The Chondroprotective Role of Erythromycin in a Murine Joint Destruction Model

Abstract

Objective

Inflammation is a major player in the joint destruction process. Macrolide antibiotics have recently been found to have a novel anti-inflammatory function, but their effects on the joint are unknown. Our objective was to investigate the effect of macrolide antibiotic erythromycin on cartilage gene expression under inflammatory conditions as well as on joint pathology in an in vivo inflammatory joint destruction model.

Design

In our in vivo studies, mouse knee joints were injected with monosodium iodoacetate (MIA), a chemical that inhibits glycolysis and causes joint inflammation and matrix loss. Erythromycin was administered by daily intraperitoneal injection. Changes in joint cartilage and synovium were evaluated by histological analysis. In our in vitro studies, primary bovine articular chondrocytes were treated with erythromycin in the presence of pro-inflammatory cytokine IL-1β or lipopolysaccharide (LPS), and cartilage gene expression analysis was performed.

Results

Regional differences in cartilage matrix destruction along the medial-lateral axis were observed in joints of MIA-injected mice. Erythromycin treatment inhibited cartilage matrix loss and synovitis in these joints. In addition, erythromycin inhibited IL-1β and LPS-induced expression of MMPs and iNOS, as well as the positive regulatory loop between IL-1β and Toll-like receptor 4 (TLR4) in articular chondrocytes. Furthermore, erythromycin prevented LPS-induced NF-κB activation, a key mediator of TLR4-mediated cartilage destruction process.

Conclusions

Erythromycin has the ability to inhibit catabolic gene expression mediated by IL-1β and TLR4 in chondrocytes in vitro and maintains cartilage matrix levels in experimental inflammatory joint destruction in vivo, suggesting that it possesses a chondroprotective activity.

Keywords

erythromycin antibiotics osteoarthritis chondrocytes IL-1β TLR4

Introduction

The joint is an important structure that supports the body and facilitates movement. However, it is also the target of arthritis, a disease that causes widespread disability. The most common form of arthritis is osteoarthritis (OA), which is characterized by cartilage matrix loss, synovitis, and other pathological changes of joint tissues.^1-3 Yet, there are few treatment options for OA due to the limited understanding of OA development and heterogeneity in its pathology.⁴ Local inflammation is widely considered to be associated with OA development and cartilage degradation.^5-7 In fact, interleukin-1-beta (IL-1β) is a prominent pro-inflammatory cytokine that causes cartilage matrix loss. Antagonizing IL-1β directly in the joint leads to matrix protection in posttraumatic experimental OA.^8-10 However, a recent report indicated that while IL-1β in human synovial fluid was up-regulated in the early phase of OA, its level was much lower in late OA.¹¹ Furthermore, systemic IL-1β inhibition in mice did not lead to a reduction in joint inflammation and cartilage destruction in experimental OA, or symptom relief in a human clinical trial,¹² suggesting that other catabolic factors in the joint may be promoting continued cartilage damage in OA. It is now established that Toll-like receptors (TLRs), especially TLR2 and TLR4, can be induced by IL-1β and are upregulated in OA.^13-16 Among the TLRs, TLR4 seems to be especially important in joint maintenance, as local IL-1β-driven joint pathology was seen to be dependent on TLR4 activation rather than TLR2 activation.¹⁷ Downstream of these inflammatory stimuli lies the nuclear factor-kappa B (NF-κB) signaling pathway, a key mediator whose activation leads to the induction of cartilage matrix degrading enzymes such as matrix metalloproteinases (MMPs) and aggrecanases, resulting in subsequent matrix degradation.⁶

Recently, certain macrolide antibiotics such as erythromycin (EM), clarithryomycin, and roxithromycin were shown to have anti-inflammatory activities in chronic inflammatory conditions such as diffuse panbronchiolitis, rheumatoid arthritis and in debris-induced inflammatory osteoclastogenesis in bone.^18-20 A derivative of erythromycin without antibiotic function still suppresses NF-κB activity in airway epithelial cells,²¹ suggesting that its anti-inflammatory activity is independent of its antibiotic function. Among the macrolide antibiotics, only EM has been tested in a human OA clinical trial, which showed EM enhanced the effect of acetaminophen in reducing joint effusion and pain over the course of 4 months.²² However, whether joint cartilage integrity was improved or whether EM alone has any effects on OA progression was not assessed in this study. Therefore, the role of EM on articular cartilage maintenance is still not known.

In this study, we utilized the monosodium iodoacetate (MIA)–induced murine inflammatory joint destruction model to assess the effect of EM on joint cartilage integrity in vivo. As a specific antagonist of glyceraldehyde 3-phosphate dehydrogenase (GAPDH), MIA inhibits glycolysis and ATP production, which are changes that have been found in OA.^23-26 Injecting MIA into the joint is known to result in cell death, joint cartilage matrix loss and synovitis.^23-26 As MIA induces robust inflammation that subsequently subsides within days, it recapitulates the inflammatory aspect of joint destruction seen in patients after traumatic joint injury.²⁷ Consistently, it has been reported that administering MIA to chondrocytes resulted in increased MMP13 expression in vitro.²⁸ In this study, we evaluated the effect of EM treatment on cartilage matrix loss and synovitis in the MIA model in vivo and investigated the biochemical mechanisms by which EM acts in primary articular chondrocytes in vitro.

Materials and Methods

Experimental Animals

All animal care and experimental procedures were approved by the Institutional Animal Care and Use Committees at Tufts University. Wild-type CD1 male mice were purchased from Charles River Laboratories. Mice were caged in groups under the standard conditions with a 12-hour light/dark cycle and standard chow diet. 8 week old mice were anesthetized by isoflurane/O₂ inhalation and received a single intra-articular injection of MIA (Sigma) using a 30G 0.5-inch needle with 62.5 μg MIA dissolved in 5 μL of sterile phosphate buffered saline (PBS). This concentration was because it was the lowest of the 3 concentrations tested (250, 125, and 62.5 μg) that elicited matrix loss (data not shown), which is much lower than the amount used for pain analysis (0.5-1 mg).^29,30 EM (Sigma) was delivered into the mice by daily intraperitoneal injection at 400 μg/day, starting from 1 day before MIA injection through 10 days after MIA injection when mice were euthanized. This dosage of EM was chosen based on the amount of EM commonly used to treat inflammation (e.g., in diffuse panbronchiolititis) in humans (400-600 mg/day), the body weight ratio of human and mouse (around 2000:1) and the half-life of EM (39 minutes in the mouse, 3.4 hours in the human).^29,30 PBS-injected contralateral knees served as a negative control. A total of 20 mice were used in all experiments, including those injected with MIA or PBS for histology and mice used for gene expression analysis. Five knee joints per group were used.

