A Passive Ankle Dorsiflexion Testing System for an In Vivo Model of Overuse-induced Tendinopathy

Summary

This protocol presents a testing system used to induce quantifiable and controlled fatigue injuries in a rat Achilles tendon for an in-vivo model of overuse-induced tendinopathy. The procedure consists of securing the rat's ankle to a joint actuator that performs passive ankle dorsiflexion with a custom-written MATLAB script.

Abstract

Tendinopathy is a chronic tendon condition that results in pain and loss of function and is caused by repeated overload of the tendon and limited recovery time. This protocol describes a testing system that cyclically applies mechanical loads via passive dorsiflexion to the rat Achilles tendon. The custom-written code consists of pre- and post-cyclic loading measurements to assess the effects of the loading protocol along with the feedback control-based cyclic fatigue loading regimen.

We used 25 Sprague-Dawley rats for this study, with 5 rats per group receiving either 500, 1,000, 2,000, 3,600, or 7,200 cycles of fatigue loads. The percentage differences between the pre- and post-cyclic loading measurements of the hysteresis, peak stress, and loading and unloading moduli were calculated. The results demonstrate that the system can induce varying degrees of damage to the Achilles tendon based on the number of loads applied. This system offers an innovative approach to apply quantified and physiological varying degrees of cyclic loads to the Achilles tendon for an in vivo model of fatigue-induced overuse tendon injury.

Introduction

As tendons connect muscle to bone and experience daily repetitive motions throughout their lifetime, they are highly prone to overuse injuries that are painful and limiting and result in impaired mechanical function, affecting 30-50% of the population¹. Tendinopathies are chronic conditions considered overuse injuries due to repetitive fatigue motions and inadequate healing to pre-injury levels. Both upper and lower extremities are commonly affected, including the rotator cuff, elbow, Achilles tendon, and patellar tendon^2,3,4,5. Achilles tendinopathy is common in activities involving running and jumping, especially athletes involved in track and field, middle- and long-distance running, tennis, and other ball sports, affecting 7-9% of runners^6,7. Injuries from running and jumping may also cause limited ankle dorsiflexion, which is a risk factor for Achilles and patellar tendinopathies^8,9,10. Thus, there is a need for a better assessment and characterization of tendinopathy, which this study can provide as a rat model of passive ankle dorsiflexion for overuse Achilles tendon injuries.

Previous work using small animal models has been aimed at studying the development and markers of tendinopathy. These include treadmill exercise, repetitive reaching, direct tendon loading, collagenase injections, surgery, and in vitro studies^{11,12,13,14,15,16}. Although the literature has benefited from the identification of damage markers from employing these tendinopathy models, limitations include loading the tendon in non-physiologically relevant joint motions, as in the case of direct loading of the tendon, not directly measuring applied loads, such as for treadmill studies, and not using physiological overuse, as in the case for collagenase injections, among others. To that end, this study aimed to develop a system that noninvasively applies quantified loads to the Achilles tendon with the application for overuse-induced tendinopathy studies to fill the gaps in previously developed small animal models for tendinopathy. We performed a pilot study to demonstrate that the system induces reproducible changes in mechanical properties over a range of loading cycles. This system enables physiologically relevant motion and loading to induce overuse while simultaneously quantifying and measuring the forces applied to and experienced by the tendon during the loading regimen.

Protocol

This study was conducted per Institutional Animal Care and Use Committee (IACUC) approval at Beth Israel Deaconess Medical Center. Animals were anesthetized using 5% isoflurane for induction and 2.5% for maintenance, and care was taken to avoid hypothermia.

1. Setting up the testing system

1. Control passive ankle rotation by a stepper motor to apply consistent rotation and torque. Control the stepper motor with a microcontroller. Use the inputs from the 3D position and orientation system to mark the degrees of rotation. Use the outputs from the torque sensor to provide feedback control for increasing the angle of dorsiflexion if the limit to threshold is not reached.
2. To begin, connect the microcontroller, torque sensor, 3D electromagnetic positioning, and orientation system to a computer and the power supply. Control the custom-built system using an in-house developed MATLAB code (Figure 1). Download the MATLAB code files from GitHub and follow specific instructions on running the code from the GitHub page instructions (https://github.com/Nazarian-Lab/PassiveAnkleDorsiflexionSystem).
3. Open MATLAB with the code files. Open the PDImfc software to connect the 3D electromagnetic positioning and orientation system to the MATLAB program. Click Connect | Continuous P&O | StartSockExport(). Keep the application open in the background.

