1. Osteoclasts are recruited to the subchondral bone in naturally occurring post-traumatic equine carpal osteoarthritis and may contribute to cartilage degradation 
A. Bertuglia, M. Lacourt, C. Girard, G. Beauchamp, H. Richard, S. Laverty 
Osteoarthritis and Cartilage 
Volume 24, Issue 3, Pages (March 2016)
DOI: /j.joca
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

2. Fig. 1
Study design. The intercarpal surface of equine third carpal bones (C3) from independent horses was scored for articular cartilage macroscopic changes and classified as post-traumatic osteoarthritis affected surface (PTOA: n = 10) or control (n = 5). One cylindrical osteochondral core was cut from the radial facet of C3 close to the dorsal border of the bone and notched for orientation purposes and decalcified. Sections were stained with Hematoxylin, Eosin, Phloxin and Safran (HEPS) and SOFG. Immunohistochemistry was performed with Cathepsin K and RANKL. All the sections were digitalized and the hyaline/calcified cartilage were digitally separated from the subchondral bone and blindly evaluated. A computer analysis was employed to generate a bone mask of the subchondral bone. The subchondral bone mask and the corresponding cartilage were subdivided in 1-mm-width ROIs for a subregional histomorphometric assessment. This yielded n = 42 controls and n = 83 PTOA ROIs for subregional analyses (dependent observations). Parameters assessed included hyaline cartilage modified Mankin score, calcified cartilage microcracks and cartilage pits, osteoclast density, histomorphometry and RANKL expression.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

3. Fig. 2
Comparisons between Control and PTOA specimens: Hyaline articular cartilage score, osteoclast density and osteoclast location in bone. A. Total hyaline cartilage scores in either complete sections (n = 15 independent observations) or subregional ROI analyses (n = 125 dependent observations) in control and PTOA groups. A Wilcoxon (complete sections) or a Cochran–Mantel–Haenszel test (ROIs) was employed to detect differences between independent or dependent observations respectively in groups (control and PTOA). B. Osteoclast density (Oc/TA cells/mm2) scores in either complete sections (n = 15 independent observations) or subregional ROI analyses (n = 125 dependent observations) in control and PTOA groups. A Wilcoxon test (complete sections) was employed to detect differences between independent observations in groups (control and PTOA). A mixed linear model (using horse ID as a random factor) was employed to detect differences between dependent observations (ROIs) in groups (control and PTOA). C. Osteoclast location in mineralized tissues. Osteoclast density in the three zones of the mineralized tissues (calcified cartilage, subchondral plate and subchondral trabecular bone) in control and PTOA groups (n = 15 horses; three zones for each specimen). A Wilcoxon test was employed in each zone to determine significant differences between independent observations in groups (control and PTOA).
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

4. Fig. 3
HEPS, SOFG and Cathepsin K immunohistochemical stained sections revealing osteoclasts. A. A basic multicellular units with arrow pointing to three osteoclasts in a cutting cone. B & C. Numerous osteoclasts are visible in this section. Dotted square magnified in C. Arrow reveals large osteoclasts. D, E and F. Cathepsin K stained sections with intracellular Cathepsin K in osteoclasts. Insert from E is magnified (F) to reveal morphology. G, H and I. Additional examples of osteoclast morphology observed in the subchondral bone. Bar = 100 μm.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

5. Fig. 4
SOFG and RANKL immunohistochemical stained osteochondral sections. A. SOFG stained section revealing intact hyaline cartilage structure with loss of staining in the superficial zone of cartilage. Microcracks are visible in the calcified cartilage. B. RANKL staining in the superficial zone of the hyaline cartilage and patchy staining in the subchondral bone. C. Close up of dotted square in B revealing superficial diffuse staining. Patchy artefactual staining within the hyaline cartilage. D. Loss of staining in all the hyaline cartilage. Coalescence of microcracks in the subchondral bone. E. RANKL staining (intracellular and pericellular) in deep zone of the hyaline cartilage and calcified cartilage. Marked staining in the subchondral bone matrix. F. Close up of dotted square in E. G. Loss of staining principally in the hyaline cartilage. Collapse of the calcified cartilage into the subchondral bone on the left of the image. H. RANKL staining in the intercellular zone in the matrix of the hyaline cartilage and in the subchondral bone matrix. I. Close up of dotted square in H. J. Loss of up to 50% of hyaline cartilage matrix and loss of Safranin O staining. K. & L. Intense focal uptake of RANKL stain in the calcified cartilage. M. Focal loss of hyaline cartilage structure and Safranin O staining in lesion top the right of the image. N. & O. RANKL staining inversely proportional to loss of Safranin O staining. P. Focal loss of hyaline cartilage structure and Safranin O staining in lesion top the right of the image. Q. & R. RANKL staining both intracellularly and in the matrix. Key: RANKL = receptor-activator of Nuclear Factor kappa-β ligand. A, B, D, E, G, H, J, K, M, N, P, Q: Bar = 100 μm. C, F, I, L, O, R: Bar = 1 mm: 500 μm.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

