Joint cytokine quantification in two rodent arthritis models: kinetics of expression, correlation of mRNA and protein levels and response to prednisolone treatment

Abstract

Biomarker quantification in disease tissues from animal models of rheumatoid arthritis (RA) can help to provide insights into the mechanisms of action of novel therapeutic agents. In this study we validated the kinetics of IL-1β, TNF-α and IL-6 mRNA and protein expression levels in joints from DBA/1OlaHsd murine collagen-induced arthritis (CIA) and Lewis rat Streptococcal cell wall (SCW)-induced arthritis by real-time polymerase chain reaction (PCR) TaqMan® and Enzyme-linked immunosorbent assay (ELISA). Prednisolone was used as a reference to investigate any correlation between clinical response and cytokine levels at selected time-points. To our knowledge this is the first report showing a close pattern of expression between mRNA and protein for IL-1β and IL-6, but not for TNF-α, in these two models of RA. The kinetics of expression for these biomarkers suggested that the optimal sampling time-points to study the effect of compounds on both inflammation and cytokine levels were day 4 postonset in CIA and day 3 after i.v challenge in SCW-induced arthritis. Prednisolone reduced joint swelling through a mechanism associated with a reduction in IL-1β and IL-6 protein and mRNA expression levels. At the investigated time points, protein levels for TNF-α in arthritic joints were lower than the lower limit of detection of the ELISA, whereas mRNA levels for this cytokine were reliably detected. These observations suggest that RT-PCR TaqMan® is a sensitive technique that can be successfully applied to the quantification of mRNA levels in rodent joints from experimental arthritis models providing insights into mechanisms of action of novel anti-inflammatory drugs.

INTRODUCTION

Animal models of rheumatoid arthritis (RA), while not fully representative of human disease, are widely used to gain insights into the pathogenic mechanisms underlying the initiation and perpetuation of joint inflammation. Since the earliest preclinical stages of RA are not easily accessible to investigation, a diversity of experimental arthritis models are considered as valuable tools not only in delineating mechanisms of inflammation and autoimmune phenomena, but also at the preclinical phase in the development process of new therapeutic agents. In this context, two rodent models of both acute and chronic destructive arthritis: rat SCW-induced arthritis and mouse CIA, respectively, have become industry standards by which potential therapeutics targets are evaluated.

CIA is a polyarthritis induced by sensitization of susceptible strains of animals with type II collagen. Both humoral and cellular immune responses to collagen II are seen in sensitized animals and both components are involved in disease progression. There are certain similarities with the human condition including linkage of disease to genes residing in the histocompatability locus [1], mononuclear cell infiltration, pannus development, fibrin deposition, erosion of cartilage and bone and autoreactive T and B cells [2,3]. In the reactivation model of SCW-induced arthritis in rats, inflammation and synovial hyperplasia are the main features of the disease in which an intrarticular (i.a) injection of SCW antigen followed by intravenous (i.v) challenge results in a T cell-mediated monoarticular arthritis [4,5]. In comparison with some other animal models, in which the arthritic response develops gradually and unpredictably, in SCW arthritis the inflammatory response develops synchronously, allowing precise definition of the regulatory mechanisms underlying pathology.

Cytokines such as interleukin-1β (IL-1β), tumour necrosis factor alpha (TNF-α) and interleukin-6 (IL-6) have been shown to display potent pro-inflammatory actions that are thought to contribute to the pathogenesis of RA [6]. TNF-α and IL-6 are involved in inflammation, differentiation and proliferation of T and B cells and bone resorption [7,8]. IL-1β is shown to have physiological functions such as induction of inflammation, modification of immune response and activation of osteoclasts [9,10]. These pro-inflammatory cytokines, which are increased in RA synovium, have also been shown to contribute to the development of arthritis in some arthritis models including CIA and SCW-induced arthritis in different species [11–14]. Furthermore, clinical efficacy with anticytokine therapies, has also been shown in rat and mouse models of RA [13,15,16]. Thus, in order to meaningfully interpret the results obtained from animal models when investigating the mechanism of action of potential antirheumatic drugs there is a significant need to assess cytokines and other effector pathways that contribute to disease pathogenesis. In the current study, we characterized the kinetics of cytokine expression in murine CIA and in the reactivation model of SCW-induced arthritis in Lewis rats. Although several kinetics studies describing different pattern of cytokine expression in animal models of RA have previously been published [11,17] a comparative study between mRNA and protein expression levels in arthritic joints from these two particular models of experimental arthritis has not been fully described.

