Thermal Preconditioning Prevents Peritendinous Adhesions and Inflammation

Abstract

Adhesion formation is one of the foremost obstacles to a reliably good outcome in tendon and joint surgery. Thermal preconditioning has been found to reduce the inflammatory response through the induction of molecular chaperone expression, a recently described family of cytoprotective intracellular proteins. The authors analyzed the effect of thermal preconditioning on the inflammatory response to surgery, on tendon healing, and on the formation of peritendinous adhesions in 16 New Zealand White rabbits. Very significant decreases in adhesion formation and in the gliding and dimensions of tendons in animals that had thermal preconditioning were found. Tendons from these animals also showed a decreased level of adhesion formation and a significantly diminished inflammatory response on histologic examination with no biomechanically significant deleterious effect on the strength of tendon healing on testing load to failure. These findings are consistent with induction of heat shock proteins by hyperthermic pretreatment. Such prevention of peritendinous adhesions and the inflammatory response to injury and surgery without compromising healing are findings that have significant implications for tendon surgery and all surgery involving joints and soft tissues.

List of Abbreviations Used: SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis, IgG1 immunoglobulin G1

The problem of postoperative clinical care of peritendinous adhesions is complex, involving the type and exact site of injury, the quality of surgery, inflammatory responses, and local biomechanical factors. It generally is accepted that adhesions probably represent the foremost factor compromising the results of tendon surgery and repair, with recent work focusing on the importance of the inflammatory response to injury in promoting adhesions. ^15,16,35 Loss of movement after any procedure involving joints or soft tissues is a major problem encountered by all orthopaedic surgeons.

Postoperative adhesions can cause decreased range of movement, poor function, and permanent deformity. Various investigators in tendon surgery using materials, drugs, and improved surgical techniques have addressed this problem. ^{8,11,25,34,39} In these studies substances such as Teflon, polyethylene, and Silastic were used with little benefit and results using steroids, antihistamines, dimethyl sulfoxide, and pulsed electromagnetic fields also have been disappointing. The development of most benefit probably has been improved surgical techniques, but unfortunately, even this has not solved the problem.

Preconditioning as mentioned above refers to the process whereby exposure of a cell to significant stress is associated with reduced susceptibility to the traumatic consequences of a subsequent stress. The heat shock proteins are a family of intracellular cytoprotective proteins expressed in cells subjected to various stresses including trauma, ischemia, and hyperthermia. ^2–4,19 The heat shock proteins have become known as molecular chaperones because they interact with other proteins facilitating protein folding, formation, and transport, thus serving to protect the cells and organism from additional traumatic, ischemic, or inflammatory damage. ^10,12

Although heat shock proteins have been shown to have cytoprotective properties, to decrease the local inflammatory response, and to stabilize vascular endothelia, their expression in and effects on injured tendon tissue have not been examined previously. ^3,31,37 The current authors hypothesized that preconditioning would protect against the inflammation associated with tendon surgery by upregulation of heat shock proteins in an established model of tendon surgery. This first required validation of a model of heat shock protein induction in rabbit tendons, which was done using Western immunoblotting. Macroscopic and microscopic analyses then were used to assess the effectiveness of heat shock protein induction in preventing peritendinous adhesion formation and inflammation. The effect on tendon healing was measured using biomechanical measures of load to failure of preconditioned and control tendons.

MATERIALS AND METHODS

Thermal Preconditioning

Twenty-four adult New Zealand White rabbits were used for this part of the study and were maintained according to standard guidelines under license of the department of health. They were allowed to acclimatize for 1 week before the start of the study and their mean weight at that point was 2.9 kg. A previously described model of reproducible tendon injury and adhesion formation was used in the study, which involved tenotomy of the flexor digitorum longus in isolation 2 cm proximal to the calcaneus. ³⁵

To examine the effects of induction of molecular chaperones, 12 animals randomly were selected to have thermal preconditioning and 12 were selected to receive no pretreatment. The latter group acted as controls. All rabbits were anesthetized according to the institution protocol. This involved an initial slow intravenous bolus injection, into an ear vein, of 30 mg/kg sodium pentobarbital. Anesthesia was maintained throughout the procedures with halothane and oxygen.