6. Fig. 5
RANKL expression in Control and PTOA osteochondral sections. Total RANKL scores in either complete sections (n = 14 independent observations) or subregional ROI analyses (n = 118 dependent observations) in control and PTOA groups. A Wilcoxon (complete sections) or a Cochran–Mantel–Haenszel (ROIs) test was employed to detect differences between independent or dependent observations, respectively, in groups (control and PTOA).
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

7. Fig. 6
Correlations between Osteoclast density and total hyaline cartilage score, hyaline cartilage structural & Safranin O stain scores and calcified cartilage microcracks in complete sections and on subregional analysis. A. Correlation of osteoclast density and total hyaline cartilages score in both complete sections (n = 15 independent observations) and ROIs (n = 125 dependent observations). B. Correlation of osteoclast density and hyaline cartilage structure score in both complete sections (n = 15 independent observations) and ROIs (n = 125 dependent). C. Correlation of osteoclast density and hyaline cartilage Safranin O stain score in both complete sections (n = 15 independent observations) and ROIs (n = 125 dependent observations). D. Correlation of osteoclast density and calcified cartilage microcracks in both complete sections (n = 15 independent observations) and ROIs (n = 125 dependent observations). A–D. Spearman rank correlation and Pearson's correlation test were employed to compare osteoclast density with total hyaline cartilage scores, hyaline cartilage structure score, hyaline cartilage Safranin O stain score, and calcified cartilage microcracks in the complete sections. A mixed linear model (using horse ID as a random factor) was employed to assess the association between the osteoclast density and total hyaline cartilage score, hyaline cartilage structure score, and hyaline cartilage Safranin O stain score, and calcified cartilage microcracks in the ROIs. Key: TA = total area.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

8. Fig. 7
Osteoclast density correlated with the hyaline articular cartilage RANKL expression. Correlations between osteoclast density and RANKL hyaline cartilage score in complete sections (n = 14 independent observations) and ROIs (n = 118 observations). Spearman rank correlations were employed to compare osteoclast density and RANKL hyaline score in the complete sections. A mixed linear model (using horse ID as a random factor) was employed to assess the association between the osteoclast density and RANKL hyaline cartilage score in the ROIs. Key: TA = total area.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

9. Fig. S1
Sections stained immunohistochemically with either RANKL or Cathepsin K antibodies and controls. A. Osteochondral section stained with RANKL. Increased staining can be observed at the surface of the cartilage and also in the subchondral bone matrix. B. Negative control: Adjacent section to A stained with PBS-BSA. C. Cathepsin K stained section with increased staining at the surface of hyaline cartilage and also visible in the bone matrix. D. Negative control: Adjacent section to A stained with rabbit antiserum. Key: RANKL = receptor-activator of Nuclear Factor kappa-β ligand. Bar = 500 μm
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions

10. Fig. S2
Articular calcified cartilage RANKL score correlated with hyaline cartilage histological scores and microcracks in ROIs. A. Total hyaline cartilage score correlated with the calcified cartilage RANKL score. B. Hyaline cartilage structure score correlated with the calcified cartilage RANKL score. C. Hyaline cartilage Safranin O score correlated with the calcified cartilage RANKL score. A–C. A Cochran–Mantel–Haenszel test for ordinal variables and stratified by horse, to take into account the non-independence of the multiple observations on ROIs for each horse, was used to test the association between the hyaline cartilage score, the hyaline cartilage structure score, hyaline cartilage Safranin O stain score and the calcified cartilage RANKL score. D. Microcracks in calcified cartilage correlated with calcified cartilage RANKL score. A mixed linear model (using horse ID as a random factor), to take into account the non-independence of the multiple measurements obtained for each horse, tested the association between the microcracks in calcified cartilage and the RANKL score. Key: RANKL = receptor-activator of Nuclear Factor kappa-β ligand.
Osteoarthritis and Cartilage  , DOI: ( /j.joca )
Copyright © 2015 Osteoarthritis Research Society International Terms and Conditions