Intervertebral disc and macrophage interaction induces mechanical hyperalgesia and cytokine production in a herniated disc model in rats

Authors


Abstract

Objective

The expression of proinflammatory factors such as tumor necrosis factor α (TNFα), interleukin-6 (IL-6), IL-8, and prostaglandin E2 (PGE2) is significantly correlated with the symptoms of herniated disc disease. Among the different types of immune cells, macrophages are frequently noted in the herniated disc tissue. We undertook this study to clarify the interaction of the intervertebral disc (IVD) and macrophages with regard to the production of TNFα, IL-6, IL-8, and PGE2.

Methods

We developed 2 animal models to assess the interactions of IVDs with macrophages in terms of TNFα, IL-6, IL-8, and PGE2 production and pain-related behavior. We also cocultured IVDs and macrophages to assess the role of TNFα in IL-6, IL-8, and PGE2 production.

Results

IVD autografts induced TNFα, IL-6, IL-8, and cyclooxygenase 2 (COX-2) messenger RNA (mRNA) up-regulation; macrophage infiltration was seen shortly after the autograft was implanted. A significant decrease was noted in the mechanical threshold of the ipsilateral paw following the up-regulation of TNFα, IL-6, IL-8, and COX-2 mRNA. Only IVD and macrophage cocultures resulted in IL-8 and PGE2 up-regulation. TNFα up-regulation was maximized before that of IL-6 and IL-8. TNFα neutralization attenuated production of IL-6 and PGE2, but not that of IL-8. Neutralization of TNFα and IL-8 significantly increased the paw withdrawal mechanical threshold in the IVD autograft and spinal nerve ligation model.

Conclusion

IVD–macrophage interaction plays a major role in sciatica and in the production of TNFα, IL-6, IL-8, and PGE2. TNFα is required for IL-6 and PGE2 production, but not for IL-8 production, during IVD–macrophage interaction. Neutralization of TNFα and IL-8 can be a valuable therapy for herniated disc disease.

Herniated disc disease is a common disorder, affecting at least 1% of the total population (1). In 2005 in Japan, the treatment costs for discopathy were 9.2% of the total costs for all musculoskeletal disease treatment and 22.3% of all costs for spinal disease treatment (2). The clinical symptoms of herniated disc disease are induced by nerve root compression and inflammation around the herniated nucleus pulposus. However, while nerve root compression is commonly observed in patients with herniated disc disease, 37–52% of normal volunteers also show asymptomatic herniated discs on myelography (3), computed tomography (4), or magnetic resonance imaging (5). The severity of the clinical symptoms of herniated disc disease (such as paresis, muscle weakness, muscle wasting, impaired reflexes, sensory deficits, and positive straight leg raising) is not correlated with the size or position of the herniated disc (6), which indicates nerve compression. Clinical observation of herniated disc disease patients with sciatica shows that C fibers, which are unmyelinated and sensitive to inflammation, are more affected than Aδ fibers, which are myelinated and sensitive to compression (7). Therefore, inflammation of the herniated disc tissue is thought to be important in the pathologic analysis of herniated disc disease.

Herniated disc tissues express a variety of inflammatory factors such as interleukin-1α (IL-1α) (8), IL-1β (8), IL-6 (9–15), IL-8 (8–10, 13, 16), IL-10 (8), leukotriene B4 (14), monocyte chemotactic protein 1 (MCP-1) (16), nitric oxide (11, 12), prostaglandin E2 (PGE2) (9, 10, 15, 17, 18), RANTES (13), transforming growth factor β (8), thromboxane β2 (14), and tumor necrosis factor α (TNFα) (11, 12, 19, 20). Among these, TNFα, IL-6, IL-8, and PGE2 have unique pathologic characteristics in herniated disc disease. TNFα is always present in the nucleus pulposus and is thought to induce pain-related behavior in animal models (21, 22). High levels of IL-6, IL-8, and PGE2 are produced in human herniated disc tissue (9, 10, 12, 13, 15, 19), and their levels show significant correlations with preoperative symptoms of herniated disc disease (8, 10, 17). However, the mechanisms by which TNFα, IL-6, IL-8, and PGE2 contribute to herniated disc disease symptoms, especially sciatica, are unclear.

