Clinical and Experimental Rheumatology 2009; 27: 475-482. Chronic arthritis directly induces quantitative and qualitative bone disturbances leading to compromised biomechanical properties J. Caetano-Lopes1, A.M.Nery1,2,3, R. Henriques4,5, H. Canhão1,6, J. Duarte7,8, P.M. Amaral2,3, M. Vale1,9, R.A. Moura1, P.A. Pereira1, P. Weinmann1, S. Abdulghani1, M. Souto-Carneiro8, P. Rego9, J. Monteiro9, S. Sakagushi10, M.V. Queiroz6, Y.T. Konttinen11, L. Graça7,8, M.F. Vaz2,3, J.E. Fonseca1,6 1Rheumatology Research Unit, Instituto de Medicina Molecular, Faculdade de Medicina da Universidade de Lisboa, Lisbon, Portugal; 2Instituto de Ciência e Engenharia de Materiais e Superfícies, Instituto Superior Técnico, Lisbon, Portugal; 3Departamento de Engenharia de Materiais, Instituto Superior Técnico, Lisbon, Portugal; 4Unidade de Biologia Celular, Instituto de Medicina Molecular, Faculdade de Medicina da Universidade de Lisboa, Lisbon, Portugal; 5Instituto de Sistemas e Robótica, Instituto Superior Técnico, Lisbon, Portugal; 6Serviço de Reumatologia e Doenças Ósseas Metabólicas, Hospital de Santa Maria, Lisbon, Portugal; 7Cellular Immunology Unit, Instituto de Medicina Molecular, Faculdade de Medicina, Universidade de Lisboa, Lisbon, Portugal; 8Instituto Gulbenkian de Ciência, Oeiras, Portugal; 9Serviço de Ortopedia, Hospital de Santa Maria, Lisbon, Portugal; 10Department of Experimental Pathology, Institute for Frontier Medical Sciences, Kyoto University, Kyoto 606-8507, Japan; 11University of Helsinki, ORTON Orthopaedic Hospital of the Invalid Foundation, Helsinki, COXA Hospital for Joint Replacement, Tampere, Finland. Abstract Objectives Rheumatoid arthritis (RA) is associated with an increased risk of fragility fractures. In RA patients, the direct effect of inflammation on bone is difficult to study because their skeleton is also affected by medication with corticosteroids and other drugs as well as aging and menopause, which contribute to bone fragility. This study used an animal model of chronic arthritis to evaluate the direct impact of chronic inflammation on biomechanical properties and structure of bone. Methods In the SKG mouse chronic arthritis model three point bending tests were performed on femoral bones and compression tests on vertebral bodies. Collagen structure was analysed using second-harmonic generation (SHG) imaging with a two-photon microscope, ultramorphology by scanning electron microscopy (SEM) coupled with energy dispersive x-ray spectroscopy (EDS) and bone density using water pycnometer. Results Arthritic bones had poor biomechanical quality compared to control bones. SHG, SEM and pycnometry disclosed variable signs of impaired collagen organization, poor trabecular architecture and low bone density. Conclusion Present data demonstrate for the first time that chronic inflammation per se, without confounding influence of drugs and aging, leads to impairment of bone biomechanics in terms of stiffness, ductility and ultimate strength (fracture). Key words Rheumatoid arthritis, SKG mice, osteoporosis, bone, mechanical tests, multiphoton microscopy. ----!@#$NewPage!@#$---- 476 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. Joana Caetano-Lopes, PhD student* Ana Margarida Nery, MSc* Ricardo Henriques, PhD student Helena Canhão, MD, PhD Joana Duarte, PhD student Pedro M. Amaral, PhD Mário Vale, MD Rita A. Moura, PhD student Patrícia A. Pereira, MSc Pamela Weinmann, PhD Saba Abdulghani, PhD Margarida Souto-Carneiro, PhD Paulo Rego, PhD Jacinto Monteiro, MD, PhD Shimon Sakagushi, PhD Mário Viana Queiroz, MD, PhD Yrjö T. Konttinen, MD, PhD Luís Graça, MD, PhD Maria Fátima Vaz, PhD João Eurico Fonseca, MD, PhD *Both authors contributed equally to this study. This work was supported by the PTDC/ SAU-BEB/65992/2006 grant from Fundação para a Ciência e Tecnologia. Please address correspondence and reprint requests to: João Eurico Fonseca, Rheumatology Research Unit, Instituto de Medicina Molecular, Edifício Egas Moniz, Faculdade de Medicina da Universidade de Lisboa, Av. Professor Egas Moniz, 1649-028 Lisboa, Portugal. E-mail: jefonseca@netcabo.pt Received on November 19, 2008; accepted in revised form on March 3, 2009. © Copyright CLINICAL AND EXPERIMENTAL RHEUMATOLOGY 2009. Conflict of interest: Dr Y.T. Konttinen has received support from the Sigrid Jusélius Foundation, evo