Histological Analysis of Mouse Knee Sections

Mice were sacrificed 10 days after MIA injection, and their knee joints were fixed with 4% paraformaldehyde overnight, decalcified with 0.33 M ethylenediaminetetraacetic acid and paraffin embedded. The medial half of the knee joint was sectioned sagittally at 5-μm thickness. Sections were stained with 0.1% Safranin O and counterstained with hematoxylin and fast green. Percentage matrix loss was quantified by calculating the surface areas that exhibit Safranin O loss over total surface area, as delineated by ImageJ. OARSI (Osteoarthritis Research Society International) scores were determined based on the percentage loss of Safranin O staining using the established OARSI scoring system ^31,32. Cellularity was evaluated by the DAPI (4′,6-diamidino-2-phenylindole) staining in comparison with bright field images. A loss of DAPI staining in the lacuna (as shown in the bright field) indicated cell loss. The severity of synovitis was evaluated according to the commonly used scoring system ^33,34, which has the following three criteria: enlargement of the synovial cell layer (0-3 points), cellularity of resident cells (0-3 points) and inflammatory infiltration (0-3 points). When the points are added up, a full score (9 points) indicates severe synovitis and a lower score indicates less synovitis. All scorings were performed blinded.

Primary Bovine Articular Chondrocytes In Vitro Culture

Primary bovine articular chondrocytes (BACs) were isolated from the articular surface of the fetlock joints as previously described.³⁵ Briefly, cartilage pieces were digested with 1 mg/mL bovine hyaluronidase (Sigma) for 15 minutes followed by 0.25% trypsin (Sigma) digestion for 30 minutes and 2 mg/mL collagenase (Sigma) digestion for 15 hours. Single cell suspension was obtained by passing the cells through a 40 μm cell strainer (BD Biosciences). Cells were cultured in DMEM supplemented with 10% fetal bovine serum (Hyclone) and antibiotic-antimycotic (Invitrogen) and used at passage 1 for the study. For in vitro studies, chondrocytes were cultured on tissue culture dishes, and were pretreated with EM (Sigma; 0.2 or 1 μg/mL) for 3 hours prior to treatments of 1 ng/mL IL-1β (Peprotech) or 100 ng/mL lipopolysaccharide (LPS; Sigma) for 4 days.

Cytotoxicity Assay of Erythromycin

Chondrocytes were seeded on tissue culture dishes and treated with EM (0.2, 1, 10, and 100 μg/mL) for 4 days before being subjected to Live/Dead Cell Viability/Cytotoxicity assay (Invitrogen) according to the manufacturer’s instruction. The ratios of dead cells/live cells were then determined. Data are reported as mean with 95% confidence intervals.

RT-qPCR Analysis

For experiments involving cultured primary chondrocytes, cells were grown for 4 days prior to RNA isolation using the RNeasy mini kit (Qiagen). For mouse articular cartilage isolation, the tibial plateau and femoral chondyle cartilage were isolated under the dissection microscope at 7 days post-MIA injection, and were cut into small pieces, and subsequently dissolved in the lysis buffer that accompanies the RNeasy mini kit. There were three mice/group (MIA and PBS injection). RNA was extracted from each joint individually, and for each knee joint, 100 to 200 ng of total RNA was obtained. RNA was reverse transcribed using the M-MLV reverse transcriptase (Invitrogen). Quantitative polymerase chain reaction (qPCR) was carried out using iTaq universal SYBR Green supermix on the iQ5 Real Time PCR Detection System (BioRad). Relative PCR values were calculated based on the values of the gene of interest and the reference gene using associated software. All primers were first tested for efficiency, and only melt curves resulting in single peaks with adequate and linear amplification were selected for subsequent chondrocyte gene expression analysis. Primer sequences are available on request. At least three independent experiments were performed in triplicate.

Western Blot Analysis

Cells were pretreated with erythromycin at 1 μg/mL for 3 hours followed by treatment of 100 ng/mL LPS for 30 minutes, 1 hour or 2 hours. Total protein was isolated using the standard RIPA buffer. Nuclear and cytoplasmic lysates were isolated using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce) according to the manufacturer’s instruction. Protein concentrations were determined using the DC Protein Assay (Bio-Rad). Total protein (10 μg), cytoplasmic protein (10 μg) and nuclear protein (10 μg) were loaded. Primary and secondary antibodies are: rabbit anti-NF-κB/p65 (E498), mouse anti-IκBα (L35A5) (Cell signaling), mouse anti-β-actin (Abcam) and goat anti-rabbit or mouse IgG, (H + L) HRP conjugate (Millipore/Chemicon). Western Blot images were taken by EPSON Perfection scanner V500, and band intensity was quantified by ImageJ. At least 3 independent experiments were performed.

Immunocytochemistry Analysis

Cells were seeded at a density of 5 × 10⁴ cells per well in 12-well plates and pretreated with EM at 1 μg/mL for 3 hours followed by 100 ng/mL LPS treatment for 1 hour. Cells were fixed with 4% paraformaldehyde and blocked with goat serum for 1 hour. Cells were incubated with rabbit anti-NF-κB/p65 (E498, Cell Signaling) antibody overnight at 4°C and 594 Alexa Fluor conjugated goat anti-rabbit antibody for 1 hour at room temperature. Nuclei were counterstained with DAPI.

NF-κB Transactivation Assay

Cells were seeded at a density of 10⁵ cells per well in 12-well plates and transfected with 200 ng of a NF-κB element-driven luciferase reporter construct (generous gift from Dr. Gail Sonenshein, Tufts University) and 5 ng of Renilla luciferase plasmid (internal transfection control) using Fugene6 (Roche). 24 hours later, cells were treated with EM at 1 μg/mL for 3 hours followed by 100 ng/mL LPS treatment for 16 hours. Dual luciferase assay was conducted using the luciferase assay kit (Promega). NF-κB luciferase activity was normalized to Renilla luciferase activity. At least 3 independent experiments were performed in triplicate.

