Effect of physical training on the proportion of slow-twitch type I muscle fibers, a novel nonimmune-mediated mechanism for muscle impairment in polymyositis or dermatomyositis

Authors


Abstract

Objective

To compare muscle fiber type composition and muscle fiber area in patients with chronic polymyositis or dermatomyositis and healthy controls, and to determine whether physical training for 12 weeks could alter these muscle characteristics.

Methods

Muscle fiber type composition and muscle fiber area were investigated by biochemical and immunohistochemistry techniques in repeated muscle biopsy samples obtained from 9 patients with chronic myositis before and after a 12-week exercise program and in healthy controls. Muscle performance was evaluated by the Functional Index (FI) in myositis and by the Short Form 36 (SF-36) quality of life instrument.

Results

Before exercise, the proportion of type I fibers was lower (mean ± SD 32% ± 10%) and the proportion of type IIC fibers was higher (3% ± 3%) in patients compared with healthy controls. After exercise, percentage of type I fiber increased to 42% ± 13% (P < 0.05), and type IIC decreased to 1% ± 1%. An exercise-induced 20% increase of the mean fiber area was also observed. The functional capacity measured by the FI in myositis and the physical functioning subscale of the SF-36 increased significantly. Improved physical functioning was positively correlated with the proportion of type I fibers (r = 0.88, P < 0.01) and type II muscle fiber area (r = 0.70, P < 0.05).

Conclusion

Low muscle endurance in chronic polymyositis or dermatomyositis may be related to a low proportion of oxidative, slow-twitch type I fibers. Change in fiber type composition and increased muscle fiber area may contribute to improved muscle endurance and decreased muscle fatigue after a moderate physical training program.

INTRODUCTION

Polymyositis and dermatomyositis are chronic inflammatory muscle disorders that are clinically characterized by muscle weakness, particularly low muscle endurance or muscle fatigue mainly in proximal muscles such as thigh, shoulder, and neck muscles (1–4). Most patients with polymyositis or dermatomyositis experience at least a partial improvement of muscle function with glucocorticoid treatment and other immunosuppressive agents (5). However, many patients are left with longstanding low muscle endurance and reduced quality of life (6, 7). The pathophysiologic background to the sustained low muscle endurance and muscle weakness in patients with chronic polymyositis or dermatomyositis is not known.

In healthy individuals there is a strong relationship between skeletal muscle function and muscle characteristics such as muscle fiber type composition and cross-sectional muscle fiber area. Thus, speed of muscle contractions relies mainly on fiber type composition and muscle strength on cross-sectional muscle fiber area (8–10). The different fiber types, fast- or slow-twitch fibers, are classified based on their content of contractile proteins, fast or slow myosin heavy-chain (11). The relative frequency of fast- and slow-twitch fibers in a muscle determines its functional property with respect to strength/power or endurance capacity. Type I fibers (slow-twitch) contain exclusively slow myosin and possess higher oxidative capacity compared with type IIA and IIB fibers (fast-twitch), which contain fast myosin (12–16). Type IIC fibers are intermediate fibers because they coexpress both fast and slow myosin, can develop into either type I or type II fibers (15, 16), and are infrequent in normal adult muscle. Whether patients with chronic dermatomyositis or polymyositis and persistently reduced muscle function possess aberrant muscle fiber characteristics that could contribute to the clinical symptoms has not been investigated previously.

From this perspective, the reported beneficial effects of exercise in patients with polymyositis or dermatomyositis are interesting (17–21). Until recently exercise was controversial in patients with inflammatory myopathies due to fear of exacerbation of muscle inflammation. Now several studies have demonstrated that moderate exercise in combination with immunosuppressive therapy is safe and also beneficial to muscle function. The physiologic explanation for the improved muscle function after training in patients with myositis has not yet been addressed. In a recent study, we reported improved muscle endurance with a training program in a group of patients with chronic polymyositis or dermatomyositis (18). These patients were subject to muscle biopsy before and after the exercise program, and these biopsy findings made it possible for us to address the question of whether molecular changes such as muscle fiber characteristics correlated to improved muscle performance. Therefore, the first goal of our study was to investigate if muscle fiber type composition and muscle fiber area were different in patients with chronic polymyositis or dermatomyositis with persisting muscle weakness as compared with age-matched healthy individuals. A second goal was to investigate whether the previously reported improved physical function after a 12-week exercise program was related to any changes in muscle fiber characteristics (18).