grants, Finska Läkaresällskapet and the ORTON Foundation; the other co-authors have declared no competing interests. Introduction Rheumatoid arthritis (RA) is a chronic inflammatory disease, of unknown ae- tiology, which affects around 1% of the world-population (1). Bone erosions can develop in the first few months of the disease. These bone erosions are caused by the inflammatory synovial membrane, which attaches to the joint surface (pannus) and invade cartilage and bone (2). Recent evidences suggest that bone remodelling disturbances also contribute to bone erosions and in par- ticular to the development of secondary osteoporosis (OP). In fact, RA patients have an increased risk of vertebral frac- tures, which is independent from bone mineral density and corticosteroid use (3). In addition, an elevated hip frac- ture risk was also noted in RA patients not exposed to corticosteroids (4). Thus, RA seems in itself to predispose to fractures although the underlying mechanisms and their effects on bone are not completely understood yet. The connection between OP and RA can be partially explained by the ac- tivation of the Receptor Activator of Nuclear Factor κB (RANK) by RANK Ligand (RANKL) (5) produced by activated T lymphocytes (6) and os- teoblasts (7). This activates signalling leading to pre-osteoclast priming and osteoclast activation, which contribute to the formation of bone erosions and reduction of the bone mineral density. RANK/RANKL interaction is antago- nised by osteoprotegerin (OPG) (5), a soluble decoy receptor. In RA the RANKL/OPG ratio is increased (8). Osteoclasts dissolve mineralized bone matrix (mainly hydroxyapatite crys- tals) and destroy the exposed non-min- eralized organic bone matrix (mainly type I collagen) (9). However, it is not clear if, how, and to what extent the bone structure itself is affected by in- flammation although it is known that the arrangement of bone trabeculae and the orientation of collagen fibres and hydroxyapatite crystals in relation to load play an important role for bone strength (10). The main objective of the present work was to study to what extent the biomechanical properties of bone are influenced by the destructive arthritic changes. Due to multiple confounding factors involved, like the use of glucocorticos- teroids and other drugs as well as aging and hormonal changes, it is difficult to specifically study the effect of arthritis per se on the structure and biomechan- ics of bone in RA. The use of a mouse model of chronic arthritis would re- duce such variability and allow access to bones for ex vivo testing. Animal models are already widely used in the development of new drugs for OP (11). The recently described SKG mouse arthritis model resembles in many as- pects human RA. These mice develop a rheumatoid factor positive, erosive chronic polyarthritis that affects both large and small joints and present some of the systemic features of RA. The SKG mouse has a BALB/c background (12) with a single recessive point muta- tion in the ZAP-70 gene, a G→T substi- tution that alters codon 163 from tryp- tophan to cystein. This gene encodes a protein, which has an important role for T cell signal transduction and its muta- tion leads to changes in the thymic se- lection threshold. As a result SKG mice maintain otherwise negatively selected autoreactive T cell clones. Unlike their normal BALB/c counterparts, SKG mice are genetically prone to spontane- ously develop chronic polyarthritis, the onset of which can be precipitated by zymosan or by other dectin-1 agonist (12, 13). This experimental arthritis model was used to study the direct ef- fect of chronic arthritis on bone struc- ture and biomechanics. Material and methods SKG arthritis model and BALB/c controls Twenty-five female SKG and sixteen female BALB/c mice were bred and maintained under pathogen free condi- tions. All experiments were conducted according to the guidelines from the Animal User and Institutional Ethics Committee. A single intraperitoneal injection of 2mg of zymosan (Sigma- Aldrich Co, USA) was administered at 2 months of age. Joint swelling was monitored by inspection as follows: 0, no joint swelling; 0.1, swelling of one finger joint; 0.5, mild swelling of wrist or ankle; and 1.0, severe swelling of ----!@#$NewPage!