In addition to the inflammatory factors, nucleus pulposus cells with matrix and a variety of immune cells are also present in the herniated disc tissue, and, among these, macrophages play a role in herniated disc disease. Approximately 36–66% of surgically obtained herniated disc specimens show abundant macrophage infiltration (23–25). Macrophages could be an important source of inflammation in herniated disc disease. We have shown that IL-6 production is induced when intervertebral discs (IVDs) and macrophages are cocultured (26). However, the mechanisms of production of TNFα, IL-8, and PGE2 in herniated disc disease are not clearly understood. Therefore, this study was designed to clarify 1) the contribution of IVDs and macrophages to herniated disc disease in terms of TNFα, IL-6, IL-8, and PGE2 production and pain-related behavior and 2) the roles of TNFα, IL-6, IL-8, and PGE2 in IVD–macrophage interaction.

MATERIALS AND METHODS

Animals.

All experimental protocols were approved by the Institutional Animal Care and Use Committee of Kobe University. Male Sprague-Dawley rats (body weight 257–300 grams; CLEA) were used in this study.

Autologous IVD autograft and spinal nerve ligation model.

To compare the effects of spinal nerve compression with and without herniated disc–induced inflammation on noxious stimulus–evoked responses, we devised an autologous IVD autograft and spinal nerve ligation model. The right L5 spinal nerve was approached via a midline incision, and the transverse process was removed under sodium pentobarbital anesthesia (50 mg/kg, administered intraperitoneally [IP]) under sterile conditions. Streptomycin (20 mg/kg) was administered IP to all rats for the first 3 days after surgery. The animals were divided into 3 groups. In the first group, one-third to one-half of the right L5 spinal nerve was ligated with a 10-0 nylon suture (Mani) just outside the foramen, and the coccygeal IVD was removed (nerve compression group; n = 6). In the second group, spinal nerve ligation was performed in a manner similar to that in the nerve compression group, and a coccygeal IVD was autografted onto the ligated spinal nerve (disc plus nerve compression group; n = 6). In the third group, the right L5 spinal nerve was exposed and the coccygeal IVD was removed (sham-operated group; n = 6).

Behavioral test—examinations of mechanical and thermal hyperalgesia.

Paw withdrawal thermal latency and paw withdrawal mechanical threshold of the hind limbs were determined before and after surgery. To examine paw withdrawal thermal latency, we used the hot plate method (27). The rats were individually placed on the surface of a hot aluminum plate (Tokyo Glass Kikai), which was maintained at a constant temperature (mean ± SD 54 ± 0.5°C), and the latency period until the rat showed the first noxious stimulus–evoked response (e.g., licking the hind paw, stamping, or jumping) was noted. For the examination of paw withdrawal mechanical threshold, we used the previously reported 50% paw withdrawal mechanical threshold method (28) with the von Frey hair (Stoelting). The changes in paw withdrawal thermal latency and paw withdrawal mechanical threshold were calculated as percentage differences before and after surgery: PWTLpost/PWTLpre × 100 and PWMTpost/PWMTpre × 100, where PWTL = paw withdrawal thermal latency and PWMT = paw withdrawal mechanical threshold.

Intramuscular autologous IVD autograft model.

To determine macrophage infiltration, cytokine production, and PGE2 expression in IVD autografts, 4 coccygeal IVDs were obtained from a rat tail and autografted into its back muscle (n = 42). We euthanized 6 rats on postoperative days 0, 1, 3, 5, 7, 11, and 14 and collected 4 IVD autografts. Three of these IVDs (mean ± SD 45 ± 2.2 mg) were frozen in liquid nitrogen and preserved at −80°C for reverse transcription–polymerase chain reaction (RT-PCR), and 1 was preserved at −30°C for immunohistochemical analysis.

Quantitative determination of macrophages in disc autografts.

The interstitial CD68+ macrophages in IVD autografts, which had been collected from the back muscles of the rats, were immunohistochemically stained, counted in 10 consecutive high-power fields (hpf; 400×), and expressed as cells per hpf.

Quantitative real-time RT-PCR.