The rabbits in the preconditioning group were anesthetized as per routine, wrapped completely (excluding only the snout) in a standard patient warming blanket, and placed on a padded warming table, both preset to 41.5° C. The core temperature of each animal was raised, with monitoring using a rectal probe, from 38° C to 41.5° C, which then was maintained for 20 minutes. Control rabbits had a sham procedure that involved general anesthesia as above, but no warming, and were recovered after 20 minutes anesthesia. The animals were recovered after conditioning. Sixteen had their surgical procedure 18 hours later, with eight being sacrificed for Western immunoblotting heat shock protein analysis. The procedures and Western immunoblotting were done at 18 hours because previous studies showed that the beneficial effects of heat shock protein expression are maximal at this time. ^{3,12,26,31,33}

Western Immunoblotting for Heat Shock Protein 72 Expression

To confirm that the proposed model of inducing molecular chaperone production was effective, specimens were obtained from four rabbits that were sacrificed 18 hours after hyperthermia and from four normal, unconditioned controls. These animals did not participate in the main study. Heat shock protein 72 was analyzed specifically because the protective effects of the induction of heat stress have been shown to occur via these proteins in a transgenic mouse model. ³⁶ The tissue samples were taken from the Achilles tendon, soleus and gastrocnemius muscles, and kidneys (the latter via midline laparotomy incision). They were snap frozen in liquid nitrogen and stored at −70° C pending testing for heat shock proteins. The samples then were defrosted and homogenized in cold phosphate buffered saline (1 mL) and centrifuged at 4100 g for 30 minutes at 4° C. The supernatants were collected and protein concentrations were quantified using a Coomassie protein assay reagent (Pierce, Rockford, IL).

The final concentration of samples was diluted to 50 μg/10 μL. The protein was denatured at 100° C for 10 minutes and aliquots containing equal amounts of proteins then were suspended in sodium dodecyl sulfate glycerol loading buffer (pH 6.8, 62.5 mmol/L Tris, 2% sodium dodecyl sulfate, 10% glycerol, 5% mercaptoethanol, 0.01% bromophenol blue) and proteins were separated by SDS-PAGE (ExcelGel, Pharmacia, Sweden) with 75 mg of total protein loaded per lane. Proteins were transferred to a nitrocellulose membrane (Sigma Chemical Co, St Louis, MO) and labeled with a primary monoclonal antibody, mouse antihuman IgGl, specific for heat shock protein 72 (StressGen, Victoria, British Columbia, Canada). After the secondary monoclonal antibody was added, sheep antimouse IgGl conjugated with alkaline phosphatase (Serotic, Oxford, England), the protein was observed using bromochloroindolyl phosphate and nitro blue tetrazolium (BCIP/NBT, Sigma Chemical Co).

Operative Procedures

The left hind leg of all the rabbits was shaved and prepared with povidone-iodine antiseptic solution. One intravenous prophylactic dose of cephradine was administered preoperatively to all animals. After sterile draping, a standard 2-cm incision was made over the Achilles tendon, 2 cm proximal to the calcaneus. The Achilles tendon in the rabbit is composed of three separate bundles, the gastrocnemius and soleus tendons and the flexor digitorum longus tendon. ²⁸ The tendon complex was exposed, the paratenon was opened, and a transverse tenotomy was done in the flexor digitorum longus tendon. This minimized surgical trauma as the flexor digitorum longus is the most superficial of the tendon complex. The wound was irrigated with sterile saline and the skin wound was closed with interrupted 4/0 nylon. A simple gauze dressing was applied, with no casting and the animals were recovered with routine administration of intramuscular buprenorphine (5 μ/kg) for analgesia.

Clinical Assessment of Movement and Adhesions

The animals from all studies were sacrificed 21 days postoperatively and blindly assessed for wound healing, ability of the tendon to slide against overlying skin, quantity of adhesions, and the ability of the tendons to glide against the surrounding tissue and fascia. Digital images (Kodak 2010, Rochester, NY) of all the tendons were obtained for records and later correlation of adhesion scores. Adhesions and gliding were assessed macroscopically and scored according to established criteria in a blinded fashion by separate observers. ^13,35 Gliding was assessed directly by palpation and by observation when the foot was put through a full range of movement. All scores were on a scale of 0 to 4, where 0 was no adhesions or free tendon gliding against skin or surrounding tissues; 1 was minimal adhesions with the tendon easily exposed by blunt dissection or there was minimal impairment of free gliding; 2 was moderate adhesions but blunt dissection still was possible or there was moderate impairment of gliding; 3 was severe adhesions but the tendon could be exposed by sharp dissection or there was grossly limited gliding; and 4 was maximum adhesions and the inability to clearly distinguish the tendon from the surrounding tissues on dissection or no gliding was possible. Tendons from both groups were removed en bloc and sectioned for histologic analysis of inflammation, adhesion formation, and tendon repair.

Histologic Evaluation

Tendons for histologic evaluation were fixed immediately in 10% formaldehyde solution and sectioned transversely for sectioning and mounting. Sectioning was done through the healed flexor digitorum longus and whole tendon complex at the level of the previous tenotomy. Staining was done in the standard fashion with hematoxylin and eosin. The pathologist who reviewed the samples had no knowledge of the nature of the treatment given to the animals for each specimen and reviewed the samples for tendon healing, inflammation, and degree of adhesion formation between tendon and surrounding structures. These observations then were correlated separately with the clinical gross morphologic findings.