@#$---- 477 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. wrist or ankle, with a range of the total sum score varying from 0 to 5 (12). At five months the SKG and BALB/c mice were sacrificed and femoral bones and vertebrae were dissected free of soft tissues and stored at -20°C. Immedi- ately before testing, the samples were defrosted at room temperature. Mechanical testing All mechanical tests were performed using a universal testing machine (In- stron 5566™, Instron Corporation, Canton, USA) with a load cell of 500 N. The biomechanical strength vari- ables were displayed in stress-strain curves by the Bluehill 2 Software (In- stron, Copyright 1997-2007) and ana- lysed using MatLab 7.1 software (R14 SP3, The Mathworks, Inc, Copyright 1984-2006). The software has the abil- ity to build stress-strain representations from load-displacement points, once initial dimensions are provided for each specimen. Femoral bones were submitted to three- point bending tests (Fig. 1A) and verte- brae to compression tests (Fig. 1B). For three-point bending tests the span between the outer loading points was 5 mm with the load being applied to the centre of the femoral shaft at a cross- head speed of 0.01 mm/s. The param- eters analysed from the stress-strain curves were Young’s modulus (E), yield stress (σy), ultimate stress (σu), ul- timate strain (εu) and work/energy until ultimate stress (Wu) (Fig. 1C). The second (L2) and fourth (L4) lum- bar vertebrae were used for compres- sion tests. In these tests a cross head speed of 0.01 mm/s was used, and then obtained Young’s modulus (E), yield stress (σy), maximum stress (σM), work until yield stress (Wy) and work until maximum stress (WM) (Fig. 1D). Assuming that femurs behave like cyl- inders (14), stress-strain curves can be built according to the following equa- tions (1 and 2): Fig. 1. Experimental setup and stress-strain representations obtained after mice femurs bending and vertebrae compression. Images of a three point bending test setup (A), and a vertebra standing on the inferior plate of a compression test setup (B). Femurs (arthritic and control) bending curves (C) were obtained in order to determine the analysed parameters directly from the graphic: Yield point=(yield strain, yield stress) and Ultimate point=(ultimate strain, ultimate stress), or after calculations: Young’s modulus – slope of the curve between the origin and the yield point; Energy until ultimate point – area under the graphic from the origin until the ultimate point, calculated using trapezoidal numerical integration. Vertebrae (arthritic and control) compression curves (D) were obtained from the mechanical test in or- der to determine the analysed parameters directly from the graphic: First yield point=(first yield strain, first yield stress), Second yield point = (second yield strain, second yield stress) and Densification point = (densification strain, densification stress), or after calculations: Young’s modulus – slope of the curve between the origin and the first yield point; Energy until first yield, or densification points – area under the graphic from the origin until the first yield point or until densification point, calculated using trapezoidal numerical integration. (1) (2) ----!@#$NewPage!@#$---- 478 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. Where σ - stress (Pa); L = load (N); s = support span (mm); df = femoral outer diameter (mm); ε = strain (%); Δl = displacement (mm). All the calculations were done for the middle cross-section area, applied to the centre of the femoral shaft, by the upper load support. From these curves the mechanical parameters were read, except for the Young’s modulus, which was obtained from the slope of the elas- tic regime of the stress-strain curves, and the energies absorbed until deter- mined points that were calculated us- ing trapezoidal numerical integration, with using MatLab 7.1 software. For vertebrae, compression tests were made (14). Equations (3 and 4) show how stress-strain curves can be ob- tained, assuming that a cylinder, of homogenous material, is being com- pressed: Where dv = vertebral diameter (mm); h0 = initial vertebral height (mm); Δh = compression suffered by the vertebra (displacement) (mm). As for femurs, the biomechanical pa- rameters were read from the stress strain curves and the Young’s modulus as well as the energies