A limitation when quantifying multiple cytokines at the protein level is that biological tissue samples available to be analysed from rodent joints are often too small to allow a full analysis. Furthermore, the sensitivity of the ELISA method is not always high enough to detect cytokine levels in supernatants from biological samples. In the current study, a more sensitive approach to study cytokine induction has been applied by measuring mRNA levels. A recently described strategy was used to perform quantitative real-time PCR for cytokine mRNA quantification using specific TaqMan® probes on an ABI Prism® 7900HT Sequence Detection System (TaqMan®; PE Applied Biosystems, UK) [18]. This technique allows the direct monitoring of amplicon accumulation during the PCR process using fluorogenic probes, providing fluorescence kinetics that accurately reflect the amounts of cytokine mRNA levels in biological samples [19].

Glucocorticoids such as dexamethasone and prednisolone, despite having a number of side-effects, are potent and commonly used anti-inflammatory agents in human RA. They are known to down-regulate proinflammatory cytokine production, such as IL-1 and TNF-α, normally produced by macrophages and monocytes [20]. This inhibition is due, among other mechanisms, to enhanced expression of IκBα protein synthesis that retains NF-κB in inactive cytoplasmic complexes so that the expression of genes involved in the pathogenesis of the immune response is reduced [21]. In the present study, prednisolone was used to investigate the effect of steroid treatment on joint cytokine production. The possible correlation between reduction of inflammation and inhibition of cytokine mRNA and protein levels in joints after Pred-nisolone treatment, was also assessed at selected sampling time-points.

To our knowledge, this is the first report describing the comparative kinetics of cytokine protein and mRNA expression in joints from the rodent models of both acute and chronic destructive arthritis: rat SCW-induced arthritis and mouse CIA, respectively, using quantitative real-time PCR TaqMan® analysis and ELISA. The data shows that both models demonstrate a close relationship between cytokine mRNA and protein expression for IL-1β and IL-6 (but not TNF-α) in inflamed joints and that prednisolone decreases both cytokine message and protein for IL-1β and IL-6 in accordance with reduced joint swelling.

MATERIALS AND METHODS

Reagents

Bovine type II Collagen was purchased from Elastin Products Company, Inc. (Owensville, MO, USA). Heat-killed Mycobacterium tuberculosis strain H37Ra was purchased from Difco (Detroit, MI, USA). The 100p fraction of streptococcal cell wall was purchased from Lee Laboratories (Grayson, GA, USA). All reagents required for reverse transcription PCR were from PE Applied Biosystems (Warrington, UK). Forward and reverse primers were purchased from Invitrogen™ Life Technologies (UK). TaqMan® probes were synthesized by PE Applied Biosystems (Warrington, UK). RiboGreen, used to quantify RNA, was obtained from Molecular Probes, Inc (the Netherlands) and RNA 6000 Nano LabChip Kit® from Agilent Technologies Inc (UK). Prednisolone (P6004) was purchased from Sigma-Aldrich (Dorset, UK).

Animals

All in vivo studies were undertaken in certified dedicated in vivo experimental laboratories at the GlaxoSmithKline Medicines Research Centre, Stevenage, UK. The studies complied with national legislation and with local policies on the Care and Use of Animals and with related codes of practice. Male Lewis rats and male DBA/1OlaHsd mice obtained from Harlan, UK at age 6–8 weeks were housed under standard conditions and received food and water ad libitum. Animals were habituated to the holding room for a minimum of one week before experimental procedures. Mice were individually identified by subcutaneous transponders (Avid, Uckfield, UK) implanted under isofluorane anaesthesia.

Induction and assessment of SCW-induced arthritis

SCW-induced arthritis was induced in 6–8 week-old male Lewis rats (125–150 g) (Harlan, UK) following a similar method to the previously described by Esser et al. [4]. A SCW preparation (100p fraction) was suspended in phosphate-buffered saline (PBS) and 10 µl of the suspension containing 5 µg Peptidoglycan-polysaccharide from Streptococcus pyogenes (PG-PS) was injected into the right ankle joint (day −14). Animals from control groups were injected similarly with 10 µl of PBS. Reactivation of the arthritic inflammation was induced 14 days after intra-articular injection (designated day 0) by intravenous injection of 200 µg of PG-PS. This resulted in monoarticular arthritis involving the joint originally injected with PG-PS [13]. Ankle swelling was measured using a caliper at different time points. The inflammatory response is expressed as change in ankle diameter relative to the starting diameter. For the time course study, 5 animals injected with PG-PS or PBS were sacrificed at different time points and whole ankle joints were dissected, snap frozen into liquid nitrogen and stored at −80°C for subsequent analysis. In another separate experiment, prednisolone (1 mg/kg) or vehicle (0·5% methylcellulose) were administered bid, p.o. from day 0 to day 2 giving the last dose 2 h before sacrificing the animals on day 3 after i.v injection of SCW.

