Effect of Amniotic Membrane to Reduce Postlaminectomy Epidural Adhesion on a Rat Model

Abstract

Objective

Epidural fibrosis and adhesion are the main reasons for post-laminectomy sustained pain and functional disability. In this study, the authors investigate the effect of irradiated freeze-dried human amniotic membrane on reducing epidural adhesion after laminectomy on a rat model.

Methods

A total of 20 rats were divided into two groups. The group A did not receive human amniotic membrane implantation after laminectomy and group B underwent human amniotic membrane implantation after laminectomy. Gross and microscopic findings were evaluated and compared at postoperative 1, 3 and 8 weeks.

Results

The amount of scar tissue and tenacity were reduced grossly in group of rats with human amniotic membrane implantation (group B). On a microscopic evaluation, there were less inflammatory cell infiltration and fibroblast proliferation in group B.

Conclusion

This experimental study shows that implantation of irradiated freeze-dried human amniotic membrane reduce epidural fibrosis and adhesion after spinal laminectomy in a rat model.

INTRODUCTION

Failed back surgery syndrome (FBSS) is a condition characterized by persistent back pain with or without leg pain after lumbar spine surgery. FBSS occurs approximately about 5-40% in the literature^7,14,28). Multiple factors may contribute to develop FBSS such as inadequate surgical decompression, spinal instability, recurrent disc herniation, and epidural nerve fibrosis¹⁵⁾. However, many clinicians consider epidural fibrosis is one of major causes of persistent pain after lumbar spinal surgery^{8,20,27,28,35,40)}. The postoperative epidural fibrosis can cause extra-dural compression or dural tethering, which results persistent back and leg pain^7,31,33). Therefore, various methods and researches were done to prevent or reduce the amount of scar formation. Implanting materials including free or pedicle fat graft, synthetic membrane, fibrin foam and Adcon-L have been tried^{13,17,18,21,24,29,32,40)}. However, results with these methods are variable and complications have been reported^{19,23,25,26,30)}.

Recently, many studies have been reported the effect of human amniotic membrane on reducing adhesion scar in various areas of surgical practices. Davis¹⁰⁾ first reported the use of an amniotic membrane on skin transplantation on 1910. Since then it has been used to treat variable diseases such as skin ulcers⁴⁾, vaginal atresia³⁷⁾, and plastic repair of conjunctiva defects¹²⁾ and to reduce surgical adhesions in abdominal surgery and otolaryngology surgery^38,43,44). Amniotic membrane may reduce postoperative adhesion by inhibit vascularization, reduce inflammation and suppress apoptosis in epithelial cells^6,11,39).

The purpose of this experimental study was to assess the effect of implanting a human amniotic membrane around the nerve root to reduce epidural adhesion in a rat model.

MATERIALS AND METHODS

Animals

Six to eight weeks old Spraque-Dawley male rats (400-500 g) were acclimated to a housing facility (25±2℃ room temperature, 50% room humidity, 12 hours light-dark cycle) for 1 week under the permission of our animal experimental centre according to Korean National Veterinary Research and Quarantine Service Policy.

Amniotic membrane harvesting

Normal human amniotic membrane (AM) was obtained from one pregnant donor who agreed to informed consent of use of AM to animal experiment. The AM was obtained after cesarean section screened for HIV, hepatitis B and syphilis and washed with normal saline. The AM was separated from the placenta and attached to nitrocellulose paper and stored for one week under -80℃ in a mixed solution with Dulbecco modified eagle medium and glycerol (1 : 1 vol/vol). AM was melted on a room temperature before implantation. All AM were irradiated with 2.5 kGy radiation.

Grouping of rats

Total of 20 rats were used in this experiment. These rats were grouped into two groups. Groups A was the group of rats which AM was not implant after laminectomy (n=10) and implantation of freeze-dried AM was done after laminectomy in group B (n=10).

Surgical procedure and method

General anesthesia was induced with an intra-muscular injection of ketamine hydrochloride (0.1 g/kg). The experiment was done on thoraco-lumbar junction of rat, since thoraco-lumbar junction is largest spine in a rat model and has a least motion. A 3 cm skin incision was done on rat's thoraco-lumbar junction and the muscles dissected. Unilateral laminectomy was done to expose nerve root. Immediate muscle and skin closure were done using Dexon 3-0 and Nylon 3-0 after hemostasis around the exposed nerve root in group A. In group B, size of 0.4×0.8 cm freeze-dried AM was implanted before muscle and skin closure. All procedures were performed using loupe magnification of ×5.

