Tumor necrosis factor α– and interleukin-1β–dependent induction of CCL3 expression by nucleus pulposus cells promotes macrophage migration through CCR1




To investigate tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β) regulation of CCL3 expression in nucleus pulposus (NP) cells and in macrophage migration.


Quantitative reverse transcription–polymerase chain reaction and immunohistochemistry were used to measure CCL3 expression in NP cells. Transfections were used to determine the role of NF-κB, CCAAT/enhancer binding protein (C/EBPβ), and MAPK on cytokine-mediated CCL3 promoter activity. The effect of NP-conditioned medium on macrophage migration was measured using a Transwell system.


An increase in CCL3 expression and promoter activity was observed in NP cells after TNFα or IL-1β treatment. Treatment of cells with NF-κB and MAPK inhibitors abolished the effect of the cytokines on CCL3 expression. The inductive effect of p65 and C/EBPβ on the CCL3 promoter was confirmed through gain-of-function and loss-of-function studies. Notably, cotransfection with p50 completely blocked cytokine- and p65-dependent induction. In contrast, c-Rel and RelB had little effect on promoter activity. Lentiviral transduction with short hairpin RNA for p65 (shp65) and shIKKβ significantly decreased the TNFα-dependent increase in CCL3 expression. Analysis of degenerated human NP tissue samples showed that CCL3, but not CCL4, expression correlated positively with the grade of tissue degeneration. Importantly, treatment of macrophages with conditioned medium of NP cells treated with TNFα or IL-1β promoted their migration. Pretreatment of macrophages with an antagonist of CCR1, the primary receptor for CCL3 and CCL4, blocked cytokine-mediated migration.


Our findings indicate that TNFα and IL-1β modulate the expression of CCL3 in NP cells by controlling the activation of MAPK, NF-κB, and C/EBPβ signaling. The CCL3–CCR1 axis may play an important role in promoting macrophage infiltration in degenerated, herniated discs.

The intervertebral disc (IVD) is a unique tissue that permits rotation, as well as flexion and extension, of the spine. It consists of a gel-like nucleus pulposus (NP) surrounded circumferentially by a fibrocartilaginous annulus fibrosus (AF). NP cells secrete a complex extracellular matrix that contains fibrillar collagens and the proteoglycan aggrecan. The initial phase of disc degeneration is characterized by increased expression of catabolic enzymes, decreased proteoglycan synthesis, and an overall shift toward synthesis of a fibrotic matrix and events that compromise the structural integrity of the tissue (1–4). Structural failure of the NP and AF leads to herniation of NP tissue, which is often followed by an inflammatory phase, characterized by invasion of immune cells in the tissue (2, 5, 6).

It has been reported that during degeneration, resident NP and AF cells produce high levels of the cytokines tumor necrosis factor α (TNFα) and interleukin-1β (IL-1β) (7, 8). These cytokines stimulate the production of nerve growth factor, brain-derived neurotrophic factor, and vascular endothelial growth factor, which are molecules associated with nerve ingrowth into the NP and angiogenesis (9). Moreover, in response to high cytokine levels, disc cells also produce chemoattractive proteins such as monocyte chemotactic protein 1 and IL-8 (10). However, mechanisms that control the expression of these chemokines during disc degeneration have received little attention.

Chemokines and their receptors have been shown to be involved in many inflammatory diseases, including rheumatoid arthritis (RA) and osteoarthritis (11, 12). Of the chemokine receptors, CCR1 is directly linked to the pathogenesis of RA. Moreover, a recent study showed that the inflammatory cytokine IL-1β induced the expression of CCL3 and CCL4 in human chondrocytes (13). High levels of CCR1-expressing macrophages and the chemokines CCL3 and CCL4 have been identified in RA synovial fluid and tissue (14–17). In vitro migration studies have shown that CCR1-mediated monocyte migration induced by RA synovial fluid can be blocked with either a CCR1-blocking antibody or a small-molecule CCR1 antagonist (18). A clinical study using a specific CCR1 antagonist in patients with RA has confirmed the potential of this approach (15).

While CCL3 has been reported to be expressed in herniated IVDs (10), it was noted that reactivity was associated with fibroblasts, endothelial cells, and infiltrating macrophages in the granulation tissue. Aside from that study, little is known about the expression and regulation of CCL3 in NP cells during disc degeneration. Since disc cells are known to mount a robust inflammatory response, we advance the notion that secretion of chemokines such as CCL3 by NP cells in response to inflammatory cytokines promotes tissue infiltration of macrophages and T cells. Herein, we show for the first time that TNFα and IL-1β control CCL3 transcription in NP cells in an MAPK-, NF-κB–, and CCAAT/enhancer binding protein β (C/EBPβ)–dependent manner. Importantly, our results show that CCL3, through its receptor CCR1, may play an important role in promoting the cytokine-dependent migration of macrophages into the disc and exacerbation of the inflammatory state.


Reagents and plasmids.

