Determination of two fluoroquinolones and their combinations with hyaluronan effect in in vitro canine cartilage explants

Canine cartilage explant culture

Normal (no OA lesions) canine articular cartilage samples were prepared from the stifle and carpal joints of canine cadavers (3–5-year-old specimens weighing ∼10–20 kg), which had suffered a car accident and did not show any indication of muscular system disorders or diseases. Cadaver specimens were obtained from the Veterinary Cadaveric Unit, Faculty of Veterinary Medicine, Chiang Mai University, Chiang Mai, Thailand. According to the Animals for Scientific Purposes Act, B.E. 2558 (2015), since a part of this experiment was performed on an animal carcass, no ethical approval was required for this study and this was confirmed by the Animal Ethics Committee, Faculty of Veterinary Medicine, Chiang Mai University (License number U1006312558). The collection of cartilage samples and canine cartilage explant cultures has been previously described before (Siengdee et al., 2016). Briefly, within 6 h after collection, the joints were opened aseptically in the laminar flow hood. Cartilage disks were then sliced from the condyles of the carpal and tarsal joints approximately 6–7 mm in diameter and 1–2 mm in depth. All explants were soaked once with 70% ethanol for 30 s to 1 min and thrice with sterile phosphate-buffered saline, containing 10 × of a routine antibiotic dose (a final concentration of 1,000 units/mL of penicillin, 1,000 µg/mL of streptomycin, and 2.5 µg/mL of amphotericin B). The cartilage disks were weighed, and three disks (total, ∼30–35 mg/well) were placed into the same sterile 24-well plates. The cartilage disks were then incubated in 1 mL of 5% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) with a routine antibiotic dose (100 units/mL of penicillin, 100 µg/mL of streptomycin, and 0.25 µg/mL of amphotericin B) and were maintained at 37 °C under 5% CO₂ until the start of the experiment.

Treatment reagents

Recombinant human IL-1 β was obtained from Bio-Techne/R&D Systems (Minneapolis MN, USA) and used at a final concentration of 20 µg/mL. The final concentration of 1.5 mg of medium-molecular-weight HA for intra-articular injection (∼500–730 kDa) was obtained from TRB Chemedica (Bangkok, Thailand). The doses of Mar and Enro were determined based on previous studies in vitro both at 200 µg/mL (Lim et al., 2008), and 400 and 1,000 µg/mL of Mar and Enro were used as positive control (Beluche et al., 1999; Peters et al., 2002).

Experimental design

Figure 1 shows the experimental design of this treatment, and Table 1 presents the codes for the treatment conditions. The canine cartilage explants were divided into three treatment groups under IL-1 β-stimulated conditions: (1) non-IL-1 β group, cartilage explants exposed to FQs with/without HA; (2) pre-IL-1 β group, cartilage explants pre-incubated with recombinant human IL-1 β for 48 h to activate cell inflammation before exposure to FQs with/without HA; and (3) with IL-1 β group, cartilage explants co-treated with IL-1 β and FQs with/without HA at the same time. Treatment groups were inoculated with 5% FBS-supplemented Dulbecco’s modified Eagle’s medium (DMEM) containing 200 ng/mL of Enro or Mar alone and Enro or Mar co-treated with 1.5 mg HA, the negative control for each treatment group was exposed to 5% FBS-supplemented DMEM at equal volume without drugs, and the positive control was treated with 400 and 1,000 µg/mL of each drug through T5 without replacing the growth medium. All cultures were maintained at 37 °C under 5% CO₂ and examined in three replications. The conditioned medium was collected and replaced every 48 h and stored at −20 °C for further monitoring of sulfate glycosaminoglycan (s-GAG) release. The conditioned cartilage explants, which were assayed for cartilage matrix degeneration assay and examined for histopathology investigation and gene expression, were harvested at days 4 and 8 after treatment (T3 and T5; Fig. 1). Except for the positive control group, the conditioned medium and conditioned explants were collected only once at T5. All cultures were maintained at 37 °C under 5% CO₂ and examined in three replications.

Determination of s-GAG content

The s-GAG content in the conditioned medium was attributed to the s-GAG accumulation derived from both new s-GAG production and s-GAG degradation from the cartilage disks. The s-GAG content was measured using a colorimetric dye: dimethylmethylene blue (DMMB) binding assay. Subsequently, 16 mg of DMMB was dissolved in 5 mL ethanol, and the volume was adjusted to 1 L with deionized water containing 0.2% sodium formate buffer to adjust the dye solution (pH 3.5) and stored at room temperature in a bottle shielded from light. Chondroitin 6-sulfate from shark cartilage (CS-C; Sigma-Aldrich, St. Louis, MO, USA), which varies from 0 to 60 µg/mL in DMEM, was used as a standard. Subsequently, 50 µL of CS-C or conditioned medium was pipetted into 96-well plates, followed by 200 µL of DMMB dye solution, and the plate was shaken for 5 s. Thereafter, the absorbance was immediately measured at 540 nm using a microplate reader. The serum-free medium was set as blank. The CS-C standard curves were used to calculate the apparent s-GAG levels in each treatment group and calculated as a percentage of s-GAG content relative to the negative control group.