Macroscopic and microscopic evaluations of epidural fibrosis were done on 1, 3, and 8 weeks post-operatively. Three rats were sacrificed at post-operative 1 and 3 week and 4 rats were sacrificed 8 weeks in both groups.

Both macroscopic and microscopic evaluation was done by one pathologist who was unaware of the operative details in a blinded manner to minimize the bias.

Macroscopic analysis

Macroscopic examination was done on a space between the dura mater and surrounding soft tissues. Quality of wound healing, possible adverse effects, and epidural adhesion were examined. Gross analysis of epidural adhesion was carried out by peel-off manually and the adhesion tenacity was scored using a visual 4-point qualitative scale proposed by Einhaus et al.¹³⁾ (Table 1). The scale was consisted of four grades : 1) Grade 0 : no adhesion between dura, 2) Grade 1 : slight adhesions to dura and easily detached, 3) Grade 2 : moderate adhesions and detachable by moderate traction and 4) Grade 3 : tenacious adhesions and detachable on by sharp dissection.

Table 1.

Gross analysis scoring system

Histological analysis

The nerve and surround soft tissues were excised in 0.5 cm apart and fixed in 10% formalin for 48 hours and then decalcified for 1 hour. An axial slice of each nerve specimen with surrounding tissue was prepared on slides. For histological evaluation, the slides were stained with Hematoxylin-eosin (H&E). Masson-trichrome stain was done for evaluation of surrounding connective tissues.

For evaluation of differentiation of myofibroblast, an immunohistochemistry staining with alpha smooth muscle actin (α-SMA) (Dako, Denmark) antibody was carried out. For immunohistochemistry stain, avidin-biotin-peroxidase detection system was used with DAKO Envision Kit (Dako, Denmark). First, de-paraffin was done with 10% xylene and then dehydrated in 100%, 95%, 70% alcohol was done in sequence before rinsed on running water for 10 minutes. Peroxidase blocking was done using 0.3% hydrogen peroxide (H₂O₂) prior to primary antibody incubation to reduce nonspecific background staining. For blocking non-specific antigen, the specimens were incubated in 10 mmol/L citrate buffer (pH 6.0) and irradiated on microwave for 10 minutes. Primary antibody reaction was carried out with anti-mouse anti-serum α-SMA actin (Dako, Denmark) and secondary antibody reaction was done with avidin-biotin-peroxidase complex for one hour at room temperature. The slides were counter stained with Mayers hematoxylin and examined with microscope.

Grade of adhesions

Histology grading method was used to standardize the quantity of adhesion between two groups. For slides that stained with H&E method, the number of inflammatory cells was counted per ×400 magnification field by a pathologist at post-operative 1, 3, and 8 weeks for both groups and classified into 4 groups : 1) scanty : less than 10 cells, 2) mild : 11-20 cells, 3) moderate : 21-30 cells, 4) severe : more than 31 cells.

The evaluation of the extend of myofibroblast differentiation was also performed under ×400 microscope magnification and scored into 4-points system : 1) point 1 : less than 10%, 2) point 2 : 11-50%, 3) point 3 : 51-75%, 4) point 4 : more than 76%.

Statistic analysis

Statistical difference of macroscopic and histological results between two groups was tested with Mann-Whitney test and statistically significant values were defined as p<0.05.

RESULTS

Macroscopic analysis

All specimens demonstrated epidural fibrosis and adhesion after re-exploration, especially on post-operative 3 and 8 weeks. Base on 4-point scale score system of amount and tenacity of epidural fibrosis and adhesion, the average score of AM not implant group (group A) was 1.8 and 1.5 at post-operative 1 week. However, the average score of AM implant group (group B) was 0 for both amount and tenacity of fibrosis and adhesion. The average score for group A was 2, 1.7 and 2, 2 at post-operative 3 and 8 weeks consequently for amount and tenacity of adhesion, whereas 0.5, 0.3 and 0.7, 1.0 at post-operative 3 and 8 weeks for group B. All results show statistically significant differences for scoring system of amount of fibrosis and tenacity of epidural adhesion at all post-operative week between two groups (p<0.05).