Human CCL3 promoter constructs were a kind gift from Dr. Linda Sandell (Washington University, St. Louis, MO). The pCMX-IκBαM (catalog no. 12330), RelA/p65-cFLAG-pcDNA3 (catalog no. 20012), and p50-cFLAG-pcDNA3 (catalog no. 20018) were a gift from Dr. Inder Verma (Salk Institute of Biological Sciences, La Jolla, CA), pCMV-FLAG-LAP-2 (catalog no. 15738) and pCMV-FLAG-LIP (catalog no. 15737) were a gift from Dr. Joan Massague (Memorial Sloan-Kettering Cancer Center, New York, NY), and psPAX2 (catalog no. 12260) and pMD2G (catalog no. 12259) were a gift from Dr. Didier Trono (School of Life Sciences, EPFL, Lausanne, Switzerland). RelB-cFLAG-pcDNA3 (catalog no. 20017) and c-Rel-cFLAG-pcDNA3 (catalog no. 20013) were obtained from the Addgene repository. Plasmids short hairpin RNA for p65 (shp65) and shIKKβ in a lentiviral FSVsi vector that coexpresses yellow fluorescent protein (YFP) were kindly provided by Dr. Andree Yeremian (University of Lleida, Lleida, Spain). As an internal transfection control, vector pRL-TK (Promega) containing the Renilla reniformis luciferase gene was used. The transfection methodology has been optimized for rat NP cells (19). Wild-type and p65-null cells were a kind gift from Dr. Denis Guttridge (Ohio State University, Columbus, OH). Antibodies against phospho-p38, p38, ERK-1/2, and phospho–ERK-1/2 were from Cell Signaling Technology. Antibodies against CCL3, CCL4, and CD11b were from Abcam, β-tubulin was from the Developmental Studies Hybridoma Bank, University of Iowa, and GAPDH was from Novus Biologicals. TNFα and IL-1β were purchased from PeproTech.

Harvesting of rat disc tissue, isolation of NP cells, and treatments.

NP and AF tissue and NP cells were isolated from adult Wistar rats (weighing 350 gm) using a method previously described by Risbud et al (19). Rat NP cells were maintained in Dulbecco's modified Eagle's medium (DMEM) and 10% fetal bovine serum (FBS) supplemented with antibiotics. To investigate the effect of cytokines, cells were treated with IL-1β (5–20 ng/ml) and TNFα (25–100 ng/ml) for 24 hours.

Human tissue collection and grading.

Human lumbar IVD tissue was obtained either from patients undergoing surgery or from donors at postmortem examination, with informed consent of the patient or relatives (Sheffield Research Ethics Committee [09/H1308/70]). Eight IVDs were recovered postmortem from 2 donors. They consisted of intact IVDs within the complete motion segment from which the IVDs were removed. Thirty-one IVD tissue samples were obtained from patients undergoing microdiscectomy procedures for the treatment of low back pain and root pain as caused by prolapse of the IVD (details regarding the tissue samples are available from the corresponding author upon request). NP tissue was divided into 2 sections, one of which was fixed in 10% neutral buffered formalin and processed for histologic and immunohistochemical examination. The remaining tissue was used for RNA isolation. Hematoxylin and eosin–stained sections were used to score the degree of morphologic degeneration according to previously published criteria (7). Briefly, tissue sections were scored for the presence of cell clusters, the presence of fissures, loss of demarcation, and loss of hematoxophilia. A score of 0–3 indicates a histologically normal (nondegenerated) IVD, and a score of 4–12 indicates evidence of degeneration. Tissue samples were also examined for the presence of infiltrating cells based on both cell morphology and CD11b staining. Multiple sections throughout each tissue block were examined for infiltrating cells and classified as infiltrated if these were observed on any section. Gene expression study samples were classified as nondegenerated (grade 0–3), degenerated but free of infiltrating cells (grade 4+), or infiltrated based on histologic examination. Grading was performed independently by 2 researchers, and the grades were averaged.

Real-time quantitative reverse transcription–polymerase chain reaction (RT-PCR).

Human tissue samples were processed as previously described (7). Extracted RNA was treated with DNase (Qiagen) and purified using a Qiagen MinElute Cleanup kit prior to complementary DNA (cDNA) synthesis using Moloney murine leukemia virus reverse transcriptase (Bioline) and random hexamers. Real-time PCR analysis was performed using predesigned, FAM-labeled TaqMan Gene Expression Assays (Applied Biosystems). A total of 35 IVDs were used for this component of the study (details are available from the corresponding author upon request). These included 6 nondegenerated samples (2 samples obtained postmortem and 4 samples obtained during surgery) from patients with a mean age of 37 years (range 23–45 years), 16 degenerated samples (all obtained during surgery) from patients with a mean age of 39 years (range 25–52 years), and 13 infiltrated samples (all obtained during surgery) from patients with a mean age of 38 years (range 20–63 years). For rat tissue or cultured NP cells, 1–2 μg of total DNA-free RNA was used to synthesize cDNA using a SuperScript III cDNA synthesis kit (Invitrogen). Reactions were set up in triplicate in 96-well plates using cDNA and the appropriate PCR Master Mix (Applied Biosystems). Reactions were performed in a StepOnePlus real-time PCR system (Applied Biosystems) according to the manufacturer's instructions. Predesigned human primer/probe mixtures (GAPDH, 18S, CCL3, and CCL4) were purchased from Applied Biosystems. Rat gene primers were synthesized by Integrated DNA Technologies.