Quantification of uronic acid

The carbazole colorimetric assay was modified according to the method of Cesaretti et al. (2003) to measure the uronic acid (UA) component in the cartilage matrix. Briefly, treated cartilage disks were digested with 200 µL of papain (2 units) at 60 °C for 48 h. Subsequently, papain cartilage-digested media were diluted with distilled water at 1:5 ratio. Glucuronolactone was used as the standard reagent, and 50 µL of diluted digested sample, or standard, was placed in a 96-well plate, followed by the addition of 200 µL of sodium tetraborate/sulfuric acid reagent (0.025 M Na₂B₄O₇). Thereafter, the 96-well plate was heated at 100 °C for 15 min. After the plate was cooled down to room temperature, 50 mL of carbazole solution was added (0.125% crystallized carbazole in purified absolute ethanol). The sample was heated at 100 °C for 15 min and cooled down to room temperature again. Subsequently, the quantity of UA in the sample was measured using a microplate reader at a wavelength of 540 nm, and the UA content (% µg/mg) was calculated and normalized to the dry weight of each cartilage explant.

Safranin-O staining of GAGs in the cartilage matrix

The cartilage disks were fixed in 10% neutral buffer formalin fixative solution for approximately 12–24 h at room temperature, embedded in paraffin block, and afterward, the cartilage disks were serially sectioned into approximately 6 µm in depth. The sections of each group were stained with safranin-O to determine the GAG content (Brown, Crawford & Oloyede, 2007; Cook et al., 2010). The cartilage sections were imaged with a Biocom microscope using an AxioCam camera (Carl Zeiss, Oberkochen, Germany) and were performed by processing and calculating the stained image using ImageJ (Király et al., 1996) 1.47 morphometric computer software (Wayne Rasband, National Institute of Health, USA).

RNA extraction and quantitative real-time PCR (qPCR)

Total RNA was isolated from conditioned explants harvested at T3 and T5 using an innuPREP DNA/RNA Mini Kit (Analytik, Jena, Germany) according to the manufacturer’s instructions. Total RNA was eluted using 20 µL of RNase-free water. The quality and quantity of the RNA samples were checked and measured using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Total (DNase-treated) complementary DNA was synthesized using 200 U of Tetro reverse transcriptase enzyme (Bioline, Taunton MA, USA) and 10 mM of oligo (dT) primer in a total reaction volume of 20 µL. The qPCR was performed according to the manufacturer’s recommendations in 10 µL reactions using 2 × SensiFAST SYBR^® No-ROX Mix (Bioline) and 20 ng cDNA/reaction well on a LightCycler^® 480 (Roche, Basel, Switzerland). qPCR was used to quantify the expression levels of the selected inflammatory, protease, and ECM component genes. Ten pair primer sequences for all genes were designed using Canis lupus familiaris cDNA information published on the Ensembl database, and Primer3 software was used for primer design. The sequences of the primer pairs are listed in Table 2. The qPCR reaction was performed at 95 °C for 10 min, followed by 40 cycles for 20 s and 30 s at 95 °C and 60 °C, respectively. All reactions were run in duplicate using RNA isolated from three bio-replicates of the samples. HPRT1 and GAPDH were used as reference control genes. Relative quantification was performed using the 2^−ΔΔCT threshold cycle method (Livak & Schmittgen, 2001).

Statistical analysis

Differences between groups were tested by one-way analysis of variance, followed by Tukey’s post-hoc test. Values of P < 0.05 were considered statistically significantly different. Pearson correlation coefficient (r) tests were used to assess the presence of bivariate correlations between percentages of s-GAG levels of the treatment conditions. In addition, a correlation heat map was generated by hierarchical cluster analysis using MeV version 4.9.0 (Saeed et al., 2006). The software package SPSS 14.0 for Windows (SPSS, Armonk, NY, USA) was used for computations. Data were expressed as the mean ± standard deviation (SD) of separate experiments (n = 3, where n represents the number of independent experiments).

