OBJECTIVE This study was designed to investigate whether 1,25-dihydroxyvitamin D3 (1,25-(OH)2D3), produced by activated synovial fluid macrophages, promotes its own catabolism by upregulating vitamin D-24-hydroxylase (24-OHase) in synovial fibroblasts through a vitamin D receptor (VDR) mediated mechanism.
METHODS Synovial macrophages and fibroblasts were derived from patients with rheumatoid arthritis. Expression of VDR and 24-OHase mRNAs was determined using in situ hybridisation. Vitamin D hydroxylase activity was determined by incubating cells with [3H]-25-(OH)D3, or [3H]-1,25-(OH)2D3, and metabolite synthesis quantified using high performance liquid chromatography.
RESULTS 1,25-(OH)2D3increased expression of mRNA for both VDR and 24-OHase in fibroblasts by approximately threefold over 24 hours. 1,25-(OH)2D3 increased fibroblast 24-OHase activity, yielding 24-hydroxylated, and more polar, metabolites. In co-culture, fibroblasts were able to catabolise macrophage derived 1,25-(OH)2D3.
CONCLUSIONS 1,25-(OH)2D3is produced by macrophages in vitro at biologically relevant concentrations and can increase its own catabolism by synovial fibroblasts; this effect is probably mediated via upregulation of both synovial fibroblast VDR and 24-OHase.
- vitamin D-24 hydroxylase
- in situ hybridisation
Statistics from Altmetric.com
The active metabolite of vitamin D3, 1α,25-dihydroxyvitamin D3(1,25-(OH)2D3), has immunoregulatory, pro-differentiating, and anti-proliferative properties that have been shown to inhibit the progression of a number of autoimmune diseases, including experimental autoimmune encephalomyelitis1 and autoimmune diabetes.2 3 A recent paper by Cantornaet al 4 has shown that 1,25-(OH)2D3 can both prevent the development and halt the progression of arthritis in murine models of the human disease. Oral 1,25-(OH)2D3 treatment has also been shown to improve human psoriatic arthritis,5 and its analogue 1α-hydroxyvitamin D3 inhibits type II collagen induced arthritis in rats.6
Exogenous administration of 1,25-(OH)2D3 may be a useful therapeutic approach for arthritis, however substrate dependent endogenous production of 1,25-(OH)2D3can occur within the arthritic joint in vivo.7 8 The source of the extra-renal 1,25-(OH)2D3 is likely to be the synovial macrophage, as we have demonstrated in vitro that macrophages isolated from the synovial fluids of patients with inflammatory arthritis express vitamin-D-1α-hydroxylase (1-OHase) activity,9 which can be regulated by locally encountered factors such as cytokines, for example, interferon gamma (IFNγ), bacterial lipopolysaccharide, and eicosanoids.10-14
Another vitamin D hydroxylase enzyme, namely the 24-hydroxylase (24-OHase), is responsible for the catabolism of 1,25-(OH)2D3, and is expressed by synovial fibroblasts.13 1,25-(OH)2D3 and its receptor (VDR) bind to the vitamin D response elements (VDREs) of target genes to control their transcription and, as the promoter region of the human 24-OHase gene contains VDREs,15 it appears that the activity of the synovial fibroblast 24-OHase might be under the control of the local concentration of macrophage derived 1,25-(OH)2D3.
