Abstract
Objective. To test the hypothesis that increased concentrations of prostaglandin D2 (PGD2) correlate with bone remodeling. Studies using isolated bone cells indicate that PGD2 may be implicated in the regulation of bone homeostasis, with a positive influence on bone anabolism. We studied patients with traumatic fractures and age- and sex-matched healthy controls as an in vivo model of increased bone remodeling.
Methods. Thirty-five patients with bone fracture and matched controls were recruited. Urine and sera samples were collected. Urinary 11ß-PGF2α, a PGD2 metabolite, and PGE2 metabolites (PGEM), serum lipocalin-type PGD2 synthase (L-PGDS), bone alkaline phosphatase (bone ALP), and crosslinked C-telopeptides of type I collagen (CTX) were measured.
Results. At 5–6 weeks post-fracture, 11ß-PGF2α, L-PGDS, bone ALP, and CTX were significantly increased in the fracture patients compared to controls. PGEM levels were not different between groups. Levels of 11ß-PGF2α and bone ALP were positively correlated, suggesting that PGD2 may be implicated in fracture repair.
Conclusion. These results support our working hypothesis that PGD2 could be implicated in the control of bone anabolism in humans.
- BONE
- PROSTAGLANDIN D2
- PROSTAGLANDIN SYNTHASE
- FRACTURE REPAIR
- BONE REMODELING
Fracture repair is a physiological process where inflammation, angiogenesis, and bone remodeling play important roles. A number of local and systemic factors influence the fracture repair process: these include cytokines, growth factors, and lipid mediators, which tightly regulate the action of cells responsible for bone healing1–3.
Prostaglandins (PG) strongly influence the process of fracture repair, as demonstrated by delayed repair in mice null for cyclooxygenase-2 (COX-2)4 and by animal studies where COX inhibitors were used5,6. However, despite the common use of nonsteroidal antiinflammatory drugs (NSAID) post-surgery and post-fracture, the effects of COX inhibition, if any, remain unclear during fracture repair in humans7. Some studies suggest that COX inhibition may delay fracture healing or increase the rate of non-unions8,9, but prospective and controlled clinical trials are still lacking. PG may thus play a role in human bone metabolism, remodeling, and repair.
PG act on osteoblasts and osteoclasts, the cells responsible for bone formation and resorption, respectively10. PGD2 is mostly known for its effects on the immune system and sleep11,12, but recent evidence suggests new roles for this PG in bone physiology. We demonstrated that human osteoblasts produce PGD2, and express the functional PGD2 receptors D protanoid receptor type 1 (DP) and chemoattractant receptor-homologous molecule expressed on T helper cell type 2 (CRTH2). Activation of both receptors can lead to a different osteoblast phenotype, as DP receptor activation decreases osteoprotegerin (OPG) secretion, while CRTH2 increases osteoblast migration and decreases expression of RANKL (receptor activator of nuclear factor-κB ligand)13. Moreover, PGD2 and its metabolites are potent inducers of collagen synthesis by human osteoblasts in culture14,15. Recently, we showed that PGD2 inhibits both human osteoclast differentiation and mature osteoclast activity, reducing bone resorption in vitro16. These results suggest that PGD2 may have a physiological role in bone remodeling, favoring bone anabolism.
PGD2 is produced by 2 terminal PGD2 synthases: hematopoietic-type synthase (H-PGDS) or lipocalin-type (L-PGDS) synthase. H-PGDS is strongly expressed by immune cells, while L-PGDS is mostly expressed in the brain and heart17. In addition to PGD2 synthesis, L-PGDS also acts as a transporter for small lipophilic ligands such as retinoids and thyroid hormones. Altered expression of L-PGDS has been related to detached retinas18, atherosclerosis19, sleep disorders20–22, and impaired renal function in diabetes23,24. However, no evidence in the literature links this enzyme to human bone metabolism.
Our study is based on the hypothesis that PGD2 is an important anabolic autacoid implicated in bone repair, as suggested by the in vitro studies using isolated osteoblasts and osteoclasts. Should this hypothesis be true, we would expect PGD2 production to be increased in vivo during situations of increased bone remodeling. We studied patients with traumatic fractures and compared them to healthy age- and sex-matched controls as a model of increased bone turnover to test this hypothesis. Specifically, we determined the levels of a stable PGD2 metabolite and of L-PGDS in patients within 5–6 weeks post-fracture, a period of active bone remodeling, and compared these levels to healthy controls.
MATERIALS AND METHODS
Patients
Patients with traumatic bone fracture were recruited 5–6 weeks post-fracture at the Centre Hospitalier Universitaire de Sherbrooke (CHUS) in collaboration with the Division of Orthopaedic Surgery during their routine visit at the hospital. Patients gave informed consent and the study was approved by the institution's Ethics Review Board. Age-, sex-, and weight-matched healthy controls were recruited at the same center and Bishop's University (Sherbrooke, QC, Canada). Patients and controls were excluded from the study if they used NSAID or corticosteroids less than 7 days before enrolment, if they had ever received calcitonin, bisphosphonates or chemotherapy treatments, or if they suffered from kidney failure, diabetes or asthma.