PATIENTS AND METHODS

Patients and healthy controls.

A total of 8 women and 1 man with chronic polymyositis or dermatomyositis, according to the Bohan and Peter criteria (1, 2), who had all participated in a resistive exercise study were included in the present study (18). Two of the original 11 patients were excluded due to insufficient muscle biopsy material. Five patients were classified as having polymyositis (all women) and 4 as having dermatomyositis (3 women and 1 man) (1). None of the patients had clinical or histopathologic features compatible with inclusion body myositis (22). None had diabetes mellitus or uncontrolled thyroid disease. Clinical and laboratory data have previously been presented in detail (18). Inclusion criteria were as follows: clinically stable; inactive polymyositis or dermatomyositis with a history of remaining muscle weakness; and a duration of immunosuppressive treatment ≥12 months, with no change in immunosuppressive therapy during the 3 months before inclusion in the study (18). Inactive disease was defined as absence of inflammatory infiltrates in muscle biopsy samples and absence of sign of inflammation on magnetic resonance imaging (MRI) of thigh muscles. Demographic data and clinical characteristics at the initiation of the study are presented in Table 1. Clinical investigations, muscle biopsy sampling, and MRI were carried out before and after 12 weeks of a physical exercise program (18). The disease was inactive at both investigations as assessed by muscle biopsy samples and MRI. Inflammatory infiltrates were absent in all biopsy samples but one, in which one small isolated perivascular infiltrate was present without any other histopathologic aberrations (18).

Table 1. Summary of patients' clinical features at first biopsy*
PatientDiagnosisAge/sexDisease duration, monthsTreatmentDuration with prednisolone, yearsDuration after prednisolone was stopped, years
  • *

    DM = dermatomyositis; PM = polymyositis. Mean age 52 years, median disease duration 44 months.

ADM44/M49Methotrexate 7.5 mg/week32
BDM52/F44022
CDM50/F34Prednisolone 5 mg every second day3Ongoing
DDM59/F40Prednisolone 3 mg every second day, azathioprine 100 mg/day4Ongoing
EPM56/F102Azathioprine 50 mg/day27
FPM54/F56Prednisolone 5 mg, azathioprine 100 mg/day3Ongoing
GPM60/F38012
HPM56/F18Prednisolone 2.5 mg/day1Ongoing
IPM37/F44Prednisolone 5 mg/day, azathioprine 100 mg/day3Ongoing

Muscle tissue from vastus lateralis muscle of 11 healthy individuals, matched for sex (6 women and 5 men, median age 59 years, range 42–78 years), with no clinical symptoms of muscle weakness and normal muscle biopsy results served as controls for analyses of fiber type composition and fiber cross-sectional area. Due to the limited amount of control muscle tissue from healthy individuals, muscle biopsy specimens from 8 other healthy controls (1 man and 7 women) with a median age of 50 years (range 43–60 years) were used as control samples for analyses of regeneration markers.

All patients and controls gave their informed consent to participate in the study. The study was approved by the local ethics committee at Karolinska University Hospital.

Methods.

Physical training.

The training program used in the present study was based on performing 10 repetitions in several different muscle groups with the intention to improve muscle endurance. Each patient was individually trained to perform a standardized 15-minute home exercise program 5 days a week for 12 weeks followed by a 15-minute walk as previously reported (18). The program contained a warm-up followed by climbing up and down a stool, exercise for shoulder mobility and grip strength with a pulley apparatus, strength exercises for quadriceps and hip muscles, sit-ups without support of the neck, and careful stretching. For those with small or moderate reductions of muscle function, weights (0.25–2.0 kg) were added as well as upper limb strengthening exercises.

Assessments.

Muscle function.

The Functional Index (FI) in myositis was used for evaluation of muscle strength and muscle endurance (23). The FI is a disease-specific index comprising 14 functional tests, which include recordings of the maximum numbers of repetitive movements that can be performed in different muscle groups. The FI also includes a recording of the number of observed transfers from side to side and up to a sitting position and peak expiratory flow (23). The FI is scored from 0 (no performance) to 64 (full performance).

Perceived health.