calculated using MatLab 7.1 software. Even though, in this paper, all the ver- tebrae tested refer to the same lumbar position (L2 and L4) and the long bones were all femurs, there was always vari- ability in dimensions associated with the fact that each bone comes from a different animal. In order to normalize this parameter, and allow all samples to be compared against each other, instead of load (L) – displacement (d) values, we used stress (σ) – strain (ε) represen- tations. Stress is the load applied on a sample, per area unit (a) (equations 1 and 3), and the strain is the deforma- tion suffered, when compared with the initial dimension of the bone (equa- tions 2 and 4). In this study it was assumed that both femurs and vertebrae were similar to cylinders, but with different orienta- tions and dimensions, according to the mechanical test (three point bending for femurs and compression for verte- brae) performed in each case. In this way, the results from the control and arthritic groups could be compared between each other, even if the dimen- sions were different, because this effect was already contemplated in the stress- strain analysis. Second-harmonic generation and two-photon excitation microscopy Mice vertebrae and femurs were de- calcified, embedded in paraffin, cut to 7μm sections using a microtome and deparaffinised to be inspected in a Zeiss LSM510 META laser scan- ning microscope featuring a Coherent Mira 900 femtosecond multiphoton excitation laser. Multicolour nonlin- ear microscopy of collagen was done through second-harmonic generation (SHG) using two-photon excitation (TPE). The signal was acquired by two opposing detectors: 1) the META spec- tral detector configured for bandpass detection between 390-430 nm collect- ing light from a Zeiss Fluar 20x/0.75 objective; this signal was associated with a green look up table (LUT) and is referred to as backward-SHG and 2) the non-descanned detector (NDD) collecting light from a Zeiss 0.8 nu- merical aperture (NA) condenser and filtered by a 390-430 nm bandpass fil- ter; this signal was associated with the blue LUT and is referred to as forward- SHG. The backward-SHG channel detects the backscattered SHG signal and the forward-SHG channel the SHG receives the photons that are transmit- ted through the sample (15). The wave- length of the laser was set to 820 nm. The Jefferies’ method (16) was applied to all acquired images to quantify the amount of trabecular area occupied by collagen detected by the forward-SHG channel (essentially mature polymer- ized collagen) and by collagen detected by the backward-SHG channel (mainly immature collagen fibril segments, in- dicating ongoing fibrillogenesis) (17). Image analysis was done using NIH ImageJ software (W. Rasband, Na- tional Institute of Health, Bethesda, USA) by applying a 40-255cps inten- sity threshold for each channel to cover only the extracellular collagen matrix. All the trabecular area was covered us- ing a lower 20-255cps threshold. Scanning electron microscopy Bone samples were inserted into an epoxidic and transparent resin and mounted in a mixture of Résine Mec- aprex MA2 (04008) and Durisseur pour résine Mecaprex MA2 (Presi SA. Tavernolles, 38320 Brie & Angonnes, France) in 100:12 ratio. After 24 hours the surfaces of the vertebrae samples were polished using grid papers with 1000, 800, 600, 320 and 200 μm gran- ulometries. After deposition of a uni- form and thin gold layer the samples were observed with an electron beam energy of 25 keV in a scanning elec- tron microscope (Hitachi model S2400 SEM. Hitachi, Ltd. Tokyo, Japan). The images obtained were analysed with the software ImageJ, already mentioned for the MPM images. In this case the region of interest was the ver- tebral body, where pictures were taken and parameters such as the area occu- pied by trabeculae (%), inter-trabecu- lar distance (μm), trabecular thickness (t¯, μm) and cortical thickness (c¯, μm) were determined. These measurements were made according to Heyn’s and Jefferies’ methods (16). Energy dispersive x-ray spectroscopy The energy dispersive x-ray spectros- copy (EDS) analysis was performed to measure the proportion of calcium and phosphorus in the bone samples using an analytical SEM with Rontec stand- ard EDS detector (Hitachi S2400, To- kyo, Japan). Density measurements The densities of both femoral bones and vertebrae were measured prior to me- chanical testing with a water pycnom- eter. The pycnometer was filled with water, with corrections being made to the ambient temperature following the Archimedes’ principle. Bone density was calculated using equation (5): (3) (4) (5) ----!