Microscopy Analysis

Bright-field and fluorescent images for histological and immunofluorescent analysis were taken with the Olympus IX71 inverted microscope and the Olympus DP70 digital camera.

Statistical Analysis

For parametric data, data are reported as mean with 95% confidence intervals and statistical analysis was performed using a paired Student’s t test or a 1-way analysis of variance followed by a post hoc Tukey test with P values <.05 considered significant. For nonparametric data (i.e., OARSI [Osteoarthritis Research Society International] and synovitis scoring), data are reported as a dot plot with median and statistical analysis was performed using the Kruskal-Wallis test followed by Mann-Whitney U test with Bonferroni correction. P values <.05 are considered significant. All statistical analyses were conducted using Prism 5 (GraphPad).

Results

Erythromycin Inhibits Cartilage Matrix Loss In Vivo

We hypothesized that EM has an anti-inflammatory activity that slows down joint destruction. To test this hypothesis in vivo, we used the MIA inflammatory joint destruction mouse model and delivered EM by daily intraperitoneal injection.^25,36,37 Ten days after MIA injection, knee joints were evaluated by histological analysis. As different regions of the joint may be differentially affected in human OA,^38,39 we divided the medial half of the knee joint into 3 zones along the medial-lateral axis ( Fig. 1A ).

Figure 1.

Erythromycin (EM) prevents monosodium iodoacetate (MIA)–induced articular cartilage destruction in the outer region of medial condyle of the mouse knee joint. (A) Schematic diagram of the 3 zones in the medial condyle of tibia and femur in the mouse knee joint based on their medial-lateral positions. Histological analysis of the outer region (zone 1) is shown. (B) Safranin O staining with fast green and hematoxylin counterstaining on sagittal sections of mice knee joint 10 days after MIA injection. Scale bar = 400 μm. Arrows indicate areas with loss of Safranin O staining. (C-E) Percentage loss of Safranin O staining at articular surface of medial femur (C), tibia (D), or femur and tibia combined (E). (F-H) Degree of cartilage matrix destruction quantified according to the OARSI scoring system.

In the most external region (zone 1), MIA treatment caused a significant loss of glycosaminoglycan, as evidenced by the loss of Safranin O staining ( Fig. 1B ). EM treatment in the control joint did not cause visible changes to cartilage matrix expression, but EM treatment in the MIA-injected joint largely diminished MIA-induced matrix loss ( Fig. 1B ). We quantified the extent of cartilage matrix loss using the established OARSI system, which categorizes OA severity into a scale of 1 to 6 based on percentage surface cartilage matrix loss.³¹ We found that both the femur and the tibia showed significant articular cartilage matrix loss, while co-treatment with EM reduced such loss by half ( Fig. 1C-H ). In the middle region (zone 2), MIA-induced femoral cartilage damage was less severe compared to the external region (zone 1), but MIA-induced tibial cartilage damage was much more severe ( Fig. 2A-C ). EM co-treatment significantly reduced percentage cartilage matrix loss in this zone as well ( Fig. 2 ). However, when the percentage loss of surface cartilage was assigned an OARSI score, which rounds percentage loss into ordinal numbers (such as a loss of 26% to 49% to be given a score of 3), the differences between the treatment groups were reduced; as a result, the effect of EM on the femur was not statistically significant ( Fig. 2F ). EM inhibited cartilage matrix loss in the most internal region (zone 3) as well, although much less cartilage matrix is normally present as compared with other zones, and MIA-induced damage was also less severe ( Fig. 3 ).

Figure 2.

Erythromycin (EM) prevents monosodium iodoacetate (MIA)–induced articular cartilage destruction in the middle region of medial condyle of the mouse knee joint. (A) Schematic diagram of the 3 zones in the medial condyle of tibia and femur in the mouse knee joint based on their medial-lateral positions. Histological analysis of the middle region (zone 2) is shown. (B) Safranin O staining with fast green and hematoxylin counterstaining on sagittal sections of mice knee joint 10 days after MIA injection. Scale bar = 400 μm. Arrows indicate areas with loss of Safranin O staining. (C-E) Percentage loss of Safranin O staining at articular surface of medial femur (C), tibia (D), or femur and tibia combined (E). (F-H) Degree of cartilage matrix destruction quantified according to the OARSI scoring system.

Figure 3.

Erythromycin (EM) prevents monosodium iodoacetate (MIA)–induced articular cartilage destruction in the inner region of medial condyle of the mouse knee joint. (A) Schematic diagram of the 3 zones in the medial condyle of tibia and femur in the mouse knee joint based on their medial-lateral positions. Histological analysis of the inner region (zone 3) is shown. (B) Safranin O staining with fast green and hematoxylin counterstaining on sagittal sections of mice knee joint 10 days after MIA injection. Scale bar = 400 μm. Arrows indicate areas with loss of Safranin O staining. (C-E) Percentage loss of Safranin O staining at articular surface of medial femur (C), tibia (D), or femur and tibia combined (E). (F-H) Degree of cartilage matrix destruction quantified according to the OARSI scoring system.

Erythromycin Inhibits Cell Loss and Synovitis In Vivo

Chondrocyte cell loss is a characteristic change in OA.⁴⁰ Treatment with EM alone did not result in cell death, indicating minimum cell toxicity at this dosage. However, there was significant cell loss in MIA-injected knees (around 30%) ( Fig. 4A and B ). Significantly, treatment of EM reduced the percentage of cell loss to only 10% in MIA-injected knees, indicating that EM inhibits MIA-induced cell loss ( Fig. 4A and B ). In addition to cell loss, we also evaluated the synovium, since the synovium is an integral part of the joint and synovitis is also present in joint degeneration in OA.^33,34 As shown in Figure 4C , the synovium of the PBS control knee exhibited a normal appearance consisting of 1 to 2 layers of cells, with a lower cellularity. However, the synovium in the MIA-injected knee clearly had more than 3 layers of cells with significantly increased cellularity, indicating synovium hyperplasia, even though clusters of inflammatory cells were rarely observed ( Fig. 4C and D ). EM injection alone did not result in any changes in the synovium ( Fig. 4C and D ). However, EM co-treatment with MIA substantially prevented synovium hyperplasia in the MIA knees ( Fig. 4C and D ). These results were further quantified in Figure 4D using the established synovitis quantification system.^33,34

Figure 4.

