Osteoclast function, bone turnover and inflammatory cytokines during infective exacerbations of cystic fibrosis

Journal of Cystic Fibrosis 9 (2010) 93 – 98 www.elsevier.com/locate/jcf

Original Article

Osteoclast function, bone turnover and inflammatory cytokines during infective exacerbations of cystic fibrosis Elizabeth F. Sheada , Charles S. Haworthb , Helen Barkerb , Diana Biltonc , Juliet E. Compstond,⁎ a

Department of Haematology, Addenbrooke's Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK b Adult Cystic Fibrosis Centre, Papworth Hospital, Cambridge, UK c Department of Cystic Fibrosis, Royal Brompton Hospital, London, UK d Department of Medicine, University of Cambridge, Cambridge, UK Received 21 September 2009; received in revised form 19 November 2009; accepted 23 November 2009 Available online 14 December 2009

Abstract Background: Raised levels of pro-inflammatory, pro-resorptive cytokines during pulmonary infection may contribute to osteoporosis in cystic fibrosis (CF). We assessed osteoclast number and activity during infective exacerbations and examined their relationship to serum inflammatory cytokines and bone turnover markers. Methods: Serum samples from 24 adults with CF were obtained before, during and after treatment of infection. Osteoclastic cells were generated from peripheral blood mononuclear cells and their number and activity assessed. Serum osteocalcin, type 1 collagen cross-linked N-telopeptide (NTx), interleukin-6 (IL-6), tumour necrosis factor alpha (TNFα), receptor activator of NFkB ligand (RANKL) and osteoprotegerin (OPG) were measured. Results: Osteoclast number and activity were increased at the start of exacerbation and decreased with antibiotic therapy. Significant correlations were demonstrated between osteoclast formation and serum TNFα, OPG, osteocalcin and NTx and between osteoclast activity and serum IL-6 and NTx. Conclusions: The systemic response to infection is associated with increased bone resorptive activity in patients with CF. © 2009 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Cystic fibrosis; Osteoporosis; Osteoclasts; Cytokines; Bone turnover

1. Introduction Cystic fibrosis (CF) is associated with reduced bone mineral density (BMD) and increased risk of fracture [1–4]. The aetiology is multifactorial and includes vitamin D insufficiency, malnutrition, glucocorticoid use, reduced levels of physical activity and hypogonadism [5]. However, the most consistent correlate of low bone mass is the severity of disease, as defined by lung function and nutritional parameters [6]. The mechanism of this association remains only partially identified, but may relate to effects of the systemic inflammatory response to pulmonary infection on osteoclast function and formation [7–9]. Thus many of the cytokines released during infection also have stimulatory effects on osteoclast development and activity, resulting in increased ⁎ Corresponding author.

bone resorption; these include interleukin-6 (IL-6), tumour necrosis factor alpha (TNFα) and receptor activator of NFkB ligand (RANKL) [10–12]. Inflammation and associated alterations in cytokine levels have been implicated in the pathogenesis of bone loss in postmenopausal osteoporosis [13], rheumatoid arthritis [14–16], Paget's disease of bone, multiple myeloma [17], periodontitis [18] and inflammatory bowel disease [19,20]. In patients with CF, increased production of a range of proinflammatory, pro-resorptive cytokines has been reported during infective exacerbations and, in some studies, BMD or rates of bone loss have been shown to correlate with circulating levels of these cytokines [21–23]. Based on the hypothesis that bone resorption is increased during infection as a result of increased cytokine production, we studied osteoclast number and activity during infective exacerbations in patients with CF and investigated their relationship to changes in serum cytokine levels and biochemical markers of bone formation and resorption.