@#$NewPage!@#$---- 479 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. Where ρ - bone density; ρL - water den- sity at ambient temperature; m0 - mass of the empty pycnometer; m1 - mass of the pycnometer and the bone sample; m2 - mass of the pycnometer, the bone sample and water; m3 - mass of the pycnometer filled with water. Statistical analysis Results were represented by mean and standard deviation values. According to their distribution either t test for in- dependent-samples or non–parametric Mann-Whitney test were used to com- pare continuous variables. Significance level was set at 0.05. Statistical analy- ses were performed using the Statisti- cal Package for Social Sciences (SPSS) Manager software (SPSS, Inc, Chicago, IL, USA). Results Female SKG mice develop chronic arthritis Twenty-five female SKG mice, injected with a single intraperitoneal injection of 2mg of zymosan at two months of age, developed chronic arthritis (in small and large joints) after one week and reached peak activity by the sixth week. By the time of sacrifice (5 months of age), the SKG mice had approximately 3 months of disease activity and all had a clinical score of 5. Inflammatory joint infiltration was severe and destructive (Fig. 2). The sixteen control BALB/c mice, after the 2mg zymosan injection, did not develop arthritis. These two strains of mice are identical, differing only in a point mutation in the ZAP-70 protein that only affects the selection of T lymphocytes in the thymus. BALB/c mice are widely used as a control for SKG mice. A significant reduction in body weight of the SKG mice was observed in comparison to the BALB/c mice (18.60±1.04g and 24.08±0.81g respec- tively, p<0.001). Chronic arthritis reduces mechanical properties of femurs and vertebrae Mechanical 3-point bending test re- sults (Table I and Fig. 1C) showed that arthritic femurs have a significantly lower: elastic (Young’s) modulus (re- flecting reduced stiffness), yield stress (less force was needed to cause the first microfractures and to start a plastic and definitive deformation of bone), ulti- mate stress (reflecting the maximum strength of the bone at fracture) and en- ergy until ultimate stress (reflecting the energy required to cause fracture, thus the toughness of the bone) as compared to control femurs. In the case of vertebrae compression tests (Table II and Fig. 1D), significant differences were found between arthrit- ic and control L2 and L4 vertebras with regards to elastic (Young’s) modulus, yield stress, ultimate stress and energy until yield stress, as compared to con- trol vertebrae. Bone collagen content is preserved during chronic arthritis The SHG technique with MPM was not appropriate for the study of corti- cal bone as the images produced were homogenous and did not allow us to characterize the presence and organi- zation of collagen molecules. How- ever, images obtained in vertebral trabecular bone were very informa- tive. Five control and 5 arthritic verte- brae were used to study bone collagen Fig. 2. SKG mice developed a severe destructive polyarthritis after zymosan injection. Joint swelling was monitored by inspection and scored as follows: 0: no joint swelling; 0.1: swelling of one finger joint; 0.5: mild swelling of wrist or ankle; and 1.0: severe swelling of wrist or ankle. Scores for all fingers and toes, wrists and ankles were totalled for each mouse. Graph (A) shows that the scores evalu- ation started at week zero and peak in the arthritis group at week 6 (squares and triangles represent median scores). BALB/c mice did not develop arthritis (C, E) and were used as controls to SKG mice (B, D), which showed evident swelling of the ankle joints and an erosive synovitis at 5 months of age (haematoxylin and eosin staining, magnification ×400). ----!@#$NewPage!