Preserved IVD autografts were used for messenger RNA (mRNA) quantification. Total RNA was isolated from the IVD autograft using the RNeasy kit (Qiagen); complementary DNA was synthesized using MultiScribe reverse transcriptase (Applied Biosystems), and quantitative real-time RT-PCR was performed using the PRISM 7500 sequence detector (Applied Biosystems). The levels of TNFα, IL-1β, IL-6, IL-8, cyclooxygenase 2 (COX-2), nerve growth factor (NGF), and MCP-1 mRNA were quantified relative to the mRNA level of GAPDH (Applied Biosystems). Each sample was analyzed in triplicate, and each data point was calculated as the average of the 3 findings. Quantification was performed using the 2math image technique (29). The relative mRNA level was compared with the level of mRNA in IVD autografts on postoperative day 14.

Tissue coculture of IVDs and macrophages.

A 35-ml solution of thioglycolate medium (BD Diagnosis System) was injected IP into male Sprague-Dawley rats (n = 24). Four days after the injection, the rats were euthanized, and the peritoneal space was rinsed twice with 60 ml phosphate buffered saline (PBS) in order to collect exudate macrophages. Immunohistochemical staining showed that >90% of these macrophages were resident macrophages (data not shown). Coccygeal IVDs were harvested at the same time. Coccygeal IVDs (20 mg) were cocultured with macrophages (106/ml) in 2 ml serum-free Opti-MEM (Invitrogen) supplemented with Penicillin–Streptomycin Mixed Solution (20 units penicillin, 20 μg streptomycin) (Nacalai Tesque) in a 24-well cell culture cluster (Corning). The samples were incubated in a humidified atmosphere of 5% CO2 at 37°C.

Twenty-four–hour coculture study.

We cocultured IVDs and macrophages for 24 hours. Macrophages (106/ml) cultured alone and subcutaneous fat tissue (20 mg) cultured with or without macrophages (106/ml) were used as controls. Supernatants of the culture medium and precipitates were collected and preserved at −20°C and −80°C, respectively.

Time course study.

To determine the diurnal levels of TNFα, IL-6, IL-8, and PGE2 in the supernatants of IVD and macrophage coculture media, we collected the supernatants after 0, 3, 6, 12, 18, and 24 hours of incubation. Each supernatant sample was collected from a different coculture well. The production levels were compared between the 0 hour and given time points.

TNFα neutralization.

To clarify the role of TNFα in the IVD and macrophage coculture, we added a neutralizing anti-TNFα polyclonal antibody (Innogenetics) to the culture medium and incubated it for 24 hours before collecting the supernatant.

Immunohistochemical staining procedure.

To clarify the localization of CD68+ macrophages and IL-8 or COX-2 in culture precipitates and IVD autografts, the precipitates and autografts were cryosectioned, yielding 5 μm–thick sections. Nonspecific binding was blocked by treatment with 10% normal goat serum, and the sections were incubated for 3 hours with the following antibodies: 1 μg/ml mouse monoclonal anti-rat CD68 antibody (Serotec) and 0.5 μg/ml rabbit anti-rat IL-8 antibody (Panafarm) or 5 μg/ml goat anti-rat COX-2 polyclonal antibody (Santa Cruz Biotechnology). After incubation with the primary antibodies, the sections were washed 3 times with PBS and incubated for 1 hour with fluorescence-labeled secondary antibodies (10 μg/ml goat anti-mouse IgG antibody or 10 μg/ml goat anti-rabbit IgG antibody or 10 μg/ml donkey anti-goat IgG antibody; all from Molecular Probes).

Conventional RT-PCR for IL-8 and COX-2.

To confirm the production of IL-8 and COX-2 in the IVD and macrophage coculture, preserved culture precipitates were used for comparing the mRNA levels using conventional procedures. The sequences of the primers used for PCR were as follows: for IL-8, sense 5′-CCCCATGGTTCAGAAGATTG-3′ and antisense 5′-TTGAACGACCATCGATGAAA-3′ (270-bp product); for COX-2, sense 5′-GCAAATCCTTGCTGTTCCAATC-3′ and antisense 5′-GGAGAAGGCTTCCCAGCTTTTG-3′ (335-bp product); for GAPDH, sense 5′-ACCACAGTCCATGCCATCAC-3′ and antisense 5′-TCCACCACCCTGTTGCTGTA-3′ (451-bp product). The PCR conditions were 30 cycles of denaturation at 94°C for 1 minute; annealing at 55°C for 45 seconds (IL-8), 60°C for 30 seconds (COX-2), and 50°C for 1 minute (GAPDH); and extension at 72°C for 1 minute. The PCR products were separated by electrophoresis in 2% agarose gels and stained with Vista green (Amersham Pharmacia Biotech). The gels were observed using the LAS-3000mini (Fujifilm).

