Article Text
PDF
Abstract
Background It has previously been shown that dual activation of the Liver X Receptors (LXRα and LXRβ) by the agonist, GW3965, enhances pathology in a murine model of collagen-induced arthritis.
Objective To determine whether LXRα or LXRβ have discrete roles in driving articular inflammation.
Methods Arthritis was induced in male C57BL/6 wild-type (WT), LXRα−/−, LXRβ−/− and LXRα/β double KO mice by injection with type II collagen and treated with 30 mg/kg of the LXR agonist GW3965 or vehicle control. The mice were monitored for articular inflammation and cartilage degradation by scoring for clinical signs of arthritis and by histological examination of the joints.
Results Administration of 30 mg/kg GW3965 significantly increases the severity of arthritis in WT but not LXRα−/−, LXRβ−/− or LXRα/β KO mice as assessed by an increase in the clinical score, paw thickness and articular histological analysis.
Conclusion The proinflammatory effects associated with the administration of GW3965 are mediated specifically through LXRs. The absence of increased disease severity in the LXRα−/− and LXRβ−/− GW3965-treated groups shows for the first time that agonism of both LXRα and LXRβ is required to drive proinflammatory pathways in vivo.
Statistics from Altmetric.com
Introduction
Rheumatoid arthritis (RA) is a chronic autoimmune inflammatory condition of unknown aetiology associated with progressive functional decline but also with increased mortality. The latter is largely attributable to excess cardiovascular risk that is unexplained by conventional risk factors.1 Vascular physiological parameters including carotid intima-media thickness are dysregulated in RA.2 3 Moreover, a pro-atherogenic plasma lipid profile is evident up to 10 years before the clinical onset of RA and a decrease in ‘protective’ high-density lipoprotein-cholesterol negatively correlates with C-reactive protein.4,–,6 Such studies suggest that lipid dysregulation might directly or indirectly modify RA-associated inflammation (reviewed by Sattar and McInnes7). In support of this we have previously shown that agonism of molecular pathways endogenously activated by cholesterol derivatives significantly increased the severity of articular inflammation in a murine model of collagen-induced arthritis (CIA).8
The liver X receptors (LXRα and LXRβ) are members of the nuclear receptor superfamily of ligand activated transcription factors. LXRs complex with the nuclear receptor retinoid X receptor (RXR) to form an LXR/RXR heterodimer.9 In the absence of ligand, LXR/RXR heterodimers are conformationally inert and are bound by transcriptional co-repressors rendering them transcriptionally inactive. Transcriptional activation of LXRs is achieved by the binding of specific oxidised cholesterol derivatives. Oxysterols induce a conformational change of the LXR/RXR heterodimer complex, promote release of co-repressors and thereby allow binding of co-activator complexes, and subsequently the transcription of LXR target genes—for example, apolipoprotein E, ATP-binding cassette A1 (ABCA1) and Toll-like receptor-4.10,–,12
We have previously shown that administration of the LXR agonist, GW3965, in murine CIA was associated with a significant dose-dependent increase in disease severity.8 Similarly, treatment with GW3965 of human macrophages co-cultured with cytokine-activated T cells significantly increased the secretion of multiple proinflammatory cytokines and chemokines. Together these data suggest a predominantly proinflammatory role for LXR activation in RA. However, as GW3965 simultaneously activates both LXRα and LXRβ we were previously unable to distinguish whether the proinflammatory effects were mediated through LXRα, LXRβ, or both. Therefore, we generated LXR-deficient mice to determine the individual roles of LXRα and LXRβ in driving arthritis disease pathology in vivo.
Materials and methods
Reagents
GW3965 was dissolved in 5% cremophore/phosphate-buffered saline (Sigma-Aldrich, UK).
Mice
Experiments were carried out according to UK Home Office guidelines. All mice were bred in a pathogen-free environment at the University of Glasgow's biological service facility and fed a chow diet. LXRα−/−, LXRβ–/– and wild-type (WT) littermates on the C57BL/6 background were supplied by Lexicon Pharmaceuticals. LXRα–/– and LXRβ–/– were crossed to generate LXR heterozygotes (LXRα±/β±), which were self-crossed to generate LXRα/β double knockout (LXRα/β KO) mice (data not shown).
