Resolvin E1 and chemokine-like receptor 1 mediate bone preservation

Abstract

The polyunsaturated ω-3 fatty acid eicosapentaenoic acid-derived resolvin E1 (RvE1) enhances resolution of inflammation, prevents bone loss, and induces bone regeneration. Although the inflammation-resolving actions of RvE1 are characterized, the molecular mechanism of its bone-protective actions are of interest. To test the hypothesis that receptor-mediated events impact bone changes, we prepared transgenic mice overexpressing the RvE1 receptor chemokine-like receptor 1 (chemR23) on leukocytes. In zymosan-initiated peritonitis, neutrophil polymorphonuclear leukocyte infiltration in response to RvE1 was limited requiring log order lower doses in chemR23tg mice. Ligature-induced alveolar bone loss was diminished in chemR23tg mice. Local RvE1 treatment of uniform craniotomy in the parietal bone significantly accelerated regeneration of the bone defect. In in vitro bone cultures, RvE1 significantly enhanced expression of osteoprotegerin (OPG) without inducing change in receptor activator of NF-κB ligand levels, whereas the osteogenic markers alkaline phosphatase, bone sialoprotein, and Runt-related transcription factor 2 remained unchanged. These results indicate that RvE1 modulates osteoclast differentiation and bone remodeling by direct actions on bone, rescuing OPG production and restoring a favorable receptor activator of NF-κB ligand/OPG ratio, in addition to known anti-inflammatory and proresolving actions.

Figure 1

Characterization of ChemR23tg mouse. A. ChemR23 transgene construction. The full-length hChemR23 cDNA was cloned downstream of the hCD11b promoter in pCDNA3 plasmid. The 2.9 kB KpnI-NotI fragment was used for pronuclear injections. B. Genotyping of ChemR23tg × FVB offspring with PCR. Genomic DNA was amplified with primers for mouse ChemR23 (316 base pairs, upper band) and human transgenic ChemR23 (253 bp, lower band) in the same reaction. Double bands in lanes 2, 3, 4 and 6 indicate transgenic mice (Tg). Single bands correspond to WT mice. Std: size standards. C. Leukocyte RNA was reverse-transcribed and amplified with quantitative real-time PCR using human chemR23-specific primers. CT values represent the PCR cycle number at which an arbitrary threshold of PCR product accumulation is achieved. Lower CT values indicate higher gene expression. Transgenic lines Tg#3 and Tg#33 express ChemR23 mRNA abundantly in comparison to WT samples. Positive control: purified transgene DNA. D. Flow cytometry analysis of peritoneal leukocyte expression of ChemR23 in transgenic and wild-type mice. Leukocytes were isolated from ChemR23 transgenic mice (dark line) and WT controls (grey line) and incubated with PE-conjugated antibodies for ChemR23. Non-permeabilized ChemR23tg leukocytes show increased cell-surface ChemR23 immunoreactivity compared to WT leukocytes.

Figure 2

Leukocyte infiltration in zymosan-induced peritoneal inflammation with RvE1 in WT and ChemR23tg mice. RvE1 (0–100ng) was injected intraperitoneally simultaneously with 0.2 mg zymosan in 1 mL PBS. 24 hours later, leukocytes were harvested from the peritoneum and quantified with immunocytochemistry and flow cytometry. RvE1 (1 and 10 ng) decreased total leukocyte (A) and neutrophil (B) recruitment in chemR23tg mice while in WT mice only the 10ng dose was effective. Macrophage recruitment is increased with 100 ng RvE1 (C). In ChemR23tg mice, the RvE1 inhibitory dose response curve for PMN is shifted to the left (D). (mean±SEM, n=3 each data point, †: P<0.05 between WT and chemR23tg, two-way ANOVA, *: P<0.05 for doses indicated in pairwise comparisons).

Figure 3

Periodontal disease induced by molar ligation. A 9-0 silk suture was tied around the second maxillary molar to induce periodontal disease and alveolar bone loss. White arrow points to the cementoenamel junction – alveolar bone crest (CEJ-ABC) distance on WT (A) and chemR23tg (B) samples. CEJ-ABC distance is similar in the non-ligated (control) side of WT and ChemR23tg mice, however, ligation induces significantly less bone loss in chemR23tg mice (C)(mean ± SD, *P<0.05, t-test, n= 8 each group).

Figure 4

RvE1 significantly enhances bone healing in vivo. A: A 1 mm wide circular bone defect was created in the parietal bone of WT and chemR23tg mice and treated with subperiosteal injections of RvE1 (100ng in 20 µL) every other day for 2 weeks. Bone healing is expressed as percentage of original defect. RvE1 significantly enhanced bone healing in both WT and chemR23tg (Tg) mice (Mean + SEM, P<0.05, t-test, n=16 for each group). No significant difference between WT and Tg was found. B: Histological section across a healing calvarial bone defect (Masson’s trichrome staining).

Figure 5

ChemR23 is expressed in calvarial osteoblast cultures. A: RT-PCR of a 357 bp fragment of the chemR23 transcript (Std.: 100bp size standards, d1: 1-day, d10: 10-day old cultures). B: Immunohistochemistry of chemR23 expression in 10-day old calvarial cultures. Cell nuclei were labeled with DAPI (4',6-diamidino-2-phenylindole). Isotype ctr: Negative control staining with a non-specific antibody.

Figure 6

RANKL and OPG measurements in primary osteoblast cultures. Inflammatory phenotype was induced with administration of IL-6 (10nM) and IL-6 receptor (IL6R, 10nM) for 48 hours. Culture supernatants were assayed for RANKL and OPG with ELISA. RANKL expression was not significantly altered by RvE1 (10 ng) when administered with IL-6 (10ng) and IL-6R (10ng) (A). RvE1 (1–100nM) rescued OPG expression that was suppressed by administration of IL-6 and IL-6R (B). EPA and chemerin did not impact OPG levels under these conditions (B) (mean±SEM, *: P<0.05, ANOVA, n=4).