The Swedish version of the Medical Outcomes Study Short Form 36 (SF-36) was used to evaluate perceived health-related quality of life (24). The SF-36 is a generic, self-administered questionnaire composed of 8 subscales: physical functioning, role physical, bodily pain, general health, vitality, social functioning, role emotional, and mental health. Each subscale is scored from 0 to 100, where 100 indicates good health.

Muscle biopsies.

Muscle biopsy specimens were obtained from musculus vastus lateralis in the patients, with the second biopsy specimen, after training, obtained from the contralateral side. The specimens were obtained under local anesthesia using a semi-open technique (25). At least 2 tissue samples were taken from each biopsy site. The samples were frozen in isopentane precooled with dry ice and stored at −80°C. All muscle biopsy evaluations were performed on coded slides and the evaluators were blinded to diagnosis as well as order of the biopsy samples. One of the 2 samples from each biopsy site was investigated for presence of histopathologic changes including muscle fiber degeneration, regeneration, atrophy, number of fibers with central nuclei, and inflammatory infiltrates by an experienced neuropathologist. A second biopsy sample was used for analyses of muscle fiber characteristics as described below.

Determination of muscle fiber type composition and muscle fiber cross-sectional area.

Histochemical myofibrillar ATPase stainings at pH 4.3, 4.6, and 10.3 were used to classify muscle fibers into type I, IIA, IIB, and IIC subtypes (9, 26). The relative number of the different fiber types was determined after classification of the fiber types in ∼500 fibers in each muscle tissue section.

Fiber type characteristics of type I and type II fibers were also determined by a standard immunohistochemistry protocol using monoclonal antibodies directed against different isoforms of myosin: anti-slow myosin heavy-chain antibodies (Clone WB-MHCs; Novocastra, Newcastle, UK) for identification of type I fibers and anti-fast myosin heavy-chain antibodies (Clone MY-32; Sigma, St. Louis, MO) for type II fibers (11, 13). In this case the relative numbers of type I and type II fibers were calculated out of ∼400 fibers depending on the smaller tissue samples with cross-sectional fibers.

The validity of the ATPase staining was tested by performing a regression analysis between the type I and type II muscle fiber proportion determined by the ATPase staining versus determined by the immunohistochemistry protocol (type I percentage: r2 = 0.76, P < 0.05 and type II percentage: r2 = 0.76, P < 0.05). Cross-sectional muscle fiber areas of type I and type II were established by calculating the mean value out of the individual area of 20 fibers of each fiber type per biopsy. This was achieved by applying computerized image analysis (Leica system, BX 60, Tokyo, Japan; digital camera, Sony CDK-500, Tokyo, Japan) from an NADH staining (27).

Determination of regenerating fibers.

Immunohistochemistry was also used to determine signs of regenerating muscle fibers. We used antibodies toward different markers known to be expressed in different phases of regeneration of the muscle fibers (anti-CD56 antibody [Clone T 199; Dako, Glostrup, Denmark] [28], anti-vimentin antibody [Clone V9; Dako] [29], and anti-neonatal myosin heavy-chain antibody [Clone Wb-MHCn; Novocastra] [30, 31]) using a standard protocol for immunohistochemistry (32). The whole tissue sections were assessed and the percentage of positively stained fibers per tissue section was estimated.

Statistical analysis.

Unless otherwise stated, values in the text are the mean ± SD. P values less than 0.05 were accepted as statistically significant. Student's group t-test and Mann-Whitney test were applied to test the difference between groups for the muscle characteristic variables before training (basal). Wilcoxon's signed rank test for paired observations was applied to test training response for the clinical variables FI and physical functioning in the patient group. Student's t-test for paired observations was applied to test training response for the different fiber types, CD56 percentage, and vimentin percentage. Regarding cross-sectional muscle fiber area, a 2-factor analysis of variance (ANOVA; fiber type and time) was applied to test the difference between fiber type in response to time. Single regression analyses were applied to determine the relationship between training-induced changes in physical functioning versus training-induced changes in proportion of type I fibers and type II cross-sectional muscle fiber area.

RESULTS

Clinical data.

Following the 12-week physical training program, muscle endurance, recorded by the FI score, increased significantly (mean ± SD FI score 52 ± 12 before and 58 ± 7 after; P < 0.05) (Figure 1A). Health-related quality of life as assessed by the SF-36 physical functioning subscale was also significantly improved by the exercise program in 8 of 9 patients (mean ± SD 58 ± 19 before and 69 ± 22 after; P < 0.05) (Figure 1B). The other subscales did not change significantly.