Biomechanical Testing

Whole Achilles tendon complexes were obtained for biomechanical testing. ¹³ The tendons were sectioned distally at the level of the calcaneus and proximally at the level of the musculotendinous junction. They were wrapped in saline moistened swabs, sealed in airtight containers, and frozen at −20° C pending testing. The process of freezing has been shown not to significantly affect the mechanical properties of tendons in similar studies. ^17,21 Before testing, the tendons were thawed at room temperature and kept moist with saline during the testing process. The tendons were mounted on an Instron materials testing machine (Instron Corporation, Canton, MA), using a modification of the tendon mounting technique described by Loitz et al. ²¹ Uniform tightening of the tendon grips was done with a torque wrench and a small preload of 1 N was applied before testing. Load to failure then was tested using a velocity of 1 cm/minute.

Statistics

All of the above data were accumulated, and having passed a test of normality in every category, the results were analyzed statistically using Student’s t tests. The SigmaStat statistical program (version 2.0, Jandel Scientific, Chicago, IL) was used to do the tests on a personal computer (Compaq Presario 1245, Compaq Computer Corporation, Houston, TX).

RESULTS

None of the animals in the study had any complication of the surgical procedures, with no wound infections or postoperative complications. The mean weight of all animals at the end of the experiment was 3.1 kg (range, 2.85–3.25 kg) with no difference between control and conditioned groups.

Western Immunoblotting for Heat Shock Protein 72 Expression

Qualitative analysis of the Western blots revealed that heat shock protein 72 was uniformly strongly expressed in the tendons of the animals sacrificed 18 hours after thermal preconditioning, but was not expressed at all in control tendons (Fig 1). Because a demonstration of heat shock protein expression in tendon tissue could not be found in previous studies, the current authors verified the model described here with the samples from kidney and muscle as described above. These confirmed heat shock protein 72 expression after the conditioning process described here, as would have been expected based on previous reports using similar models of inducing heat shock protein. ^9,25

Fig 1.: (A) A Western blot showed heat shock protein 72 expression in renal tissue (B), muscle (D), and tendons (F) after thermal preconditioning, but was not expressed in control tendons (E). There was minimal expression in control kidneys (A) and moderate expression in control muscle (C).

Gross Morphologic Features

There was a significant difference between groups with respect to the weight of retrieved tendons (p = 0.005) as shown in Table 1. In each case the whole Achilles tendon complex from the level of the calcaneus to the musculoskeletal junction was weighed to ensure that all adhesions and fibrous tissue were included in this measurement. Conditioned tendons were lighter than controls, with control tendons and paratenon being visibly thicker with opaque fibrous tissue.

TABLE 1: Dimensions of Retrieved Tendons

Macroscopic assessment of adhesion formation showed a highly statistically significant difference (p < 0.001) for the conditioned groups versus the control groups (Table 2). Tendons in the control group all had adhesions develop, some of which were dense (Fig 2) whereas there were two tendons in the conditioned group that did not have any adhesions develop (Fig 3). In all cases with adhesion formation, these were present between tendon and paratenon and between paratenon and surrounding peritendinous tissues. Adhesions around preconditioned tendons when present typically were filmy and easily broken down by blunt dissection whereas adhesions in the control group were more substantial usually requiring sharp dissection for tendon exposure.

TABLE 2: Results of Tendon Assessments

Fig 2.: A control tendon shows dense adhesions at the level of previous surgery to peritendinous structures and surrounding skin.

Fig 3.: A tendon of a molecular chaperone conditioned animal shows an absence of peritendinous adhesions.

Assessments of tendon gliding against the skin before retrieval and against the surrounding tissue and fascia once the tendon was exposed also were done sequentially in all cases. There was a significant difference in gliding against skin (p = 0.002) and against the surrounding tissue and fascia (p = 0.002) when comparing conditioned animals with controls (Table 2).

Histologic Evaluation

There was histologic evidence of tendon healing in both groups, with the appearance of fewer adhesions in the thermally preconditioned group. The histologic findings correlated well with the gross morphologic clinical observations in the two groups. Preconditioned tendons showed a more mature stage of healing tendon tissue and decreased cellularity and volume of inflammatory involvement (Figs 4, 5). To see a higher power view of a control tendon see Figure 4B in CORRONLINE (http://www.corronline.com). In terms of inflammation, the preconditioned tendons also showed a lesser degree of involvement, with confinement of the inflammatory process. There also was less histologic adherence between tendon, paratenon, muscle, and surrounding structures in the preconditioned group, which again correlated directly with the independently observed macroscopic adhesion formation.

Fig 4.: A photomicrograph of a transverse section of control tendon shows disorganized tendon architecture and a large number of inflammatory cells throughout the area of healing at the tenotomy site (Stain, hematoxylin and eosin; magnification, ×20). See http://www.CORRONLINE.com for a higher power view.