Transfections and dual-luciferase assay.

Rat NP cells were transferred to 48-well plates at a density of 2 × 104 cells/well 1 day before transfection. To investigate the effect of NF-κB and C/EBPβ/liver-enriched transcriptional activator protein 2 (LAP-2) on CCL3 promoter activity, rat NP cells were cotransfected with 50–150 ng of IκBαM, p65, p50, p65 plus p50, RelB, c-Rel, LAP-2, or dominant-negative (DN) C/EBPβ/liver-enriched inhibitory protein (LIP) with or without appropriate backbone vector and 175 ng CCL3 reporter and 175 ng pRL-TK plasmid. In some experiments, rat cells were transfected with 250 ng of CCL3 reporter with 250 ng pRL-TK and treated with inhibitors of NF-κB (SM7368; 10 μM), p38 (SB203580; 10 μM), ERK (PD98059; 10 μM), or JNK (SP60025; 10 μM) (all from Calbiochem) in the presence of absence of TNFα or IL-1β. Lipofectamine 2000 (Invitrogen) was used as a transfection reagent. Forty-eight hours after transfection, the cells were harvested, and a Dual-Luciferase Reporter Assay system (Promega) was used for sequential measurements of firefly and Renilla luciferase activities. At least 3 independent transfections were performed, and all analyses were carried out in triplicate.

Lentiviral particle production and viral transduction.

HEK 293T cells were seeded in 10-cm plates (1.3 × 106 cells/plate) in DMEM with 10% heat-inactivated FBS 2 days before transfection. Cells were transfected with 2.5 μg of control shRNA, shp65, or shIKKβ plasmids along with 1.875 μg psPAX2 and 0.625 μg pMD2.G. After 16 hours, transfection medium was removed and replaced with DMEM with 5% heat-inactivated FBS and penicillin/streptomycin. Lentiviral particles were harvested 48 hours and 60 hours after transfection. Human NP cells were plated in DMEM with 5% heat-inactivated FBS 1 day before transduction. Cells in 10-cm2 plates were transduced with 5 ml of conditioned media containing viral particles along with 6 μg/ml Polybrene. After 24 hours, conditioned media were removed and replaced with DMEM with 5% heat-inactivated FBS. Cells were harvested for protein extraction 5 days after viral transduction.

Protein extraction and Western blotting.

Rat and human cells were placed on ice immediately following treatment and washed with ice-cold Hanks' balanced salt solution. All wash buffers and the final resuspension buffer included 1× protease inhibitor cocktail (Pierce), NaF (5 mM), and Na3VO4 (200 μM). Total cell proteins were resolved on 8–12% sodium dodecyl sulfate–polyacrylamide gels and transferred to PVDF membranes by electroblotting (Bio-Rad). The membranes were blocked with 5% nonfat dry milk in TBST (50 mM Tris, pH 7.6, 150 mM NaCl, 0.1% Tween 20) and incubated overnight at 4°C in 3% nonfat dry milk in TBST with p38, phospho-p38, ERK-1/2, and phospho–ERK-1/2 (all 1:1,000) and anti–β-tubulin (1:3,000). Immunolabeling was detected using ECL Reagent (Amersham Biosciences).

Immunohistochemical analysis.

Immunohistochemistry was used to confirm and localize production of CCL3 and CCL4 in 30 human IVDs, including 8 samples obtained postmortem and 22 samples obtained during surgery from patients with a mean age of 48 years (range 25–66 years) (further details are available from the corresponding author upon request). Tissue sections measuring 4 μm were dewaxed and rehydrated, and endogenous peroxidases were quenched and, following heat-mediated antigen retrieval, blocked in goat serum. Sections were incubated overnight at 4°C with rabbit polyclonal antibodies against human CCL3 (1:4,000), CCL4 (1:200), and CD11b (1:50). Preimmune rabbit IgG (Abcam) was used as a negative control. After washing, sections were incubated with biotinylated goat anti-rabbit antiserum (1:500 dilution; Abcam), and binding was detected by the formation of streptavidin–biotin complex (Vector) with 3,3′-diaminobenzidine tetrahydrochloride solution (Sigma-Aldrich). Sections were counterstained with Mayer's hematoxylin (Leica Microsystems), dehydrated, cleared, and mounted in Pertex (Leica Microsystems). Sections were visualized as images captured using an Olympus BX60 microscope and QCapture Pro software version 8.0 (Media Cybernetics). A total of 200 NP cells were counted in each section, and the number of immunopositive cells was expressed as a percentage of the total count. Linear regression analysis was performed to investigate correlations between the measured percentage immunopositivity and the histologic grade of degeneration.