Histological analysis for cartilage cellularity and synovitis. (A) DAPI and bright field overlaid images of mouse knee joints at day 10 after monosodium iodoacetate (MIA) injection. Scale bar = 400 μm. Arrows: cell loss. (B) Quantification of cellularity from images shown in A. (C) Safranin O staining with fast green and hematoxylin counterstaining on sagittal sections of mouse knee joints harvested 10 days after MIA injection. The synovium attached to the anterior horn of the meniscus is shown in each condition. Scale bar = 200 μm. Rectangles denote areas magnified in the subsequent subfigure. Arrows indicate areas with increased cellularity of resident cells in the synovium. Arrowheads indicate increased enlargement of synovial cell layers. Scale bar = 100 μm. (D) Quantification of the severity of synovitis, with 0 = no synovitis, and 9 = severe synovitis.

Erythromycin Inhibits IL-1β- and TLR4-Mediated Catabolic Gene Induction in Chondrocytes

We next determined how EM may exert a protective effect on joint cartilage. RT-PCR analysis on articular cartilage isolated from the MIA-injected joints showed increased levels of pro-inflammatory cytokines IL-1β and TNFα as well as inflammatory mediator TLR4 ( Fig. 5A ). Since these pro-inflammatory stimuli are known to cause catabolic changes in chondrocytes,⁶ we used them to perform in vitro experiments to test the effect of EM on chondrocytes. As very few primary adult murine articular chondrocytes could be obtained for culturing, primary bovine articular chondrocytes were used, which is a widely established in vitro culture system and allows large quantities of cells for qPCR and Western blot analysis.^35,41 To determine the suitable dosage for EM treatment in vitro, we performed live/dead assays on chondrocytes treated with varying concentrations of EM to avoid potential complications of cell toxicity. As 0.2 and 1 μg/mL EM treatment resulted in minimal cell death, we proceeded to use these 2 dosages in future experiments ( Fig. 5B and C ). As expected, IL-1β inhibited the expression of cartilage matrix genes Col-II and Aggrecan, and induced MMP1, MMP13 and iNOS expression ( Fig. 5D-H ). EM pretreatment strongly inhibited IL-1β-induced expression of MMPs and iNOS, but had no effect on IL-1β-induced reduction of Col-II and Aggrecan ( Fig. 5D-H ). Because IL-1β has been shown to induce the expression of TLR4,⁶ we also assessed TLR4 expression. As shown in Figure 5I , TLR4 was clearly induced by IL-1β, and this induction was abolished by co-treatment of EM.

Figure 5.

Erythromycin (EM) inhibits interleukin-1β (IL-1β)-induced catabolic gene expressions in bovine articular chondrocytes. (A) Quantitative polymerase chain reaction (qPCR) analysis of articular cartilage was isolated from the femoral condyle and tibial plateau of mice at 7 days after MIA or phosphate buffered saline (PBS) injection (3 mice/treatment). Data are reported as fold inductions using values obtained from PBS-injected control. To increase the accuracy of data analysis for in vivo sample treated with or without MIA, TBP, and GAPDH were used as reference genes. (B and C) Live-dead assay on chondrocytes treated with EM (0.2, 1, 10, 100 μg/mL) for 4 days. The percentages of dead cells among all cells were quantified. (D-I) qPCR analysis of bovine articular chondrocytes pretreated with EM (0, 0.2, and 1 μg/mL) overnight followed by treatment of IL-1β (1 ng/mL) for 4 days. Data are reported as mean with 95% confidence intervals from at least 3 independent experiments, each performed in triplicates. TBP served as a reference gene.

To determine the effect of EM on TLR4-mediated catabolic gene expression, we utilized TLR4 receptor-specific agonist, LPS.⁴² Since 1.0 μg/mL EM rather than 0.2 μg/mL showed a slightly stronger inhibitory effect on IL-1β activity in the previous experiments ( Fig. 5 ), 1.0 μg/mL EM was used for subsequent experiments. Similar to IL-1β, LPS inhibited the expression of cartilage matrix genes Col-II and Aggrecan, and strongly induced the expression of catabolic genes MMP1, MMP13 and iNOS ( Fig. 6A-F ). EM was also not able to inhibit LPS-mediated Col-II and Aggrecan induction ( Fig. 6A and B ), but significantly inhibited the expression of MMP1, MMP13 and iNOS induced by LPS ( Fig. 6C-E ). In addition, EM treatment reduced LPS-induced IL-1β expression, as well as IL-1β-induced TLR4 expression, suggesting that EM inhibits a positive regulatory loop between IL-1β and TLR4 ( Fig. 6F and G ).

Figure 6.

Erythromycin (EM) inhibits Toll-like receptor 4 (TLR4)–mediated catabolic gene expressions in bovine articular chondrocytes. (A-F) Bovine articular chondrocytes were pretreated with EM (1 μg/mL) for 3 hours followed by treatment of lipopolysaccharide (LPS; 100 ng/mL) for 4 days. Indicated chondrocyte gene expressions were determined by quantitative polymerase chain reaction (qPCR). Data are reported as mean with 95% confidence intervals from at least 3 independent experiments, each performed in triplicates. TBP served as a reference gene for normalization.

