muscle necrosis was not present at day 10 in the rats that continued to be dosed, indicating that dosing for only 7 days can result in muscle necrosis but not before at least 10 days have elapsed.

As noted previously, there is a continuum of muscle fibre types in adult mammalian skeletal muscles from the slow twitch type I to the fastest twitch type IIB: I ↔ IC ↔ IIC ↔ IIA ↔ IIAD ↔ IID ↔ IIDB ↔ IIB (pure fibres shown in bold—Staron et al., 1999). The type I and IIB fibres represent the 2 metabolic extremes with a continuum of metabolic properties through these fibre types (Rivero et al., 1998). Type I fibres have slow contraction times, are oxidative in metabolism, have a high content of mitochondria and myoglobin, and have low quantities of glycogen and myosin AT-Pase. In contrast, type IIB fibres have fast contraction times, are glycolytic in metabolism, have a low content of mitochondria and myoglobin, and have high quantities of glycogen and myosin ATPase (Goll et al., 1977). Type I fibres are considered slow oxidative fibres, type IIB fast glycolytic fibres, and types IIA and IID fast oxidative/glycolytic fibres.

In all our investigations with simvastatin and cerivastatin there was no evidence in the soleus muscle of an induced necrosis and no ultrastructural changes. The soleus is a slow twitch muscle containing predominantly slow oxidative type I fibres and the more oxidative of the fast fibres IIC and IIA. Other muscles that were totally or relatively spared were the tongue, region of flexor carpi ulnaris, masseter, and diaphragm. All samples of these muscles were devoid of fast glycolytic type IIB fibres. For hind limb muscles that showed moderate and reproducible induced necrosis with simvastatin and cerivastatin our histochemical studies (with a selection of the muscles) and published data (Ariano et al., 1973) show that all contain a substantial proportion of fast glycolytic IIB fibres. These observations elaborate previous findings that fast twitch muscles are primarily affected (Smith et al., 1991; Waclawik et al., 1993; Reijneveld et al., 1996).

We can also report that in these muscles with a high content of fast glycolytic fibres there can be quite severe necrosis of the type II fibres with essentially total sparing of the type I fibres, which also show no ultrastructural changes, supporting reports from other investigators (Smith et al., 1991). In addition, our results illustrate that when muscles showing acute multifocal single fibre necrosis are assessed by pH-dependent myosin ATPase staining there is a differential sensitivity to statin-induced necrosis between the different type II fibres. Although the interpretation of such staining in early necrotic fibres is more problematic than in viable fibres, these necrotic fibres do retain some myosin ATPase activity and when assessed alongside fibre diameter a good evaluation of fibre type can be made in serial sections. The differential sensitivity seen favours the survival of type IIA fibres relative to type IID fibres, and the survival of both type IIA and IID relative to type IIB fibres.

So, our studies have shown that muscles showing the most severe necrosis following administration of simvastatin or cerivastatin contained a substantial proportion of type IIB fibres, and in these muscles it is the type IIB fibres that become necrotic first. In other muscles with areas containing a mixture of fibres including IIA and IID, but not IIB, it is the IID fibres that become necrotic first, and these muscles are relatively spared of necrosis. The soleus muscle that has no type IIB or IID fibres is totally spared. Taken together, these observations illustrate a continuum of fibre sensitivity to statins that mirrors their oxidative/glycolytic metabolic properties: Least sensitive →I ↔ IIA ↔ IID ↔ IIB ← Most sensitive.

There is, therefore, a clear relationship between the metabolic nature of individual muscle fibres and their sensitivity to statin-induced necrosis. As mitochondria play a central role in metabolism they may be important to this process. Ragged red fibres have been reported in humans (Phillips et al., 2002) following administration of statins. This change indicates the presence of subsarcolemmal accumulations of abnormal mitochondria. Further studies in rats have shown a close relationship between muscle necrosis and ultrastructural abnormalities in mitochondria although there is some discrepancy concerning the time of earliest mitochondrial change. Waclawik (1993) reported abnormal membranous organelles by day 5 whereas Schaefer et al. (2004) saw changes only at day 15. Ultrastructural changes in rabbit (Nakahara et al., 1998) and rat (Schaefer et al., 2004) muscle have been shown to consist of swollen mitochondria with disrupted cristae generally accumulating in the subsarcolemmal areas. However, they are often reported in association with other subcellular changes including disruption of myofibrils and derangement of Z bands. Because of this, Nakahara et al. (1998) concluded that the necrosis in rabbit muscle was not a consequence of muscle mitochondrial respiratory dysfunction, but that the mitochondrial abnormalities were secondary changes. Also, Schaefer et al. (2004) noted that mitochondrial injury does not appear to be the primary cause of skeletal muscle necrosis induced by cerivastatin but are likely secondary to necrosis or degeneration.

Our findings do not support these conclusions. Although we have also mainly detected mitochondrial changes in muscles that showed frank fibre necrosis, and ultrastructural mitochondrial changes and the development of frank fibre necrosis have a close temporal relationship, we did observe increased mitochondrial myelinoid bodies in the muscles of 2 rats dosed simvastatin but showing no necrosis by histology. Also, in some (nonnecrotic) glycolytic fibres from muscles showing very early multifocal single fibre necrosis, the only subcellular alterations present in isolation of any other changes were in the mitochondria. They were characterised by an increased incidence of mitochondrial vacuolation and further mitochondrial degeneration resulting in myelinoid and vesicular bodies, which tended to accumulate in the subsarcolemmal areas. These findings occurred in fibres with no concurrent changes to the contractile elements, endoplasmic reticulum, or other subcellular components.

It may be proposed that these accumulations result from segmental fibre necrosis; that they have arisen in a necrotic portion of the fibre. This seems doubtful in the absence of any other subcellular changes; swelling of the endoplasmic reticulum in particular. Mitochondrial abnormalities and subsequent muscle fibre necrosis are clearly associated with mevalonate reduction following inhibition of HMG-CoA reductase as mevalonate supplementation abolishes all changes. The HMG-CoA pathway not only supplies the cell with mevalonate and cholesterol but also other important molecules such as ubiquinone and the isoprenoids farnesol and geranylgeraniol. It has been suggested that starvation of mevalonate by inhibition of HMG-CoA reductase might lead to reduced supply of ubiquinone, which is required in mitochondrial electron transport. Supplementation of muscle cell cultures with geranylgeraniol (precursor of ubiquinone but not cholesterol) abrogates the in vitro myotoxicity of statins (Flint et al., 1997a). In vitro studies also show that inhibition of cholesterol synthesis alone via inhibition of squalene synthase does not induce myotoxicity (Flint et al., 1997b). These findings indicate that depletion of products of the HMG-CoA pathway other than cholesterol including ubiquinone play an important role in statin muscle necrosis. The mitochondrial changes seen in our studies may be the first observable morphological alterations resulting from such depletion. However, the precise nature of the early mitochondrial changes and why they are specific to glycolytic fibres still needs to be determined.