Changes in pennate muscle architecture after gradual tibial lengthening in goats



The purpose of this investigation was to examine the changes in unipennate muscle architecture after distraction osteogenesis. Nine adult goats underwent 20% tibial lengthening in one of the hind limbs. Immediately after distraction, lengthened and contralateral (untreated) tibialis caudalis (TC) muscles were harvested. Lengths of the muscle belly, muscle fiber (FL), sarcomere (SL), tendon (TL), and superficial aponeurosis, as well as muscle mass, pennation angle (PA), and physiological cross-sectional area (PCSA), were compared between the treated and contralateral sides. Lengthened TC muscle demonstrated 20.8% increase in belly length, 4.39% increase in TL, and 36.7% increase in FL, while PA decreased by 37.2% (P = 008). Muscle length increase was mainly due to lengthening of muscle belly, which resulted both from FL increase and 15.3% length increase in the aponeurosis component of muscle belly, without significant effect of the PA decrease. The FL increase was due to SL increase, not to sarcomere neogenesis, while mass and PCSA did not change. We concluded that although muscle architecture can be adversely affected by distraction because of deficient sarcomere neogenesis, PCSA can remain unchanged, giving false impression of preserved function. Change in PA plays only minimal role in muscle adaptation to distraction. Anat Rec, 2007. © 2007 Wiley-Liss, Inc.

Efficiency of the soft tissue adaptation to bone lengthening during distraction osteogenesis is a major determinant of the treatment outcome. Failure of the soft tissues, especially the skeletal muscles, to keep pace with the gradual increase in bone length gives rise to such complications as joint contracture, subluxation, or dislocation, and axial deviation of bone segments due to bone regenerate bending or fracture (Paley, 1990; Hantes et al., 2001).

Previous investigations demonstrated that muscles increase in length after bone distraction (Lee et al., 1993; Simpson et al., 1995; Day et al., 1997b; Kanbe et al., 1998; De Deyne et al., 2000; Fink et al., 2000b, 2000c; Castano et al., 2001; Lindsey et al., 2002). Some reports demonstrated that distraction increased fiber length (FL), mainly by generating new sarcomeres in series (Williams et al., 1998; Lindsey et al., 2002). However, the muscle fiber length increase during distraction was often less than the increase in bone length (Williams et al., 1998; Lindsey et al., 2002; Samchukov et al., 2003). Changes in other parameters, such as the muscle fiber pennation angle and muscle aponeurosis length, should also be considered during limb lengthening. Pennation angle (PA) is the angle between muscle fibers and the long axis of muscle movement (Gans and de Vree, 1987; Gans and Gaunt, 1992).

It is an important architectural parameter that allows different arrangements of muscle fibers and therefore affects skeletal muscle function (Muhl, 1982; Gans and de Vree, 1987; Gans and Gaunt, 1992). Both FL and the cosine of PA are components of the standard formula used to calculate the muscle physiological cross-sectional area (PCSA) (Sacks and Roy, 1982; Lieber, 2002), which is the only muscle architectural parameter that is directly proportional to the maximum tetanic tension of the muscle, hence to its ability of force production (Lieber, 2002). Although numerous studies evaluated how the PA affects normal muscle anatomy and physiological function (Muhl, 1982; Gans and de Vree, 1987; Lieber and Blevins, 1989; Gans and Gaunt, 1992; Scott et al., 1993; Van Eijden et al., 1997; Ledoux et al., 2001; Maganaris et al., 2001), little is known about the effect of gradual distraction on muscle fiber PA and PCSA.

The purpose of this study was to explore the changes in muscle architecture, especially the PA and PCSA, after distraction and the relative contribution of PA and FL changes to the overall increase in muscle belly length and the change in PCSA. Our first hypothesis was that gradual stretching of a pennate muscle would result in reorientation of muscle fibers so they would become more in line with the long axis of the muscle. The second hypothesis was that the increase in muscle belly length would be totally explained by the effect of FL increase and PA decrease. Our third hypothesis was that PCSA would decrease due to an increase in muscle fiber length and decrease in muscle mass and fiber pennation angle.


Animal Model

Nine skeletally mature, neutered male Spanish cross goats underwent 20% lengthening of one randomly assigned tibia using the distraction osteogenesis technique, with the contralateral (untreated) limb serving as matched control. The housing, care, and experimental protocols were in accordance with the guidelines established by the Institutional Animal Care and Research Advisory Committee at the authors' institution.