In this study we wanted to investigate how induction of synovial fibroblast 24-OHase by 1,25-(OH)2D3 may, in turn, regulate the local concentration of 1,25-(OH)2D3 in the synovial joint. In theory, this could be done via several mechanisms. Firstly, it may promote catabolism of existing 1,25-(OH)2D3 to 1,24,25-trihydroxyvitamin D3(1,24,25-(OH)3D3), the start of the catabolic cascade for 1,25-(OH)2D3. While 1,24,25-(OH)3D3 has some biological activity, it is less potent than the precursor 1,25-(OH)2D3.16 Secondly, the 24-OHase may lessen the pool of precursor 25-(OH)D3available to macrophages for 1-hydroxylation (which is known to be substrate dependent), by diverting 25-(OH)D3 into the synthesis of 24,25-dihydroxyvitamin D3(24,25-(OH)2D3). An increase in 24,25-(OH)2D3 concentration may itself provide a third mechanism to reduce entry of freshly synthesised 1,25-(OH)2D3 into the system; we have demonstrated that 24,25-(OH)2D3 can be converted to 1,24,25-(OH)3D3 by synovial fluid macrophages (unpublished data), and will therefore compete with 25-(OH)D3 as a substrate for 1-hydroxylation.
If 1,25-(OH)2D3 (or one of its analogues) is to be used therapeutically for inflammatory arthropathies it is important to understand its breakdown in vivo, firstly for the administration of an effective dose and secondly to ensure that the concentration of 1,25-(OH)2D3 can be controlled locally, thus preventing unwanted systemic side effects and possible juxta-articular osteoporosis. The aim of this study was to investigate further the regulation of the synovial fibroblast 24-OHase by 1,25-(OH)2D3 in vitro, at both the molecular and functional level, and to determine whether this effect is mediated via VDR. Expression of synovial fibroblast VDR and 24-OHase mRNAs was determined using in situ hybridisation. The effects of exposure to 1,25-(OH)2D3 on the metabolism of [3H]-25-(OH)D3 and [3H]-1,25-(OH)2D3 by synovial fibroblasts, and the effects of fibroblasts on the metabolism of 1,25-(OH)2D3 produced by macrophages when the two cell types were co-cultured, were also investigated using a hydroxylase assay.
SOURCE OF SYNOVIAL CELLS
Synovial fibroblasts (designated F733), derived from an explant culture from a patient with rheumatoid arthritis, were kindly supplied by Dr David Taylor. Cells were routinely grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 1 × 10-3 M sodium pyruvate (Gibco), 2 × 10-3 M L-glutamine, 50 IU/ml penicillin, 50 μg/ml streptomycin, and 10% v/v fetal bovine serum (Imperial) and passaged every 1–2 weeks using 0.25% trypsin, 0.02% w/v EDTA in phosphate buffered saline (Imperial). Cells were used at passage 10 in these experiments.
A primary culture of synovial fluid macrophages was established from the synovial fluid of a patient with rheumatoid arthritis using the method of Hayes et al.12Briefly, the synovial fluid sample was diluted in RPMI-1640 culture medium containing 2% v/v fetal bovine serum, 4 × 10-4 M L-glutamine, 10 IU/ml penicillin, and 10 μg/ml streptomycin. From experience, we have found that incubating the sample in 2% v/v fetal bovine serum, rather than 10% v/v fetal bovine serum, encourages the culture of macrophages, but not fibroblasts, thus allowing fibroblast free cultures of macrophages to be established. Any macrophage culture contaminated by fibroblasts can be readily identified, because fibroblasts proliferate in culture and can be seen as colonies.
The sample was incubated at 37°C in a humidified atmosphere of air with 5% carbon dioxide for 2–3 days; after this time macrophages had adhered to the culture plate surface and non-adherent cells were washed away. All cells from the sample stained positive for non-specific esterase activity, a marker of monocyte-macrophages.