Enzyme-linked immunosorbent assay (ELISA) for TNFα, IL-6, IL-8, and PGE2.

The coculture supernatant was collected and ELISA was performed using commercially available kits for TNFα (BioSource International), IL-6 (R&D Systems), IL-8 (Panafarm), and PGE2 (Techne) according to the manufacturers' instructions.

Administration of neutralizing anti-TNFα antibody and/or neutralizing anti–IL-8 antibody in autologous IVD autograft and spinal nerve ligation model.

To clarify the importance of TNFα and IL-8 in herniated disc disease, we administered a neutralizing anti-TNFα antibody and/or a neutralizing anti–IL-8 antibody in the autologous IVD autograft and spinal nerve ligation model (n = 32). On the day after the surgery, the animals were divided into 4 groups, and saline or the following antibodies were gently injected around the IVD autograft: 3.1 μg/ml anti-TNFα antibody (Abcam), 10 μg/ml anti–IL-8 antibody (R&D Systems), or anti-TNFα antibody/anti–IL-8 antibody cocktail. Antibody concentration was determined by the 24-hour IVD and macrophage coculture experiment and the median neutralization dose of the neutralizing antibody. The paw withdrawal mechanical threshold was measured 3, 5, 7, 11, and 14 days after the surgery. The percentage of paw withdrawal mechanical threshold was compared between the group receiving saline and the groups receiving antibodies at the given time points.

Statistical analysis.

In the behavioral test, all data were expressed as the mean ± SEM. The statistical differences within a group were analyzed by Friedman's test, followed by the Wilcoxon matched pairs signed rank test. The statistical differences between groups were analyzed by two-way repeated-measures analysis of variance, followed by Scheffe's test. The difference between groups at a given postoperative testing time point was further compared by the Mann-Whitney U test. P values less than 0.05 were considered significant. For quantitative determination of macrophages in an autograft model, quantitative real-time PCR and ELISA were performed, and the statistical significance of differences between data points was determined by Student's t-test. All computations were performed using PASW Statistics 18 (SPSS).

RESULTS

Effects of IVD autografting and spinal nerve compression demonstrated by behavioral tests.

The disc plus nerve compression group and the nerve compression group showed significant decreases in paw withdrawal mechanical threshold over time. Compared with baseline values, the rats in the disc plus nerve compression group showed a significant and prolonged decrease in paw withdrawal mechanical threshold of the ipsilateral paw from postoperative days 1 to 14, which was indicative of mechanical hyperalgesia. The disc plus nerve compression group showed marked allodynia compared with that in the nerve compression group from postoperative days 1 to 5. The mechanical threshold of the disc plus nerve compression group was the lowest on postoperative day 5, but it continuously increased thereafter (Figure 1). In contrast, none of the 3 groups showed any differences in paw withdrawal thermal latency from postoperative days 1 to 14 (data not shown). These results indicate that IVDs autografted on the ligated spinal nerve induce more mechanical hyperalgesia than nerve root ligation alone. The mechanical threshold of the disc plus nerve compression group was minimal on postoperative day 5.

Figure 1.

Changes over time in percentage paw withdrawal mechanical threshold (PWMT) of the operated paw in the sham-operated (diamonds), nerve compression (squares), and disc plus nerve compression (triangles) groups. The mechanical threshold of the disc plus nerve compression group was the lowest on postoperative day 5. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01 versus preoperative (pre Op) threshold. † = P < 0.05; †† = P < 0.01 versus nerve compression group.

IVD autograft and macrophage infiltration.

The in vivo IVD autograft model showed extensive macrophage infiltration into the IVD autograft from postoperative days 1 through 14. The number of macrophages peaked on postoperative day 7 (Figure 2A).

Figure 2.

