All animal and tissue harvesting protocols were approved by Cornell University’s Institutional Animal Care and Use Committee (Protocol Number: 2011-0027).
Cell isolation and culture
Bone marrow was collected from the sternebrae of 31 standing, sedated Quarter Horse, Warmblood and Thoroughbred horses (n = 10 female, 2 intact male and 19 castrated male) ranging in age from 3 to 12 years old for primary BMSC isolation as previously described . Briefly, bone marrow biopsy needles (Jamshidi, VWR Scientific, Bridgeport, NJ, USA) were used to aspirate 60–180 mL of sternebral bone marrow into one to three 60 mL syringes containing heparin (APP Pharmaceuticals, LLC, Schaumburg, IL, USA) at a final concentration of 1,000 units/mL. BMSCs were isolated via selective tissue culture plastic adherence after plating bone marrow 1:1 in Dulbecco’s modified Eagles’ media; 1,000 mg/L glucose (Gibco-Life Technologies, Grand Island, NY, USA) supplemented with 1 ng/mL bFGF (Gibco, Invitrogen, Camarillo, CA, USA), 25 mM HEPES (Gibco-Life Technologies, Grand Island, NY, USA), 100 units/mL penicillin-streptomycin, and 10% fetal calf serum. Non-adherent cells were removed via media changes every other day. Following colony formation, adherent cells were passaged and replated at 10–12,000 cells/cm2. Cells were thawed following cryopreservation, cultured for 72 h, collected in lysis buffer and frozen at ˗80 °C for subsequent RNA isolation for constitutive galectin gene expression analysis. Bone marrow was collected from the sternebrae of an additional six standing, sedated Thoroughbred-type horses (n = 2 female and 4 castrated male) ranging in age from 1 to 8 years old and culture expanded as described above for use in cytokine gene and protein expression, confocal imaging and migration experiments. All cells were stored cryopreserved prior to experiments.
Adipogenic, osteogenic and chondrogenic differentiation potential of primary BMSC lines was evaluated using standard protocols following induction with StemPro® differentiation kits (Thermo Fisher Scientific, Waltham, MA, USA). Adipogenic and osteogenic differentiation were induced in confluent monolayer cultures, and chondrogenic differentiation was induced in 5 μL cell pellets. Differentiated cells and non-induced controls were stained with Oil Red O, Alizarin Red, or Toluidine Blue between 7 and 14 days and photographed with a × 20, 0.7 NA objective using an Olympus IX73 microscope equipped with an Olympus DP80 CCD camera (Olympus, Tokyo, Japan).
Synovial membrane (n = 27) and articular cartilage (n = 16) tissues were aseptically harvested from the shoulder, stifle, carpal and fetlock joints of Thoroughbred, Warmblood and Quarter Horse cadavers (n = 9 female, 6 intact male and 12 castrated male) ranging in age from 1 to 8 years old immediately post-euthanasia. All horses were healthy without gross signs of OA at dissection. Cells were isolated, pooled and culture expanded as previously described for synoviocytes  and chondrocytes . Briefly, synovial lining was digested in 0.15% collagenase (Worthington Biochemical, Lakewood, NJ, USA) and 0.015% DNAseI (Roche, Indianapolis, IN, USA) for 3 h at 37 °C, followed by filtration and centrifugation at 250 × g for 10 minutes. Synoviocytes were cultured in Dulbecco’s modified Eagles’ media; 4500 mg/L glucose (Gibco-Life Technologies, Grand Island, NY, USA) supplemented with 25 mM HEPES, 100 units/mL penicillin-streptomycin, and 10% fetal calf serum. Articular cartilage was digested in 0.075% collagenase overnight at 37 °C, followed by filtration and centrifugation at 250 × g for 10 minutes. Chondrocytes were cultured in Ham’s F12 medium (Corning Inc., Corning, NY, USA) supplemented with 50 μg/mL ascorbic acid, 30 μg/mL α-ketoglutarate, 300 μg/mL L-glutamine, 25 mM HEPES, 100 units/mL penicillin-streptomycin and 10% fetal calf serum.