1569-1993/$ - see front matter © 2009 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jcf.2009.11.007

94

E.F. Shead et al. / Journal of Cystic Fibrosis 9 (2010) 93–98

2. Methods 2.1. Patients Patients were recruited from Papworth Adult Cystic Fibrosis Centre, UK using the following inclusion criteria: CF confirmed by gene analysis or abnormal sweat test, age ≥18 years, forced expiratory volume in 1 s (FEV1) b 75% predicted, at least one course of intravenous (IV) antibiotic therapy for respiratory exacerbation in the preceding year and primarily infected with Pseudomonas aeruginosa. Patients were excluded if they had received oral glucocorticoids in the 3 months prior to recruitment or during the study period, had received bisphosphonates, were pregnant during the study period, had renal dysfunction (elevated serum urea and/or creatinine) or if they were post-transplantation. Six healthy controls were recruited (3 male, 3 female) from laboratory staff with a mean age of 23.3 [1.2] years. The study was approved by the Local Research Ethics Committee and informed written consent was obtained from all patients and controls. In 9 patients peripheral blood samples were taken at baseline (clinically stable, defined as no exacerbation requiring treatment for ≥ 1 month), day 1 (start of exacerbation prior to intravenous antibiotic therapy, exacerbation defined by Fuchs criteria [24]), day 14 (end of intravenous antibiotic therapy and patients clinically stable with clinical evidence of resolution) and day 42 (follow-up clinic appointment). However, due to an insufficient number of patients experiencing exacerbations after baseline samples had been taken, the protocol was changed so that blood samples in the subsequent 15 patients were taken at just three time points during exacerbation: day 1, day 14 and day 42 as defined above. All 24 patients had exacerbation during the study period and in all cases were clinically stable at day 42. Clinical evaluation of infective status before, during and after treatment was performed by a senior clinician (CSH or DB), based on clinical examination, lung function tests and blood tests. For practical reasons it was not possible to obtain fasting samples, although most samples were obtained during the morning.

tartrate-resistant acid phosphatase (TRAP) stain, which is specific for osteoclastic cells. Positive staining cells were quantitatively assessed by microscopy. Osteoclast activity was assessed by measurement of the area resorbed on Osteologic™ slides [25]. The total area resorbed was quantified using image capture software or Adobe® Photoshop® (Adobe systems, Inc.) to calculate total area resorbed as a percentage of total area.

2.4. Enzyme linked immunosorbant assays (ELISA assays) ELISA kits used were IL-6 (900-K16), TNFα (900-K25), sRANKL (900-K142) from Peprotech Ltd. (Europe) and OPG (DY805) from R&D systems. Serum osteocalcin was measured using the Biosource (France) human osteocalcin (KAP1381) ELISA. Serum type-I collagen cross-linked N-telopeptide (NTx) was measured using the Osteomark (Unipath Ltd., UK) competitive inhibition human NTx (9021) ELISA. Intra- and inter-assay variability were calculated using standards and inhouse controls and, expressed as coefficients of variation, were 4.3% and 5.8% for IL-6, 6.1% and 8.8% for sRANKL, 8.1% and 3.4% for OPG, and 9.3% and 10.1% for TNFα. The intraassay variation for osteocalcin and NTx was 3.5% and 4.9% respectively.

2.5. Statistical analysis An ANOVA was performed as a global assay to detect variance over the time course of an exacerbation. A non-paired Tukey's multiple comparison test was used to compare the baseline measurements obtained in 9 patients with normal controls and to examine changes over the time course of the study in all 24 patients. Correlation analysis was performed using a two-tailed non-parametric Spearman's rank test with paired data points. All graphs box and whisker plots show the median and interquartile range (IQR). For all tests p b 0.05 was considered significant.

2.2. Serum separation

3. Results

Peripheral blood was collected into serum-gel monovettes (Sarstedt, UK), mixed well and allowed to clot. Serum was frozen at − 80 °C within 2 h of separation.