S-GAG release

s-GAG concentrations in the conditioned medium indicated the effect of FQs on conditioned explants. Here, we calculated the s-GAG concentration as a percentage relative to the negative control group. Our results were significant differences of the percentage of s-GAG levels (P ≤ 0.05) in HA treatment groups. The s-GAG levels were decreased in T3, T4, and T5 compared with the T2 and T3, i.e., non + HA (T3–T5), pre + HA (T2–T3 and T2–T5), pre + HA + Enro (T2–T4), pre + HA + Mar (T3–T5), with + HA + Enro (T3–T4), and with + HA + Mar (T2–T5). Pre + Enro was the only group with significant differences between the time period without HA treatment, which was found between T2–T3, T2–T4, and T2–T5 (P = 0.019 to P = 0.043). The increase in the s-GAG level in the pre-IL-1 β-induced group was higher at T2, whereas the increase in the s-GAG level in the with-IL-1 β-induced group was higher at T3. Of these, the highest percentage of s-GAG content was found at T3 in the with + HA + Mar group (148.97 ±7.19%), which was also higher than Enro or Mar treated alone in the positive control groups at T5 (Enro 400 = 147.26% ±3.39%, Enro 1,000 = 147.43% ±3.63%, Mar 400 = 148.63% ±4.84%, and Mar 1,000 = 142.64% ±3.15%). However, the s-GAG levels decreased in all treatments after removing IL-1 β in both pre- and with-IL-1 β-induced explants.

The heat map of the clustered correlation matrix was applied (MeV 4.9.0, Pearson correlation, HCL support tree) to the mean percentages of s-GAG (Fig. 2). The samples were clustered after unsupervised hierarchical clustering. A different pattern of percentages of s-GAG under different IL-1 β-stimulated conditions was observed. A high correlation coefficient for s-GAG levels (0.685–0.984, P ≤ 0.05) between adjacent clusters of pre-IL-1 β-stimulated groups was found to increase from T1 to T2 and increased again at T3–T5, and the negative control group (non, −FQs), non + HA, non + Mar, and non + HA + Mar were also clustered in this pattern. Although the explants treated with IL-1 β-stimulated groups were found closely clustered together (r = 0.679–0.969, P ≤ 0.05), the percentages of s-GAG increased from T1 to T3 and increased again from T4 to T5. The Enro treatments (non + Enro and non + HA + Enro) were clustered in this pattern.

UA remaining in the explants

The UA content represents the GAG level (excluding keratan sulfate, which contains galactose instead of hexuronic acid) that remains in the explants after treatment. The DMB assay was used to measure the UA content remaining in the cartilage explant disks, and the results are shown in Fig. 3. At T3, the percentages of UA content changes were observed in FQs combined with HA treatment groups, and a slight increase was observed in the single HA-treated groups compared with the negative control. Co-treatment of HA and FQs in normal explants could be attributed to UA accumulation, and non + HA + Enro and non + HA + Mar groups significantly increased the UA content (P = 0.034 and 0.019, respectively).

Safranin-O staining

Safranin-O staining also identifies the GAG content of the cartilage explants combined with HA shown at different time periods of Enro and Mar treatment in Figs. 4 and 5. We found that the decrease of safranin-O intensity (faded) was more prominent in the superficial area than the deep zone in nearly every treatment group, including the negative control and HA groups at both T3 and T5 (Figs. 4 and 5). At T3, the safranin-O staining in most Enro treatments was stronger than Mar, except when treated on normal explant with HA combination. The highest intensity was found in the non + HA + Mar treatment at 120.72% compared with the negative control group (P = 0.022; Fig. 4Q). A significant decrease in safranin-O intensity was observed in the pre, −FQ group and single Mar-treated explants (non + Mar and with + Mar) compared with the negative controls. HA combined with Mar treated in both normal and IL-1 β-stimulated explant conditions performed rather efficiently, and the treatments with FQs were successful in enhancing the proteoglycan content in the matrix of the explant cultures in the pre-IL-1 β-stimulated explant conditions at T3 treatment. At T5, the safranin-O staining in a single Enro treatment group (Figs. 5C, 5I and 5M) was stronger than that in the other groups in every explant condition, particularly in the IL-1 β-stimulated explants (Figs. 5I and 5M). The decrease of proteoglycan staining was significantly found in HA combined with Enro treatment in every explant condition, and the major decrease in intensity was found in the non + HA + Enro group (decrease to 91.60%) and was lower than that in the non + Enro treatment (P < 0.0001; Fig. 5U) .

Effects of FQs with and without HA on articular cartilage genes

The effects of FQs on mRNA expression in conditioned explants of some selected genes that code for cytokines and enzymes that respond to inflammatory pathways and the ECM component are shown in Figs. 6 and 7.