Equine galectin gene expression analysis
For constitutive galectin expression, passage 1 to 3 equine BMSCs (n = 31), synovial fibroblasts (n = 27) and chondrocytes (n = 16) were cultured for 72 h after thawing in appropriate growth media as discussed above. Cells were collected in lysis buffer and frozen at ˗80 °C for subsequent RNA isolation using a purification kit (5 Prime Inc., Gaithersburg, MD, USA). Synovial membrane and cartilage tissue obtained from healthy equine carpal joints at the time of surgery or euthanasia (n = 23) from a previous study  that had been stored at ˗80 °C was thawed for RNA isolation and purification using the same purification kit. RNA purity and concentration were assessed using UV microspectrophotometry (NanoDrop 2000, Thermo Fisher Scientific, Waltham, MA, USA). Gene expression was quantified through the use of quantitative real-time polymerase chain reaction (qRT-PCR) using the ABI PRISM 7900 sequence detection system (Applied Biosystems, Foster City, CA, USA), with all samples analyzed in duplicate using primers and a dual-labeled fluorescent probe (6-FAM™ as the 5’ reporter label and Iowa Black® FQ as the 3’ quenching label). Primers and probes were generated using the equine galectin-1 and -3 sequences as described above and were designed using Primer Express software (Primer Express v2.0b8, Foster City, CA, USA):
(Galectin-1 Fwd: 5’- CAAGGCAGACCTGACCATCA -‘3,
Galectin-1 Rev: 5’- TCACGGCCTCCAGGTTGA -3’,
Galectin-1 Probe: 5’-/56-FAM/CTGCCGGAT/ZEN/GGCTACTCGTTCAAGTTC/3IABkFQ/-3’,
Galectin-3 Fwd: 5’- TAAATTTCAACAGAGGGCATGATG -3’,
Galectin-3 Rev: 5’- CAATGACTCTCCTGTTGTTCTCGTT -3’
Galectin-3 Probe: 5’-/56-FAM/TGCCTTCCA/ZEN/CTTTAACCCGCGCTT/3IABkFQ/-3’).
Total copy number of mRNA was determined from a validated standard curve using serial dilutions of E. coli-expressed equine galectin-1 and -3 as standards for absolute quantitation, and copy number was normalized to the housekeeping gene 18S.
For inflammatory cytokine treatments, passage 3 BMSCs (n = 3) were plated in duplicate in 24-well plates at a concentration of 2 × 104 cells/cm2. BMSCs remained in serum-containing MSC growth media for 24 h prior to cytokine treatments. Serum-containing media was replaced with serum-free Opti-MEM (Invitrogen, Grand Island, NY, USA) 4 h prior to stimulation. BMSCs were stimulated with recombinant equine IL-1β (IBI Scientific, Peosta, IA, USA) at 5 ng/mL and 10 ng/mL, recombinant equine TNF-α (IBI Scientific, Peosta, IA, USA) at 25 ng/ml and 50 ng/ml or LPS from E. coli 055:B5 (Sigma-Aldrich, St. Louis, MO, USA) at 0.1 μg/mL, 1 μg/mL, 10 μg/mL and 50 μg/mL. BMSCs remained in serum-free Opti-MEM as the control condition. Media supernatants were collected at 4, 8, 20 and 30 h post-treatment, frozen and stored at ˗80 °C for galectin quantification via custom ELISA. Cells were lysed at 4, 8, 20 and 30 h after treatment for RNA isolation, and gene expression was determined using qRT-PCR for galectin-1 and galectin-3 mRNA with 18S used as a housekeeping gene. In parallel, BMSCs were lysed at 4, 8, 20 and 30 h after treatment with ice-cold RIPA buffer containing protease inhibitors. Cell lysates were stored at -80 °C for immunoblotting analysis.