3.1. Patient demographics

2.3. Osteoclast generation Peripheral blood mononuclear cells were plated at 3.0 × 105 cells/ml on to Osteologic™ slides (BD Biosciences, UK) or 96-well plates and cultured in α-Minimal Essential Medium (MEM) containing RANKL (50 ng/ml) (Insight Biotechnology) and macrophage-colony stimulating factor (25 ng/ml) (RnD Systems). Cultures were incubated at 37 °C in 5% CO2-air for up to 10 days (Osteologic™ slides) or 14 days (plastic) before analysis of osteoclast formation and resorption [25]. The number of osteoclastic cells was assessed using a

24 patients (14 male, 10 female) were recruited with a mean age [SD] of 27.0 [6.0] years. All had pancreatic insufficiency. 22 were F508 del homozygous, one F508 del/Q220X and one G551D/unknown. Mean values [SD] at recruitment for FEV1 and BMI were 46.7% [15.4] and 21.1 [3.2] kg/m2, respectively. Demographic data in the 9 patients with true baseline samples were not significantly different from those in the remaining 15 patients. For the 9 patients that had baseline samples, the mean (SD) number of days between these baseline samples being taken and day 1 of their next exacerbation was 162 (157) days. For the 15 patients without baseline samples, the mean (SD) number of days between their day 1 exacerbation sample and their last exacerbation was 125 (122) days.

E.F. Shead et al. / Journal of Cystic Fibrosis 9 (2010) 93–98

3.2. Osteoclast formation and bone resorption No significant difference was seen in the number of TRAP positive cells formed in controls (median 39.1, IQR 45.1–54.6) and CF patients at baseline (median 29.6, IQR 37.0–53.8). Over the course of an infective exacerbation there was a significant variation in the number of TRAP positive cells formed in culture (p = 0.013; Fig. 1A). A significant increase in the number of TRAP positive cells was seen at day 1 of exacerbation (median 48.3, IQR 52.8–59.1) and day 14 (median 52.9, IQR 48.9–58.8) compared to baseline (p b 0.05). A significant decrease was seen by day 42 (median 44.3, IQR 39.4–48.6) compared with the day 1 and day 14 values, to a level comparable to baseline (p b 0.05) and normal controls. No significant difference was seen between resorption area in CF patients at baseline (median 7.9%, IQR 6.9–11.8) and normal controls (median 6.8%, IQR 4.7–9.4). Over the course of an infective exacerbation there was a significant variation in the resorption area (expressed as a % of total area) (p = 0.005; Fig. 1B). A significant increase in resorption area was seen at day 1 (median 12.9%, IQR 11.6–17.0) and day 14 (median 12.0%, IQR 7.4– 15.4) compared to baseline (p b 0.05). Levels of resorption at day 42 were comparable to baseline and normal controls.

95

p b 0.05). No significant variation in osteocalcin levels was seen over the time course of an infective exacerbation (Fig. 2A). Levels of NTx in CF patients at baseline (median 375.0 BCE/l, IQR 322.4–424.5) were not significantly different to those seen in normal controls (median 373.7 BCE/l, IQR 296.3–469.0). Significant variation in serum levels of NTx was of an seen over the time course of an infective exacerbation (p = 0.027; Fig. 2B). A significant increase in NTx levels compared to baseline was seen at day 14 (median 613.2 BCE/l, IQR 476.8–691.4, p b 0.05). By day 42 levels of NTx had decreased (median 512.5 BCE/l, IQR 393.1–697.7), although had not returned to baseline. 3.4. Serum cytokine levels

Serum osteocalcin levels in CF patients at baseline (median 7.3 ng/ml, IQR 3.9–11.9) were significantly lower than those seen in normal controls (median 12.5 ng/ml, IQR 9.3–16.9;

Serum levels of sRANKL in CF patients at baseline (median 39.9 pg/ml, IQR 16.3–48.2) did not differ significantly from levels in normal controls (median 38.7 pg/ml, IQR 19.3–49.1). The median level at day 1 was 36.9 pg/ml (IQR 816.4–2683). Significant variation in serum levels of sRANKL was seen over the time course of an infective exacerbation (p = 0.05) with significantly increased levels at day 14 (median 132.9 pg/ml, IQR 37.6–729.1, p b 0.05) compared to baseline. Serum levels decreased by day 42 (median 48.4 pg/ml, IQR 18.3–63.5) to a level comparable to baseline and normal controls. Serum levels of OPG in CF patients at baseline (median 1664 pg/ml, IQR 1545–1740) were significantly lower than those in normal controls (median 2494 pg/ml, IQR 1805–3066, p b 0.05). At day 1 the median level was 1560 pg/ml (IQR 8.7– 317.8). Significant variation in serum levels of OPG was seen