Figure 1.

Individual changes in patients with chronic myositis after the 12-week exercise program. A, Changes in Functional Index (FI) score. Sixty-four corresponded to the maximal score of the FI. B, Changes in physical functioning, a subscale of the Short Form 36 Health Survey. A score of 100 corresponds to full health. C, Individual changes in the proportion of type I and type IIC fibers. D, Individual changes in muscle fiber cross-sectional area of type I and type II fibers (μm2). There are missing data for 1 subject in A and B.

Muscle characteristics at baseline.

Fiber type composition by ATPase staining.

At baseline, the relative proportion of type I fibers was significantly lower and the relative proportions of type IIB and IIC fibers were higher in patients compared with age-matched controls (Table 2). The relative proportion of type IIA did not differ between the groups.

Table 2. Fiber type distribution and fiber type areas in vastus lateralis muscle in patients versus controls before training*
 Patients (n = 9)Controls (n = 11)
  • *

    Values are the mean ± SD unless otherwise indicated.

  • Indicates difference between patients and controls at the statistical level of P < 0.05.

  • n = 6.

  • §

    n = 1.

  • n = 5.

Sex, no.  
 Female86
 Male15
Type I percentage32 ± 1047 ± 16
Type IIA percentage39 ± 1235 ± 10
Type IIB percentage26 ± 1015 ± 14
Type IIC percentage3 ± 31 ± 1
Type I fiber area, μm2 females4,467 ± 9164,694 ± 1,122
Type II fiber area, μm2 females3,538 ± 7983,441 ± 1,301
Type I fiber area, μm2 males5,398§4,979 ± 923
Type II fiber area, μm2 males4,614§4,559 ± 909

Cross-sectional fiber area.

In female patients, type I and type II cross-sectional fiber area did not differ significantly from the female controls (Table 2). The male patient's cross-sectional muscle fiber area of type I and II was within the range of the controls' muscle fiber areas.

Muscle fiber regeneration.

Three markers were used to identify regenerating fibers: CD56, vimentin, and neonatal myosin heavy-chain. At baseline, only a limited number of fibers expressed any of these markers and there was no significant difference between patients and controls. CD56 was expressed in 3.4% ± 3.7% of the fibers of the patients and in 3.8% ± 6.0% of the fibers of the healthy controls. The mean ± SD percentage of vimentin positive fibers did not differ between patients and controls (0.6% ± 0.9% and 0.4% ± 0.5%, respectively). The percentage of fibers expressing neonatal myosin heavy-chain was <0.5% in patients as well as in healthy controls.

Muscle characteristics after training.

Fiber type composition.

After the 12-week training program, the relative proportion of type I fibers was 10% higher (P < 0.05) and the relative proportion of type IIC fibers was 2% lower (P < 0.05) compared with before the exercise program (Figure 1C and Table 3). The relative proportion of type IIA and IIB fibers did not change with training.

Table 3. Distribution of muscle fiber type and cross-sectional area in type I and II in patients with chronic myositis before and after training*
 Type I%Type IIA%Type IIB%Type IIC%Type I area, μm2Type II area, μm2
  • *

    Values are the mean ± SD unless otherwise indicated.

  • Indicates difference between before and after training at the statistical level of P < 0.05.

Before training32 ± 1039 ± 1226 ± 103 ± 34,570 ± 9123,658 ± 827
After training42 ± 1334 ± 1126 ± 101 ± 14,929 ± 1,2164,555 ± 1,330
Interaction: type vs time    P < 0.05

Cross-sectional fiber area.

When applying a repeated-measures analysis (ANOVA), a fiber type difference was revealed regarding exercise-induced changes in cross-sectional muscle fiber area (interaction term: fiber type × time; P < 0.05) (Figure 1D and Table 3). A 25% increase of type II cross-sectional area (P < 0.05) was observed, whereas the type I cross-sectional area did not increase significantly.

Markers of regeneration.

The percentage of CD56 positive fibers was unchanged (mean ± SD 3.4% ± 3.7% versus 3.9% ± 2.5%) whereas the percentage of vimentin positive fibers was higher after the training program (0.6% ± 0.9% versus 1% ± 1%; P < 0.05). The training program did not change the percentage of fibers expressing neonatal myosin heavy-chain.