Collection of conditioned medium and cell migration assay.

Rat NP cells cultured in 6-cm plates (5 × 105/plate) were treated with TNFα (50 ng/ml) or IL-1β (10 ng/ml) for 24 hours in serum-free medium. Following the cytokine treatment, conditioned medium was collected, centrifuged at 2,000 revolutions per minute for 10 minutes to remove any cells and debris, and used immediately for cell migration assays. Macrophage migration was measured using a 24-well cell culture insert system (VWR Scientific), in which the top and bottom compartments were separated by a filter with 8-μm pores. RAW264.7 cells were plated in the upper chamber of 24-well plate inserts at a density of 2 × 105 cells/well. The next day, RAW264.7 cells were exposed to control medium or conditioned medium from NP cells that had been treated for 24 hours with or without TNFα or IL-1β in the presence or absence of the CCR1 antagonist J113863 (Tocris Bioscience). After 24 hours, RAW264.7 cells were removed from the upper side of the membranes, and cells were measured on the lower side of the membrane and in the bottom chamber by MTT assay to determine the migration rate.

Statistical analysis.

All measurements were performed in triplicate, and data are presented as the mean ± SEM. Differences between groups were analyzed by Student's t-test and analysis of variance. P values less than 0.05 were considered significant. RT-PCR expression data for human tissue samples was found to be nonparametric in distribution, so the Kruskal-Wallis test combined with post hoc analysis by Conover-Inman test was used to determine whether 2math image values from nondegenerated, degenerated, and infiltrated samples were significantly different. The 2-sample proportion test was used to determine whether the differences in the proportions of samples in each study group that exhibited expression of each target gene were significant.


Regulation of CCL3 expression in rat NP cells by TNFα and IL-1β.

Expression of CCL3 in mature rat tissue was studied using real-time PCR analysis. The basal expression of CCL3 messenger RNA (mRNA) in healthy mature rat NP as well as AF tissue was very low (Figure 1A). To investigate whether cytokines regulate CCL3 expression, we treated rat NP cells with TNFα or IL-1β. Figures 1B and C show that treatment with TNFα or IL-1β resulted in a peak increase in CCL3 mRNA expression at 4 hours, although expression remained significantly elevated, as compared to untreated control, at 24 hours. A similar but smaller increase in CCL4 expression was also seen when cells were treated with either of the cytokines (results are available from the corresponding author upon request).

Figure 1.

Expression and cytokine dependency of CCL3 in nucleus pulposus (NP) cells. A, Real-time reverse transcription–polymerase chain reaction (RT-PCR) analysis showing very low CCL3 expression in adult rat NP and annulus fibrosus (AF) tissue. HPRT-1 = hypoxanthine guanine phosphoribosyltransferase 1. B, and C, Real-time RT-PCR analysis of CCL3 expression by rat NP cells treated with 50–100 ng/ml tumor necrosis factor α (TNFα) (B) or 10–20 ng/ml interleukin-1β (IL-1β) (C) for different time periods. Induction was maximal at 4 hours and declined thereafter, although it remained elevated until 24 hours. D, Schematic representation of CCL3 promoter constructs of different lengths used in the study. Putative NF-κB and CCAAT/enhancer binding protein β (C/EBPβ) elements are shown. E and F, Dose-dependent increase in CCL3 promoter activity in rat NP cells transfected with a 1.4-kb CCL3 reporter construct and treated with increasing doses of TNFα (E) or IL-1β (F). G, Significant induction of activity of CCL3 promoter fragments in rat NP cells transfected with a 300-bp and a 140-bp reporter construct and treated with TNFα or IL-1β. Bars show the mean ± SEM of 3 independent experiments. ∗ = P < 0.05.

To investigate if the CCL3 regulation occurs at the transcription level, we measured the cytokine-dependent activity of the 1.4-kb CCL3 promoter (Figure 1D). The CCL3 promoter contains 3 putative NF-κB binding motifs (20); their sequences, locations in the promoter relative to the transcription start site, and relative JASPAR database scores are as follows: AAAATTTCCC, −80/−71 bp, 78; GGGACTGACT, −289/−280 bp, 76.4; and GGGAAATCAA, −1300/1291 bp, 76.1. In addition, several C/EBPβ binding sites are present in the promoter. Figures 1E and F show that the cytokines significantly increased promoter activity. Moreover, IL-1β–dependent expression was dose dependent (Figure 1F). Cytokine treatment also induced activity of the shorter 300-bp and 140-bp promoter fragments (Figure 1G).

TNFα and IL-1β promote CCL3 expression by activating NF-κB signaling.