To further determine the mechanism by which EM inhibits TLR4-mediated catabolic gene expression, we assessed the canonical NF-κB pathway, which mediates TLR4 signaling.⁶ Typically, canonical NF-κB signaling activation involves IκBα degradation, resulting in subsequent translocation of NF-κB (subunits p65 and p50) into the nucleus.⁶ To determine the timing of NF-κB activation in our system, we treated the chondrocytes with LPS for 30, 60, and 120 minutes, and assayed IκBα degradation by Western blot ( Fig. 7A and B ). The IκBα protein level was significantly reduced at 60 min and recovered to basal levels by 120 minutes ( Fig. 7A ). Consistently, a significantly increased NF-κB/p65 protein level in the nucleus was observed at 60 minutes ( Fig. 7B ). Therefore, we selected 60 minutes as the time frame for our subsequent studies. EM alone did not affect IκBα and p65 protein levels ( Fig. 7C ), but diminished LPS-induced p65 nuclear localization without affecting IκBα degradation ( Fig. 7C ). The effect on NF-κB nuclear localization was confirmed by immunocytochemistry, which showed reduced nuclear p65 localization in cells co-treated with EM and LPS, as compared with LPS alone ( Fig. 7D ). These observations were further confirmed using an NF-κB reporter construct, which showed much reduced luciferase activity when cells were co-treated with EM and LPS ( Fig. 7E ). The effects of EM on IL-1β and LPS-induced catabolic gene expressions in chondrocytes are shown in Figure 8 .

Figure 7.

Erythromycin (EM) inhibits Toll-like receptor 4 (TLR4)–mediated NF-κB/p65 activation in bovine articular chondrocytes. (A-B) Western Blot analysis showing IκBα and NF-κB/p65 protein levels in bovine articular chondrocytes at various time points after 100 ng/mL of lipopolysaccharide (LPS) treatment. p65 (NE): p65 in the nuclear extract. p65 (CE): p65 in the cytoplasmic extract. β-actin served as a loading control. (C) Western blot analysis on bovine articular chondrocytes pretreated with EM (1 μg/mL) 3 hours followed by LPS treatment (100 ng/mL) for 60 minutes. p65 (NE): p65 in the nuclear extract. p65 (CE): p65 in the cytoplasmic extract. β-actin served as a loading control. Band intensities were quantified by using ImageJ. (D) Immunocytochemistry (ICC) analysis on nuclear localized NF-κB subunit p65 in bovine articular chondrocytes after pre-treatment with EM (1 μg/mL) overnight followed by LPS treatment (100 ng/mL) for 60 minutes. Arrows indicate cells with NF-κB/p65 in the nucleus. Percentage of cells with nuclear NF-κB/p65 staining among the total number of cells in the ICC analysis was quantified. (E) NF-κB transactivation assay in bovine articular chondrocytes. Cells were transiently transfected with NF-κB/p65 reporter construct for 24 hours and pretreated with EM (1 μg/mL) overnight followed by LPS treatment (100 ng/mL) for 16 hours. Renilla luciferase construct was co-transfected as an internal control for normalizing NF-κB luciferase activity. Data are reported as mean with 95% confidence intervals from at least 3 independent experiments, each performed in triplicates.

Figure 8.

Schematic diagram of the effects of erythromycin (EM) on interleukin-1β (IL-1β) and lipopolysaccharide (LPS)–induced catabolic gene expressions in chondrocytes. IL-1β signaling and Toll like receptor 4 (TLR4) signaling form a positive regulatory loop activate the NF-κB pathway and the expression of catabolic gene such as matrix metalloproteinases (MMPs) or inducible nitric oxide synthase (iNOS). EM inhibits the nucleus translocation of NF-κB/p65 protein, and MMPs and iNOS expression as well as the positive loop between IL-1β and TLR4.

Discussion

Prior to this study, the tetracycline class of antibiotics (including doxycycline and minocycline) has been investigated as potential OA therapeutics. These antibiotics have the ability to chelate zinc at the active site of metalloproteinases, and were shown to reduce the levels of collagenase activity in rheumatoid arthritic (RA) joints.⁴³ In OA, doxycycline and a chemically modified tetracycline (CMT-7) were found to reduce cartilage loss in vivo in the dog and the guinea pig, respectively.^44,45 Moreover, doxycycline treatment seemed to reduce joint space narrowing in human OA trials; however, in these trials, symptom relief such as pain and swelling was minimal.^46,47 As a result, tetracycline-like antibiotics were not pursued further for treating OA.

In this study, we show that EM can inhibit catabolic gene expression in primary chondrocytes as well as joint degeneration in vivo. This is the first report of any macrolide antibiotic on maintaining cartilage stability in experimental OA. While OA in the MIA model is artificially induced, this model exhibits the typical inflammatory aspects of traumatic injury–induced joint destruction, including cell death, matrix loss and synovitis. As there is a robust early inflammatory phase that eventually subsides, this model is not considered a systemic rheumatoid arthritis model, but rather a localized inflammatory joint destruction model.^36,37,48 Indeed, we have found increased mRNA levels of IL-1β, TNFα and TLR4 in the articular cartilage after 7 days of MIA injection. However, the extent of their induction was modest, which may be related to the relatively late timing of tissue harvest. Another feature of human joint destruction in OA is the regional differences of cartilage loss within the joint.^38,39 Yet, this feature was rarely discussed in reported studies using animal OA models.

We carried out a thorough study analyzing the medial condyle of the femur and the medial plateau of the tibia in areas along the medial-lateral axis. In our normal knee joint samples, there was less Safranin O staining in the innermost region (zone 3), which is least covered by the meniscus. Consistent with this observation, Little et al.⁴⁹ also reported less matrix staining in this zone. MIA injection caused much more tibial cartilage damage in the middle region (zone 2) and the least in the inner region (zone 3). We also observed more femoral cartilage damage in the outer region than the inner zones. These are also consistent with the magnetic resonance imaging analysis of human OA cartilage,^38,39 supporting the validity of this animal model. Such topographical differences could result from differential mechanical stress and the different proteoglycan synthesis of the chondrocytes in these regions.^50-52 Remarkably, EM inhibited cartilage loss in all these regions, which is conceivable as EM is delivered systemically through intrapertoneal injections. Since EM was introduced into the mice one day before MIA treatment, it is not clear whether EM inhibits the initiation or the progression of joint destruction. It will be interesting to apply EM after MIA treatment to distinguish these possibilities. A limitation of this study is that the MIA inflammatory model normally develops joint destruction at a much faster pace, and that we have used very young mice and have not followed a longer time course in this inflammatory joint destruction model. As OA develops over a long period of time and is more commonly found in older individuals, it is important to perform follow-up experiments using older mice and evaluate them at later time points. While MIA mimics the inflammatory aspects of injury-induced OA, it does not involve a direct physical injury to the knee joint. Thus, it will be important to follow up with surgery-induced OA models (such as destabilization of the medial meniscus, and anterior cruciate ligament transection). Furthermore, it will be interesting to compare the effect of EM with other antibiotics within the macrolide class of antibiotics or those in the tetracycline class.