Induction and assessment of CIA

Bovine type II collagen was diluted with 0·1 m acetic acid to a concentration of 4 mg/ml and was emulsified in equal volumes of complete Freund's adjuvant (CFA) (4 mg/ml). Eight to 12-week-old male DBA/1OlaHsd mice (Harlan, UK) were immunized intradermally at the base of the tail with 50 µl of emulsion containing 100 µg of collagen. At day 21 the animals were boosted with an i.p. injection of 200 µg bovine type II collagen dissolved in 0·45 m NaCl/0·05 m Tris buffer. Disease onset was usually 4–14 days after boost and was considered to occur on the day that swelling or erythema in the paws was observed (day 1). Clinical score, a composite index of disease severity and the number of limbs affected, was assessed daily to monitor disease progression. The scoring used was as follows: 0 = normal and 1–4 = slight, prominent, severe or extreme oedema and erythema, respectively. For the time course study, mice were sacrificed at different days of disease (covering onset and progression of CIA) and whole paws were dissected, snap frozen into liquid nitrogen and stored at −80°C for subsequent analysis. A group of nonimmunized mice was used as control. In a separate experiment, prednisolone (0·3, 1, 3 or 9 mg/kg) or vehicle (0·5% methylcellulose) were administered daily, p.o. from day 1 to day 4, last dose being given 3 h before sacrificing the animals on day 4.

Joint cytokine measurements (ELISA)

For protein quantification frozen tissues (whole joints including synovium, adjacent tissues and bones) were pulverized using a mortar and pestle filled with liquid nitrogen. Due to the small size of the samples, whole mouse joint tissues from different animals were required for protein and mRNA measurements. In contrast, individual rat joint tissues were split after pulverization, allowing quantification of mRNA and protein levels in the same sample. Tissue was transferred to 15 ml tubes, placed on dry ice and resuspended in 1 ml PBS/200 mg of tissue (containing Complete™ Protease Inhibitor Cocktail Tablets, one tablet per 25 mls solution) and homogenized using a Polytron tissue homogeniser (Kinematica Inc, Switzerland) for 20 s. Mouse joint homogenates were centrifuged for 10 min at 500 g at 4°C. Supernatants were transferred to 1·5 ml eppendorf tubes, centrifuged at 15 000 g for 5 min and collected for cytokine analysis using R & D ELISA Duoset® kits according to the manufacturers instructions. The detection limit of the assays was 15·6 pg/ml for IL-1β, TNF-α and IL-6. Data are expressed as pg cytokine/g tissue (data was multiplied by a dilution factor for conversion from pg/ml to pg/g). Rat joint homogenates were transferred to 1·5 ml eppendorf tubes and centrifuged for 5 min at 15 000 g at 4°C. Supernatants were collected for cytokine analysis using R & D ELISA Quantikine® kits according to the manufacturers instructions. The detection limit of the assays was 12·5 pg/ml for TNF-α and 31·2 pg/ml for IL-1β and IL-6. Data are expressed as pg cytokine/ml.

Total RNA isolation from mouse and rat joints

Frozen tissues (whole joints including synovium, adjacent tissues and bones) were pulverized in liquid nitrogen using a mortar and pestle and total RNA was isolated from individual homogenized joints (nonpooling strategy used) using RNeasy® Mini-kits (Qiagen Ltd, UK) and following manufacturer's instructions. In order to ensure no cross-contamination with genomic DNA samples were treated for 15 min with 10 units of RNase-free DNase (Qiagen Ltd, UK) at room temperature. RiboGreen® RNA Quantification Kit was used to determine the total RNA concentration of the samples. RNA 6000 Nano LabChip Kit® (Agilent 2100 Bioanalyser) was used to assess the total RNA integrity.

Quantitative real-time PCR (TaqMan®)

Quantitative real-time PCR was performed using the ABI Prism 7900 Sequence Detector System® (PE Applied Biosystems, UK) as previously described [22]. For cDNA synthesis 600 ng of total RNA were reverse transcribed using TaqMan® reverse transcription reagents (PE Applied Biosystems, UK) in a MJ Research PTC-200 PCR Peltier Thermal Cycler. TaqMan® probes and primers for the genes of interest were designed with primer design software Primer Express™ (PE Applied Biosystems, UK) (Table 1) and optimized for use. TaqMan® probes were labelled at the 5′ end with the reported dye molecule FAM and at the 3′ end with the quencher dye molecule TAMRA. cDNA samples (25 ng in a total volume of 25 µl) were mixed with primers, probe and TaqMan® Universal PCR Master Mix as described in the manufacturer's directions (PE Applied Biosystems). The PCR was conducted using the following parameters: 50°C for 2 min, 95°C for 10 min and 40 cycles at 95°C for 15 s and 60°C for 1 min. The fluorescent signal at each cycle generated by the release of fluorophore (FAM) from the quencher (TAMRA) by the 5′-exonuclease activity of AmpliTaq polymerase was plotted versus the cycle number. Standard curves for each individual target amplicon were constructed using sheared mouse genomic DNA (Promega Ltd, UK) or rat genomic DNA (BD Clontech, UK). All PCR assays were performed in duplicate and results are represented by the mean values of copy No/50 ng cDNA. Ubiquitin [23] was used as housekeeping gene against which all the samples were normalized for differences in the amount of total RNA added to each cDNA reaction and for variation in the reverse transcriptase (RT) efficiency among the different cDNA reactions.