A, Number of interstitial CD68+ macrophages in intervertebral disc (IVD) autografts in 10 consecutive high-power fields (hpf) (400×). Extensive macrophage infiltration into the IVD autograft is seen from postoperative day 1. Values are the mean ± SEM. ∗∗ = P < 0.01. BH, Quantitative real-time reverse transcription–polymerase chain reaction analysis showing relative quantities (Rel. Quant.) of mRNA for tumor necrosis factor α (TNFα) (B), interleukin-1β (IL-1β) (C), IL-6 (D), IL-8 (E), cyclooxygenase 2 (COX-2) (F), nerve growth factor (NGF) (G), and monocyte chemotactic protein 1 (MCP-1) (H) in IVD autografts. TNFα, IL-1β, IL-8, COX-2, and NGF levels were maximal on postoperative day 1, and IL-6 and MCP-1 levels were maximal on postoperative day 3. Values are the mean ± SEM. ∗∗ = P < 0.01 versus postoperative day 14.

Cytokine and COX-2 mRNA expression in IVD autografts.

To gain a better understanding of the IVD autograft–induced “inflammatory” reaction, we used the real-time RT-PCR method; mRNA expression levels were compared with those on postoperative day 14. Healthy untreated IVDs expressed COX-2 mRNA (68% of that on postoperative day 14) and IL-8 mRNA (15% of that on postoperative day 14), while TNFα, IL-1β, IL-6, NGF, and MCP-1 mRNA were undetectable. Coccygeal IVD autografts showed marked up-regulation of TNFα, IL-1β, IL-6, IL-8, COX-2, and NGF mRNA from postoperative days 1 through 3. Compared with postoperative day 14, on postoperative day 1 TNFα mRNA was up-regulated 37.2-fold (Figure 2B), IL-1β mRNA was up-regulated 4.1-fold (Figure 2C), IL-8 mRNA was up-regulated 47.5-fold (Figure 2E), COX-2 mRNA was up-regulated 92.4-fold (Figure 2F), and NGF mRNA was up-regulated 3.2-fold (Figure 2G). Compared with postoperative day 14, on postoperative day 3 IL-6 mRNA was up-regulated 18.7-fold (Figure 2D). MCP-1 mRNA expression in the IVD autografts was maximal on postoperative day 3, but it was not significantly different from the expression on postoperative day 14 (Figure 2H).

The in vivo results can be summarized as follows. Extensive macrophage infiltration into the IVD autografts was observed from postoperative day 1. A significant decrease was noted in the mechanical threshold of the ipsilateral paw following the up-regulation of TNFα, IL-6, IL-8, and COX-2 mRNA.

Effects of IVD–macrophage interaction on TNFα, IL-8, and PGE2 production.

To reconfirm the IVD–macrophage interaction in herniated disc disease, we used an IVD and macrophage coculture model. To clarify the differences between IVDs and other tissues, we compared macrophage cocultures with IVDs and macrophage cocultures with other tissues. We have previously reported significantly higher IL-6 up-regulation only in IVD–macrophage coculture and not in macrophage cocultures with any other tissues (26). In the present study, we found that coculturing of IVDs and macrophages for 24 hours up-regulated IL-8 and PGE2 to levels greater than those found by culturing IVDs alone, macrophages alone, or fat tissue either alone or with macrophages (Figures 3A and B). TNFα levels did not differ among the groups (data not shown). Conventional RT-PCR results showed greater IL-8 and COX-2 mRNA up-regulation in IVD and macrophage coculture than in cultures of IVDs alone or macrophages alone (Figure 3C). These and our previous results indicate that only IVDs and macrophages induce significant production of IL-6, IL-8, and PGE2, but not of TNFα, when cocultured for 24 hours.

Figure 3.

A and B, Interleukin-8 (IL-8) (A) and prostaglandin E2 (PGE2) (B) production in the intervertebral disc (IVD) and macrophage coculture medium supernatants. IVD and macrophage coculturing resulted in significant production of IL-8 and PGE2. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01. C, Conventional reverse transcription–polymerase chain reaction for IL-8 and cyclooxygenase 2 (COX-2) mRNA levels in the IVD and macrophage coculture precipitates.