Induction of arthritis and scoring
CIA was induced in male LXRα–/–, LXRβ–/–, LXRα/β KO or WT littermates at 8–11 weeks by injection of 200 μg chicken collagen/complete Freund's adjuvant.13 Mice were monitored for clinical signs of disease from day 21 onwards by micro-calliper measurements of paw thickness and a clinical score of disease severity: 1 = swollen digit(s), 2 = erythema, 3 = swollen paw/ankle and 4 = loss of function, allowing a maximum score of 16 per animal. For histological assessment, rear paws were fixed, decalcificied and 7 µM paw sections were stained with haematoxylin and eosin. RNA was purified and analysed as previously described using the following primers; mouse TBP – Mm00446973_m1 and mouse ABCA1 – Mm00442646_m1 (all Applied Biosystems, Paisley, UK).8
Statistical analysis
Results are displayed as mean±SEM. All statistical analysis was by the Student t test or two-way analysis of variance using the Graph Pad Prism 4 software.
Results
To evaluate potentially discrete roles for LXRα or LXRβ-dependent signalling in vivo, arthritis was induced in male LXRα–/–, LXRβ–/–, LXRα/β KO or WT mice. Vehicle (5% cremophore/phosphate-buffered saline) or 30 mg/kg GW3965, a dose we have previously found to be optimal,8 was administered by daily intraperitoneal injection from 1 day before the induction of arthritis to day 37. To confirm activation of LXRs by GW3965 the expression of the reporter gene ABCA1 in the jejunum was measured by TaqMan QRT-PCR (figure 1). The expression of ABCA1 was significantly higher (13-fold) in WT GW3965-treated mice than in vehicle controls. Similarly, administration of GW3965 induced a seven- and threefold increase in the expression of ABCA1 in LXRα–/– and LXRβ–/–, respectively. As expected the expression of ABCA1 was not significantly higher in LXRα/β KO mice treated with GW3965 than in respective vehicle controls (p=0.235). However, basal ABCA1 expression in the vehicle-treated LXRα/β KO group was significantly increased compared with the WT vehicle-treated control group (p<0.001).
Liver X receptor (LXR) reporter gene expression. Collagen-induced arthritis was initiated by injection of type II collagen/complete Freund's adjuvant. RNA was extracted from the jejunum of arthritic mice to measure the induction of the LXR reporter gene ATP-binding cassette A1 (ABCA1) by intraperitoneal injection of 30 mg/kg GW3965 relative to vehicle (5% cremophore/phosphate-buffered saline) control groups in wild-type (WT), LXRα−/−, LXRβ−/− and LXRα/β double knockout (LXRα/β KO). Unpaired Student t test; **p<0.01, ***p<0.001, GW3965 vs vehicle within each genotype. n = 9–13/group.
We monitored clinical parameters of arthritis from day 21 after type II collagen immunisation. Clinical signs of arthritis were evident in approximately 50% of mice regardless of genotype or treatment group (figure 2A). The incidence and severity of disease in the WT vehicle-treated mice was comparable to previously published data in C57BL/6 mice using this protocol.13 14 Administration of 30 mg/kg GW3965 in WT mice significantly increased disease severity assessed by increased paw swelling and clinical scores compared with the WT vehicle-treated control group (figure 2B,C). In contrast, administration of GW3965 did not significantly increase the clinical severity of disease in LXRα–/–, LXRβ–/– or LXRα/β KO mice compared with their respective vehicle control groups. Paradoxically, LXRα/β KO mice treated with vehicle were associated with a significantly higher clinical score (p=0.0375) and paw swelling (p=0.011) than the WT vehicle group. A similar trend was also evident in the LXRα–/– and LXRβ–/–vehicle groups. However, there was no difference in the concentration of serum inflammatory cytokines and anti-collagen IgG2a antibodies between any of the groups at this time point (data not shown).
GW3965 increases the severity of collagen-induced arthritis specifically through liver X receptors (LXRs). Arthritis was induced in male wild-type (WT), LXRα–/–, LXRβ–/– and LXRα/β double knockout (LXRα/β KO) mice by injection with type II collagen/complete Freund's adjuvant and treated with vehicle (5% cremophore/PBS) or 30 mg/kg GW3965 by daily intraperitoneal injection. (A) Incidence of disease assessed by first signs of clinical disease (n=18–27 mice/group). For the incident mice the (B) clinical score and (C) paw swelling was measured (n=9–13/group). For clarity the transgenic groups are shown in comparison with the WT groups. Two-way analysis of variance: (A) *p<0.05, WT vehicle vs WT GW3965 (left panel) and WT vehicle vs LXRα/β double KO (right panel). (B) **p<0.01 WT vehicle vs WT GW3965 (left panel) and *p<0.05 WT vehicle vs LXRα/β double KO vehicle (right panel). The results are pooled from three independent experiments.