@#$---- 480 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. content and structure. Quantitative im- age analysis of the collagen content showed no statistical differences be- tween BALB/c control mice and SKG arthritic mice (Fig. 3A and 3B). The calculated percentage of collagen oc- cupied areas depicted by forward-SHG channel in SKG and control vertebrae were 40.87±6.89% and 47.00±14.26%, respectively. This percentage was very similar to the one obtained by using the backward-SHG channel (52.81±6.05% and 53.90±11.03% for arthritic and control samples, respectively) (Fig. 3C). These percentages have to be calculated separately, as the channels in this imaging technique are indi- vidually analysed, so there is no di- rect correlation between values found for each group from the two channels. Furthermore, the ratio of mature col- lagen to immature collagen (the per- centage obtained in the forward-SHG channel dividing by the one obtained in the backward -SHG channel) was of 0.775±0.179 for arthritic animals and 0.907±0.363 for control bones (p- value of 0.372). As mature polymerised collagen has a signal present predominantly in the forward-SHG channel and amor- phous collagen fibril segments in the backward-SHG channel one can as- sume that the relative distribution of these two types of collagen and the total amount of collagen per bone sur- face are equivalent in the arthritic and control groups. However, qualitative analysis of mice vertebrae using SHG microscopy revealed structural differ- ences in the mature collagen organi- zation in the arthritic group images in comparison to controls. In fact, in the SKG mice (arthritic group) areas of a nodular pattern in the collagen organi- zation were observed among the lon- gitudinal fibrils of the mature collagen (Fig. 3A and 3B). Vertebrae from arthritic mice have a higher inter-trabecular distance and a decreased trabeculae thickness In arthritic SKG mice vertebrae the trabeculae thickness was reduced giv- ing rise to an increased inter-trabecu- lar distance in comparison to BALB/c mice vertebrae, as highlighted by SEM images (Table III and Fig. 4). Moreo- ver, vertebral cortical thickness was significantly lower in the SKG bone as compared to the controls. However, the trabeculae area, as assessed by SEM, was similar between the 2 groups. Chronic arthritis reduces bone density The density measurements were lower in the arthritic group as compared to the control group (Table IV). These results were statistically different for femurs and L2 vertebrae. Mineral bone composition is not affected by arthritis The results from the EDS analysis did not reveal any significant difference in the bone mineral composition in terms of calcium and phosphorous propor- tion between the arthritic and control bone samples, both in femurs (arthritic femur: calcium = 73% (72-75%) and phosphorus = 27% (25-28%); control femur: calcium = 73% (73-74%); and phosphorus = 27% (26-27%); no statis- tically significant differences between groups) and in vertebrae (arthritic ver- tebrae: calcium = 72% (42-79%) and phosphorus = 28% (21-58%); control vertebrae: calcium = 71% (52-79%); Table I. Mean and standard deviation values calculated from mechanical tree-point bend- ing test of mice female femurs. SKG mice BALB/c mice p-value (n=9) (n=8) Diameter (mm) 1.10 ± 0.03 1.04 ± 0.02 0.001** Stiffness (N/mm) 147.65 ± 18.87 197.84 ± 27.55 <0.001** Yield stress (MPa) 108.68 ± 18.12 221.35 ± 62.72 0.001** Ultimate stress (MPa) 171.57 ± 33.06 289.12 ± 48.28 <0.001** Young’s modulus (GPa) 6.79 ± 1.16 11.36 ± 1.96 0.001* Ultimate strain (%) 5.69 ± 1.40 5.80 ± 1.50 0.564* Energy until ultimate stress (N.mm/mm3) 5.16 ± 2.32 8.30 ± 2.58 0.019** Values are mean ± standard deviation. *Mann-Whitney test. **Independent-samples t-test. Table II. Mean and standard deviation values calculated from mechanical compression stress-strain curves of female mice vertebrae L2 and L4. L2 L4 SKG mice BALB/c mice p-value SKG mice BALB/c mice p-value (n=7) (n=6) (n=7) (n=5) Height (mm) 2.82 ± 0.39 3.28 ± 0.46 0.077* 3.23 ± 0.25 3.67 ± 0.30 0.015* Diameter (mm) 2.99 ± 0.15 2.97 ± 0.14 0.824* 2.82 ± 0.14 2.96 ± 0.07 0.063* Young’s Modulus (MPa) 20.17 ± 4.74 42.60 ± 18.70 0.007** 16.81 ± 2.33 30.29 ± 16.28 0.138* Yield stress (MPa) 2.14 ± 0.61 3.85 ± 1.39 0.029* 2.06 ± 1.08 4.45 ± 1.83 0.028** Energy until yield stress (Nmm/mm3) 0.13 ± 0.08 0.22 ± 0.10 0.109* 0.12 ± 0.09 0.30 ± 1.15 0.028** Maximum stress (MPa) 4.73 ± 0.92 6.89 ± 1.58 0.015** 3.35 ± 1.09 6.13 ± 0.87 0.005** Energy until maximum stress (Nmm/mm3) 1.13 ± 0.68 1.10 ± 0.54 0.668** 0.66 ± 0.15 0.57 ± 0.17 0.268** Values are mean ± standard deviation. *Mann-Whitney test; **Independent-samples t-test. ----!