Fig 5.: A photomicrograph of a transverse section, at the level of the tenotomy, of a tendon from a conditioned animal shows significantly fewer cells within the tendon structure, with a more mature phase of healing and relatively normal architecture (Stain, hematoxylin and eosin; magnification, ×20).

Biomechanical Testing

The mean load to failure of control tendons was 213.8 N (range, 145–280 N) and for thermally preconditioned tendons it was 170.8 N (range, 115–255 N). This difference in the mean load to failure between the groups was not statistically significant (p = 0.085). All tendons in both groups failed in the substance of the tendon and there were no failures at the fixation points to the grips of the machine.

DISCUSSION

There is little doubt based on clinical experience and the literature that the formation of adhesions, with subsequent stiffness and deformity, is one of the foremost obstacles in achieving predictably good results in tendon surgery and all surgery around joints. ^15,20 Although tendon healing seems to be a multifactorial process, major proposed elements of healing include the concept that adhesions are part of the natural healing process, supplying the fibroblasts and vascularity necessary for repair. ²⁹ It also has been shown that fibroblastlike cells can be generated from synovial fluid and from acellular canine allografts. ³⁰ Alternatively, it has been shown that extrinsic repair cells may not be essential, because intrinsic tenocytes have been shown to have the capability to divide and produce collagen. ^22,23

The concept that the inflammatory process is necessary for adequate healing led to concerns with the goal of the current study, which was to decrease adhesions using the essentially antiinflammatory mechanism of molecular chaperone induction. Previous studies using extrinsic antiinflammatory treatments, particularly steroid agents, have led to decreased strengths of healed tendons versus controls. ^13,14,27,39 The biomechanical testing of the tendons in the current study, however, showed that molecular chaperone expression induction, although effective in suppressing the inflammatory reaction, did not have a deleterious effect on the strengths of the healed tendons, which were comparable with normal strengths as described in previous studies. ^13,39

The current authors showed that heat shock proteins are induced in response to thermal stress in tendons. This occurs by preferential gene transcription and subsequent translation during a period of normothermic recovery, as allowed by the 18-hour recovery period as described previously. ^36,38 This is the first time molecular chaperones have been used, or analyzed in the attempted prevention of soft tissue and in particular tendon adhesions.

Although heat is well-described as an effective therapeutic modality for patients with stiff joints and tendons and for inflammation, heat always has been used as a local symptom-based treatment after surgery using repeated, topical, local modalities, with no long-term curative role. ^1,18,40 The treatment of animals by induction of molecular chaperone expression using thermal conditioning, however, involves an entirely different approach. By raising core temperature on one occasion by at least 2.5° C before any intervention, the current authors showed that it is possible to prevent adhesions, excess peritendinous inflammation, and fibrous tissue in vivo. Although a cause-and-effect relationship between heat shock protein expression and this observed reduction in inflammation and adhesions was not established, the data are consistent with the conclusion that the hyperthermic induction of heat shock proteins was responsible for these effects. There were significantly fewer adhesions in preconditioned animals and conditioned tendons were lighter than controls, with weight previously having been described as an indicator of tendon size and fibrous tissue formation. ^5,13 The assessments of tendon gliding against the skin before retrieval and within the surrounding tissues give another indirect measure of adhesion formation. ^35,39 Again the current authors showed significant benefits in this regard for the preconditioned animals versus controls.

Heat has been shown to be one of the most effective inducers of heat shock proteins, quantitatively and qualitatively, and as mentioned previously, the protective effects of heat stress in this regard have been shown to specifically occur via the heat shock protein 70 family in a transgenic mouse model. ³⁶ Despite this, heat is a relatively nonspecific stress and investigations of the possibility of alternative selective stimuli, such as a pharmacologic agent, are ongoing. ⁷ Notwithstanding such investigations, the current study showed that heat is a simple, relatively physiologic, well-tolerated modality, without any evident adverse effects in the population studied. Additional investigations will involve isolated heating of the extremity or part to be operated on as such an approach would be more realistically applicable to human subjects.

The finding of adequate healing histologically and biomechanically indicates that a marked inflammatory response and scar formation is unnecessary for healing, as has been well-documented in the fetus. ^6,24,32 This reflects findings by others that although cells involved in the inflammatory response may be rendered quiescent by heat shock protein induction, many specialized cells can retain their function, therefore leading to healing without some of the adverse consequences. ^26,33

The results of the current study showed that it is possible to reduce and eliminate postoperative, posttraumatic inflammation, and peritendinous adhesions without detrimental side-effects on healing and tendon strength. This was done using a simple and well-tolerated method of thermal preconditioning. These observations have significant implications in not only tendon surgery, but also for all surgery in which preservation of joint and soft tissue mobility is desired.