In our in vitro studies, we showed that EM has the ability to inhibit catabolic activities in chondrocytes. EM inhibited the positive regulatory loop between IL-1β and TLR4 both in terms of activity and gene expression. Although multiple studies showed that intra-articular injection of an IL-1β antagonist inhibited injury-induced OA,^8,9 no results of OA manipulation on IL-1β knockout animals have been demonstrated. It was reported that the knockout of TLR2 and TLR4 mildly protect against meniscectomy-induced OA,⁵³ suggesting that TLRs may be required for OA development. However, the lack of a complete rescue in OA progression in the TLR KO indicates that other factors are also involved in this process. A limitation of our study is that we did not use natural TLR4 ligands in our study, although we did use a TLR4-specific agonist (i.e., LPS).¹⁶ Among the natural ligands of TLR4 are low molecular weight hyaluronan and high mobility group box chromosomal protein 1 (HMGB1), although they are not necessarily specific to TLR4.⁵⁴ Therefore, future experiments would involve testing whether EM can inhibit TLR4 signaling induced by these natural ligands in chondrocytes. Our in vitro analysis also indicated that EM inhibits NF-κB activation, providing a plausible mechanism for its inhibitory effects on the activities of inflammatory stimuli. As NF-κB activation takes place as early as 30 minutes following IL-1β treatment, applying EM at the same time of IL-1β treatment might not allow sufficient time for EM to inhibit NF-κB nuclear localization. Thus, EM was applied before IL-1β administration. While EM prevented NF-κB/p65 nuclear localization and activity, it did not seem to affect IκBα levels. It is possible that since Western blot analysis is semiquantitative, subtle changes in IκBα proteins were not detected. It is also possible that EM regulates NF-κB nuclear localization without affecting IκBα. In other cell types, several compounds were found to inhibit p65 nuclear localization, but were not found to cause IκBα degradation.^55,56 However, in all these cases, the underlying mechanisms are still not known. It is possible that EM affects nuclear shuttling of p65 and NF-κB regulators,⁵⁷ which would be our next step of investigation. Interestingly, EM did not inhibit NF-kB activity in synovial cells from rheumatoid arthritis patients, although it did inhibit Cox-2 expression, suggesting that EM may have different effects on different cell types in the joint.⁵⁸

The direct target of EM on chondrocytes has not been established. Previous research showed that EM acts as an agonist to the motilin receptor, a receptor involved in gastrointestinal movement.⁵⁹ However, motilin and its receptors are pseudogenes in rodents^60,61; thus, the effect from our mouse experiments cannot possibly be due to its binding to the motilin receptor. Identifying the direct target of EM would be a key area in our future study. Despite these unanswered questions, in this study, we thoroughly analyzed the effect of EM on cartilage matrix loss and synovitis in an in vivo model of OA, providing insights into devising strategies for inhibiting cartilage loss and promoting cartilage repair. While prolonged systemic antibiotic use poses risks to both the individual patient as well as public health, such concerns may be addressed with EM delivery in a localized fashion or evaluating the efficacy of a non-antibiotic derivative of EM in OA. This derivative has already demonstrated anti-inflammatory activity in the lung,²¹ and in following with current research here, may offer great potential as a therapeutic for halting OA progression.

Footnotes

Acknowledgments and Funding

We thank Dr. Gail Sonenshein (Tufts University) for providing us with the NF-κB luciferase reporter construct. We thank Daisy Nakamura, Averi Gibson, and Carrie Hui Mingalone (Tufts University) for critically reviewing this article. This work has been supported by grants to LZ from the National Institutes of Health (AR054611), the National Science Foundation (CBET-0966920), and a grant from the Tufts Clinical Translational Science Institute (UL1TR001064). The funding agencies had no role in the design, analysis or interpretation of the data or in the writing of the article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical Approval

Ethical approval for this study was obtained from the Institutional Animal Care and Use Committees at Tufts University (APPROVAL NUMBER/ID B2014-11).

Animal Welfare

The present study followed international, national, and/or institutional guidelines for humane animal treatment and complied with relevant legislation.

References

Kim

Jeon

Shin

Won

Lee

Kwak

. Regulation of the catabolic cascade in osteoarthritis by the zinc-ZIP8-MTF1 axis. Cell. 2014;156(4):730-43.

Loeser

Goldring

Scanzello

Goldring

. Osteoarthritis: a disease of the joint as an organ. Arthritis Rheum. 2012;64(6):1697-707.

Thysen

Luyten

Lories

. Loss of Frzb and Sfrp1 differentially affects joint homeostasis in instability-induced osteoarthritis. Osteoarthritis Cartilage. 2015;23(2):275-9.

Bitton

. The economic burden of osteoarthritis. Am J Manag Care. 2009;15(8 Suppl):S230-5.

Daghestani

Pieper

Kraus

. Soluble macrophage biomarkers indicate inflammatory phenotypes in patients with knee osteoarthritis. Arthritis Rheumatol. 2015;67(4):956-65.

Goldring

Otero

Plumb

Dragomir

Favero

El Hachem

. Roles of inflammatory and anabolic cytokines in cartilage metabolism: signals and multiple effectors converge upon MMP-13 regulation in osteoarthritis. Eur Cell Mater. 2011;21:202-20.

Berenbaum

. Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Osteoarthritis Cartilage. 2013;21(1):16-21.

Furman

Mangiapani

Zeitler

Bailey

Horne

Huebner

. Targeting pro-inflammatory cytokines following joint injury: acute intra-articular inhibition of interleukin-1 following knee injury prevents post-traumatic arthritis. Arthritis Res Ther. 2014;16(3):R134.

Elsaid

Zhang

Shaman

Patel

Schmidt

Jay

. The impact of early intra-articular administration of interleukin-1 receptor antagonist on lubricin metabolism and cartilage degeneration in an anterior cruciate ligament transection model. Osteoarthritis Cartilage. 2015;23(1):114-21.