Table 1. Oligonucleotides for real time RT-PCR TaqMan® analysis.

mRNA targets	Species	Primer type†	Oligonucleotides (5′-to 3′)
IL-1β	Rat	FP	CACCTCTCAAGCAGAGCACAG
		RP	GGGTTCCATGGTGAAGTCAAC
		P	6-FAM-TGTCCCGACCATTGCTGTTTCCTAGG-TAMRA
IL-6	Rat	FP	CCAAGACCATCCAACTCATCTTG
		RP	CACAGTGAGGAATGTCCACAAAC
		P	6-FAM-TCGGCAAACCTAGTGTGCTATGCCTAAGCA-TAMRA
TNF-α	Rat	FP	CCAGGTTCTCTTCAAGGGACAA
		RP	CTCCTGGTATGAAATGGCAAATC
		P	6-FAM-CCCGACTATGTGCTCCTCACCCACA-TAMRA
Ubiquitin	Rat	FP	CGAGAACGTGAAGGCCAAGA
		RP	GGAGGACAAGGTGCAGGGTT
		P	6-FAM-CCCCTGACCAGCAGAGGCTCATCTTTG-TAMRA
IL-1β	Mouse	FP	TCGCTCAGGGTCACAAAGAAA
		RP	CCATCAGAGGCAAGGAGGAA
		P	6-FAM-CATGGCACATTCTGTTCAAAGAGAGCCTG-TAMRA
IL-6	Mouse	FP	CTATACCACTTCACAAGTCGGAGG
		RP	TGCACAACTCTTTTCTCATTTCC
		P	6-FAM-TTAATTACACATGTTCTCTGGGAAATCG-TAMRA
TNF-α	Mouse	FP	TCTCTTCAAGGGACAAGGCTG
		RP	ATAGCAAATCGGCTGACGGT
		P	6-FAM-CCCGACTACGTGCTCCTCACCCA-TAMRA
Ubiquitin	Mouse	FP	CGAGAACGTGAAGGCCAAGA
		RP	CATCTTCCAGCTGTTTGCCG
		P	6-FAM-CCCCTGACCAGCAGAGGCTGATCTTTG-TAMRA

^†FP, RP and P indicate forward and reverse primers and probes, respectively. The final optimized concentration of FP, RP and P for all the target genes was 900 nm, 900 nm and 100 nm, respectively. The fluorogenic probes are oligonucleotides with a reporter (FAM) at the 5′ end and a quencher (TAMRA) at the 3′ end.

Statistical analysis

Data was analysed using Statistica software (StatSoft®). The results are presented as mean ± SEM; n represents the number of animals. The level of statistical significance was determined on log-transformed data by analysis of variance (anova) followed by Dunnett's t-test for multiple comparisons, except for clinical score data, which was analysed using Kruskall-Wallis test for nonparametric data. Gene expression data was normalized against the housekeeping gene Ubiquitin [23] and log transformed for subsequent statistical analysis.

RESULTS

Time course of inflammation in the SCW-induced arthritis model

Intra-articular (i.a) injection of SCW resulted in increased ankle swelling that peaked between 24 and 72 h after injection (days −13 to −10) followed by a gradual reduction by day 0. At this time point, i.v. challenge with SCW led to the reactivation of the inflammatory response that peaked 72 h thereafter (day 3). Animals injected i.a. with PBS were used as control group at each specific time point. Another group of naïve animals (noninjected rats) was used to discard possible inflammatory responses due to the i.a. injection on itself (Fig. 1). No statistical differences on the inflammatory response between the PBS injected animals and the naïve control group was observed during the time course study.

Fig. 1.

Time-course of inflammation in the reactivation model of SCW-induced arthritis in male Lewis rats. Animals injected with SCW (filled squares) suspension or PBS (open squares) were sacrificed on the days indicated, for cytokine measurements. A group of naïve (filled trangles) noninjected animals was used as control. The inflammatory response is expressed as the increase on ankle diameter (mm) after the initial injection of SCW. Data represent means ± S.E.M. (n = 5 animals/group). Analysis of variance was performed on log-transformed data; *P < 0·05, **P < 0·01.