Fig. 1. Variations in osteoclast formation and activity during an infective exacerbation. (A) Formation of TRAP positive cells and (B) resorption activity both increased at day 1 and day 14 compared to baseline (*p b 0.05).

Fig. 2. Variations in markers of bone metabolism during infective exacerbation. (A) Serum levels of osteocalcin at baseline were significantly lower than normal control levels. (B) Serum levels of NTx were significantly increased at day 14 compared to baseline (*p b 0.05).

3.3. Serum markers of bone turnover

96

E.F. Shead et al. / Journal of Cystic Fibrosis 9 (2010) 93–98

over the time course of infective exacerbation (p = 0.03) with a significant increase at day 14 (median 2207 pg/ml, IQR 1869– 2895, p b 0.05) compared to baseline. The level had decreased by day 42 (median 1739 pg/ml, IQR 1486–1994, p b 0.05) to a level comparable to baseline. Serum levels of IL-6 in CF patients at baseline (median 456.0 pg/ml, IQR 233.0–891.9) were significantly higher than levels in the controls (median 115.0 pg/ml, IQR 95.2–199.7). Significant variation in serum levels of IL-6 was seen over the time course of infective exacerbation (p = 0.005) with a significant increase at day 1 (median 756.0 pg/ml, IQR 175.2–1253.0) compared to baseline. A decrease in levels compared to baseline was seen by day 14 (median 125.0 pg/ml, IQR 68.1–195.0, p b 0.05) and day 42 (median 56.0 pg/ml, IQR 33–82.4, p b 0.01). Serum levels of TNFα in CF patients at baseline (median 32.4 pg/ml, IQR 26.7–73.9) did not differ significantly from levels in normal controls (median 61.9 pg/ml, IQR 53.7–115.8). The median value at day 1 was 34.7 pg/ml (range 10.3–96.1). Significant variation in serum levels of TNFα was seen over the time course of infective exacerbation (p = 0.003) with significant increases at day 14 (median 113.8 pg/ml, IQR 48.3–137.6) and day 42 (median 113.0 pg/ml, IQR 82.5–136.3) compared to baseline. 3.5. Correlation between osteoclast formation, activity and serum levels of bone turnover markers and cytokines The number of TRAP positive cells in culture showed positive correlations with resorption area (r = 0.40, p = 0.02; Fig. 3A), serum NTx (r = 0.48, p = 0.014) and serum TNFα (r = 0.37, p = 0.03) and a negative correlation with serum osteocalcin (r = − 0.38, p = 0.04) and OPG (r = − 0.36, p = 0.05). Positive correlations were also seen between resorption area and serum NTx (r = 0.84, p = 0.005; Fig. 3B) and IL-6 (r = 0.41, p = 0.02). 4. Discussion The results of our study add new evidence to support the role of infection in the pathogenesis of CF-related low BMD. In particular, our results show evidence of increased production and activity of osteoclasts during infective exacerbations in association with increases in circulating levels of pro-resorptive cytokines and serum NTx, a marker of whole body bone resorption. These findings extend our previous work showing an increase in the production of potential osteoclast precursors in the peripheral blood during CF infective exacerbations [9]. The significant correlations observed in the present study between serum NTx and both TRAP positive cell production and resorption area add further evidence to the proposed association between infection and increased bone resorption. Our results are broadly consistent with and extend those reported in two previous studies in adults with CF. In one of these [8], serum samples were obtained at the start and end of a 2 week course of antibiotics for infective exacerbation. At the start of therapy, serum levels of NTx, IL-1β and IL-6 were