DETERMINATION OF mRNA EXPRESSION USING IN SITU HYBRIDISATION
F733 cells were pre-treated for five or 24 hours with 5 × 10-8 M 1,25-(OH)2D3 (Hoffman La Roche) or vehicle (0.125% v/v ethanol) in glass chamber slides (Lab-Tek) and fixed in 10% v/v formaldehyde. Expression of VDR and 24-OHase mRNA was determined using in situ hybridisation with [35S] labelled riboprobes. The 2.1 kb human VDR cDNA probe was kindly supplied by Dr G Williams already subcloned into the ampicillin resistant plasmid pBluescript II KS (Stratagene). The 1.8 kb partial length human 24-OHase cDNA probe was kindly supplied by Drs J Omdahl and B May, cloned into the Pst I site of Bluescript KS+. The in situ hybridisation technique used in this study has previously been described by Mee et al.17Controls were sense riboprobe and RNase A treated cells (Sigma; 1 mg/ml in water for one hour at 37°C) to ensure that probe binding to cells was RNA specific. Probe hybridisation was detected by dipping slides in photographic emulsion (Ilford K5) diluted 1:1 with distilled water and developed after seven days at 4°C. The radiation produces silver grains within the emulsion that can be visualised as “black dots” and, as the path length of [35S] is 15–20 μm, the positions of the black dots localise the areas of probe binding. Cells were counterstained using eosin and Harris haematoxylin and examined using light field microscopy. Probe hybridisation was quantified with a Quantimet 600 image analyser linked to a Leica DM microscope (Leica UK Ltd),18 allowing the percentage cell area covered by silver grains to be calculated, corrected for background levels of probe binding.
DETERMINATION OF VITAMIN D HYDROXYLASE ACTIVITY
Effect of 1,25-(OH)2D3 on synovial fibroblast vitamin D hydroxylase activity
To study the effects of 1,25-(OH)2D3on F733 vitamin D hydroxylase activity, cells were pre-treated with 1,25-(OH)2D3 (10-7 to 10-11 M), or vehicle (ethanol, 0.25% v/v), in DMEM containing 6.25 μg/ml insulin, 6.25 μg/ml transferrin, 6.25 ng/ml selenous acid, 1.25 mg/ml bovine serum albumin (BSA), 5.35 μg/ml linoleic acid (ITS+; Becton Dickinson), 2 × 10-3 M L-glutamine, 50 IU/ml penicillin, and 50 μg/ml streptomycin. Cells were incubated for 24 hours in a humidified atmosphere of 95% air, 5% carbon dioxide, then washed and left for three hours in ITS+ supplemented DMEM; this allowed time for any pre-treatment 1,25-(OH)2D3 to be metabolised by cells before incubation with substrate. Enzyme activity was then determined by incubating cells for 1.5 hours in fresh DMEM containing either 25-(OH)D3 or 1,25-(OH)2D3 (358 pg/ml and 43.5 pg/ml respectively), supplemented with 5 × 104 dpm tritiated substrate (25-hydroxy [26,(27)-methyl-3H] cholecalciferol ([3H]-25-(OH)D3), specific activity about 15.0 Ci/mmol or 1α,25-dihydroxy [26,27-methyl-3H] cholecalciferol ([3H]-1,25-(OH)2D3), specific activity about 173 Ci/mmol; Amersham) 0.5% v/v BSA (Orthodiagnostic Systems Ltd) and 2 × 10-3 M L-glutamine, using the hydroxylase assay described previously by Hayes et al.12 A blank incubation (no cells) was also included in the assay. Metabolites were extracted using chloroform: methanol, 3:2 v/v and quantified using high performance liquid chromatography (HPLC) and liquid scintillation counting. Briefly, the chloroform layer was extracted and analysed on a manual injection normal phase HPLC system developed with a mobile phase of n-hexane:isopropanol:methanol, 115:3.3:4 (v/v) at 2 ml/min and a Zorbax-Sil column (4.6 mm × 25 cm). Eluant fractions were collected, the radioactivity measured, and the known specific activity of substrate used to calculate the rate of synthesis of vitamin D3 metabolites in each sample. An aliquot of the upper aqueous phase of the extraction mixture was also counted as a measure of more polar metabolites.