Kinetics of cytokine mRNA and protein expression levels in rat joints

Ankle joints from SCW or PBS-injected animals were taken at different time points covering both the pre- and postreactivation phases of disease. Cytokine levels were measured in supernatants from homogenized joints by ELISA. Total RNA was extracted from joint tissues as described above and IL-1β, TNF-α and IL-6 mRNA expression levels were quantified by real-time RT-PCR TaqMan® analysis. As shown in Fig. 2, protein and mRNA levels for IL-1β and IL-6, and mRNA levels for TNF-α, peaked 24 h after i.a. of SCW (day 13) and levels decreased gradually thereafter. In the postreactivation phase, IL-1β, IL-6 and TNF-α protein and mRNA levels peaked 1 day after i.v. injection of SCW, preceding the peak inflammatory response, which occurred on day 3. It should be noticed that in this model the concentration of TNF-α protein detected in rat joints, at the different analysed time points, was lower than the lower detection limit of the ELISA (detection limit = 12·5 pg/ml). A significant correlation between mRNA and protein levels for the pro-inflammatory cytokines IL-1β (r = 0·78, P < 0·0001) and IL-6 (r = 0·88, P < 0·0001) was observed demonstrating a similar pattern of expression during the disease progression. The quantitative analysis for protein and mRNA showed that in this model IL-1β was the most highly expressed cytokine followed by IL-6 and TNF-α. Since the inflammatory response peaked on day 3 when mRNA levels for the three cytokines and protein levels for IL-1β and IL-6 were still statistically significantly raised, we concluded that 3 days after i.v. injection of SCW suspension could be an optimal time point to study the possible relationship between reduction of inflammation and IL-1β and IL-6 production at both, mRNA and protein levels. Thus, day 3 after i.v challenge was chosen in our following studies to investigate the effect of prednisolone on joint swelling and cytokine levels.

Fig. 2.

Kinetics of cytokine protein and mRNA expression levels in joints from the reactivation model of SCW-induced arthritis in rats. Animals injected with SCW (——) or PBS (-----) were sacrificed on the days indicated and joints taken for subsequent analysis. Protein (filled squares) and mRNA (open squares) levels for (a) IL-1β, (b) IL-6 or (c) TNF-α were measured using ELISA and real-time RT-PCR TaqMan®, respectively. Data represent means ± S.E.M. (n = 5 animals/group). Protein data are expressed as pg protein/ml and mRNA results, normalized against the housekeeping gene Ubiquitin, are expressed as copy No/50 ng cDNA. Statistical analysis of log-transformed data was performed using anova. *,+P < 0·05, **,++P < 0·01, compared to the PBS injected group of rats at the corresponding time point.

Time course of inflammation in the CIA model

Disease onset was seen to occur between days 4–14 after boost with type II collagen. Animals were entered into the study when the first sign of slight swelling or erythema in the paws was observed (day 1). As shown in Fig. 3, the clinical score reached a maximum between day 4 and day 11 and decreased gradually thereafter. Due to the small size of the mouse joints, two separate time course experiments were required to obtain enough tissue samples to quantify both, cytokine protein (Fig. 3, Exp. A) and mRNA levels (Fig. 3, Exp. B). As illustrated in Fig. 3, no significant differences in clinical score between the joints used in both experiments were observed, allowing a quantitative comparison of mRNA and protein levels in joints alongside the time course of disease progression.

Fig. 3.

Time-course of clinical score in mouse joints in the collagen-induced arthritis model. Animals (n = 6/group) with 1, 4, 8, 11 or 15 days of inflammation were sacrificed and whole paws were processed for protein (Exp. A, filled squares) and mRNA (Exp. B, open squares) quantification as described in Materials and Methods. A group of naïve unimmunized mice (N) was used as control. Data represent means ± S.E.M. (n = 6 paws from different animals/group). Statistical analysis of clinical score data (compared to control group) was performed using the Kruskal–Wallis test for nonparametric data. **,++P < 0·01.