Surgical Technique and Distraction Protocol

Under general anesthesia, a standard 120 mm diameter circular external fixator (IMEX Veterinary, Longview, TX) was applied to the tibia (Fig. 1A). The preassembled frame consisted of a proximal ring and a distal double-ring block, which were temporarily connected anteriorly and posteriorly by two threaded rods. After alignment, the proximal ring was secured to the tibia with three medially inserted 4 mm diameter half-pins. The distal double-ring block was attached to the tibia by three 1.6 mm diameter cross-wires. This wire/half-pin configuration minimized the risk of muscle injury by the bone fixation elements. Once the frame was secured to the bone, a 2 cm longitudinal incision was made on the medial surface of the tibia at the level of mid-diaphysis. The periosteum was elevated and the medial, anterior, and posterior cortices were osteotomized using an oscillating saw. Osteotomy was completed by dividing the lateral cortex using a sharp osteotome. After completion of the bone division and soft tissue closure, the two temporarily threaded rods were replaced by four telescopic distraction rods.

Figure 1.

Stabilization frames for the goat's hind limb: (A) circular distraction frame and (B) monolateral frame replacing the circular frame at necropsy.

Following a 5-day latency period, distraction was commenced at the rate of 0.25 mm three times per day and continued 59–62 days until the desired 20% limb lengthening was achieved. During the distraction, all animals were allowed unrestricted ambulation. Passive flexion and extension movements of the stifle (knee) and hock (ankle) joints were performed daily. The animals were euthanized immediately after the completion of distraction using an intravenously administered barbiturate.

Muscle Harvesting and Preparation

At necropsy, the circular frame was replaced by a monolateral external fixator (Fig. 1B), which was attached to the proximal and distal tibia by two pairs of medially inserted half-pins and extended proximally and distally. The extended monolateral fixator was attached proximally to the femur and distally to the metatarsal bone by two additional pairs of medially inserted half pins so that the knee and ankle joints were stabilized at the neutral position obtained before necropsy. A similar fixator was attached to the contralateral extremity, stabilizing the knee and ankle joints at the same angles as those on the lengthened limbs.

After the removal of skin and underlying subcutaneous tissues, both hind limbs were disarticulated at the hip joint and immediately immersed in 10% buffered formalin. Subsequently, the tibial muscle compartments were opened, allowing the formalin solution to penetrate into the deeper muscles.

Tibialis Caudalis Muscle

The M. tibialis caudalis (TC; Fig. 2) is a unipennate muscle that lies in the posterior compartment of the goat's tibia under the gastrocnemius and superficial digital flexor muscles (Constantinescu, 2001). The muscle originates from the posterolateral aspect of the lateral tibial condyle and consists of a short spindle-shaped belly and a long slender tendon, which joins the tendons of the flexor digitorum lateralis and medialis muscles superiorly and inferiorly to the level of the ankle joint, respectively. The common tendon, flexor digitorum profundi tendon, continues distally toward the digits beneath the tendon of the flexor digitorum superficialis muscle. At the level of the metatarso-phalangeal joints, the common tendon is divided into two tendons, which pass through the splits in the divided tendon of the flexor digitorum superficialis muscle and are inserted at the base of the distal phalanxes. The muscle fibers of tibialis caudalis extend from the deep proximal aponeurosis to the relatively short and narrow superficial distal aponeurosis, which gradually transforms into the muscle tendon. Functionally, this muscle is involved in the flexion of the digits. Muscle nomenclature has been adopted from Nomina Anatomica Veterinaria (International Committee on Veterinary Gross Anatomical Nomenclature, 2005).

Figure 2.

Schematic of the tibialis caudalis muscle showing the superficial aspect of the muscle as well as a sagittal section parallel to the plane of muscle fiber (MF) orientation. SA, superficial aponeurosis; DA, deep aponeurosis; T, muscle tendon.

Muscle Architectural Measurements

After 2 weeks of fixation, the origin and insertion of all tibial muscles were exposed, the muscle origin-to-insertion length and the muscle belly length were measured, and the tendon length was calculated. The muscles were then dissected, harvested, and the wet weight was measured using a digital scale (XP-3000; Denver Instrument, Arvada, CO), labeled, and immersed back in formalin.

Under a 10× magnification on a dissecting microscope (AO Scientific Instruments, Buffalo, NY), three full-length fiber bundles consisting of 10–20 equal-length muscle fibers were isolated from the middle portion of the muscle. The muscle fiber (bundle) length was measured from the proximal to the distal myotendinous junctions using a digital caliper (Mitutoyo, Tokyo, Japan). The sarcomere length was identified by the laser diffraction analysis technique previously described (Lindsey et al., 2002).