Histological analysis

H&E stain

At postoperative 1 week, predominant lymphocyte infiltration was shown in group A and predominant neutrophil infiltration was shown in group B (Fig. 1). Mixed lymphoplasma cells were shown at post-operative 3 weeks and small amount of lymphocytes were observed at post-operative 8 week in both groups.

Fig. 1.

In the human amniotic membrane in group without implant shows severe amount of acute inflammatory cell (A) and moderate inflammatory cell infiltration are shown in the human amniotic membrane implant group (B) at post-operative 1 week (H&E stain; ×400).

At post-operative 1 week, moderate number of inflammatory cells was shown in group B and severe number of inflammatory cells in group A. Mild and scanty number of inflammatory cells was shown at 3 and 8 weeks consequently in group B. However, moderate number of inflammatory cells was observed at post-operative 3 weeks and mild for 3 specimens and scanty for one specimen at post-operative 8 weeks in group A (Table 2).

Table 2.

Cell component and inflammatory cell count on H&E stain

AM : amniotic membrane

Masson-trichrome stain

Muscle fibers and fibroblast stained in red, and collagen tissues or fibrosis were stained in blue in Masson-trichrome stain. Group B shows only scanty amount of collagen tissues, in contrast of group A which showed profuse amount of collagen tissues at post-operative 1, 3 and 8 weeks (Fig. 2).

Fig. 2.

Dense and thick fibrotic band (stained in blue) around dura are show in group without in group with human amniotic membrane (AM) implantation (A). In contrast, thin and loose fibrotic adhesions around the dura are shown in human AM implant group at post-operative 8 weeks (Masson-trichrome stain, ×400) (B).

Immunohistochemistry staining with α-SMA

The score for evaluation of extend of myofibroblast differentiation was two at post-operative one week and one at post-operative 3 and 8 weeks in each specimens of group B. However, each specimens score was 4 at post-operative one week and 4, 3, 3 point for each specimens at post-operative 3 weeks and 1, 2, 2, 1 point for each specimens at post-operative 8 weeks (Table 3) (Fig. 3). These results also show statistically significant differences (p<0.05).

Table 3.

Smooth Muscle Actin Score

AM : amniotic membrane

Fig. 3.

At an alpha smooth muscle actin immunohistochemical stain, section from human amniotic membrane in group without implant shows strong positive reaction around the nerve root at post-operative 1 week (A) and reduced immnunohistochemical reaction at post-operative 3 weeks (B). In human amniotic membrane implant group shows weak positive reaction under amniotic membrane at post-operative 1 week (C) and also reduced immnunohistochemical reaction at post-operative 3 weeks (D).

DISCUSSION

The formation of epidural fibrosis and adhesion is an inevitable sequel of spinal laminectomy. Persistent pain, peresthesia and neurological deficit such as motor weakness can occur after lumbar laminectomy due to epidural fibrosis and adhesion^27,31,35). Although there are multiple factors that may cause FBSS, epidural fibrosis plays a major role in as many as 24% of cases³³⁾. Persistent or recurrent symptom and sign by epidural adhesion are due to direct mechanical tethering of nerve root or the dura, a powerful mediator of inflammation, phospholipase A₂, prostaglandin E1 and E2, leukotriene B, and impaired axoplasmic transport and nutritional or conduction disturbance as well as restricted arterial supply and venous return^1,7,31,34). Smith et al.³⁶⁾ reported that that the nerve roots of this region moved 0.5 to 5 mm distally and laterally during a straight leg raising test, depending on vertebral level on human cadavers study. Thus, a postoperative anterolateral epidural and periradicular scar situated in critical proximity to a lumbar nerve root might induce dynamic neural tension in a patient's everyday activities. Additionally, chronic compression of nerve root by epidural fibrosis may lead to markedly increase mechanical sensitivity^1,24).