To determine the sources of IL-8 and COX-2, we performed immunohistochemical staining of the IVD and macrophage coculture precipitates. Macrophages were predominantly deposited in the peripheral areas of the nucleus pulposus (Figures 4A and D). IL-8 and COX-2 were mainly coexpressed by the peripherally deposited macrophages (Figures 4C and F). We previously found that IL-6 is mainly produced by macrophages (26). The results of these immunohistochemical analyses indicate that peripherally deposited macrophages are the source of IL-8 and PGE2 in the IVD and macrophage cocultures.

Figure 4.

Immunohistochemical staining of interleukin-8 (IL-8) (A–C) and cyclooxygenase 2 (COX-2) (D–F) in the intervertebral disc and macrophage coculture and in the nucleus pulposus. Peripherally deposited macrophages on the surface of the nucleus pulposus were visualized as fluorescent red using an anti-rat CD68 antibody (A and D). IL-8 (B) and COX-2 (E) were abundant on peripherally deposited cells and nucleus pulposus cells. Colocalization images of CD68 with IL-8 (C) or COX-2 (F) showed that IL-8 and COX-2 were mainly coexpressed by the peripherally deposited macrophages. Bars = 10 μm.

Time courses of TNFα, IL-6, IL-8, and PGE2 production in IVD and macrophage coculture.

Compared with TNFα levels 0 hours after coculture, TNFα levels were significantly up-regulated for the first 6 hours; however, TNFα levels diminished thereafter (Figure 5A). IL-6 and IL-8 levels were up-regulated over time (Figures 5B and C). PGE2 levels were up-regulated for the first 6 hours and then maintained at the same level until 24 hours (Figure 5D). These results indicate that the levels of TNFα peak before those of IL-6, IL-8, and PGE2 in IVD and macrophage coculture.

Figure 5.

Production of tumor necrosis factor α (TNFα), interleukin-6 (IL-6), IL-8, and prostaglandin E2 (PGE2) in the coculture experiments. TNFα (A), IL-6 (B), IL-8 (C), and PGE2 (D) were measured in the supernatant 24 hours after the coculture. TNFα levels peaked before those of IL-6, IL-8, and PGE2 in the intervertebral disc and macrophage coculture medium. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01 versus the 0 hour time point.

Neutralizing effect of anti-TNFα on IVD and macrophage coculture.

To elucidate the pivotal role of TNFα in more detail, we added neutralizing anti-TNFα antibody to IVD and macrophage coculture. The neutralizing anti-TNFα antibody diminished the expression of IL-6 and PGE2 in a dose-dependent manner (Figures 6A and C). However, it did not affect IL-8 expression in IVD and macrophage coculture (Figure 6B).

Figure 6.

AC, Modulation of interleukin-6 (IL-6) (A), IL-8 (B), and prostaglandin E2 (PGE2) (C) production in the intervertebral disc (IVD) and macrophage coculture by administration of neutralizing anti–tumor necrosis factor α (anti-TNFα) antibody. IL-6, IL-8, and PGE2 are expressed as percentages in the supernatant at 24 hours of IVD and macrophage coculture. Values are the mean ± SEM. ∗∗ = P < 0.01 versus no anti-TNFα antibody. D, Percentage paw withdrawal mechanical threshold (PWMT) of the ipsilateral paw in groups receiving anti-TNFα antibody/anti–IL-8 antibody cocktail (×), anti-TNFα antibody alone (▴), anti–IL-8 antibody alone (▪), and saline (♦). Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01 versus saline-treated group. pre Op = preoperative.

Neutralizing effect of anti-TNFα antibody and/or anti–IL-8 antibody in the autologous IVD autograft and spinal nerve ligation model.

From 3 to 14 days after surgery, the paw withdrawal mechanical threshold in the group that received the anti-TNFα antibody/anti–IL-8 antibody cocktail was significantly higher than that in the group that received saline. The paw withdrawal mechanical threshold was higher in the anti–TNFα antibody–treated group than in the saline-treated group at 5 and 7 days after surgery. The paw withdrawal mechanical threshold of the group that received the anti-TNFα antibody/anti–IL-8 antibody cocktail surpassed that of the anti-TNFα antibody–treated group. The paw withdrawal mechanical threshold of the group that received the anti–IL-8 antibody was not significantly different from that of the saline-treated group (Figure 6D).