Kinetics of cytokine mRNA and protein expression levels in mouse joints

Joints from mice with 1, 4, 8, 11 and 15 days of inflammation were taken and processed for mRNA and protein quantification by real-time RT-PCR TaqMan® and ELISA, respectively. A nonimmunized group of animals was used as control. In this model, IL-1β protein expression levels, preceded by IL-1β mRNA, reached a peak on day 4 and after that levels gradually decreased, similar to the kinetics of the inflammatory response (Fig. 4a). A rapid transient peak of mRNA and protein expression was observed for the pro-inflammatory cytokine IL-6, which clearly peaked on day 1 and remained significantly elevated at days 4 and 8 compared to nonimmunized animals (Fig. 4b). Despite cytokine mRNA and protein quantification not being performed on identical joint tissue, a good correlation between both parameters was observed for IL-1β (r = 0·95, P = 0·003) and IL-6 (r = 0·89, P = 0·01). Analogous to the rat SCW-reactivation model of RA, IL-1β was shown to be the most highly cytokine expressed followed by IL-6. Although statistically significant compared to control, very low levels for TNF-α protein were detected in mouse joints (values close to the detection limit of the ELISA assay) from day 4 through to day 15. The expression of TNF-α mRNA preceded the protein synthesis for this cytokine, without showing a clear peak of expression. Taking into account all these observations, day 4 was selected as a possible optimal time point to study the effect of the steroid prednisolone on both, clinical score and cytokine production in joints.

Fig. 4.

Kinetics of (a) IL-1β, (b) IL-6 and (c) TNF-α protein (red filled squares) and mRNA (blue open squares) levels in mouse joints in the collagen-induced arthritis model. Animals with 1, 4, 8, 11 or 15 days of inflammation were sacrificed and whole paws processed for protein and mRNA quantification by ELISA and real-time RT-PCR TaqMan®, respectively. A group of naïve unimmunized mice (N), showing basal cytokine expression levels in normal joints, was used as control. Data represent means ± S.E.M. (n = 6 paws from different animals/group). Protein data are expressed as pg protein/g of tissue and mRNA results, normalized against the housekeeping gene Ubiquitin, are expressed as copy No/50 ng cDNA. Statistical analysis of log-transformed data was performed using anova. +P < 0·05, **, ++P < 0·01, compared against nonimmunized mice (N).

Effect of prednisolone in the reactivation model of SCW-induced arthritis in rats

Prednisolone (1 mg/kg) or vehicle (0·5% methylcellulose) were orally administered twice daily from day 0 to the day when the maximal inflammatory response occurred (day 3 after i.v challenge). At the dose assayed, prednisolone completely suppressed ankle swelling from treated animals (Fig. 5a). A highly significant reduction for IL-6 mRNA (P < 0·01) and protein (P < 0·01) expression levels in rat joints was observed (Fig. 5c). Similar results were obtained for IL-1β (Fig. 5b), although protein levels for this cytokine were reduced to a lesser but still significant (P < 0·05) extent. Again, corroborating the previous results observed in the time-course study, TNF-α protein concentration in joints was lower than the lower detection limit of the ELISA and despite mRNA levels being detected by real-time RT-PCR TaqMan® analysis, no effect on mRNA levels was observed at the dose of prednisolone assayed.

Fig. 5.

Effect of prednisolone on (a) inflammation and cytokine (protein and mRNA) levels in rat joints in the SCW-induced arthritis model. Animals were sacrificed on day 3 after i.v challenge. Naïve (non injected animals), PBS (animals injected with PBS), Vehicle (animals injected with SW and administered methylcellulose), Pred (animals injected with SCW and daily administered prednisolone 1 mg/kg, p.o). The inflammatory response is expressed as the increase on ankle diameter (mm) after the initial injection of SCW suspension. Protein (□) and mRNA (▪) levels for (b) IL-1β, (c) IL-6 or (d) TNF-α were measured by ELISA and real-time RT-PCR TaqMan®, respectively. Data represent means ± S.E.M. (n = 6 animals/group). Protein data are expressed as pg protein/ml and mRNA results, normalized against the housekeeping gene Ubiquitin, are expressed as copy No/50 ng cDNA. *P < 0·05, **P < 0·01 (for inflammation and protein data);+P < 0·05, ++P < 0·01 (for mRNA data), compared to vehicle treated group of animals.