The main reasons for epidural fibrosis formation are epidural fat destruction, hematoma, and paraspinal muscle fiber invasion^22,29). The extent of fibrosis depends primarily on the extent of the surgical procedure, especially on the degree of hemostasis³⁾. Preventing the migration of the fibroblasts into the exposed dura in the early healing phase may be the most important step to reduce epidural fibrosis formation. Thus, the interposition of a physical barrier to limit cell migration is considered an effective strategy to reduce scar formation.

Various biological and synthetic materials used to prevent scar formation have been evaluated, such as hemostatic sponges, free fat grafts, silastic, hyaluronic acid, polyactic acid, carboxymethylcellulose gels, a mixture of dextran sulfate and gelatin (Adcon-L, Adba), 5-fluorouracil, cyclosporine, and radiation therapy^{5,13,18,21,40-42)}. But, the results show only limited success. Some complications from use of Adcon-L have been reported such as cerebrospinal fluid leakage and infection^19,23). Also, seroma formation, dimpling of the scar, and cauda equina syndrome were reported as complications from free fat graft^25,30).

In this study, we assess the ability of human amniotic membrane to reduce postlaminectomy epidural adhesion in a rat model. Amniotic membrane has been used clinically in prevention of ocular disease¹²⁾, vaginal reconstruction surgery³⁷⁾, pelvic and abdominal adhesions³⁸⁾, repairing omphaloceles¹⁶⁾, and skin lesions including burn, ulceration and trauma^2,4,9). AM is a thin tissue sized about 0.2-0.5 mm that creates the inner layer of the amniotic sac. AM consists of a single layer of ectodermally derived columnar cells firmly fixed to an underlying layer of mesenchyme which contains large amounts of collagen. AM protects fetus from maternal infection and immune-reaction. AM modulates levels of cytokine and growth factor levels, and have also been shown to have other key properties such as pain reduction, antifibroblastic activity, and anti-bacterial effect⁹⁾. The membrane also contains a host of growth factors (epidermal growth factor, hepatocyte growth factor, nerve growth factor), anti-inflammatory cytokines (interleukin-6) and antivasculogenic factors (thrombospondin and tissue inhibitors of metalloproteases) which promote wound healing and suppress inflammation and neovascularization^11,12). Moreover, AM does not express HLA-A, B, C, or DR antigen or beta-2 microglobulin which immunological reaction after implantation does not occur^11,12).

In this study, we used irradiated freeze-dried AM to reduce immune reaction and infection. Our experiment shows that implanting of a human AM after laminectomy significantly reduces the epidural fibrosis on macroscopic evaluation. On histological evaluation, less amount of inflammatory cell infiltration was shown in group of human AM implantation on H&E stain. There were definite differences between two groups in which group A showed profound amount of collagen tissue staining compare to group B on Masson-trichrome staining. This implies that the group without AM implantation formed more fibrosis between dura and surrounding tissues. Also, less extent of myofibroblast differentiation was observed in group B compared to group A in immunohistochemistry staining with α-SMA.

These results are due to anti-inflammatory reaction by down-regulation of the cytokine such as interleukin-1, 8 and polymorphonuclear neutrophil elastase, suppression of epithelial apoptosis and antifibrotic effect by inducing a down-regulation of transforming growth factor-β which prevent differentiation of fibroblast into myofibroblast^11,12). In summary, our experiment suggest that AM can be used to reduce post-laminectomy epidural adhesion by inhibiting inflammation, antifibrotic effect and mechanical barrier between nerve root and surrounding structures.

In the previous clinical use of human AM in skin ulcers and ocular diseases, low incidences of clinical complications have been reported. Even though there is no clinical report on use of human AM on spine operation, it may be a potentially attractive material to prevent epidural adhesion after spine operation. Since mass production of human AM in tissue bank is possible, the price is affordable compared to alternative anti-adhesion products. It also may be used in patients with very thin body or undergoing long level spinal decompression instead of using fat graft. However, this experimental study has limitation in that a small number of animals have been used. In macroscopic evaluation, using visual 4-point scale system is subjective rather than objective and can lead to bias. Before clinical use of AM, more number with longer period animal experimental study and larger animal model study should be done.

CONCLUSION

This experimental study demonstrates that human amniotic membrane is an effective material to reduce epidural fibrosis and adhesion after spinal laminectomy in a rat model and also suggests the potential use of human amniotic membrane as an anti-adhesion technique in clinical practice.