To ascertain if cytokine-induced CCL3 expression requires NF-κB signaling, we pretreated rat NP cells with the inhibitor SM7368. Pretreatment caused a significant suppression of TNFα- or IL-1β–induced expression of CCL3 mRNA levels (Figures 2A and B). When rat cells were treated with the NF-κB inhibitor SM7368 (Figures 2C and D) or cotransfected with DN-NF-κB/IκBαM (Figures 2E and F), the cytokine-mediated induction of CCL3 promoter activity was completely abolished. In contrast, cotransfection with p65 resulted in a dose-dependent increase in CCL3 promoter activity (Figure 3A). Surprisingly, cotransfection with p50 blocked the inductive effect of p65 on CCL3 promoter activity (Figure 3B). The addition of p50 alone had a small suppressive effect on promoter activity (Figure 3B). Notably, Figures 3C and D show that p50 completely suppressed the inductive effect of the cytokines on the CCL3 promoter. We also determined if the NF-κB subunits RelB and c-Rel control CCL3 expression. Neither RelB (Figure 3E) nor c-Rel (Figure 3F) had a significant effect on CCL3 promoter activity. To determine if RelA controlled CCL3 promoter activity in a cell type–specific manner, we measured CCL3 promoter activity in wild-type and p65-null rat fibroblasts. The promoter activity was induced by treatment with cytokines in the wild-type cells only (Figures 3G and H).

Figure 2.

Modulation of cytokine-dependent CCL3 expression by NF-κB signaling in rat NP cells. A and B, Real-time RT-PCR analysis of CCL3 expression by rat NP cells following TNFα (A) or IL-1β (B) treatment for 24 hours with or without the NF-κB inhibitor SM7368 (SM; 10 μM). Inhibition of NF-κB signaling resulted in significant blocking of cytokine-dependent induction of CCL3 mRNA expression. C–F, CCL3 promoter activity following TNFα (C and E) and IL-1β (D and F) treatment with or without SM7368 (C and D) or cotransfection with dominant-negative NF-κB/IκBαM (E and F). Cytokine-mediated induction of promoter activity was completely blocked by inhibition of NF-κB signaling. Bars show the mean ± SEM of 3 independent experiments. ∗ = P < 0.05. See Figure 1 for other definitions.

Figure 3.

NF-κB regulation of CCL3 expression. A, Dose-dependent increase in CCL3 promoter activity in rat NP cells transfected with p65. B, Promoter activity in rat NP cells transfected with p65, p50, or both p65 and p50. Transfection with p65, but not p50, resulted in increased promoter activity. Cotransfection with p50 significantly blocked the p65-mediated induction of CCL3 promoter activity. C and D, CCL3 promoter activity after cotransfection with p50 and TNFα (C) or IL-1β (D). Cotransfection with p50 completely abolished both the TNFα– and the IL-1β–mediated induction of CCL3 promoter activity. E and F, Effect of RelB (E) and c-Rel (F) on CCL3 promoter activity. In contrast to p65, RelB and c-Rel had little or no effect. G and H, Reporter activity in wild-type (WT) and p65-null mouse cells transfected with CCL3 reporter and treated with IL-1β (G) or TNFα (H). Only wild-type cells showed an increase in CCL3 reporter activity. Bars show the mean ± SEM of 3 independent experiments. ∗ = P < 0.05. NS = not significant (see Figure 1 for other definitions).

To further validate the role of p65/RelA in controlling CCL3 expression, we silenced the expression of p65 and its upstream controller IKKβ and measured CCL3 expression in human NP cells. Silencing was achieved using a lentivirus coexpressing YFP and shp65 or shIKKβ. Figure 4A shows that there was robust expression of YFP by the virally transduced cells, indicating a high level of transduction efficiency and transgene expression. As expected, in cells transduced with shp65 and shIKKβ there was a significant decrease in the expression of p65 and IKKβ, respectively, compared to cells transduced with control shRNA (Figure 4B). Suppression of individual NF-κB signaling components blocked the inductive effect of TNFα on CCL3 mRNA expression (Figure 4C).

Figure 4.

A, Immunofluorescence detection of yellow fluorescent protein (YFP) in human NP cells transduced with lentivirus coexpressing YFP and NF-κB pathway–specific short-hairpin RNAs (shRNAs) (LV-shp65 and LV-shIKKβ), showing high transduction efficiency. Original magnification × 20. B, Real-time RT-PCR analysis of cells transduced with a lentivirus expressing control shRNA (LV-shC), LV-shp65, and LV-shIKKβ. Expression of p65 and IKKβ was significantly suppressed by the corresponding shRNAs when compared to cells transduced with a lentivirus expressing control shRNA. C, Real-time RT-PCR analysis of CCL3 expression in cells infected with control, LV-shp65, and LV-shIKKβ following TNFα treatment. Note that the TNFα-dependent induction of CCL3 mRNA levels was significantly blocked by suppression of components of the NF-κB signaling pathway. Bars show the mean ± SEM of 3 independent experiments. ∗ = P < 0.05. See Figure 1 for other definitions.

Control of CCL3 promoter activity in rat NP cells by C/EBPβ and MAPK signaling.