Effect of prednisolone in the mouse CIA model

Mice were orally administered with vehicle (0·5% methylcellulose) or prednisolone (0·3, 1, 3 or 9 mg/kg), daily, from day 1 to day 4. Two hours after receiving the last dose of steroid, animals were sacrificed and joints were taken for protein and mRNA quantification. Due to limited sample availability the effect of prednisolone on cytokine mRNA levels in joints was only investigated at two of the four doses assayed (1 and 9 mg/kg). As shown in Fig. 6, prednisolone treatment reduced the clinical score in a dose-dependent manner showing an effective dose 50% (ED₅₀) of 1·1 mg/kg. In order to study the anti-inflammatory mechanism of action of prednisolone in joints from mouse CIA, cytokine concentrations were quantified using ELISA. IL-6 protein levels in joints were highly significantly reduced, compared to the vehicle group, at the four doses of steroid assayed, with an ED₅₀ of 0·7 mg/kg (Fig. 6c). IL-1β was also reduced by prednisolone in a dose-dependent manner (ED₅₀ = 1·5 mg/kg) (Fig. 6b), although no effect was observed at the lower dose assayed. Similar to protein, mRNA expression levels for IL-6 and IL-1β were found to be highly significantly reduced after prednisolone treatment at the doses of 1 and 9 mg/kg (P < 0·01). Since day 4 was selected as an optimal time point to further investigate the effect of anti-inflammatory drugs on clinical score and IL-1β and IL-6 levels in joints, we also looked at the effect of steroid on TNF-α protein and mRNA expression levels at the same time point. As previously shown in the CIA time-course study (Fig. 4c), TNF-α levels in joints from mice with 4 days of arthritis, despite significantly up-regulated compared to nonimmunized mice, were very low and close to the lower limit of detection of the ELISA. Moreover, a poor therapeutic window to investigate the effect of anti-inflammatory drugs on this cytokine expression level was also found at the analysed time point. As shown in Fig. 6d, TNF-α mRNA levels were only reduced at the higher dose of steroid tested. Unexpected results were found for TNF-α protein levels in joints, which were increased after treatment with 1 or 3 mg/kg of prednisolone.

Fig. 6.

Effect of prednisolone on (a) clinical score and on (b) IL-1β, (c) IL-6 and (d) TNF-α protein and mRNA levels in mouse joints in the CIA model. Animals with 4 days of arthritis were sacrificed and whole paws collected for protein (□) and mRNA (▪) quantification. N (naïve mice), V (Vehicle), 0·3, 1, 3 and 9 mg/kg (doses of prednisolone assayed), N.D. (not determined). Data represent means ± S.E.M. (n = 6–8 paws from different animals/group). Protein data are expressed as pg protein/g of tissue and mRNA results, normalized against the housekeeping gene Ubiquitin, are expressed as copy No/50 ng cDNA. *P < 0·05, **P < 0·01 (for inflammation and protein data);+P < 0·05, ++P < 0·01 (for mRNA data), compared to vehicle treated mice.

DISCUSSION

The current study aims to offer a valuable tool to researchers who employ animal models of RA to evaluate the clinical efficacy and the mechanism of action of novel therapeutic agents in arthritis. In particular, we have focused on the quantitative analysis of the expression of different biomarkers in joints from CIA in DBA/1OlaHsd mice and from the reactivation SCW-induced arthritis model in Lewis rats. Despite the extensive use of these models, the kinetics of mRNA and protein quantification of pro-inflammatory cytokines in arthritic joints has not been fully characterized and accordingly, the optimal points for biomarkers production have not yet been reported. This in turn has significant bearing on compound evaluation and efficacy.

In the present study, we describe a comparative analysis of cytokine protein and mRNA expression levels in rodent arthritic joint tissues from the two animal models mentioned above, applying ELISA and real-time RT-PCR TaqMan® techniques. As peripheral blood might not be the best indicator of disease activity and progression in RA and therefore does not necessarily correlate with clinical response [24], we investigated the kinetics of expression for the biomarkers IL-1β, TNF-α and IL-6 at the site of disease. These pro-inflammatory cytokines, which are elevated in human RA synovium, have also been shown to contribute to the development of arthritis in some animal models of RA [11–14]. In particular, using antibodies against rat IL-1 and TNF-α, Schrier et al. [13] reported that both cytokines are essential for the reactivation response induced by systemic challenge with SCW in Lewis rats. To date, the IL-6 dependency of the model has not been established. Our time-course study showed that in this model, the initial and the reactivated inflammatory responses which occurred after i.a and i.v injection of SCW-suspension, respectively, were associated with an up-regulation of IL-1β and IL-6 expression levels in joints, highlighting a significant correlation between mRNA and protein expression for both biomarkers. TNF-α showed a similar pattern of expression to IL-1β and IL-6, but only at the mRNA level, as protein could not be detected at either phase of disease. This result differs from published data [25] showing that TNF-α levels can be detected in lavage fluid from arthritic rat joints 24 h after SCW-challenge. A potential explanation for this discrepancy is that low levels of this cytokine might be diluted further in total joint homogenate compared to lavage fluid. Similar findings to SCW-induced arthritis were observed for the kinetics study in mouse CIA. In this model a blockade of IL-1β, TNF-α or IL-6 with anticytokine antibodies or anti-IL-6R treatment was previously shown to inhibit the development of arthritis [15,26]. It should be noted that blockade of TNF-α early after the onset of CIA but not in late disease was shown to be partially effective [15]. This observation was also confirmed in our laboratory (data not shown) and together with the low and inconsistent levels of TNF-α detected alongside the time-course, supports that although TNF-α may play a role during the onset of CIA this cytokine is not the main driver of disease in this model. Our results showed that the onset of the inflammatory response correlated well with high expression levels for IL-1β and IL-6 protein and mRNA in CIA arthritic joints and despite IL-6 presenting a sharper pattern of expression than IL-1β, levels for both cytokines remained significantly elevated at day 4. Our failure to detect TNF-α protein levels in homogenized arthritic rat or mouse joints was not due to a technical problem, since high levels were quantified applying the same technology in joints from rat and mouse carrageenan paw oedema models (data not shown). Given that mRNA precedes protein expression, our data clearly suggests that real-time RT-PCR TaqMan® analysis offers a more sensitive approach than ELISA to study cytokine induction in joints from animal models of experimental arthritis.