10.

Kamath

. Simultaneous targeting of IL-1α and IL-1β by a dual-variable-domain immunoglobulin (DVD-Ig™) prevents cartilage degradation in preclinical models of osteoarthritis. Arthritis Rheum. 2011;63(10):S693-4.

11.

Beekhuizen

Gierman

van Spil

Van Osch

Huizinga

Saris

. An explorative study comparing levels of soluble mediators in control and osteoarthritic synovial fluid. Osteoarthritis Cartilage. 2013;21(7):918-22.

12.

Chevalier

Goupille

Beaulieu

Burch

Bensen

Conrozier

. Intraarticular injection of anakinra in osteoarthritis of the knee: a multicenter, randomized, double-blind, placebo-controlled study. Arthritis Rheum. 2009;61(3):344-52.

13.

Kuroki

Stoker

Sims

Cook

. Expression of Toll-like receptors 2 and 4 in stifle joint synovial tissues of dogs with or without osteoarthritis. Am J Vet Res. 2010;71(7):750-4.

14.

Xie

Dong

. Calcitonin treatment is associated with less severe osteoarthritis and reduced toll-like receptor levels in a rat model. J Orthop Sci. 2014;19(6):1019-27.

15.

Schelbergen

Blom

van den Bosch

Slöetjes

Abdollahi-Roodsaz

Schreurs

. Alarmins S100A8 and S100A9 elicit a catabolic effect in human osteoarthritic chondrocytes that is dependent on Toll-like receptor 4. Arthritis Rheum. 2012;64(5):1477-87.

16.

Liu-Bryan

Terkeltaub

. Chondrocyte innate immune myeloid differentiation factor 88-dependent signaling drives procatabolic effects of the endogenous Toll-like receptor 2/Toll-like receptor 4 ligands low molecular weight hyaluronan and high mobility group box chromosomal protein 1 in mice. Arthritis Rheum. 2010;62(7):2004-12.

17.

Abdollahi-Roodsaz

Joosten

Koenders

van den Brand

van

Loo

van den Berg

. Local interleukin-1-driven joint pathology is dependent on toll-like receptor 4 activation. Am J Pathol. 2009;175:2004-13.

18.

Kwiatkowska

Maślińska

. Macrolide therapy in chronic inflammatory diseases. Mediators Inflamm. 2012;2012:636157.

19.

Ren

Chen

Wooley

. Erythromycin inhibits wear debris-induced osteoclastogenesis by modulation of murine macrophage NF-κB activity. J Orthop Res. 2004;22(1):21-9.

20.

Ren

Peng

Mayton

Ren

. Erythromycin inhibits wear debris-induced inflammatory osteolysis in a murine model. J Orthop Res. 2006;24(2):280-90.

21.

Desaki

Okazaki

Sunazuka

Omura

Yamamoto

Takizawa

. Molecular mechanisms of anti-inflammatory action of erythromycin in human bronchial epithelial cells: possible role in the signaling pathway that regulates nuclear factor-κB activation. Antimicrob Agents Chemother. 2004;48(5):1581-5.

22.

Sadreddini

Noshad

Molaeefard

Moloudi

Ardalan

Ghojazadeh

. A double blind, randomized, placebo controlled study to evaluate the efficacy of erythromycin in patients with knee effusion due to osteoarthritis. Int J Rheum Dis. 2009;12(1):44-51.

23.

Johnson

Svensson

Etten

Ghosh

Murphy

Powell

. Mediation of spontaneous knee osteoarthritis by progressive chondrocyte ATP depletion in Hartley guinea pigs. Arthritis Rheum. 2004;50(4):1216-25.

24.

Nishida

Kubota

Aoyama

Takigawa

. Impaired glycolytic metabolism causes chondrocyte hypertrophy-like changes via promotion of phospho-Smad1/5/8 translocation into nucleus. Osteoarthritis Cartilage. 2013;21(5):700-9.

25.

Guingamp

Gegout-Pottie

Philippe

Terlain

Netter

Gillet

. Mono-iodoacetate-induced experimental osteoarthritis: a dose-response study of loss of mobility, morphology, and biochemistry. Arthritis Rheum. 1997;40(9):1670-9.

26.

Guzman

Evans

Bove

Morenko

Kilgore

. Mono-iodoacetate-induced histologic changes in subchondral bone and articular cartilage of rat femorotibial joints: an animal model of osteoarthritis. Toxicol Pathol. 2003;31(6):619-24.

27.

Haslauer

Proffen

Johnson

Hill

Murray

. Gene expression of catabolic inflammatory cytokines peak before anabolic inflammatory cytokines after ACL injury in a preclinical model. J Inflamm (Lond). 2014;11(1):34.

28.

Woo

Joo

Jung

Cho

. Grape seed proanthocyanidin extract ameliorates monosodium iodoacetate-induced osteoarthritis. Exp Mol Med. 2011;43(10):561-70.

29.

Duthu

. Interspecies correlation of the pharmacokinetics of erythromycin, oleandomycin, and tylosin. J Pharm Sci. 1985;74(9):943-6.

30.

Kroboth

Brown

Lyon

Kroboth

Juhl

. Pharmacokinetics of single-dose erythromycin in normal and alcoholic liver disease subjects. Antimicrob Agents Chemother. 1982;21(1):135-40.

31.

Glasson

Chambers

Van Den Berg

Little

. The OARSI histopathology initiative—recommendations for histological assessments of osteoarthritis in the mouse. Osteoarthritis Cartilage. 2010;18(Suppl 3):S17-23.

32.

Uchimura

Foote

Smith

Matzkin

Zeng

. Insulin-like growth factor II (IGF-II) inhibits IL-1β-induced cartilage matrix loss and promotes cartilage integrity in experimental osteoarthritis. J Cell Biochem. 2015;116(12):2858-69.

33.

Krenn

Morawietz

Burmester

Kinne

Mueller-Ladner

Muller

. Synovitis score: discrimination between chronic low-grade and high-grade synovitis. Histopathology. 2006;49(4):358-64.

34.

Matsukura

Muneta

Tsuji

Miyatake

Yamada

Abula

. Mouse synovial mesenchymal stem cells increase in yield with knee inflammation. J Orthop Res. 2015;33(2):246-53.