Blood and urine samples
Urine and blood samples were collected at the time of recruitment. Urine samples were immediately centrifuged at 1500 ×g for 15 min and the supernatants were frozen at −80°C until analyzed. Whole blood was drawn by standard venipuncture without anticoagulants, allowed to coagulate for 30 min at room temperature, and the serum was separated by centrifugation at 1500 ×g for 15 min. Serum aliquots were frozen at −80°C until analyzed.
Biochemical markers
PGD2 production was assessed by measuring the levels of the PGD2 primary stable metabolite 11ß-PGF2α in urine samples using enzyme immunoassay (EIA) kits (Cayman Chemical Co., Ann Arbor, MI, USA) with a detection limit of 5.5 pg/ml. This assay specifically recognizes 11ß-PGF2α with less than 0.01% cross-reaction towards PGD2 or PGE2.
PGE2 production was assayed by measuring a stable PGE2 metabolite (PGEM) after chemical derivatization of PGE2 and its primary metabolites, namely 13,14-dihydro-15-keto PGE2 and 13,14-dihydro-15-keto PGA2, in the standards and urine samples to the single PGEM compound. PGEM concentrations were determined using commercial EIA kits (Cayman Chemical) following the manufacturer's protocol, with a detection limit of 2 pg/ml.
Lipocalin-type PGD2 synthase expression was determined in the serum using the Prostaglandin D Synthase (lipocalin-type; human) EIA dosage kit (Cayman Chemical; detection limit 2 ng/ml), according to the manufacturer's protocol.
Bone formation occurring during fracture repair was assessed by serum bone-specific alkaline phosphatase (bone ALP) measured by commercial ELISA (Metra® BAP, from Quidel, distributed by Medicorp Inc., Montreal, QC, Canada) on thawed serum aliquots. This ELISA detects as low as 0.7 U/l bone ALP with 3%–8% cross-reactivity with liver ALP and 0.4% towards intestine ALP.
To assess bone resorption, serum C-terminal telopeptides of type 1 collagen (CTX) were measured using the Serum CrossLaps® ELISA (Nordic Bioscience Diagnostics; Medicorp Inc., Montreal, QC, Canada). This ELISA has a detection limit of 0.020 ng/ml of CTX in a linear range up to 3.380 ng/ml.
Urine creatinine was measured (Vitros 950 System Chemistry, Ortho Clinical Diagnostics, Johnson & Johnson) and used to correct urine PG values.
Statistical analyses
Statistical analyses were performed using GraphPad Prism version 5.01. The Wilcoxon signed-rank test was used to compare patients to paired controls, while nonparametric Spearman correlations were used between markers. Data are shown as means ± standard error. Results were considered significant when p < 0.05.
RESULTS
Our study was designed to assess possible changes in the stable PGD2 metabolite 11ß-PGF2α and L-PGDS in patients during the active bone formation phase of fracture repair. A cohort of 35 patients and matched controls was enrolled at our local hospital. No demographic differences were observed between the patients and controls (Table 1).
Prostaglandin D2 pathway is specifically increased in fracture repair
To assess PGD2 production, we measured the concentration of 11ß-PGF2α in urine samples. As shown in Figure 1, patients with fractures showed higher levels of 11ß-PGF22α in urine compared to controls (77.4 ± 8.8 vs 49.5 ± 3.9 ng/mmol creatinine, respectively; p = 0.0023). Figure 2 shows that serum L-PGDS in patients with fracture was significantly higher than that in controls (0.613 ± 0.03 vs 0.536 ± 0.03 μg/ml, respectively; p = 0.027). Since increases in PGD2 and L-PGDS could have resulted from nonspecific inflammation associated with the fracture we also determined the concentration of stable PGE2 metabolite, PGEM, in our population. As shown in Figure 3, PGEM did not differ significantly between control and patient groups (98.7 ± 12.5 vs 118.0 ± 18.8 ng/mmol creatinine, respectively; p = 0.3633), therefore arguing against a nonspecific overall increase in PG production.
Bone markers
Fracture repair is a process where both bone formation and resorption occur. Serum bone ALP and CTX were measured to assess these 2 phenomena, respectively. As expected, level of serum bone ALP was significantly higher in patients compared to controls (31.3 ± 2.6 U/l vs 24.4 ± 1.7 U/l, respectively; p = 0.0033; Figure 4). A strong increase in the bone resorption marker CTX was also observed in patients compared to controls (0.568 ± 0.07 ng/ml vs 0.301 ± 0.05 ng/ml, respectively; p = 0.0002; Figure 5), thus confirming active bone remodeling in the fracture group.
Correlation of PGD2 metabolite and bone formation marker
To determine if the higher production of the PGD2 metabolite 11ß-PGF2α in the patient group would correlate to an increase in the bone remodeling process, we tested whether serum bone ALP correlated with urine 11ß-PGF2α. As shown in Figure 6, bone ALP and 11ß-PGF2α levels were positively correlated (Spearman r = 0.2946, p = 0.0133). In contrast, no correlation was observed between 11ß-PGF2α and CTX. The expression of L-PGDS and level of PGEM were not correlated with any marker measured.