As the TNF-α dependency of both SCW and CIA models has previously been reported [3,15], we cannot formally exclude the possibility that earlier or later time points not included in our studies are optimal for maximal TNF-α production although the time points studied overlaps with previous reports. Using IHC techniques, Marinova-Mutafchieva et al. [14], demonstrated that TNF-α-expressing cells were detected on day 1 of CIA, whereas IL-1β-expressing cells were not detected until day 3, thus suggesting that the expression of TNF-α precedes that of IL-1β in CIA mouse joints. This group and others also shown that the number of cells expressing IL-1β, TNF-α and IL-6 persisted and expanded during the course of CIA and identified macrophages as the major cytokine-producing cells located in the lining layers and in the marginal zone from CIA arthritic joints [14,27]. Our findings for IL-1β and IL-6 agree with previous reports, but we cannot explain the disconnect regarding TNF-α other than to conclude that the levels of TNF protein in homogenized joints are too low to detect.

The optimal time point to investigate the mechanism of action of antiarthritic compounds would be when the production of each specific biomarker of interest and inflammation are both maximal. Our results showed that the peak of production of the pro-inflammatory cytokines IL-1β and IL-6 precede maximal joint swelling and therefore a compromised time-point involving suboptimal levels of biomarkers but maximal inflammation was selected for further studies. In the rat SCW-induced arthritis model the peak of inflammation occurred on day 3 after systemic challenge when mRNA levels for IL-1β, IL-6 and TNF-α and protein levels for the first two cytokines were still significantly raised. Similar results where found at day 4 in the mouse CIA model. Thus, in our studies both time points were chosen to investigate any relationship between reduction of inflammation and cytokine production at both, mRNA and protein levels after treatment with the steroid prednisolone. It has been reported that glucocorticoids diffuse into the cell and bind with a cytoplasmic glucocorticoid receptor, which moves to the nucleus where it induces the transcription of IκBα. This action inactivates NF-κB, decreasing the proinflammatory cytokine production as well as inducing genes to inhibit cyclooxygenase-2 (COX-2), adhesion molecules and other inflammatory mediators [21]. Under our experimental conditions, prednisolone reduced joint swelling via a mechanism concordant with a reduction on joint IL-6 and IL-1β mRNA and protein levels in both RA models. At the dose assayed this reference compound did not inhibit mRNA TNF-α levels in rat joints from the SCW-induced arthritis model and only suppressed mRNA TNF-α production at the highest dose (9 mg/kg) in the mouse CIA model. This was an unexpected result, since it's known that glucocorticoids reduce TNF-α production from macrophages and monocytes in vivo[20]. As expected, and confirming our previous kinetic studies, protein levels for these cytokines were not reliable quantified, as detected levels were lower or close to the lower limit of detection of the ELISA method.

CONCLUSIONS

In this study we describe the kinetics of IL-1β, TNF-α and IL-6 mRNA and protein expression in joints from DBA/1OlaHsd mice with CIA and from Lewis rat SCW-induced arthritis applying ELISA and real-time RT-PCR TaqMan® methodologies. To our knowledge, this is the first report showing not only a comparative analysis between protein and mRNA expression, but also assessing the effect of prednisolone treatment on cytokine levels in joints from these two models of experimental arthritis. Our results suggest that optimal sampling time-points to study the effect of anti-inflammatory compounds on both, inflammation and IL-1β and IL-6 levels, are day 4 in CIA and day 3 after i.v challenge in SCW-induced arthritis. We have also shown that in these models there is a significant correlation between mRNA and protein levels in joints for IL-1β and IL-6. Under our experimental conditions, protein levels for TNF-α in arthritic joints at the selected time points are not detectable by ELISA, whereas real-time RT-PCR TaqMan® reliably detects mRNA levels for this cytokine.

These observations suggest that RT-PCR TaqMan® is a sensitive technique that can be successfully applied to the quantification of mRNA levels in mouse and rat joints from experimental arthritis models, thereby providing additional functional insight into the mechanism of action and therapeutic effect of potential anti-inflammatory drugs.