2. Ex-vivo and post-mortem

1. Euthanize six 13-week-old Sprague-Dawley rats via CO₂ inhalation and a secondary method of euthanasia via thoracotomy. Dissect the right Achilles tendon with the calcaneus and myotendinous junction intact. Freeze at -20° C to perform mechanical testing at a later time. After the tendon is thawed, fine dissected, and prepared for mechanical testing, perform tensile loading to failure to obtain the ultimate tensile strength (UTS) of the tendon (Preload to 0.1N, Preconditioning for 10 cycles from 0.1 - 1 N, Ramp to failure at a constant displacement of 0.1 mm/s). Use 15% of the UTS as an input for the system to perform preconditioning for a later step, as described in step 3.4.
2. Euthanize another group of five animals with the same procedure for moment arm and strain measurements. Perform an X-ray of the left leg with the ankle in 90° dorsiflexion next to a ruler as a reference. Open the X-ray image in Fiji, using the ruler in the image as a reference, measure the tendon moment arm from the center of rotation of the ankle joint to the back of the ankle to be used as an input in the MATLAB code to convert the input force for preconditioning described in step 2.1 to the corresponding torque value as well as conversion between outputted torque and force for data analysis.
3. Immobilize the left hindlimb by taping two splints to place the knee in full extension. Lightly dorsiflex the ankle by pushing on the toes to ensure that the ankle rotation occurs due to the isolated tendon rather than involving surrounding soft tissues and is in tension. If not in tension or if there is movement in the knee, retape the splint.
4. Expose the tendon by removing the skin around the Achilles tendon. Place glue on a 1/32-inch aluminum bead, place it on the free tendon closest to the myotendinous junction of the Achilles tendon, and use a cotton swab with saline to remove excess glue. Repeatedly apply saline to the Achilles tendon throughout the remaining procedure to ensure moistness of the tissue.
5. Measure the cross-sectional area of the tendon using a digital caliper prior to applying any load. Assume the tendon is an ellipse and measure the width and thickness in triplicates.
6. Place the rat on the full-body platform in a prone position. Secure the ankle onto the joint actuator with a zip tie around the ankle and another around the toes, and secure the knee split with two zip ties. Rotate the axle so that the ankle is at full plantarflexion.
7. Connect the digitizing pen of the 3D electromagnetic position and orientation system to the computer and turn on the power supply.
8. Run the system code (described in further detail in Step 3) for the specified number of cycles (in this study, six euthanized rats received 7,200 cycles).
9. At 0, 500, 1,000, 2,000, 3,600, and 7,200 cycles for the ex vivo strain measurements, pause the cyclic loading regimen and measure the length of the tendon from the calcaneus to the aluminum bead at increments of 5° from 0 to 40° of dorsiflexion (the limit of actuation due to physical constraints of the system) using the 3D digitizing pen in triplicates in alternating fashion.
10. Calculate the tendon strain at varying angles using the lengths obtained from step 2.9, where the initial length is at 0° dorsiflexion for each cycle number. Perform a linear fit to obtain the relationship between dorsiflexion angle and strain at each cycle number. Use this relationship to convert raw angle data to strain for data analysis.
11. Calculate the tendon cross-sectional area using a digital caliper at 40° dorsiflexion assuming incompressibility (constant volume) with the tendon length measurements at 0° and 40° and the measured cross-sectional area at 0°. Use this cross-sectional area at each number of cycles to convert force to stress for data analysis (stress = force / cross-sectional area).