Equine galectin protein expression after cytokine stimulation
Custom ELISAs were developed for the detection of equine galectin-1 and galectin-3 in BMSC media supernatants following cytokine stimulation. Equine galectin-1 (GenBank ID: KY264050) and galectin-3 (GenBank ID: KY264051) were cloned and sequenced from renal tissue obtained from a 19-year-old Thoroughbred cadaver mare as previously reported . In order to assess antibody cross-reactivity and to establish equine-specific standards for galectin ELISAs, equine galectins-1 and -3 were recombinantly expressed and purified as described for human galectins-1, -3 and -3C . All antibodies were validated to react against purified equine galectin-1 or galectin-3 using dot blots (Additional file 1). Briefly, for the custom competitive equine galectin-1 ELISA, 96-well high-binding plates (Corning Inc., Corning, NY, USA) were coated with 1 ug/mL of capture antibody (R&D Systems, Minneapolis, MN, USA; goat anti-mouse Gal-1 pAb, AF1245) in sodium carbonate buffer, pH 9.6 overnight at 4 °C. After rinsing 3× in 0.1% PBS-Tween, protein-free blocking buffer (Thermo Fisher Scientific, Rockford, IL, USA) was added for 1 h. Unlabeled recombinant equine galectin-1 standards (4000 to 15.63 ng/mL, plus 0 ng/mL) were diluted in a solution of 200 ng/mL biotinylated recombinant equine galectin-1 in 0.1% PBS-Tween. Supernatants from passage 3 BMSCs (n = 3) were collected in duplicate for all treatment conditions (control, IL-1β 5 and 10 ng/mL, TNF-α 25 and 50 ng/ml, and LPS 0.1 μg/mL, 1 μg/mL, 10 μg/mL and 50 μg/mL) at all time points (4, 8, 20 and 30 h). Supernatants were diluted 1:10 in 200 ng/mL biotinylated recombinant equine galectin-1 in 0.1% PBS-Tween. Blocking buffer was removed, and 100 μL of recombinant equine galectin-1 standards or samples were added to wells, covered, and incubated for 1 h at RT on a plate shaker. After rinsing 3× in 0.1% PBS-Tween, 100 μL of streptavidin HRP was added for 30 minutes, followed by 5× rinses in 0.1% PBS-Tween. TMB reagent (Thermo Fisher Scientific, Rockford, IL, USA) was added for 30 minutes, and the reaction was stopped with 1 N H2SO4. Absorbance was measured at 450 nm with 540 nm background subtraction, and all measurements were performed in duplicate.
For the equine galectin-3 sandwich ELISA, 96-well high-binding plates were coated with 2 ug/mL of capture antibody (Santa Cruz Biotechnology, Dallas, TX, USA; goat anti-human Gal-3 pAb, sc-19280) in sodium carbonate buffer, pH 9.6 overnight at 4 °C. After rinsing 3× in 0.1% PBS-Tween, protein-free blocking buffer was added for 1 h. Blocking buffer was removed, and serial dilutions of recombinant equine galectin-3 standards (400 ng/mL to1.56 ng/mL, plus 0 ng/mL) or 1:1 dilutions of BMSC supernatants in PBS + 2% BSA were incubated for 1 h. After rinsing 3× in 0.1% PBS-Tween, biotinylated detection antibody (R&D Systems Minneapolis, MN, USA; biotinylated goat anti-mouse Gal-3 pAb, BAF1197) was added at 200 ng/mL for 1 h. Plates were rinsed, and 100 μL of streptavidin HRP was added for 30 minutes, prior to 3× rinses in 0.1% PBS-Tween and addition of TMB reagent for 30 minutes. The TMB reaction was stopped with 1 N H2SO4, and absorbance was measured at 450 nm with 540 nm background subtraction. All measurements were performed in duplicate.
In order to measure intracellular and membrane-bound galectins after cytokine stimulation, cell lysates were thawed and heated at 95 °C in reducing SDS-PAGE loading buffer for 15 minutes. 10 μL of sample per well was loaded onto a 7.5% TGX gel (Bio-Rad, Hercules, CA, USA), and subjected to SDS-PAGE for 1 h at 100 V. Gels were transferred to PVDF membranes (EMD Millipore, Billerica, MA, USA), and immunoblotting was performed using antibodies against galectin-1 (R&D Systems, Minneapolis, MN, USA) goat anti-mouse Gal-1 pAb, AF1245), galectin-3 (Santa Cruz Biotechnology, Dallas, TX, USA; goat anti-human Gal-3 pAb, sc-19280) and actin (Santa Cruz Biotechnology, Dallas, TX, USA; goat anti-human actin pAb, sc-1615). Band intensities were quantified using NIH Fiji, with background subtraction, and galectin-1 and -3 band intensities were normalized to actin. Only monomeric galectins (approximately 17 kDa for galectin-1 and approximately 31 kDa for galectin-3) were quantified.