Synovial macrophages and fibroblasts in co-culture
A co-culture system was established to determine whether synovial macrophage derived 1,25-(OH)2D3 could be directly 24-hydroxylated by F733 in vitro to form 1,24,25-(OH)3D3. F733 were grown in sterile 12 mm diameter polystyrene well inserts (Costar), which could fit into the wells of a 12 well tissue culture dish (Corning) in which the macrophages were cultured. The inserts had a 3.0 μm porous membrane that allowed the culture medium to move freely between the chambers. Adequate synthesis of 1,25-(OH)2D3 by macrophages was ensured by pre-treatment for 24 hours with 5 × 10-9 M interferon gamma (IFNγ; Advanced Protein Products Ltd). After pre-treatment, macrophages were washed once with basal medium, then incubated alone or in combination with F733 for a further 24 hours with [3H]-25-(OH)D3 (12.5 ng/ml) using the hydroxylase assay described above.
Statistical analysis of results was carried out with Instat for MacIntosh (version 2.01), using analysis of variance (ANOVA) and Student’s t test. Non-normally distributed data were log transformed before analysis.
REGULATION OF 24-OHASE AND VDR mRNAS BY 1,25-(OH)2D3 IN F733 SYNOVIAL FIBROBLASTS
24-OHase mRNA (fig 1 (A–C))
Pre-treatment with 1,25-(OH)2D3 for either five or 24 hours increased probe binding to F733, indicating an increase in 24-OHase mRNA. At both time points some cells had very high activity while others had very low activity, possibly because of differences in the state of cell differentiation or cell cycle, or both. The extent of hybridisation was quantified and expressed as the % cell area covered by silver grains. There was a significant increase in probe binding (p < 0.01) from 2.3 (0.5)% (mean (SEM)) in cells treated with vehicle for five hours to 6.2 (1.0)% in cells treated with 1,25-(OH)2D3 for five hours (results not shown). There was a similar increase in probe binding (p < 0.05) from 1.7 (0.3)% in cells treated with vehicle for 24 hours fig 1(A)) to 5.1 (2.2)% in cells treated with 1,25-(OH)2D3 for 24 hours (fig 1(B)). Both RNase A treated control slides (fig 1(C)) and sense riboprobe control slides showed only background levels of silver grains.
VDR mRNA (fig 1 (D–F))
There was an increase in probe binding to F733 cells after 24 hours of 5 × 10-8 M 1,25-(OH)2D3treatment, but not after five hours. Percentage silver grain density for five hour vehicle, five hour 1,25-(OH)2D3treated cells, and 24 hour vehicle were 1.4 (0.2)%, 1.5 (0.3)%, and 1.5 (0.2)% respectively. Binding after 24 hour 1,25-(OH)2D3 pre-treatment (4.5 (0.5)%; fig 1(E)) was significantly increased compared with the 24 hour vehicle (fig 1(D); p < 0.01) and five hour treatment values (p < 0.01). Again, both RNase A treated control slides (fig 1(F) and sense riboprobe control slides showed only background levels of silver grains.
EXPRESSION OF VITAMIN D HYDROXYLASE ACTIVITY (FROM HYDROXYLASE ASSAYS)
Metabolism of [3H]-25-(OH)D3 by synovial fibroblasts pre-treated with 1,25-(OH)2D3
The HPLC trace in fig 2(A) shows the typical metabolism of [3H]-25-(OH)D3 (peak A) to [3H]-24,25-(OH)2D3 (peak B) and an unidentified metabolite (peak D) by F733 cells pre-treated for 24 hours with 10-8 M 1,25-(OH)2D3. Fig3 (A) shows that synthesis of [3H]-24,25-(OH)2D3 was increased in a dose dependent manner by increasing pre-treatment concentrations of 1,25-(OH)2D3, with a mean (SEM) of 73.8 (5.3) fmol/h/106 cells in controls (vehicle treated), and maximum activity of 463.1 (14.6) fmol/h/106 cells in 10-8 M pre-treated cells (p < 0.01). The unidentified metabolite (peak D in figures 2(A), and 3 (B)), which was more polar than 24,25-(OH)2D3, was also synthesised in a dose dependent manner, and was not present in blank (no cells) samples, indicating this is a metabolite and not an incubation artefact. This metabolite was not 25,26-dihydroxyvitamin D3, which elutes between 24,25-(OH)2D3 and 1,25-(OH)2D3 in our solvent system.