All muscles were sectioned longitudinally from the origin to approximately 10 mm proximal to the distal end of the muscle belly, transecting both the superficial and the deep muscle aponeuroses (Fig. 3). The plane of the muscle cut was always parallel to the orientation of muscle fibers in the sagittal plane to visualize a flat surface with layered muscle fibers. The muscle was then stretched back to the previously measured origin-to-insertion length and pinned to a rigid board with the transecting surface parallel to the horizontal plane. The length of the TC muscle superficial aponeurosis was measured from the most proximal point to the distal end of muscle belly.

Figure 3.

Cut section in two TC muscles showing the plane of PA measurement and the difference in PA between control (A) and experimental muscle (B).

For each muscle, three muscle fiber bundles located within the middle third of the muscle belly were identified on the transecting surface and photographed twice using a high-resolution digital camera (Coolpix 5000; Nikon, Tokyo, Japan). The camera was mounted on the copy stand with lens surface positioned parallel to the horizontal plane. To enhance the visualization of muscle fibers, a fine stainless steel wire was placed parallel to each designated fiber. All acquired images were exported into a computer work station and archived. The PA was identified as the angle between the designated fiber and the superficial aponeurosis (muscle tendon) and measured twice using AxioVision 3.0 image analysis software (Carl Zeiss Vision, Munich, Germany). The 12 PA measurements were averaged to give the PA for each muscle.

Physiological Cross-Sectional Area

Assuming that muscle density (ρ) remained unchanged with the same fixation time and procedure (1.112 gm/cm3) (Ward and Lieber, 2005), the PCSA was calculated using the equation (Sacks and Roy, 1982; Lieber, 2002)

equation image

where θ is pennation angle and NFL is the normalized fiber length after accounting for the variation in sarcomere length. The correction factor was set at 2.2 μ. Although that figure was based on electron microscopic measurements of actin and myosin filament length as well as data obtained from other species (Lieber, 1992), it was set as a mathematical correction factor to eliminate the effect of SL variability on FL in both the lengthened and contralateral muscles, not as an accurate value of the optimum sarcomere length of the goat muscle. The normalized fiber length (NFL) was calculated as follows: NFL = measured fiber length × 2.2/measured sarcomere length.

Geometrical Model

The pre- and postdistraction muscle belly length, FL, PA, and aponeurosis length were superimposed in a simple geometrical model (Fig. 4). The increase in muscle belly length resulting from the combined change in FL and PA (MBL) could be calculated using the equation

equation image

where FLd is the average fiber length after distraction, cos θ′ is the average cosine of the postdistraction pennation angle, FLc is the average fiber length before distraction, and cos θ is the average predistraction pennation angle.

Figure 4.

Geometrical model of two superimposed sagittal sections of pre- and postdistraction TC muscle, demonstrating the changes in muscle belly length, pennation angle, fiber length, and aponeuroses length: geometry before distraction is outlined by the dashed lines, while the solid lines represent the geometry after distraction. FLc is represented by “ad,” FLd is represented by “af,” and the increase in aponeurosis length is represented by “bg.” T, muscle tendon; MF, muscle fiber; DA and SA, deep and superficial aponeuroses, respectively. According to the model: ΔMBL = (af × cos θ′) − (ad × cos θ), where ΔMBL is change in muscle belly length resulting from FL increase and PA decrease, θ is muscle fiber pennation angle before distraction, and θ′ is muscle fiber pennation angle after distraction.

Statistical Analysis

The repeatability of the PA measurement was tested on a randomly selected muscle. Four fiber bundles were identified on the cut section, labeled with stainless steel wires, digitally photographed, and the PA was measured. This was repeated five times by the same observer. SPSS Advanced Statistics 11.5 for Windows (SPSS, Chicago, IL) software was used for statistical analysis. All measurements were repeated by the same investigator and averaged. Wilcoxon signed ranks test was used to detect differences between the lengthened and contralateral muscles. A P value of less than 0.05 was established as a statistically significant difference between the two groups.


There was no statistically significant difference in TC muscle mass between the lengthened and contralateral limbs. Muscles of the lengthened limb showed an average 36.9 ± 10.5 mm (7.6%) increase in muscle origin-to-insertion length, 20.1 ± 4.5 mm (20.8%) increase in muscle belly length, 10.7 ± 8.2 mm (15.3%) increase in muscle superficial aponeurosis length, and 16.78 ± 10.95 mm (4.39%) increase in tendon length, compared to that of the contralateral limb (Table 1). The average FL increased by 10.1 ± 2.4 mm (36.7%).