3. Mechanical loading protocol

1. For this section of the study, we used 25 female Sprague-Dawley 11-week-old rats, with 5 rats each randomly allocated to receive 500, 1,000, 2,000, 3,600, or 7,200 cycles of fatigue loading.
   NOTE: The preconditioning, initial calibration, and pre- and post-measurement take around 15 min to run, and the cyclic fatigue loading regimen takes 1 second per cycle. Thus, the longest time the rat is under anesthesia is around 2 h, which was performed under IACUC-approved protocols.
2. Connect the microcontroller, torque sensor, 3D electromagnetic positioning, and orientation system to a computer and the power supply. Control the custom-built system using the in-house developed MATLAB code (Figure 1).
3. Turn on the computer and open MATLAB with the code files. Open the PDImfc software to connect the 3D electromagnetic positioning and orientation system to the MATLAB program. Click Connect | Continuous P&O | StartSockExport(). Keep the application open in the background.
4. Induce anesthesia with 5% isoflurane through inhalation in an induction chamber. Following induction, secure the animal onto the full-body platform with a water-based heating element attached to maintain temperature and sustain anesthesia with 2.5% isoflurane via a nose cone attachment. Use a wet ointment on the eyes to prevent dryness during anesthesia.
5. Place the rat on the full-body platform in a prone position. Secure the ankle onto the joint actuator with a zip tie around the ankle and another around the toes, and secure the knee split with two zip ties. Rotate the axle so that the ankle is at full plantarflexion.
   NOTE: Ensure the zip ties do not cause constriction or lesions, take care in tightening, and if needed, place gauze in between the zip tie and the skin for a layer of protection.
6. For the following steps involving running the system's code, click Run on MATLAB for each section of the code corresponding to the specific loading test.
7. Cycle the ankle 50 times to 15% of the ultimate tensile stress based on the value of the ultimate tensile stress of the Achilles tendon from ex vivo pull to failure tests as measured based on step 2.1.
8. Perform an initial calibration of the tendon by dorsiflexing it three times to 12°. Use the slope of the linear region of the loading region of the hysteresis curve to calculate the exponential region of the curve.
9. Incrementally dorsiflex the ankle at increasing angles until either the exponential region of the curve is obtained by calculating the slope of the peak of the loading region of the curve (calculated using the in-house developed MATLAB code) or until it is rotated to 40°, whichever occurs first.
10. At the final obtained angle, perform five cyclic mechanical measurements as a preloading baseline.
11. Perform the cyclic fatigue loading regimen for a specified number of cycles (in this study, for either 500, 1,000, 2,000, 3,600, or 7,200 cycles).
12. Every 50 cycles, calculate the slope of the loading portion of the hysteresis curve (calculated through the in-house developed MATLAB code) to ensure it is still in the exponential region. Increase the dorsiflexion angle by 1° unless it is already at 40° until this exponential region is achieved.
13. After the cyclic loading regimen is complete, perform five cyclic mechanical measurements as post loading measurements at the initially chosen angle to measure tendon mechanical properties.
14. Remove the zip ties and splint. Return the animal to the recovery chamber. The animal is not left unattended until it has regained sufficient consciousness, followed by which it is returned to its cage. Monitor the animals daily for any adverse clinical signs and if present, administer buprenorphine at a dose of 1.2 mg/kg subcutaneously once every 72 h or perform early euthanasia. Euthanize the animals following 7 days of cage activity via CO₂ inhalation and a secondary means of euthanasia via thoracotomy.
    ​NOTE: The cyclic loading application and mechanical measurements were obtained with a custom-made jig consisting of a torque sensor, 3D-printed ankle joint actuator and animal bed, a 3D electromagnetic position and orientation system, and a stepper motor rotating a shaft to achieve dorsiflexion, as previously reported by our group¹⁷. This system is controlled by a MATLAB script mentioned in step 1.2. The torque sensor and the position and orientation system capture torque and position data throughout the system's loading protocol.

4. Data analysis

1. Load the pre- and post-measurement data separately into MATLAB.
2. Convert the torque to stress based on the measured moment arm from step 2.2 and the cross-sectional area measured at the specified number of loads applied obtained from step 2.11 using equations (1) and (2):
        (1)
        (2)
3. Convert the angle to strain based on the conversion obtained from step 2.10.
4. Calculate the average hysteresis (area between the loading and unloading curves), peak stress (maximum stress value of the cycle) and loading and unloading moduli (linear fit of the last 50% of the loading and the first 60% of the unloading curves) for the pre- and post- measurement cycles.
5. Calculate the percentage change in the mechanical properties from step 4.4 between the pre- and post-measurement cycles.

Results

With the increasing number of applied cycles, there was a greater reduction in in vivo tendon mechanical properties. There was a significantly lower reduction in hysteresis and the loading and unloading moduli for the 500-cycle group in comparison to the 3,600 and 7,200 cycle groups (p < 0.05) (Figure 2). While there was a significant reduction in peak stress per cycle from the 500 cycle to the 3,600 cycle group, there was no significant reduction between the 500 and 7,200 cycle groups. There was a consistent percentage decrease in hysteresis, peak stress, and loading and unloading moduli for the 3,600 and 7,200 cycle groups. Hematoxylin and eosin- and Masson's Trichrome stained images of tendon samples verified higher levels of microstructural damage with higher cycles of dorsiflexion with more rounded cells, hypercellularity, fiber disruption, and fiber crimping (Figure 3). The results in this paper are shown to demonstrate that higher cycles of dorsiflexion cause increased levels of damage to the Achilles tendon.