We then examined whether C/EBPβ/LAP-2 controls cytokine-dependent CCL3 expression in rat NP cells. We transfected rat NP cells with LIP, a functional LAP antagonist, and measured cytokine-dependent CCL3 promoter activity. Suppression of C/EBPβ function prevented TNFα (Figure 5A) and IL-1β (Figure 5B) induction of promoter activity. Similarly, LIP caused a robust decrease in basal CCL3 promoter activity (Figure 5C). In contrast, cotransfection with LAP-2 resulted in induction of promoter activity in the absence of exogenously added cytokines (Figure 5D).

Figure 5.

Modulation of cytokine-dependent expression of CCL3 by C/EBPβ and MAPK in rat NP cells. A and B, CCL3 reporter activity in rat NP cells cotransfected with liver-enriched inhibitory protein (LIP; dominant-negative C/EBPβ) and treated with TNFα (A) or IL-1β (B). LIP inhibited the cytokine-dependent induction of CCL3 promoter activity. C, Significant inhibition of basal CCL3 promoter activity by LIP. D, Significant induction of CCL3 promoter activity by liver-enriched transcriptional activator protein 2 (LAP-2), in contrast to LIP. E and F, Western blot analysis of p38 following treatment of rat NP cells with TNFα (E) or IL-1β (F). Treatment induced phosphorylation of p38 within the first 5 minutes, and levels of phospho-p38 remained elevated for 24 hours in the TNFα-treated group. No change in the expression of p38 was seen. G and H, Real-time RT-PCR of CCL3 expression by rat NP cells following 24 hours of treatment with IL-1β (G) or TNFα (H) with or without inhibitors of p38 (SB203580 [SB]; 10 μM), ERK (PD98059 [PD]; 10 μM), and JNK (SP60025 [SP]; 10 μM). MAPK inhibition resulted in blocking of the cytokine-dependent induction of CCL3 mRNA, with the exception of JNK inhibition, which had no effect on the TNFα-dependent increase in CCL3 expression. I, CCL3 promoter activity after p38 inhibition. The IL-1β–induced increase in CCL3 promoter activity was blocked by p38 inhibition. Bars show the mean ± SEM from 3 independent experiments. ∗ = P < 0.05. See Figure 1 for other definitions.

Since MAPK signaling controls the activity of both NF-κB and C/EBPβ, we determined if it is required for the cytokine-dependent induction of CCL3 in rat NP cells. We first evaluated activation of the p38 and ERK signaling pathway by the cytokines. TNFα (Figure 5E) or IL-1β (Figure 5F) treatment resulted in a rapid increase in phospho-p38 and phospho–ERK-1/2 (results not shown). Pretreatment of rat NP cells with p38 and ERK inhibitors caused a significant suppression in TNFα or IL-1β induction of CCL3 mRNA (Figures 5G and H). In contrast, JNK inhibition selectively suppressed the response to IL-1β (Figure 5H) but had a minimal effect on the response to TNFα (Figure 5H). We then measured the effect of the p38 inhibitor on cytokine-dependent CCL3 promoter activity. Again, inhibitor treatment resulted in significant suppression of the IL-1β–mediated induction of CCL3 promoter activity (Figure 5I).

Elevated CCL3 and CCL4 expression in degenerated human IVDs.

The CCL3 gene was expressed in 2 of 6 nondegenerated samples (grade 0–3), 8 of 16 degenerated samples (grade 4+), and 10 of 13 infiltrated/herniated samples (Figure 6A). The proportion of infiltrated samples expressing CCL3 was increased compared to the proportion of nondegenerated samples (P < 0.05), but was not significantly different from that of the degenerated (4+) samples (P > 0.05). Infiltrated samples expressed higher levels of CCL3 than degenerated samples (P = 0.0383), but did not express significantly different levels of CCL3 compared with the nondegenerated (grade 0–3) group (P > 0.05). We also examined the expression of CCL4. Similar to CCL3, CCL4 was present in 2 of 6 nondegenerated samples (grade 0–3), 5 of 16 degenerated samples (grade 4+), and 9 of 13 infiltrated samples. Thus, there was a significant increase in the proportion of disc samples containing infiltrated cells expressing CCL4 when compared to the degenerated group (grade 4+) (P < 0.05), but not when compared to the nondegenerated group (grade 0–3) (P > 0.05). No significant difference in the level of CCL4 gene expression was observed between the groups. Overall, CCL4 was expressed in 16 IVD samples that also expressed CCL3. The CCL3 gene was expressed independently of CCL4 in only 4 IVD samples.

Figure 6.