An aliquot of the aqueous phase was added to scintillation cocktail and [3H] dpm measured. The unidentified radioactivity present in the aqueous phase of the extraction mixture (corrected from the blank (no cells) incubation) increased with 1,25-(OH)2D3 pre-treatment concentration (fig 3(C)). This radioactivity indicates the presence of more polar products than those found in the chloroform layer, however the number and nature of metabolites present in the aqueous phase was not determined.
Metabolism of [3H]-1,25-(OH)2D3 by synovial fibroblasts pre-treated with 1,25-(OH)2D3
The HPLC trace in figure 2 (B) shows the typical metabolism of [3H]-1,25-(OH)2D3 (peak C) to [3H]-1,24,25-(OH)3D3 (peak E) and an unidentified metabolite (peak F) by F733 pre-treated for 24 hours with 10-11 M 1,25-(OH)2D3. As no radioactive tracer was available for 1,24,25-(OH)3D3, a=n ultraviolet absorbance trace of unlabelled 1,24,25-(OH)3D3 was used as a marker. 1,25-(OH)2D3 pre-treatment led to an increase in putative [3H]-1,24,25-(OH)3D3 synthesis (fig 3 (D)) with a mean (SEM) of 5.6 (0.9) fmol/h/106 cells in vehicle treated control cells (0.25% v/v ethanol) to 33.3 (2.7) fmol/h/106 cells in 10-9 M treated cells (p < 0.01). The concentration of [3H]-1,24,25-(OH)3D3 peaked with 10-9 M 1,25-(OH)2D3 pre-treatment and declined at 10-8 M and 10-7 M. It is possible that the cells pre-treated with these higher concentrations of 1,25-(OH)2D3 had started to metabolise [3H]-1,24,25-(OH)3D3 to further catabolic products, which would explain the much higher counts in the aqueous phase seen with these two treatment points (fig 3 (F)). 1,25-(OH)2D3 pre-treatment had no significant effect on the synthesis of the unidentified more polar metabolite (peak F in fig 2 (B), and fig 3 (E)).
Vitamin D metabolism in synovial macrophages and fibroblasts in co-culture
Table 1 shows the rates of [3H]-vitamin D3 metabolite synthesis by synovial fluid macrophages and fibroblasts alone and in co-culture. F733 cells synthesised 24,25-(OH)2D3 while macrophages synthesised 1,25-(OH)2D3. IFNγ stimulated macrophages synthesised increased levels of 1,25-(OH)2D3and a small quantity of putative 1,24,25-(OH)3D3—possibly because of induction of macrophage 24-OHase activity when local 1,25-(OH)2D3 levels were high. The co-culture of F733 and IFNγ stimulated macrophages showed synthesis of 1,25-(OH)2D3, 24,25-(OH)2D3 and 1,24,25-(OH)3D3. There were also increased concentrations of [3H] metabolites in the aqueous phases of the co-culture incubations than in cells incubated separately (p < 0.001).
We have demonstrated previously that 1,25-(OH)2D3 may be synthesised locally within the arthritic joint8 and have studied the regulation of synovial macrophage 1-OHase activity in vitro.9-14 We have now extended these observations to show that, in vitro at least, fibroblasts of synovial origin can regulate the local 1,25-(OH)2D3 concentration at or near its site of synthesis. Moreover, this mechanism is regulated in a dose dependent fashion by 1,25-(OH)2D3 itself, acting in a paracrine mode, with a significant effect seen at near physiological concentrations (10-10 M). The stimulatory effects of 1,25-(OH)2D3 on synovial fibroblast 24-OHase activity may be mediated via stimulation of both 24-OHase and VDR gene transcription, or by increasing mRNA stability, with increased mRNAs at five hours and 24 hours respectively. Initial upregulation of 24-OHase activity may be sufficient to catabolise raised local 1,25-(OH)2D3 levels, however if supraphysiological concentrations of 1,25-(OH)2D3 persist, an increase in VDR mRNA and protein may be required to increase the clearance rate.