Figure 1: Passive ankle dorsiflexion testing system. (A) Power supply, (B) microcontroller, (C) stepper motor, (D) torque sensor, (E) 3D electromagnetic positioning and orientation sensor, (F) 3D printed ankle mount, (G) 3D printed animal bed, (H) 3D printed nose cone holder. Please click here to view a larger version of this figure.

Figure 2: Representative cyclic loading stress-strain curves. Hysteresis curves at 0, 500, 1,000, 2,000, 3,600, and 7,200 cycles. The arrow indicates decreasing peak stress with an increasing number of cycles. Please click here to view a larger version of this figure.

Figure 3: Representative histologically stained images of tendon samples. Hematoxylin and Eosin (left) and Masson's Trichrome (right) stained images of tendons for 500, 1,000, 2,000, 3,600, and 7,200 cycle groups for this study demonstrated that increasing the number of cycles applied results in more rounded cells, hypercellularity (stars), fiber disruption, and fiber crimping (arrows). Please click here to view a larger version of this figure.

Discussion

This study presents a method to cyclically load the rat Achilles tendon with a passive ankle dorsiflexion system for an in-vivo overuse-induced tendinopathy model. The importance of the system lies in its ability to isolate the Achilles tendon, apply quantifiable loads without surgically accessing the tendon, and measure in-vivo tendon properties.

In 2010, Fung et al. presented a rat patellar tendon fatigue model with a custom-built testing system¹⁴. Their study presented a method of directly loading the patellar tendon by exposing the tendon. While this method also applied quantifiable fatigue loads to the tendon, the direct application of loads may introduce an additional inflammatory wound healing response to the skin incision and subsequent closure. With our method, the noninvasively applied loads ensure that any measured biological response is entirely due to the loading protocol rather than any external factors.

A critical component of this loading protocol is the feedback-control loop. By checking the slope of the hysteresis loading curve and increasing the dorsiflexion angle, if necessary, the system continuously fatigues the Achilles tendon. Knee splinting is a critical step since it ensures that the dorsiflexion only strains the tendon instead of moving the knee and other surrounding soft tissue. To check whether the splinting is done correctly, manually actuate the ankle after splinting to feel for a stiff tendon and monitor the hysteresis curves produced prior to the cyclic loading step.

One of the limitations of this study is that the strain values are relatively large. However, they are comparable to passive dorsiflexion of human Achilles tendons and could be caused by the elongation of the Achilles tendon and the gastrocnemius muscle¹⁸. Another limitation is that the conversions between torque and stress are limited to ex vivo measured average tendon cross-sectional area and moment arm around the ankle joint, which may vary between animals.

The pathology and early stages of chronic tendinopathy are yet to be elucidated. Along with age and other risk factors, overuse is a major contributing factor to the development of chronic tendinopathy. Reproducible overuse injuries can be simulated with multiple applications of fatigue cyclic loading bouts through our system. Further, the noninvasiveness of this system allows for the assessment of biological and structural changes in tendon damage and healing responses over long periods to understand critical biomarkers in tendinopathy.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1/32'' Aluminum beads
2.5% isoflurane
3D digitizing pen	Polhemus, Vermont, NH, USA
3D electromagnetic positioning and orientation sensor	Polhemus, Vermont, NH, USA
5% isoflurane
Customized device: 1) Assembly, sensors, 3D printed animal bed and ankle mount actuator			Assembled as described in manuscript
MATLAB code	MATLAB, Natick, MA, USA
Microcontroller	Ivrea, Italy	Arduino UNO, Rev3
Nose cone
Scalpel and scalpel holder			No. 11 scalpel
Sprague-Dawley rats	Charles River Laboratories, Wilmington, MA, USA		11-13 weeks old
Stepper driver	SparkFun Electronics, Niwot, CO 80503	DM542T
Stepper motor	SparkFun Electronics, Niwot, CO 80503	23HE30-2804S
Straight forceps
Torque sensor assembly	Futek Inc., Irvine, CA, USA	FSH03985, FSH04473, FSH03927
Water heating pad