Fig. 3. Correlations between osteoclast formation and activity and markers of bone metabolism during infective exacerbation. (A) Formation of TRAP positive cells and resorption activity (p b 0.05) and (B) serum levels of NTx and resorption activity (p b 0.01) were significantly correlated during infective exacerbation.

elevated compared to healthy controls and fell during antibiotic therapy, although they remained higher than the control values at the end of the 2 week study period. The values for serum IL-6 and TNFα obtained in our study were higher than those reported previously, both in patients and controls, although all values were within the normal range provided by the manufacturer. Aris et al. [7] studied 17 adults with CF at day 1 of antibiotic therapy and subsequently on day 18 and day 40. Serum IL-1β, IL-6 and TNFα values declined during antibiotic therapy and NTx also decreased, having been elevated above the normal reference range at day 1. In both of these studies, baseline samples prior to the infective exacerbation were not obtained. In contrast, in the present study a number of patients (n = 9) had blood sampling prior to infection, providing a true baseline from which to demonstrate changes attributable to infective exacerbation and allowing for a true ‘non-exacerbating’ comparison to be made with normal controls. Unfortunately, because of the low rate of infective exacerbations in patients in whom a baseline sample had been obtained, it was necessary to change the protocol and thus true baseline samples were only available in 9 patients. However, in all patients samples were obtained at day 42, when there was clinical evidence of resolution of infection in all cases and most of the measured parameters had returned to or near baseline values. Changes in serum levels of RANKL and OPG have not been previously reported in this patient group. RANKL is a major regulator of osteoclast production and activity, interacting with the RANK receptor to stimulate the formation of osteoclasts and

E.F. Shead et al. / Journal of Cystic Fibrosis 9 (2010) 93–98

inhibit apoptosis of mature osteoclasts [26]. Osteoprotegerin is a soluble decoy receptor that binds to RANKL, thereby preventing it from binding to RANK and inhibiting its proresorptive effect. Levels of serum OPG in baseline samples from patients with CF were significantly lower when compared to normal controls whereas serum RANKL levels were similar in the two groups and day 1 levels were similar to those obtained at baseline for both cytokines. The low levels of OPG, in the presence of normal serum RANKL levels may have contributed to increased osteoclastogenesis and activity at the beginning of the infective exacerbation. An increase in serum levels of both cytokines during infective exacerbation was observed in our study, although this was only statistically significant for OPG; levels of both cytokines decreased after the completion of intravenous antibiotic therapy. An increase in OPG levels has been reported to accompany increased RANKL production in other situations associated with bone loss and may represent a compensatory phenomenon aimed at protecting the skeleton from excessive bone resorption [27]. In the present study baseline values of NTx, a marker of bone resorption, were similar in CF patients and controls, whilst baseline serum osteocalcin levels were significantly lower in CF patients. These findings are consistent with our previous histomorphometric study in patients with CF, which demonstrated decreased bone formation as the predominant change, whilst bone resorption was normal in most patients, although increased in some [28]. Aris et al. [7] reported higher levels of urinary NTx and free deoxypyridinoline, both markers of bone resorption, in clinically stable CF patients when compared to controls, but similar serum osteocalcin levels. The reasons for these differences are unclear but may be related to differences in clinical characteristics of the patient groups, for example disease activity, glucocorticoid use and/or measurement-related issues (i.e. serum vs. urinary NTx assays). Taken together, the current evidence supports a role for both reduced bone formation and increased bone resorption in the pathophysiology of bone loss associated with CF, the latter being most conspicuous during infective exacerbations. Our study has several limitations. First, the number of patients studied was relatively small and the statistical power to demonstrate changes was thus limited and further compromised by our ability to obtain true baseline blood samples, prior to infection, in only 9 of the 24 patients. However, demographic characteristics and changes in cytokine and biochemical marker levels were similar in the patients with and without true baseline values. The power of the study to show significant changes may also have been reduced by differences between patients in the severity of infective exacerbation and the corresponding effect of this on the variance of some of the measured indices. The number of controls studied was also small. Secondly, levels of IL-6, osteocalcin and NTX, (but not RANKL, OPG or TNFα) exhibit circadian variations [29,30] and may also be affected by intake of food; fasting samples taken at the same time of day would therefore have been optimal. However, for practical reasons it was not possible to obtain fasting samples in most patients in this study, although in the majority sampling was performed in the morning hours. Finally, osteoclastic cells