Effects of FQs on ECM component genes

We examined which anabolic pathways of conditioned cartilage disks regulate at different days of cultivation (T3 and T5). At T3, HAS1 was highly expressed in non + HA + Mar (P = 0.014) and with + Mar (P = 0.010) compared with the negative control. The most significant increases in expression of COL2A1 were found in the with + HA + Enro treatments groups, which were up-regulated higher than the normal explant treated with HA alone (non + HA). The most significant increase in ACAN expression was found in only one treatment “non + HA + Mar,” which was significant in all Mar treatment and most of Enro treatments. At T5, HAS1 was highly expressed in the pre-IL-1 β-stimulated explant (pre, −FQs) group, non + HA + Mar, and positive control (Enro 400) compared with the negative control. Treatment with HA significantly reduced HAS1 regulation in the IL-1 β-stimulated explants, lower than the pre, −FQ group (close to the negative control group). Treatment of the explants with Enro affected COL2A1 expression, whereas Mar-treated groups did not cause any changes (no statistically significant differences compared with the negative control). Most significant increases in ACAN expression were found in Mar treatment groups and were highly expressed when in combination treatment with HA. However, a high dose of both FQ (at 400 and 1,000 µg/µL) treatments activated the explants down-regulation in the COL2A1 and ACAN expressions.

Effects of FQs on the expression of matrix metalloproteinase and their inhibitor

TIMP1 was significantly down-regulated in the Enro treatment groups at both T3 and T5, particularly in the IL-1 β-stimulated explants (both pre and with) compared with negative control. Treatments of non + Enro and with + Enro at T3 were found to significantly up-regulate MMP3 compared with all Enro-treated and control groups. At T5, Mar induced TIMP1 up-regulation higher than other treatments of Enro, particularly in the non + Mar and with + Mar groups compared with the negative control (P = 0.002 and P = 0.00026, respectively). The same trend was found in the MMP3 gene at T5, which had very low expression levels in the Enro treatment groups but highly expressed in Mar treatment. The most significant increases in MMP3 expression were found in high concentrations of Mar treatment at 400 µg/mL (positive control) up to 33-fold. Treatment of FQs on IL-1 β-stimulated explants regulated the down-expression of MMP9 at T3, except in the non + HA + Mar group at T3, which exhibited higher expression levels compared with all Enro-treated and control groups. At T5, the pre-IL-1 β-stimulated (pre, −FQs), pre + Mar, and Mar 1,000 groups were found to significantly up-regulate the MMP9 gene compared with the negative control group (P = 0.02, P = 0.003, and P = 0.04, respectively). Moreover, HA co-treatment successfully reduced MMP9 gene expression between pre + Mar and pre + HA + Mar (P = 0.001).

Effects of FQs on pro-inflammatory cytokines IL-1b and TNF

FQ treatments induced IL-1b and TNF expression up-regulation in both normal and IL-1 β-stimulated explant conditions and at different time periods. The expression level of IL-1b after treatment (at T3) was distinctively up-regulated in the pre + HA + Enro group, comparing between all treatment conditions (P ≤ 0.05). Moreover, Enro induced TNF expression up-regulation in normal and pre-IL-1 β-stimulated explants, pre + HA + Enro had a significantly higher expression compared with other groups, and the most significant was found when compared with pre + HA (P < 0.0001). The highest expression level of IL-1b at T5 was found in the non + Mar group; however, statistically significant differences were found when compared with the negative control. The expression levels of TNF switched to up-regulation in Mar-treated groups compared with Enro-treated and control groups at T5, and the expression levels were significantly increased in the with + Mar and positive control groups (Mar 1000), although the highest expression level was found in the pre, −FQs group.

Effects of FQs on PTGS1, PTGS2, and NFKB1

A slight difference was found in the expression pattern of the PTGS1 gene at T3 and T5, although treatment with Mar in normal explants (non + HA + Mar at T3 and non + Mar at T5) exhibited higher expression levels compared with all Mar-treated and control groups. Contrary effects were found for treatment with FQs for PTGS2 and NFKB1 genes at a different time period. At T3, the PTGS2 gene extremely up-regulated in the Enro-treated groups compared with all Mar-treated groups and the control group, with the highest expression level in the with + HA + Enro group (P < 0.0001) compared with all treatment groups and the negative control. Also, the NFKB1 gene significantly exhibited higher expression levels in the with + HA + Enro group compared with all treatment groups and the control group (P < 0.0001). However, at T5, the expression levels of both PTGS2 and NFKB1 in Enro-treated groups sharply decreased, to levels similar or lower than control groups. Whereas, the pre + Mar and the Mar 400 groups were there significant induced NFKB1 expression up-regulation compared with all of Mar treatment groups and the control group.