@#$NewPage!@#$---- 481 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. and phosphorus = 29% (24-48%); no statistically significant differences be- tween groups). Discussion It is well established that RA is associ- ated with secondary osteoporosis and an increased risk of fractures. This is apparently the end result of multiple factors, such as the use of corticoster- oids, aging, menopause and inflamma- tion. Thus, the direct contribution of the effect of chronic inflammation on bone in RA patients is difficult to assess. The primary aim of this study was to assess the influence of chronic inflammation in the SKG mouse chronic arthritis model on the biomechanical strength of the bone. The secondary aim was to study femoral bones and vertebral bodies, us- ing new methods to assess the effect of chronic inflammation on the collagen and mineral structure of bone tissue. This is the first report in the literature demonstrating a deleterious and direct effect of chronic inflammation per se on the biomechanical quality of bone. This may in part clarify the increased fracture risk in RA patients. Due to the anatomical qualities of the femoral bones and vertebral bodies it was necessary to use two different types of tests to measure their biomechanical strength. First, the femoral bones were tested using a three point bending test. In this test increasing force is applied to the femoral shaft using a cross-head. This leads first to fully elastic deforma- tion of the bone so that if the force is released, the bone assumes its former shape. This property of the bone is usu- ally expressed as its elastic Young’s modulus and the maximum stress still enabling such reversible changes is re- ferred to as the yield point. When more stress is applied, bone tissue starts to undergo microfractures and truly ir- reversible plastic changes occur, as it is characteristic of a ductile material. Upon release of the stress force, it can be observed that the morphological bone changes are now permanent. The maximum stress the bone can tolerate without breakage reveals its ultimate strength, after which a macroscopic fracture soon follows. In this study, it was for the first time shown that arthritic femoral bones have impaired elasticity, ductility and ultimate (fracture) strength compared to healthy control femoral bones. In accordance with these results, the compression test used to character- ize the biomechanics of the vertebral bodies also showed impairment in the arthritic mice. This could be expected by the low trabeculae area of the can- cellous vertebral bone (SEM analysis) as well as by the abnormal nodular structural organisation of its collagen Fig. 3. Multiphoton microscopy images obtained from (A) SKG mouse vertebrae and (B) BALB/c mouse vertebrae. The green colour (backward-SHG channel) corresponds to immature collagen fibril segments, indicating ongoing fibrillogenesis and the blue colour (forward-SHG channel) corresponds to the mature polymerized collagen. (C) The content of collagen in the vertebrae was not affected by arthritis. Scale bars correspond to 100μm. Fig. 4. Vertebrae from arthritic mice have a decreased trabeculae thickness and a higher inter-trabecu- lar distance. Scanning electron microscopy images from female mice vertebrae with (A) and (B) with- out arthritis, showing a higher inter-trabecular distance and a decreased trabeculae thickness. Scale bars correspond to 1mm (left) and 30μm (right). ----!@#$NewPage!@#$---- 482 Arthritis compromises bone mechanical properties / J. Caetano-Lopes et al. architecture in the vertebral bone (SHG analysis). However, differences might be even greater than the ones depicted by our results as the squared vertebral body is supported by a slowly remod- elling cortical bone shell, which, as a strong and stiff structure, protects the more vulnerable trabeculae bone inte- rior and could mask the impact of the mechanical test. Interestingly, SHG, independently both in the forward and backward mode, as well as the SEM-coupled EDS methods disclosed that the bulk collagen content and relative proportions of bone em- bedded minerals do not differ between arthritis and healthy controls. Although the