97

generated from peripheral blood mononuclear cells may not be fully representative of osteoclasts formed in vivo and peripheral blood mononuclear cells circulating at the time of infection may also have the ability to differentiate into dendritic cells or macrophages. The demonstration of a direct association between infective exacerbations in CF and increased bone resorption has potential clinical implications and suggests that measures taken to reduce the risk of infection and to treat infective exacerbations promptly and effectively may improve bone health. Our findings also provide a rationale for the use of anti-resorptive drugs in the management of bone disease in patients with CF. Acknowledgements The authors thank members of the Bone Research Group, University of Cambridge, the Haematology Department, Addenbrooke's Hospital, NHS Foundation Trust and Elaine Gunn at the Adult Cystic Fibrosis Centre, Papworth Hospital NHS Foundation Trust, for their help and guidance in this study. They also thank Kalliopi Vrotsou, MRC Biostatistics Unit, Cambridge, for statistical advice relating to this work. References [1] Haworth C, Selby P, Webb A, Dodd M, Musson H, McL Niven R, et al. Low bone mineral density in adults with cystic fibrosis. Thorax 1999;54 (11):961–7. [2] Conway S, Morton A, Oldroyd B, Truscott J, White H, Smith B, et al. Osteoporosis and osteopenia in adults and adolescents with cystic fibrosis: prevalence and associated factors. Thorax 2000;55(9):798–804. [3] Elkin S, Fairney A, Burnett S, Kemp M, Kyd P, Burgess J, et al. Vertebral deformities and low bone mineral density in adults with cystic fibrosis: a cross-sectional study. Osteoporos Int 2001;12(5):366–72. [4] Aris RM, Winders AD, Buell HE, Riggs DB, Lester GE, Ontjes DA. Increased rate of fractures and severe kyphosis: sequelae of living into adulthood with cystic fibrosis. Ann Intern Med 1998;128(3):186–93. [5] Ott SM. Osteoporosis in patients with cystic fibrosis. Clin Chest Med 1998;19(3):555–67. [6] Aris R, Guise TA. Cystic fibrosis and bone disease: are we missing a genetic link? Europ Resp J 2005;25(1):9–11. [7] Aris R, Stephens A, Ontjes D, Denene Blackwood A, Lark R, Hensler M, et al. Adverse alterations in bone metabolism are associated with lung infection in adults with cystic fibrosis. Am J Respir Crit Care Med 2000;162 (5): 1674–8. [8] Ionescu A, Nixon L, Evans W, Stone M, Lewis-Jenkins V, Chatham K, et al. Bone density, body composition, and inflammatory status in cystic fibrosis. Am J Respir Crit Care Med 2000;162(3):789–94. [9] Shead E, Haworth CS, Gunn E, Bilton D, Scott MA, Compston JE. Osteoclastogenesis during infective exacerbations in patients with cystic fibrosis. Am J Crit Care Med 2006;174(3):306–11. [10] Saidenberg-Kermanach N, Bessis N, Cohen-Solal M, De Vernejoul MC, Boissier MC. Osteoprotegerin and inflammation. Eur Cytokine Netw 2002;13(2):144–53. [11] Tanaka Y, Nakayamada S, Okada Y. Osteoblasts and osteoclasts in bone remodelling and inflammation. Curr Drug Targets Inflamm Allergy 2005;4(3):325–8. [12] Xing L, Schwarz EM, Boyce BF. Osteoclast precursors, RANKL/RANK, and immunology. Immunol Rev 2005;208:19–29. [13] Clowes J, Riggs BL, Khosla S. The role of the immune system in the pathophysiology of osteoporosis. Immunol Rev 2005;208:207–27. [14] Walsh N, Crotti TN, Goldring SR, Gravallese EM. Rheumatic diseases: the effects of inflammation on bone. Immunol Rev 2005;208:228–51.