We have previously studied the effect of 1,25-(OH)2D3 on 24-OHase activity and mRNA in a variety of rheumatoid synovial fibroblast cell lines, and on the same lines at different passage numbers, and always found a stimulatory effect on both 24-hydroxylase and mRNA levels—so much so, that fibroblast cell lines are used as reliable positive controls in house for the in situ hybridisation method. We have not studied fibroblasts from osteoarthritic synovium, however it would be interesting to see if there were differences in the expression of these molecules in the different disease states, particularly as we have previously shown a difference in endogenous 1,25-(OH)2D3 synthesis in rheumatoid and osteoarthritis.8
Once the stimulatory effect of exogenous 1,25-(OH)2D3 on fibroblast 24-OHase had been established, we examined the metabolism of 25-(OH)D3 by macrophages and fibroblasts in co-culture. We have now demonstrated that synovial fibroblasts can metabolise macrophage derived 1,25-(OH)2D3 to 1,24,25-(OH)3D3, and more polar metabolites in exactly the same way as exogenous 1,25-(OH)2D3, confirming the hypothesis that fibroblasts are able to regulate the concentration of locally synthesised 1,25-(OH)2D3 in vitro.
Our experiments also indicate that a relatively small number of macrophages (2.18 × 105 cells per incubation) have the ability to make a large amount of 1,25-(OH)2D3. The concentration of substrate 25-(OH)D3 used in the co-culture system was physiological (12.5 ng/ml or 3.125 × 10-8 M) and resulted in 1,25-(OH)2D3 concentrations of approximately 6 × 10-9 M and 1.7 × 10-8 M in control and IFNγ stimulated cultures respectively after 24 hours. As the normal serum concentration of 1,25-(OH)2D3 is in the order of 10-10 M to 10-11 M (20–50 pg/ml; from Mawer et al 19), it is theoretically possible that, unless regulated, high concentrations of 1,25-(OH)2D3 could be synthesised in the joint within a relatively short time period. The concentration of macrophage derived 1,25-(OH)2D3 was reduced to 6 × 10-9 M in the co-culture incubation (24 hour incubation with 2.18 × 105 macrophages and 1.1 × 105fibroblasts), because of the catabolic effects of the fibroblast 24-OHase. This is one third of the concentration present in the IFNγ stimulated macrophages alone, although the regulatory effect of the fibroblasts may depend on the ratio of the different cell types and incubation time. Macrophages and fibroblasts are generally the predominant cell types in the rheumatoid lesion, but their relative proportion and distribution can be highly variable, even within the same specimen.20 21 We have measured the synovial fluid concentration of 1,25-(OH)2D3 in patients with rheumatoid arthritis receiving oral vitamin D2 treatment and shown that concentrations were up to 2.5 × 10-10 M (100 pg/ml; unpublished results). We did not detect concentrations higher than this in vivo, possibly because of insufficient substrate, inactivated macrophages or local breakdown within the joint in analogous fashion to that which we have described in the co-culture in vitro.
The beneficial effects of 1,25-(OH)2D3 in a number of animal models of autoimmunity, including experimental autoimmune encephalitis and autoimmune diabetes,1-3 22have led to the suggestion that 1,25-(OH)2D3, or one of its analogues, may be a valuable treatment approach to arthritis.4 6 It would be possible, rather than giving 1,25-(OH)2D3, or its analogues, systemically (as was the case for the animal experiments), to inject it locally into the joint, as for corticosteroid treatment, thus targeting the site of inflammation directly. Our findings on the ability of 1,25-(OH)2D3 to stimulate its own catabolism within the environment of the synovial joint would be of critical importance in regulating the local concentration of 1,25-(OH)2D3 in these circumstances.
Funding: this work was supported by the Arthritis and Rheumatism Council for Research.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.