A, CCL3 and CCL4 expression in human NP cells. Relative CCL3 and CCL4 gene expression within nondegenerated discs (grade 0–3), degenerated discs (grade 4+), and discs containing infiltrating cells (I). CCL3 was seen in a greater proportion of infiltrated discs than nondegenerated discs and was expressed at a higher level in infiltrated discs than in degenerated discs. CCL4 was seen in a greater proportion of infiltrated discs than degenerated discs. Data are shown as box and whisker plots. Each box represents the 25th to 75th percentiles. Lines inside the boxes represent the mean. Whiskers represent the minimum and maximum. ∗ = P < 0.05 by proportion test; ⧫ = P < 0.05 for expression levels. B, Correlation of the expression of CCL3 by NP cells with the grade of degeneration (left). Expression of CCL4 was not correlated with the grade of degeneration (right). C, Localization of CCL3 and CCL4 staining to the NP cells within the disc. Bars = 30 μm for CCL3 and 40 μm for CCL4. D, Hematoxylin and eosin staining showing the morphology of infiltrating cells from 2 disc samples. E, CD11b staining. CD11b reactivity was absent from disc cells found in single lacuna or within clusters. Infiltrating cells displayed CD11b immunopositivity. Arrow shows a CD16+ cell. Bars = 100 μm. F, Role of CCL3 in RAW264.7 macrophage migration. NP-conditioned medium (CM) enhanced the macrophage migration rate. Conditioned medium of TNFα- or IL-1β–treated NP cells further increased the migration rate, and this increased migration was blocked by the CCR1 antagonist J113863 (J1). Bars show the mean ± SEM of 3 independent experiments. ∗ = P < 0.05. G, A proposed model of the relationship between inflammatory cytokines and CCL3 in NP cells. See Figure 1 for other definitions.

We performed immunohistochemical staining to further analyze chemokine expression and localization in disc tissue samples (Figures 6B and C). CCL3 and CCL4 staining was localized to NP cells within the IVD (Figure 6C). The percentage of cells immunopositive for CCL3 correlated positively with the grade of tissue degeneration (P = 0.0166). No correlation between CCL4 immunopositivity and grade of tissue degeneration was observed (P = 0.3355) (Figure 6B). The identity of infiltrating cells was also determined by immunostaining for CD11b. A large proportion of infiltrating cells were smaller in size (10–20 μm) and were localized as single cells at higher cell density (Figure 6D) compared to the larger resident disc cells (20–60 μm), which were observed within lacuna either singly or in clusters. These smaller infiltrating cells, together with some larger macrophage-like cells, stained positive for CD11b (Figure 6E), whereas resident disc cells within the lacuna were CD11b negative (Figure 6E).

TNFα and IL-1β promote NP-mediated migration of macrophages in a CCR1-dependent manner.

We examined the effect of conditioned medium of NP cells treated with cytokines on migration of mouse RAW 264.7 macrophages. Figure 6F shows that the conditioned medium of NP cells treated with TNFα or IL-1β promoted chemotactic migration of macrophages compared to the conditioned medium of untreated NP cells. Moreover, cytokine-dependent macrophage migration was completely blocked when pretreated with J113863, a well-characterized antagonist of CCR1.


This study is the first to demonstrate that NP cells expressed both CCL3 and CCL4, and that expression of CCL3 was regulated by the inflammatory cytokines TNFα and IL-1β through the MAPK, NF-κB/p65, and C/EBPβ signaling pathways. We showed that, in contrast to p65, NF-κB1/p50 inhibited CCL3 expression. A second major observation was that by regulating CCL3 expression by NP cells, inflammatory cytokines promoted CCR1-dependent macrophage migration. Importantly, in terms of clinical relevance, analysis of human tissue samples indicated that CCL3 expression levels correlated positively with the grade of degeneration and that expression levels were higher in herniated tissue than in degenerated but contained samples. Based on these observations, we predict that if CCR1–CCL3 activity is blocked using therapeutic agents, then macrophage infiltration into the NP and the associated inflammatory response into the herniated disc will be limited.

Expression studies showed that CCL3 was expressed in both human and rat NP tissue. Given the ability of normal NP cells to produce several cytokines and chemokines, it was not surprising to find that NP cells produced baseline levels of CCL3. The low level of expression in the rat tissue probably reflects the healthy state of the NP and suggests that CCL3 may have physiologic functions other than chemotaxis. However, our findings clearly indicated that expression was induced by TNFα as well as IL-1β, cytokines that are closely linked to degenerative disc disease. These results are consistent with those of previous studies showing that CCL3 expression is sensitive to cytokines, including TNFα and IL-1β (21–23). Moreover, promoter studies showed that regulation occurred at the transcription level and that the first 140 bases are sufficient to drive cytokine-dependent transcription.

Although the mechanism of regulation is likely to be cell type– and context-specific, there is some evidence to indicate that CCL3 transcription may be controlled by NF-κB, C/EBPβ, and MAPK signaling (13, 24). The presence of 3 putative κB motifs in the CCL3 promoter indicated that it was functionally involved in controlling transcription. Zhang et al suggested that the first 300 bases of the promoter may control its activity (13). Notably, the 2 κB motifs with the highest relative scores obtained using the JASPAR database are contained within this region. Related to this finding, a recent study has also shown the presence of c-Rel binding sites in the CCL3 promoter and demonstrated that IL-1β caused a higher induction of c-Rel mRNA than p65 in chondrocytes (13). Thus, our observation that the inductive effect is restricted to p65, and does not occur with RelB and c-Rel, further highlights the unique response of NP cells to environmental stimuli.