98

E.F. Shead et al. / Journal of Cystic Fibrosis 9 (2010) 93–98

[15] O'Gradaigh D, Bord S, Ireland D, Compston JE. Osteoclastic bone resorption in rheumatoid arthritis and the acute-phase response. Rheumatol 2003;41(11):1429–30. [16] Walsh N, Gravallese EM. Bone loss in inflammatory arthritis: mechanisms and treatment strategies. Curr Opin Rheumatol 2004;16(4):419–27. [17] Ehrlich L, Roodman GD. The role of immune cells and inflammatory cytokines in Paget's disease and multiple myeloma. Immunol Rev 2005;208: 252–66. [18] Graves D, Cochran D. The contribution of interleukin-1 and tumor necrosis factor to periodontal tissue destruction. J Periodontol 2003;74(3): 391–401. [19] Moschen AR, Kaser A, Enrich B, Ludwiczek O, Gabriel M, Obrist P, et al. The RANKL/OPG system is activated in inflammatory bowel disease and relates to the state of bone loss. Gut 2005;54(4):479–87. [20] Vestergaard P. Prevalence and pathogenesis of osteoporosis in patients with inflammatory bowel disease. Minerva Med 2004;95(6):469–80. [21] Conway S. Impact of lung inflammation on bone metabolism in adolescents with cystic fibrosis. Paed Resp Rev 2001;2(4):324–31. [22] Bell S, Bowerman AM, Nixon LE, Macdonald IA, Elborn JS, Shale DJ. Metabolic and inflammatory responses to pulmonary exacerbation in adults with cystic fibrosis. Eur J Clin Invest 2000;30:553–9. [23] Haworth C, Selby P, Webb A, L. M, JS. E, Sharples L, et al. Inflammatory related changes in bone mineral content in adults with cystic fibrosis. Thorax 2004;59(7): 613–7.

[24] Fuchs H, Borowitz DS, Christiansen DH. Effect of aerosolized recombinant human DNase on exacerbations of respiratory symptoms and on pulmonary function in patients with cystic fibrosis. The Pulmozyme Study Group. N Engl J Med 1994;331:637–42. [25] Beeton CA, Bord S, Ireland D, Compston JE. Osteoclast formation and bone resorption are inhibited by megakaryocytes. Bone 2006;39(5): 985–90. [26] Boyce BF, Xing L. Functions of RANKL/RANK/OPG in bone modeling and remodeling. Arch Biochem Biophys 2008;473:139–46. [27] Jorgensen L, Skjelbakken T, Lochen ML, Ahmed L, Bjornerem A, Joakimsen R, et al. Anemia and the risk of non-vertebral fractures: the Tromso Study. Osteoporos Int 2009 Dec 3 (Electronic publication ahead of print). [28] Elkin S, Vedi S, Bord S, Garrahan NJ, Hodson ME, Compston JE. Histomorphometric analysis of bone biopsies from the iliac crest of adults with cystic fibrosis. Am J Respir Crit Care Med 2002;166(11):1470–4. [29] Perry MG, Kirwan JR, Jessop DS, Hunt LP. Overnight variations in cortisol, interleukin 6, tumour necrosis factor a and other cytokines in people with rheumatoid arthritis. Ann Rheum Dis 2009;68:63–8. [30] Shimizu M, Onoe Y, Mikumo MH, Miyabara Y, Kuroda T, Yoshikata R, et al. Variations in circulating osteoprotegerin and soluble RANKL during diurnal and menstrual cycles in young women. Horm Res 2009;71:285–9.