Table 1. Comparison of TC muscle architectural parameters between lengthened and contralateral limbs
Muscle Mass (g)5.40 ± 0.825.58 ± 1.08p = 0.44
Total Length (mm)519.7 ± 13.9482.8 ± 15.3p = 0.008
Belly Length (mm)116.9 ± 8.596.8 ± 7.3p = 0.008
Tendon Length (mm)402.8 ± 18.77386 ± 18.26p = 0.01
Fiber Length (mm)37.61 ± 4.0927.47 ± 3.03p = 0.008
Sarcomere Length (μ)2.97 ± 0.152.18 ± 0.2p = 0.008
Pennation Angle (degree)9.10 ± 1.5814.47 ± 1.82p = 0.008
PCSA (cm2)1.73 ± 0.291.74 ± 0.28p = 0.59
Aponeurosis Length (mm)80.2 ± 869.6 ± 9p = 0.01

To validate the PA measurement method, the four fiber bundle showed an average deviation of 0.47°. Intraclass correlation coefficients showed values of 0.9715 for a single measure and 0.9942 for all five trials. Between the lengthened and contralateral muscles, the average PA decreased by 5.4 ± 2° (37.2%; Fig. 3).

However, the average SL in lengthened muscles was 0.8 μ higher than that of the contralateral muscles (36.5% increase; P = 0.008). Therefore, when FL was corrected for the difference in SL (for PCSA calculation), there was no statistically significant difference in the corrected FL (0.06 ± 2.26 mm) between experimental and contralateral muscles. Average PCSA of the lengthened TC muscles showed no significant difference compared to that of the contralateral limbs (Table 1).

The average increase in muscle belly length resulting from the combined effect of FL increase and PA decrease (Fig. 4) was 10.54 ± 2.5 mm. The other half of muscle belly length increase was accounted for by an increase in length of the connective tissue aponeurosis that forms the superficial surface of the muscle belly (by 10.7 ± 8.2 mm).


This study analyzed the architectural adaptation of a unipennate muscle to incremental stretching, focusing on the roles of PA and PCSA in such adaptation. Tibial lengthening caused a remarkable increase in TC muscle length, mostly due to lengthening in the muscle belly rather than the tendon. This pattern is consistent with previous reports (Sun et al., 1994; De Deyne et al., 2000). However, the increase in muscle belly length was not solely due to elongation of muscle fibers. In fact, almost 50% of that increase was a result of stretching of the aponeurosis component of the muscle belly, while the decrease in PA had only minimal effect.

Furthermore, after correcting for the significant increase in SL, it became clear that the recorded increase in FL was due to stretching of existing sarcomeres, with no evidence of increasing the sarcomere number in series. Although in this study, no adequate labeling was done to rule out definitely an ongoing process of sarcomere neogenesis, the amount of stretching of the sarcomeres was enough to account mathematically for the recorded increase in FL at this time point, suggesting that the number of sarcomeres in series probably did not significantly change after lengthening. This appears to be in contrast to previous reports, which indicated that distraction osteogenesis stimulated muscle growth (Day et al., 1997a; De Deyne, 2002; De Deyne et al., 2002) and induced adding new sarcomeres in series (Kim et al., 1993; Williams et al., 1998; Lindsey et al., 2002). However, those data were either obtained from a small animal model, rat or rabbit, or from an anterior compartment muscle, tibialis anterior or peroneus tertius.

In a large animal model, such as the goat, the size of the tibia allows the same distraction protocol used in human patients to be tested and maintained for the same duration. It should also be noted that the animals used in our study were skeletally mature. It has been documented that while distracted muscles in skeletally immature animals remained functionally and structurally intact after elongation, the same distraction regimen produced significant muscle damage and weakness in mature animals (De Deyne et al., 2000, 2002).

A recent report showed that all posterior group muscles in goats, including the nonpennate soleus muscle, showed lack of sarcomere neogenesis (Samchukov et al., 2003), while none of the anterior group muscles showed this problem. It has been reported that temporary partial denervation may occur in distracted muscles (Fink et al., 2000a, 2000c), which may explain the lack of sarcomere neogenesis in some muscles at the end of the distraction period. However, there is currently no evidence to support that such denervation could affect the posterior more than the anterior muscles. More research is undergoing to explore and explain the differences in muscle response to distraction between anterior and posterior group muscles.

With no change in TC muscle mass or corrected FL after distraction; the PCSA did not change in lengthened muscles. The 37.2% decrease in PA only increased its cosine from 0.97 to 0.99 (2.1%), thus was too small to cause any PCSA change. Could the increase in FL have been due to adding new sarcomeres in series, it would have caused a 27.8% decrease in PCSA (P = 0.01).