DISCUSSION

Our study led to 8 findings that contribute to our understanding of the pathomechanism of herniated disc disease. First, IVD autografting onto a ligated spinal nerve reduced the paw withdrawal mechanical threshold more than spinal nerve ligation alone. Second, macrophages infiltrated into the IVD autograft soon after the autograft had been implanted. Third, cytokine and COX-2 mRNA were expressed in IVD autografts from an early stage. Fourth, only the 24-hour coculture of IVDs and macrophages induced up-regulation of IL-8 and PGE2. Fifth, macrophages were the main source of IL-6, IL-8, and PGE2 in the IVD and macrophage coculture. Sixth, IVD and macrophage coculture induced early up-regulation of TNFα, followed by up-regulation of IL-6, IL-8, and PGE2. Seventh, anti-TNFα antibody attenuated IL-6 and PGE2 production induced by IVD and macrophage coculture, but not IL-8 production. Eighth, the anti-TNFα antibody/anti–IL-8 antibody cocktail significantly increased the paw withdrawal mechanical threshold in the autologous IVD autograft and spinal nerve ligation model. These results indicate that 1) IVD–macrophage interaction plays a major role in TNFα, IL-6, IL-8, and PGE2 production and correlates with mechanical hyperalgesia involved in the development of sciatica in herniated disc disease and 2) TNFα plays a pivotal role in IL-6 and PGE2 production, but not in that of IL-8, in IVD–macrophage coculture.

In the present study, IVD autografting onto a ligated spinal nerve produced more significant mechanical hyperalgesia than did spinal nerve ligation alone. Previous reports have shown that an IVD autografted onto a ligated nerve root produces mechanical hyperalgesia several days after the surgery (30, 31). Thus, the IVD autograft is an inflammatory material causing mechanical hyperalgesia. In the present study, IVD autografts induced the up-regulation of cytokines and COX-2 mRNA, followed by mechanical hyperalgesia. Among the cytokines, TNFα, IL-6, IL-8, and PGE2 are closely related to pain. TNFα (32–34), IL-6 (32), IL-8 (32, 34), and PGE2 (35) injection into rat hind paws evokes hyperalgesia, which is attenuated by administration of the respective antisera. TNFα application along the sciatic nerve elicits higher mean firing of C fibers than Aδ fibers, which results in a wind-up in dorsal horn neurons (33). PGE2 contributes to peripheral sensitization by binding to G protein–coupled receptors that increase the levels of cyclic AMP within nociceptors (35). PGE2 increases the excitability of rat sensory neurons, in part by shifting the voltage dependence of tetrodotoxin-resistant sodium channel activation in the hyperpolarizing direction (36). These reports indicate that IVD autografts produce TNFα, IL-6, IL-8, and PGE2 and that these cytokines may contribute to pain transduction, conduction, and modulation.

Furthermore, spinal nerve ligation induces TNF receptor type I and type II up-regulation in dorsal root ganglia, which peaks at 6 hours after ligation (37). Spinal nerve ligation also induces significant increases in PGE2 receptor (subtypes 1–3) expression in the ipsilateral lumbar dorsal horn (38). Overall, these reports suggest that IVD autograft–derived TNFα, IL-6, IL-8, and PGE2 production possibly correlates with mechanical hyperalgesia in the IVD autograft and spinal nerve ligation model.

In the healthy state, the disc tissue is avascular and isolated from the host immune system. However, in human degenerated nonherniated discs, immunocompetent CD68+ cells are frequently observed around the cleft (39). The interaction of CD68+ macrophages and IVDs can play an important role in extracellular matrix resorption via matrix metalloproteinase 3 induction (40). Thus, the interaction of CD68+ macrophages and IVDs might play an important role in disc degeneration, which can result in herniated disc disease. We autografted healthy IVDs onto ligated nerves but not onto degenerated ones, but CD68+ macrophage infiltration in IVDs provides the possibility of restaging one of the major etiologic factors of degenerative disc disease. In surgically obtained herniated disc tissue, the levels of IL-6, IL-8, and PGE2 increase with increasing exposure of the nucleus pulposus to the host immune system (10–12, 17). Our previous results and the coculture results of the present study clearly show that the interaction between IVDs and macrophages can induce IL-6 (26), IL-8 (Figures 3A and C), and PGE2 (Figures 3B and C) and that they originate from macrophages (26) (Figure 4).