Histological analysis of disease showed a high level of inflammatory infiltrate in the WT vehicle-treated group (figure 3). In addition to the inflammatory infiltrate administration of GW3965 was also associated with cartilage erosion. Similarly, an inflammatory cell infiltrate and cartilage erosion was also evident by histological evaluation of disease in LXRα–/–, LXRβ–/– and LXRα/β KO vehicle-treated or GW3965-treated mice that was comparable to that seen in WT mice treated with GW3965.
Histological assessment of collagen-induced arthritis. Arthritis was induced in male wild-type (WT), LXRα−/−, LXRβ–/– and LXRα/β double knockout (LXRα/β KO) mice and treated with vehicle or 30 mg/kg GW3965 by daily intraperitoneal injection as in figure 1. Paw sections from arthritic mice were stained with H&E; original magnification ×10. B, bone; C, cartilage; J, joint space; M, synovial membrane. Black arrows indicate areas of cartilage erosion and grey arrows indicate areas of bone loss. LXR, liver X receptor.
Discussion
Our previous study demonstrated a novel proinflammatory role for the activation of LXRs in arthritis. This study was therefore initiated to determine the individual role of LXRα and LXRβ in driving articular inflammation. In keeping with our previous findings we have shown that administration of GW3965 significantly increased the severity of arthritis in WT mice. Although GW3965 is regarded as a highly specific LXR agonist off-target effects attributable to GW3965 could not be discounted as potential drivers of inflammation in CIA. Importantly, administration of GW3965 in LXRα–/–, LXRβ–/– or LXRα/β KO was not associated with an increased severity of disease, suggesting that the modality of GW3965 action is mediated specifically through LXRs. Moreover, as there was no observed difference in the severity of disease in GW3965-treated LXRα–/– or LXRβ–/– compared with the respective vehicle controls these data further suggest that simultaneous activation of both LXRα and LXRβ is required to drive downstream proinflammatory pathways in RA. Thus taken together these data support a predominantly proinflammatory role of LXR agonism and imply that LXRα and LXRβ cooperate to drive disease pathology in RA.
LXRs are transcription factors that serve to regulate the expression of a variety of target genes. In the absence of ligand LXRs are bound by transcriptional co-repressors—for example, NCoR and SMRT, and in such context LXRs function as inhibitors of target gene expression. Deletion of LXRs results in a loss of target gene repression and therefore an increase in the basal level of target gene transcription.15 16 Similarly, in accordance with a loss of target gene repression we have shown that the basal level of ABCA1 expression is increased in LXRα/β KO mice (figure 1). The target genes regulated by LXR agonism that drive inflammation in RA are as yet unknown. However, our observations that the vehicle-treated LXRα/β KO group exhibits a significantly increased basal level of arthritis disease severity is most probably caused by loss of target gene repression, thereby increasing the activity of downstream inflammatory pathways in these mice.
A major focus in the field of rheumatology is now to understand how cardiovascular disease might contribute to the pathogenesis of RA. It is clear from our studies that LXR activation can exert proinflammatory effects in vivo and this may be a mechanism by which the atherogenic lipid profile may drive disease pathology in RA. Indeed, clinical interventions aimed at improving dyslipidaemia in RA have been shown to exert anti-inflammatory effects and recommendations are now in place as to how to manage cardiovascular risk in RA.17 18 LXR agonists that specifically target LXRβ are currently in preclinical development for the treatment of atherosclerosis. Importantly, our findings do not detract from the potential beneficial effects of specific LXRβ agonists in a metabolic context as our results suggest that activation of LXRβ alone is not sufficient to drive ongoing inflammation.
References
Footnotes
-
Funding AMM is supported by a BHF Intermediate Basic Science Research Fellowship (FS/08/035/25309). This work was financially supported by the Medical Research Council (UK) and Arthritis Research UK.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.