Correlations between exercise-induced changes in muscle fiber characteristics and clinical performance.

A positive correlation was found between increase in physical functioning (a subscale of the SF-36) and increase in the relative proportion of type I fibers in thigh muscles (r = 0.88, P < 0.01). A significant correlation was also found between increase in physical functioning and increase in type II cross-sectional muscle fiber area (r = 0.70, P < 0.05).

DISCUSSION

To our knowledge, this is the first study in which muscle fiber characteristics were investigated in patients with chronic polymyositis or dermatomyositis. Two main findings were observed. First, the patients were found to have a lower proportion of type I fibers and a higher proportion of the intermediate type IIC fibers in comparison with healthy controls. Second, after a 12-week physical training program the fiber type composition was closer to normal, the type II muscle fiber area had increased, and the changes in fiber type characteristics corresponded to clinical improvements in muscle function.

Several methodologic precautions were taken to standardize and validate the muscle biopsy data. The muscle biopsy samples were obtained from musculus vastus lateralis, which is classically involved in patients with polymyositis or dermatomyositis and because of its mixed fiber type composition, trainability, and accessibility. The control samples were also obtained from the same muscle. Because there are no differences in fiber type composition between the 2 legs in healthy individuals, the repeat muscle biopsy was performed on the contralateral muscle to avoid artifacts from the first biopsy (33, 34). To exclude a side-to-side difference as an explanation for the changed fiber type composition after training, a post hoc analysis was performed. This did not reveal any significant difference regarding the percentage of type I or type IIC fibers when comparing right and left limbs in pretraining biopsy samples (4 randomly obtained from the right leg and 5 from the left leg) or posttraining biopsy samples, respectively, in our patients. Therefore it is unlikely that a side-to-side difference could explain the observed changes of fiber type composition with training. The ATPase technique is well established for fiber typing and also allows distinguishing not only between type I and type II muscle fibers, but also between type IIA, IIB, and IIC fibers. The fiber type composition was also confirmed by immunohistochemistry. A limitation of the study is the low number of patients; nevertheless, the fiber type data and changes with exercise were consistent within the group.

The low percentage of oxygen-dependent type I fibers in comparison with healthy controls was unexpected, and there could be several explanations for this finding. One explanation could be the level of physical activity. Sedentary individuals in general demonstrate a lower proportion of type I fibers compared with physically active individuals (35). However, a sedentary lifestyle seems a less likely explanation, as low physical activity usually is accompanied by type II fiber atrophy, which was not seen in our patients. Because the controls in our study were healthy individuals with normal physical activity, we also compared the fiber type composition with published reference data for healthy, sedentary individuals, and still the patients with myositis had a lower proportion of type I fibers and a higher proportion of type IIC fibers (12, 36). Moreover, deconditioning did not affect fiber type composition in men with chronic heart failure in comparison with sedentary men when matched for aerobic capacity (37). All together, these observations imply that factors other than low physical activity could contribute to the low proportion of type I fibers in patients with chronic myositis. Effects of glucocorticoids should also be considered, because glucocorticoids are known to affect both fiber type composition and fiber area, and especially to cause type IIB fiber atrophy. All patients in our cohort had been treated with glucocorticoids over several years, and some were still receiving glucocorticoid treatment, although in low doses, at the time of study. However, in patients with rheumatoid arthritis treated with glucocorticoids, a lower percentage of type I fibers (36%) was accompanied by a reduced mean area of both type I and type II fibers, which was not the case in our patients in whom fiber area was not decreased (38–41). Age and sex are other factors that could explain aberrant muscle fiber characteristics, but a decreased percentage of type I fibers has not been reported as a consequence of age, and, moreover, women tend to have an increased oxidative phenotype (42).

A possible unexplored explanation for the low percentage of the oxidative type I fibers could be adaptation to muscle tissue hypoxia. This hypothesis is based on previous observations of a reduced number of capillaries. Furthermore, inflammation by itself may lead to tissue hypoxia (43). Interestingly, a similar low relative proportion of type I fibers was reported in patients with chronic obstructive lung disease, another condition that gives rise to local tissue hypoxia (44). Further support for local hypoxia in muscle tissue is the demonstration of low levels of ATP and phosphocreatine in muscle observed by magnetic resonance single-photon–emission computed tomography in patients with dermatomyositis or polymyositis indicating a metabolic dysfunction, which could be a consequence of hypoxia (44–46).