DISCUSSION
Bone metabolism is a complex and tightly controlled process implicating several cytokines, growth factors, and soluble mediators such as prostaglandins1,3. Inadequate control of the balance between bone production and resorption, either locally or systemically, is an important pathophysiologic sequence in different diseases such as rheumatoid arthritis25, spondylarthritis26, osteoarthritis27, periodontitis28, and osteoporosis29, to cite a few. Better understanding of the different elements implicated in this control may lead to identification of new pharmacological targets for treatment of these diseases.
Evidence in the literature from our group13,16 and others14,15 suggests that PGD2 may have a positive influence on bone metabolism, increasing bone formation and inhibiting bone resorption. Our objective was to test in vivo the hypothesis that increased levels of PGD2 would be associated with increased bone remodeling using traumatic bone fracture in humans as an experimental model.
Fracture repair is a physiological event regulated by local and systemic factors. Immediately after a fracture, the microenvironment of the damaged site is drastically changed. Formation of hematoma and inflammation are the first events to occur. During this period, leukocytes are attracted to the wound site, where they secrete various mediators modulating the inflammatory response, including cytokines and PG. While this inflammatory response lasts only a few days, PG-mediated effects may last longer as they remain actively produced by bone and surrounding cells13,30,31.
The importance of PG in the process of fracture repair is well characterized in animal models. COX-2 was shown to be essential to fracture repair4,6, while PGE2 receptors EP2 and EP4 were shown to be implicated in murine fracture repair32,33, but the role of PGD2 in fracture repair and bone remodeling remains unknown.
We investigated if L-PGDS expression and the stable urinary PGD2 metabolite 11ß-PGF2α were changed during the process of bone remodeling and if they correlated with markers of bone formation and resorption. 11ß-PGF2α is the major PGD2 metabolite produced in vivo and is considered a physiological marker of PGD2 production: it has been used to assess mast cell activation and PGD2 production during allergic and asthmatic responses34–36. We demonstrated that 11ß-PGF2α is significantly increased during fracture repair. To our knowledge, this is the first time that PGD2 has been linked to bone metabolism in humans.
To test whether this key enzyme was implicated in bone remodeling, serum L-PGDS was measured in patients with fractures. Like 11ß-PGF2α, L-PGDS was significantly increased during fracture repair. It is interesting that slight changes in serum L-PGDS can be a predictor or marker of pathologies such as coronary heart disease19, hypertension37, sleep apnea20, and diabetes23,24. We cannot rule out the possible contribution of the other PGD2 synthase, H-PGDS, in the process, but technical limitations prevent measurement of this intracellular enzyme in blood or other body fluids.
PGD2 is a key component of the inflammatory response38,39. Therefore, all patients were studied during an orthopedic followup visit 5 to 6 weeks post-fracture, a time at which inflammation is not considered to be a major component of the repair process40 and when bone remodeling starts to occur41. We found no significant difference in the stable PGE2 metabolite in patients compared to controls. This suggests that the increased 11ß-PGF2α levels observed in patients were not due to a general increase in PG synthesis associated with inflammation. Since osteoblasts were shown to produce PGD2, it is tempting to hypothesize that osteoblasts are responsible for the increased synthesis of PGD2 found in patients with fracture; this study, however, was not designed to allow identification of the cellular origin of PGD2.
Bone remodeling is often measured by biomarkers of bone turnover. During fracture repair, both endochondral and intramembranous bone formation occur, leading to formation of a bony and cartilaginous callus, which will later be resorbed in order for the bone to regain its previous integrity, resistance, and shape. Activation of osteoblasts and osteoclasts is critical for the completion of fracture repair. We measured serum levels of bone ALP, as a marker of bone formation and osteoblast activation, as well as serum CTX, to assess bone resorption. Concentrations of both markers were increased in the fracture patients compared to controls, indicating that there was significantly increased bone turnover in these patients. This confirms the premise that increased bone remodeling is occurring in the studied population. To test our hypothesis that PGD2 is increased during bone remodeling, we tested if 11ß-PGF2α correlated with levels of either bone ALP or CTX. Our analysis showed a significant and positive correlation with bone ALP and no correlation with CTX. Moreover, no correlation was found for the bone markers and PGEM, thus underlying the specificity of PGD2 production during fracture repair.
Our results show the first evidence that a specific prostaglandin pathway, PGD2 and its lipocalin-type synthase, is activated during fracture repair in humans. Moreover, we demonstrated that PGD2 production correlates with a classic osteoblast activation and bone formation marker. These findings support our hypothesis that PGD2 may be an important player in bone physiology and that it may constitute an important and biologically relevant anabolic stimulus for bone.
Acknowledgments
The authors thank the Division of Orthopaedic Surgery, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, for help in patient recruitment.
Footnotes
-
Supported by grants from the Canadian Institutes of Health Research. Dr. Gallant received a doctoral fellowship from the FRSQ. Dr. Parent and Dr. Roux receive scholarships from the FRSQ.
- Accepted for publication October 4, 2009.
REFERENCES
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.