The presence of TNFα in herniated disc tissue has long been controversial; while some studies have shown TNFα in herniated disc tissue (8, 15), others have shown opposite findings (9, 10, 12, 14). Our results indicate that TNFα appears transiently in the relatively early stage of IVD–macrophage interaction, suggesting that TNFα expression might vary at different stages of herniated disc disease. IVD–macrophage interaction provides one explanation for the changes in TNFα expression in herniated disc tissue. TNFα plays an important role in regulating the activities of other cytokines in inflammatory diseases such as rheumatoid arthritis (for review, see ref.41), Crohn's disease (for review, see ref.42), and psoriasis (for review, see ref.43). In these diseases, macrophages play a major role in the pathogenic recruitment of cytokines as a result of TNFα stimulation. Further, macrophages produce IL-6 (4) or PGE2 (44, 45) as a result of TNFα stimulation. Taken together, these observations suggest that TNFα might play a pivotal role in herniated disc disease–induced inflammation.

However, regardless of the potential importance of TNFα in IVD–macrophage interaction, a randomized clinical trial in patients with herniated disc disease showed no efficacy of recombinant anti-TNFα antibody compared with the control group after 3 months or 1 year of administration (46). One possible explanation is the existence of a TNFα-independent inflammatory cascade. Our results indicated anti-TNFα–independent IL-8 expression in the IVD and macrophage coculture (Figure 6B). In addition, the paw withdrawal mechanical threshold was higher following administration of an anti-TNFα antibody/anti–IL-8 antibody cocktail in the autologous IVD autograft and spinal nerve ligation model than following administration of anti-TNFα antibody alone or anti–IL-8 antibody alone (Figure 6D). TNFα-independent cytokine production has been indicated by several studies. Low-dose Neisseria gonorrhoeae infection of a macrophage cell line induced IL-8 production without TNFα induction (47). Ultraviolet irradiation of a keratinocyte-derived cell line induces vascular endothelial growth factor production through a TNFα-independent pathway (48). A thrombin study using human umbilical vein endothelial cells showed that IL-8 expression and E-selectin secretion were induced through a TNFα-independent pathway (49). A similar TNFα-independent IL-8 production mechanism might exist in IVD–macrophage interaction. Thus, TNFα-independent IL-8 expression in IVD–macrophage interaction may explain the ineffectiveness of an anti-TNFα antibody in treatment of herniated disc disease.

A limitation of our study is that we used only healthy IVDs, not degenerated ones. Furthermore, IVDs of rodents consist of notochordal cells, which vanish in human adult IVDs. Thus, our models cannot be used to restage the causes of herniated disc disease. In addition, we did not clarify the role of TNFα, IL-6, IL-8, and PGE2 in pain-related behavior in our autologous IVD autograft and spinal nerve ligation model. IL-1β stimulation can induce IL-6 production in herniated disc samples (19). High levels of IL-1β mRNA are expressed not only in herniated discs but also in degenerated discs (50). Our results showed IL-1β mRNA up-regulation in IVD autografts (Figure 2C). Thus, IL-1β may play an important role not only in disc degeneration but also in herniated disc disease. Further research is warranted for a better understanding of the cytokine network in IVD–macrophage interaction and the effect of its inflammatory properties on nociceptive pain.

In conclusion, this study provides new and important information on inflammation in herniated disc disease, resulting in sciatica caused by IVD–macrophage interaction. IVD–macrophage interaction results in TNFα, IL-6, IL-8, and PGE2 production, which correlates with pain-related behavior; TNFα plays a pivotal role in the production of IL-6 and PGE2 but not in that of IL-8. Administration of an anti-TNFα antibody/anti–IL-8 antibody cocktail can be a valuable therapy for herniated disc disease.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Takada 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 conception and design. Takada.

Acquisition of data. Takada.

Analysis and interpretation of data. Takada, Nishida, Maeno, Kakutani, Yurube, Doita, Kurosaka.

Acknowledgements

The authors would like to thank Ms Janina Tubby for help with the preparation of the manuscript.

Ancillary