The seemingly contradictory finding of an increased percentage of type I fibers and at the same time an increased area of type II fibers after training could possibly have 2 explanations. First, considering the extremely low percentage of type I fibers before training, there should be a high potential to increase the percentage of the oxidative fiber type exclusively by the fact that the patients trained 5 days a week at 30 minutes per session for 12 weeks. Second, the exercise program consisted of 2 different components regarding fiber type recruitment; the first component was 15 minutes of strength/weight training in the arms and legs with 10 repetitions and the second component was a 15-minute walk. The first part was based on the work load and number of repetitions. The exact load in percentage of 1 voluntary repetition maximum (VRM) was not calculated; however, the training program was defined as ∼50% of 1 VRM. The increased cross-sectional area of type II fibers and not type I fibers seems reasonable because the training program to some extent included load training, and this type of stimulus is known to increase muscle protein synthesis preferably in type II fibers.

In the present study, muscle function improved with physical training and a corresponding increase was recorded in the proportion of oxidative type I fibers and type II cross-sectional area of the thigh muscles. The change in fiber type composition resulting from the training program in our study is remarkable, because a change in muscle fiber type composition after physical training in healthy individuals is a relatively rare phenomenon (26, 47). Muscle fiber type characteristics depend on several components including levels of fast or slow myosin heavy-chain proteins, volumes of mitochondria, capillary density, and oxidative capacity (48). These characteristics are regulated by a complicated system, which involves multiple signaling pathways and transcription factors (48). The clinical relevance of the changes in fiber type composition and fiber area was supported by the improvement of physical capacity measured by both the FI test and the reported improvement of physical functioning in the SF-36 questionnaire, which correlated with the increased proportion of type I muscle fibers and the increased type II cross-sectional fiber area.

The mechanisms for the fiber type transition induced by exercise in patients with polymyositis and dermatomyositis still need to be determined. From our study we can conclude that it is less likely that the effect on muscle function and on fiber type characteristics was due to muscle fiber regeneration. A weakness of our study is that we had to include a second cohort of control individuals to compare the number of regenerating fibers with the patient data due to lack of muscle biopsy samples from the first healthy control cohort. Nonetheless, we find it unlikely that this would have affected our results regarding changes in fiber type compositions because only scattered regenerating fibers were seen in the patients' biopsy samples both before and after exercise, and a similar low frequency was seen in the controls. A possible explanation for the improved function and the adaptation of muscle characteristics could be that exercise improved microcirculation in muscle, lowered total peripheral resistance, and reduced skeletal muscle ischemia, such as has been reported as a result of exercise in patients with chronic heart failure (49).

In conclusion, skeletal muscle characteristics that are related to muscle physiology and muscle performance were altered in patients with chronic polymyositis or dermatomyositis with persisting muscle impairment. Whether this is true for patients with myositis in other phases of the disease still needs to be determined. Although the results are based on a small group of patients, we believe that low muscle endurance in patients with chronic myositis could at least in part be explained by a relatively low proportion of oxidative type I fibers. Furthermore, the beneficial clinical effects of physical exercise could at least in part be explained by molecular changes within the skeletal muscle.

AUTHOR CONTRIBUTIONS

Dr. Dastmalchi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Dastmalchi, Lundberg.

Acquisition of data. Dastmalchi, Alexanderson, Loell, Borg, Lundberg, Esbjörnsson.

Analysis and interpretation of data. Dastmalchi, Alexanderson, Ståhlberg, Borg, Lundberg, Esbjörnsson.

Manuscript preparation. Dastmalchi, Alexanderson, Lundberg, Esbjörnsson.

Statistical analysis. Dastmalchi, Esbjörnsson.

Acknowledgements

We are grateful to Ylva Friberg and Eva Lindroos for excellent technical assistance with staining procedures, to associate professor Inger Nennesmo for histopathologic evaluation of the muscle biopsy samples, and to associate professor Christer Malm for the handling/delivering of antibodies CD56. We also thank associate professor Ronald Van Vollenhoven for linguistic advice, and professor Lars Klareskog for critically reading the manuscript.

Ancillary