Another unique observation concerned the role of p50/RelA heterodimers. Lim and colleagues have shown that lipopolysaccharide (LPS) causes a rapid recruitment of p50/RelA heterodimer along with E2F1 to the CCL3 promoter, thereby enhancing gene transcription (24). In contrast to those studies, we found that in NP cells, p50 not only suppressed the inductive effect of cytokines but also inhibited the activation of the CCL3 promoter by p65. This result was unexpected, since our own recent studies have shown that p65 and p50 act synergistically to induce the expression of syndecan 4, one of the target genes of TNFα and IL-1β in NP cells (25). However, our findings are consistent with those of previous studies that demonstrated a repressive function of p50 homodimers in controlling the expression of a number of chemokines and catabolic genes, including CCL2, CXCL10, granulocyte–macrophage colony-stimulating factor, and matrix metalloproteinase 13 (26–28). Thus, formation of p50 homodimers and their binding to the κB motifs could promote recruitment of a transcriptional repressor, such as histone deacetylase 1, to the CCL3 promoter, thereby suppressing the RelA response (26). The notion that RelA/p65 signaling modulates promoter function was supported by the observation that RelA-null cells failed to induce CCL3 promoter activity, even when treated with cytokines. Moreover, silencing studies showed an inhibition of TNFα-dependent CCL3 expression following the suppression of p65 and IKKβ signaling, highlighting the importance of this pathway in controlling CCL3 gene expression.

Aside from RelA/p65, the present study indicated that C/EBPβ regulated CCL3 transcription. This finding was not unexpected, since Zhang and colleagues showed that when chondrocytes were treated with IL-1β, there was a cooperative regulation of CCL3 expression by both C/EBPβ and NF-κB (13). In a separate study, Choi et al showed that acute myeloid leukemia 1A (AML-1A) and AML-1B/RUNX-1 controlled CCL3 expression in multiple myeloma (29), while a later study demonstrated the importance of C/EBPβ in this regulation (30). Similarly, Grove and Plumb have shown that C/EBP, NF-κB, and cEts family members were important in LPS-induced expression of CCL3 and that the promoter sequence between −200 and +36 bp was critical in conferring cell-specific responses (31). With respect to cytokines, our data clearly show that in NP cells, C/EBPβ positively controlled TNFα-dependent CCL3 expression. Related to this observation, our recent investigation has clearly identified C/EBPβ as a negative regulator of cytokine-dependent ADAMTS-4 transcription, again highlighting the importance of both context and target gene specificity in the regulatory machinery of NP cells (Tian Y et al: unpublished observations).

With regard to MAPK signaling, since the activity of this pathway is responsive to both NF-κB and C/EBPβ, we investigated if it controlled CCL3 expression. We confirmed that both TNFα and IL-1β promoted MAPK activation and that the signaling pathways controlled CCL3 expression. We noted that there was differential sensitivity of CCL3 expression. Thus, while p38, ERK, and JNK positively controlled IL-1β–dependent induction of CCL3, the JNK pathway was not involved in the TNFα-mediated response. Based on these findings, it is not unreasonable to assume that by controlling the activity of MAPK signaling, cytokines regulate the expression of CCL3 through NF-κB and C/EBPβ in NP cells.

In concert with CCL3 and CCL4 expression in RA synovial fluid and tissue (12–15), our studies confirmed that CCL3 expression is elevated in degenerated human disc tissue and that staining is localized to NP as well as to the infiltrating cells. Importantly, CCL3 expression was positively correlated with the grade of tissue degeneration and was highest in tissue samples with infiltrating cells. It must be acknowledged that in vitro, NP cells exhibit efficient phagocytosis and are capable of removing apoptotic cells, a classic behavior of macrophage-like cells (32), and express CD68, a classic marker of a monocyte/macrophage (33). Thus, despite positive CD11b staining, the exact origin and identity of all of the infiltrating cells could not be determined with certainty. However, results of our in vitro cell migration studies confirmed that chemotactic factors secreted by NP cells in response to TNFα or IL-1β treatment promote macrophage migration. Notably, pretreatment of macrophages with a selective CCR1 antagonist, J113863, completely inhibited NP-induced macrophage migration, suggesting that CCR1 is critical for their migratory response. Since CCR1 is a receptor for CCL3, CCL4, and CCL5, further investigations will be necessary to elucidate the relative importance of the individual chemokines in this process. Based on these findings, it is plausible to consider that therapeutic blocking of CCR1–CCL3 activity would inhibit macrophage infiltration into the disc tissue and prevent the inflammatory response associated with degenerative disc disease.


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. Risbud 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. Shapiro, Le Maitre, Risbud.

Acquisition of data. Wang, Tian, Phillips.

Analysis and interpretation of data. Wang, Tian, Phillips, Chiverton, Haddock, Bunning, Cross, Shapiro, Le Maitre, Risbud.