BMSC nucleofection and confocal imaging
Passage 3 to 4 equine BMSCs (n = 3) were nucleofected with the fluorescent-protein tagged adhesion markers paxillin mEmerald and vinculin mApple using a human MSC nucleofector kit (Lonza, Basel, Switzerland). Cells were passaged at 48 h post-nucleofection and plated onto 12-well fibronectin-coated glass plates at a concentration of 2 × 103 cells/cm2 in serum-free media (Dulbecco’s modified Eagles’ media, 1000 mg/L glucose; 1 ng/mL bFGF; 25 mM HEPES; 100 units/mL penicillin-streptomycin). BMSCs were plated in the presence and absence of 100 mM β-lactose (Santa Cruz Biotechnology Inc., Dallas, TX, USA). At 8 h post-plating, glass plates were fixed in 4% paraformaldehyde for 10 minutes, rinsed in PBS, followed by actin staining with phalloidin Alexa647 and nuclear staining with DAPI for 15 minutes, followed by several PBS rinses. Cells were imaged with a × 100, NA 1.3 objective on an Olympus IX83 microscope equipped with a Yokagawa X1 spinning disk confocal scanning unit and an Andor Ultra 897 EMCCD camera. Images were overlaid in Adobe Photoshop CS5 (Adobe Systems Inc., San Jose, CA, USA).
Equine BMSC migration assays
Passage 3 to 4 equine BMSCs (n = 3) were plated onto 24-well tissue culture plates (Corning Inc., Corning, NY, USA) within silicone inserts containing a defined, 500 μm cell-free gap (Ibidi®, Martinsried, Germany) for migration assays. After 6 h, media was changed to either control media, media containing β-lactose (100 mM, 200 mM) or media supplemented with recombinant equine IL-1β (5 ng/mL, 10 ng/mL) or recombinant equine TNF-α (25 ng/ml 50 ng/ml). Twenty hours later, inserts were removed and media was replaced with control media or media containing β-lactose (100 mM, 200 mM). Phase contrast images were obtained at 0, 3, 8, 12, 24 and 48 hours following insert removal, using three images/well obtained with a × 10, NA 0.25 objective on an Olympus CK2 microscope with a Nikon Digital Sight DS-Fi1 CCD camera to image the entire cell-free gap. NIH Fiji was used to define the x,y coordinates of the leading edges of cells migrating across the cell-free region. Custom software (Python Software Foundation, Wilmington, DE, USA) was designed to measure the mean linear cell-free distance for each image, which was normalized to the cell-free distance at time zero for each treatment.
For constitutive galectin gene expression, a one-way ANOVA with Tukey’s post hoc tests was performed on log-transformed gene expression data for galectin-1 and galectin-3 using statistical software (GraphPad Software Inc., La Jolla, CA, USA). Summary statistics were performed on the untransformed data, and significance was set at p < 0.05. Galectin-1 and galectin-3 gene expression after stimulation data were compared using a Friedman nonparametric test, matching data by primary cell line (GraphPad Software Inc., La Jolla, CA, USA). Post hoc comparisons between all treatment groups and the control group were made using the method described by Gibbons and Chakraborti (p. 459, equation 2.13)  with a Bonferroni correction for multiple comparisons and significance set at p < 0.05.
Galectin-1 and galectin-3 immunoblotting and ELISA protein expression after stimulation data were also compared using a Friedman nonparametric test, matching data by primary cell line (GraphPad Software Inc., La Jolla, CA, USA). Post hoc comparisons between all treatment groups and the control group were made using the method described by Gibbons and Chakraborti (p. 459, equation 2.13)  with a Bonferroni correction for multiple comparisons and significance set at p < 0.05. Migration experiments were performed using three BMSC primary cell lines, and all treatment conditions were applied in duplicate. The normalized mean cell-free distances were analyzed using statistical software (JMP 11.0, Cary, NC, USA) to perform linear regression model fitting. Parameters included treatment, time, and BMSC primary cell line. Tukey’s honest significant difference (HSD) post hoc tests were performed to compare the effects of time and BMSC primary cell line, with significance set at p < 0.05. Dunnett’s post hoc tests were performed to compare the effects of each treatment condition to the control condition, with significance set at p < 0.05.