Theoretically, therefore, adding new sarcomeres in lengthened muscle fibers would negatively affect the PCSA, and hence the force production ability, of the muscle, given that muscle mass remains unchanged. Functionally, however, it is essential that the muscle adds more sarcomeres to maintain the SL, so the muscle can produce force, even at a smaller PCSA.

From the architectural aspect, a pennate muscle with short fibers and large physiological cross-sectional area is suitable for force production. In contrast, greater excursion, which would be needed after bone lengthening, is best produced by a muscle with smaller cross-sectional area and long individual fibers that are oriented more parallel to the force long axis (Lieber, 1993; Lieber and Bodine-Fowler, 1993). This can only be achieved if the increase in FL was caused by adding of new sarcomere in series. Inevitable PCSA decrease in this case can be reversed by increasing the muscle mass with rehabilitation exercises during and after distraction.

It should be noted that the time point at which the muscles were examined in this study was relatively early. Clinically, patients who suffer from impaired muscle function immediately after distraction show remarkable recovery after physiotherapy. Further research is needed to investigate the functional condition of distracted muscles after physiotherapy to increase muscle mass and PCSA, and whether sarcomere neogenesis would take place to balance the increase in FL.

Muscle weakness has frequently been reported after distraction osteogenesis (Young et al., 1993; Fitch et al., 1996; Oey et al., 1999; Fink et al., 2000b, 2000c; Hayatsu and De Deyne, 2001; De Deyne, 2002). This weakness has been attributed to temporary denervation, disuse atrophy, and myopathic changes in the muscle (Oey et al., 1999; Fink et al., 2000b, 2000c; Hayatsu and De Deyne, 2001). Although PCSA is the only architectural parameter that is directly proportional to the maximum tetanic tension of the muscle, as it represents the number of sarcomeres in parallel, its calculation is based on the assumption that fiber length consists of a series of optimal sarcomere lengths (Sacks and Roy, 1982; Lieber, 2002).

Therefore, because TC muscle did not adequately synthesize new sarcomeres during distraction, using PCSA as monitor of the muscle force production ability during and after distraction can be misleading in such case.

Numerous studies have examined how the PA affects normal muscle anatomy, architecture, and physiological function (Muhl, 1982; Gans and de Vree, 1987; Lieber and Blevins, 1989; Gans and Gaunt, 1992; Scott et al., 1993; Van Eijden et al., 1997; Ledoux et al., 2001; Maganaris et al., 2001). Previous PA studies have focused only on the PA of the muscle fibers located at the muscle belly surface (Muhl, 1982; Sacks and Roy, 1982; Lieber and Blevins, 1989; Scott et al., 1993; Van Eijden et al., 1997; Ledoux et al., 2001; Maganaris et al., 2001). The orientation of these fibers; however, does not represent the orientation of the majority of muscle fibers. In the present study, PA was measured in the actual plane of muscle fiber orientation using a consistent cut section within the muscle.

Several methods were described for in vitro PA measurement, including the direct application of a protractor (Sacks and Roy, 1982) or a goniometer (Lieber and Blevins, 1989) to the muscle surface, manual tracing of the surface PA on lead acetate paper (Van Eijden et al., 1997) or a photograph (Muhl, 1982), and the use of digital photography and image analysis software (Ledoux et al., 2001). In vivo measurement of PA during muscle contraction was carried out using ultrasonography (Maganaris et al., 2001). Finally, MRI was used to measure three-dimensional orientation of the skeletal muscle fibers (Scott et al., 1993). We have chosen the digital photo analysis method, because it had the least measurement error compared to that of goniometer and protractor measurements. Digital photography also needed less specimen exposure and manipulation time and provided backup photographs where the same angles could be archived and remeasured.

It has also been reported that PA may vary within the same normal muscle (Gans and de Vree, 1987). In this study, PA measurement was limited to the middle portion of the muscle in order to eliminate the variability at the proximal and distal ends of the muscle belly. Although not a major producer of force, tibialis caudalis is a classic example of a unipennate muscle, it crosses the level of the osteotomy, and it is safely located away from the surgical field and newly forming callus. Further research is undergoing to examine the architectural adaptation in larger muscles and the difference between muscles with different architecture and anatomic locations.

This study provided evidence that PA decreases after distraction; however, this decrease plays only a minimal role in both muscle length and PCSA changes. It also demonstrated that an unchanged PCSA after distraction does not necessarily imply preserved muscle force production ability, as the sarcomeres may be under continuous stretch, without increasing the sarcomere number. Finally, stretching of the soft tissue component of the muscle belly may explain a significant portion of muscle length increase after distraction.