reasons for the impaired quality of arthritic bone remain at present unclear, based on the current work it seems that they may relate to the mineral content and quality (architecture) of bone rath- er than to its total collagen content or the composition of the hydroxyapatite laid down in the organic bone matrix. The results of this study clearly show that chronic inflammation directly in- duces a reduction in bone density and a change in the pattern of bone organiza- tion. This is the first report to indicate that arthritis is associated with an im- pairment of the bone mechanical prop- erties, namely impaired elasticity, duc- tility and ultimate (fracture) strength. These observations encourage pharma- cological intervention studies to ana- lyze if it would be possible to counter- act these arthritis-associated and frac- ture-predisposing changes by targeting also the bone metabolism, apart from controlling the arthritis itself. References 1. ALAMANOS Y, DROSOS AA: Epidemiology of adult rheumatoid arthritis. Autoimmun Rev 2005; 4: b130-6. 2. NEUMANN E, GAY S, MULLER-LADNER U: The RANK/RANKL/osteoprotegerin system in rheumatoid arthritis: new insights from animal models. Arthritis Rheum 2005; 52: 2960-7. 3. ORSTAVIK RE, HAUGEBERG G, MOWINCKEL P et al.: Vertebral deformities in rheumatoid arthritis: a comparison with population-based controls. Arch Intern Med 2004; 164: 420-5. 4. VAN STAA TP, GEUSENS P, BIJLSMA JW et al.: Clinical assessment of the long-term risk of fracture in patients with rheumatoid arthritis. Arthritis Rheum 2006; 54: 3104-12. 5. RAISZ LG: Pathogenesis of osteoporosis: concepts, conflicts, and prospects. J Clin Invest 2005; 115: 3318-25. 6. LAMGHARI M, TAVARES L, CAMBOA N et al.: Leptin effect on RANKL and OPG ex- pression in MC3T3-E1 osteoblasts. J Cell Biochem 2006; 98: 1123-9. 7. HORWOOD NJ, KARTSOGIANNIS V, QUINN JM et al.: Activated T lymphocytes sup- port osteoclast formation in vitro. Biochem Biophys Res Commun 1999; 265: 144-50. 8. FONSECA JE, CORTEZ-DIAS N, FRANCISCO A et al.: Inflammatory cell infiltrate and RANKL/OPG expression in rheumatoid syn- ovium: comparison with other inflammatory arthropathies and correlation with outcome. Clin Exp Rheumatol 2005; 23: 185-92. 9. THEILL LE, BOYLE WJ, PENNINGER JM: RANK-L and RANK: T cells, bone loss, and mammalian evolution. Annu Rev Immunol 2002; 20: 795-823. 10. VIGUET-CARRIN S, GARNERO P, DELMAS PD: The role of collagen in bone strength. Osteoporos Int 2006; 17: 319-36. 11. PEREL P, ROBERTS I, SENA E et al.: Comparison of treatment effects between ani- mal experiments and clinical trials: system- atic review. BMJ 2007; 334: 197. 12. HIROTA K, HASHIMOTO M, YOSHITOMI H et al.: T cell self-reactivity forms a cytokine milieu for spontaneous development of IL- 17+ Th cells that cause autoimmune arthritis. J Exp Med 2007; 204: 41-7. 13. SAKAGUCHI N, TAKAHASHI T, HATA H et al.: Altered thymic T-cell selection due to a mutation of the ZAP-70 gene causes autoim- mune arthritis in mice. Nature 2003; 426: 454-60. 14. AKHTER MP, CULLEN DM, GONG G et al.: Bone biomechanical properties in prostag- landin EP1 and EP2 knockout mice. Bone 2001; 29: 121-5. 15. HELMCHEN F, DENK W: Deep tissue two- photon microscopy. Nat Methods 2005; 2: 932-40. 16. JEFFERIES Z, KLINE AH, ZIMMER EB: The determination of grain size in metals. Trans AIME 1916; 57: 594-607. 17. WILLIAMS RM, ZIPFEL WR, WEBB WW: Interpreting second-harmonic generation im- ages of collagen I fibrils. Biophys J 2005; 88: 1377-86. Table III. Trabeculae and cortical measurements of lumbar vertebrae, as determined by scanning electron microscopy. SKG mice BALB/c mice p-value (n=6) (n=4) Cortical thickness (μm) 82.40 ± 23.71 99.81 ± 27.44 <0.001* Trabeculae occupied area (%) 18.91 ± 5.39 21.63 ± 8.22 0.540** Trabeculae thickness (μm) 57.18 ± 16.78 70.91 ± 23.66 <0.001* Inter-trabecular distance (μm) 218.92 ± 88.18 188.04 ± 72.54 0.017* Values are mean ± standard deviation. *Mann-Whitney test; **Independent-samples t-test. Table IV. Density measurements for femurs and vertebra. SKG mice BALB/c mice p-value (n=6) (n=7) Femur density (g/cm3) 1.17 ± 0.02 1.35 ± 0.07 0.002** Vertebra L2 density (g/cm3) 0.82 ± 0.10 0.94 ± 0.07 0.034* Vertebra L4 density (g/cm3) 0.89 ± 0.08 1.03 ± 0.25 0.268* Values are mean ± standard deviation. *Mann-Whitney test; **Independent-samples t-test.