35.

Cairns

Uchimura

Kwon

Lee

Seufert

Matzkin

. Muscle cells enhance resistance to pro-inflammatory cytokine-induced cartilage destruction. Biochem Biophys Res Commun. 2010;392(1):22-8.

36.

Clements

Ball

Jones

Brinckmann

Read

Murray

. Cellular and histopathological changes in the infrapatellar fat pad in the monoiodoacetate model of osteoarthritis pain. Osteoarthritis Cartilage. 2009;17(6):805-12.

37.

Bove

Calcaterra

Brooker

. Weight bearing as a measure of disease progression and efficacy of anti-inflammatory compounds in a model of monosodium iodoacetate-induced osteoarthritis. Osteoarthritis Cartilage. 2003;11(11):821-30.

38.

Wirth

Benichou

Kwoh

Guermazi

Hunter

Putz

. Spatial patterns of cartilage loss in the medial femoral condyle in osteoarthritic knees: data from the Osteoarthritis Initiative. Magn Reson Med. 2010;63(3):574-81.

39.

Wirth

Hellio

Graverand

Wyman

Maschek

Hudelmaier

Hitzl

. Regional analysis of femorotibial cartilage loss in a subsample from the Osteoarthritis Initiative progression subcohort. Osteoarthritis Cartilage. 2009;17(3):291-7.

40.

Thomas

Fuller

Whittles

Sharif

. Chondrocyte death by apoptosis is associated with cartilage matrix degradation. Osteoarthritis Cartilage. 2007;15(1):27-34.

41.

Cortial

Gouttenoire

Rousseau

Ronzière

Piccardi

Msika

. Activation by IL-1 of bovine articular chondrocytes in culture within a 3D collagen-based scaffold. An in vitro model to address the effect of compounds with therapeutic potential in osteoarthritis. Osteoarthritis Cartilage. 2006;14(7):631-40.

42.

Beutler

Poltorak

. Identification of Toll-like receptor 4 (Tlr4) as the sole conduit for LPS signal transduction: genetic and evolutionary studies. J Endotoxin Res. 2001;7(4):277-80.

43.

Greenwald

. The road forward: the scientific basis for tetracycline treatment of arthritic disorders. Pharmacol Res. 2011;64(6):610-3.

44.

Greenwald

Chowdhury

Moak

Wang

Crossley

Caterson

. Long term doxycycline treatment inhibits histologic and immunologic features of spontaneous osteoarthritis in guinea pigs. Trans Orthop Res Soc. 1994;19:473.

45.

Brandt

. Modification by oral doxycycline administration of articular cartilage breakdown in osteoarthritis. J Rheumatol Suppl. 1995;43:149-51.

46.

da Costa

Nuesch

Reichenbach

Juni

Rutjes

. Doxycycline for osteoarthritis of the knee or hip. Cochrane Database Syst Rev. 2012;11:CD007323.

47.

Otterness

Brandt

Le Graverand

Mazzuca

. Urinary TIINE concentrations in a randomized controlled trial of doxycycline in knee osteoarthritis: implications of the lack of association between TIINE levels and joint space narrowing. Arthritis Rheumatism. 2007;56(11):3644-9.

48.

Guingamp

Gegout-Pottie

Philippe

Terlain

Netter

Gillet

. Mono-iodoacetate-induced experimental osteoarthritis: a dose-response study of loss of mobility, morphology, and biochemistry. Arthritis Rheum. 1997;40(9):1670-9.

49.

Little

Smith

Ghosh

Bellenger

. Histomorphological and immunohistochemical evaluation of joint changes in a model of osteoarthritis induced by lateral meniscectomy in sheep. J Rheumatol. 1997;24(11):2199-209.

50.

Bullough

Yawitz

Tafra

Boskey

. Topographical variations in the morphology and biochemistry of adult canine tibial plateau articular cartilage. J Orthop Res. 1985;3(1):1-16.

51.

Bullough

. The role of joint architecture in the etiology of arthritis. Osteoarthritis Cartilage. 2004;12(Suppl A):S2-9.

52.

Fukui

Miyamoto

Nakajima

Ikeda

Hikita

Furukawa

. Zonal gene expression of chondrocytes in osteoarthritic cartilage. Arthritis Rheum. 2008;58(12):3843-53.

53.

Liu-Bryan

. Synovium and the innate inflammatory network in osteoarthritis progression. Curr Rheumatol Rep. 2013;15(5):323.

54.

Yang

Hreggvidsdottir

Palmblad

Wang

Ochani

. A critical cysteine is required for HMGB1 binding to Toll-like receptor 4 and activation of macrophage cytokine release. Proc Natl Acad Sci U S A. 2010;107(26):11942-7.

55.

Tegeder

Niederberger

Israr

Gühring

Brune

Euchenhofer

. Inhibition of NF-κB and AP-1 activation by R- and S-flurbiprofen. FASEB J. 2001;15(1):2-4.

56.

Wong

Jiang

Fan

. Suppression of RelA/p65 nuclear translocation independent of IkappaB-alpha degradation by cyclooxygenase-2 inhibitor in gastric cancer. Oncogene. Feb 27 2003;22(8):1189-1197.

57.

Birbach

Gold

Binder

Hofer

de Martin

Schmid

. Signaling molecules of the NF-κB pathway shuttle constitutively between cytoplasm and nucleus. J Biol Chem. 2002;277(13):10842-51.

58.

Fumimori

Honda

Migita

Hamada

Yoshimuta

Honda

. Erythromycin suppresses the expression of cyclooxygenase-2 in rheumatoid synovial cells. J Rheumatol. 2004;31(3):436-41.

59.

De Smet

Mitselos

Depoortere

. Motilin and ghrelin as prokinetic drug targets. Pharmacol Ther. 2009;123(2):207-23.

60.

Irwin

Chen

Zhang

. Stepwise loss of motilin and its specific receptor genes in rodents. J Mol Endocrinol. 2010;44(1):37-44.

61.

Sanger

Holbrook

Andrews

. The translational value of rodent gastrointestinal functions: a cautionary tale. Trends Pharmacol Sci. 2011;32(7):402-9.