- Email: [email protected]
Anti-arthritic effect of methotrexate: is it really mediated by adenosine?
European Journal of Pharmaceutical Sciences 9 (2000) 333–343 www.elsevier.nl / locate / ejps
Anti-arthritic effect of methotrexate: is it really mediated by adenosine? ˚ ¨ ¨ Sven E. Andersson*, Lars-Hakan Johansson, Kristina Lexmuller, Gunilla M. Ekstrom Department of Immunopharmacology, Preclinical R& D, Astra Draco AB, Lund, Sweden Received 23 June 1999; received in revised form 23 September 1999; accepted 30 September 1999
Abstract The mechanism of action for the anti-arthritic effect of methotrexate (MTX) was investigated in rats with antigen-induced arthritis (AIA). Arthritis intensity was quantified as area under the curve (AUC) for the joint swelling. The response to MTX was in several respects similar to what is seen in the clinic. The drug reduced the AUC in a dose-dependent manner after oral weekly (2–4 mg / kg / week) or daily (0.3 mg / kg / day) dosing. This effect was not affected by supplementation with an equal dose of folate. The model thus seemed suitable for this type of study. Supplementation with folate in excess abolished the effect of MTX. A structurally similar antifolate, aminopterin, also reduced the arthritis. The effect thus seemed to be due to folate antagonism although a complete inhibition of dihydrofolate reductase (DHFR) might not be essential. Hence, it could be that the main target is a process downstream of DHFR. It has been proposed that inhibition of AICAR-transformylase induce the release of adenosine with anti-inflammatory properties. Here the adenosine antagonist R-PIA reduced the arthritis but when MTX was combined with adenosine antagonists no attenuation of the anti-arthritic effect was seen. On the contrary, three adenosine agonists (8-p-sulphophenyltheophyllamine 30 mg / kg i.p. twice daily; 3,7-dimethyl-1-propargylxanthine, p.o. 3 mg / kg / day and 8-cyclopentyl-1,3-dipropylxanthine, 1.5 mg / kg / day p.o.) potentiated MTX. The specific thymidylate synthase inhibitor 5-fluourouracil (0.3–3.0 mg / kg / day) had no anti-arthritic effect. Neither did our data support the hypotheses that syntheses of polyamines or cytokines were primary targets. It is thus possible that the mechanism of action is inhibition of a process downstream of DHFR but the release of adenosine seems not to be important. 2000 Published by Elsevier Science B.V. All rights reserved. Keywords: Adenosine; IL-6; Methotrexate; Polyamines; Arthritis; TNFa
1. Introduction Methotrexate (MTX) has become the most widely used second line drug for the treatment of rheumatoid arthritis (RA). It is relatively well tolerated and has a well documented clinical efficacy, although its ability to slow down radiological progression of the disease is uncertain (Bannwarth et al., 1994; Rau et al., 1997). It thus seems likely that knowledge of the mechanism of action for the drug would provide important information on the pathophysiology of the disease. This would possibly also enable the development of new anti-rheumatic drugs. At present, however, MTX’s main targets are largely unknown. MTX is a potent inhibitor of dihydrofolate reductase (DHFR) which subsequently inhibits de novo purine and
*Corresponding author. Department of Internal Medicine, Med. Mott., Lund University Hospital. S-221 85 Lund, Sweden. Tel.: 146-46-171000. E-mail address: [email protected] (S.E. Andersson)
pyrimidine synthesis. Based on these properties it was developed as a cytostatic agent. There is, however, much evidence indicating that the anti-arthritic effects of MTX are due more to an anti-inflammatory rather than antiproliferative or immunomodulatory effect (Bannwarth et al., 1994; Cronstein, 1996). Firstly, the dosing differs from that used for treatment of oncological disease in which MTX is administered in pulses of 20–250 mg / kg; for RA treatment it is administered in weekly doses of 0.1–0.3 mg / kg (Cronstein et al., 1993). Secondly, there are reports that RA patients can be supplemented with folate in doses as high as 27.5 mg / week without compromising the drugs efficacy (Morgan et al., 1994). Although there is no consensus over this view (Stenger et al., 1992) such data suggest that complete inhibition of DHFR is not essential in rheumatological use. Interestingly it has been reported that only antifolates with a high structural similarity to MTX have anti-arthritic effect (Galivan et al., 1986; Baggott et al., 1993). It has therefore been proposed that inhibition of other folate dependent enzymes, located downstream of DHFR, such as thymidylate synthase (TS)
0928-0987 / 00 / $ – see front matter 2000 Published by Elsevier Science B.V. All rights reserved. PII: S0928-0987( 99 )00073-1
334
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
or 5-aminoimidazole-4-carboxamide ribotide transformylase (AICAR-transformylase) are of greater importance (Baggott et al., 1993; Sandoval et al., 1995). One interesting hypothesis is that MTX exerts its anti-inflammatory effect by elevating the extracellular levels of adenosine (Cronstein et al., 1993). The line of evidence states that MTX is taken up by cells and converted to long-lived polyglutamates. AICAR-transformylase is potently inhibited by the MTX polyglutamates and subsequently the intracellular AICAR concentration will rise. High AICAR levels will inhibit adenosine degradation and hence lead to an accumulation of that mediator in the extracellular fluid (Baggott et al., 1993). Adenosine has been reported to exhibit a number of anti-inflammatory effects. Furthermore, MTX raised the adenosine and AICAR contents and inhibited the carrageenan-induced leukocyte influx in a murine air-pouch system. 3,7-Dimethyl-1-propargylxantine (DMPX), an adenosine A 2 antagonist, and a adenosine degrading enzyme reversed the MTX effect but an A 1 antagonist was ineffective (Cronstein et al., 1993). For TS, a degree of inhibition has been demonstrated in lymphocytes from MTX-treated patients. However, 5fluorouracil (5-FU), a specific inhibitor of TS is a much weaker immunosuppressant than MTX in animal models of transplantation (see: Baggott et al., 1993). Another hypothesis is that MTX interferes with polyamine synthesis (Nesher and Moore, 1990, Nesher et al., 1997). Polyamines are important in immune-mediated cellular reactions. The synthesis of polyamines is depen-
dent on availability for precursors such as methionine, which is required for the synthesis of S-adensosylmethionine (SAM). The regeneration of methionine from homocysteine is, in turn, folate dependent and will thus be inhibited by MTX (Nesher and Moore, 1990). Finally, local cytokine release in the joints is important for the propagation of the disease and it could be that MTX inhibits this production. The relevant parts of the folate metabolism are presented in Fig. 1. Mode of action studies for the anti-arthritic effect of MTX have mostly been performed in vitro or in nonarthritic in vivo models. There are only few reported attempts to confirm the findings in more complex disease models. This is somewhat surprising since there is an abundance of animal arthritis models and MTX seems to be effective in most of them: The drug ameliorates ¨ et al., collagen-induced arthritis (CIA) in rats (Ekstrom ¨ unpublished). In the same dose 1997) and mice (Ekstrom, interval (weekly dosing) as in the present study it is also reported to be efficient in rats with adjuvant arthritis (Suzuki et al., 1997), and in rabbits with antigen-induced arthritis (Noaves et al., 1996). In preliminary experiments we found that the compound also inhibited the development of antigen-induced arthritis (AIA). The general effect suggests that in these models the drug act on the same target. We wanted to investigate some of the hypotheses on the mechanism of action for MTX in an arthritis model. We chose AIA since this model has a high reproducibility and permits quantifications of the response with a continuous parameter (joint diameter).
Fig. 1. Some folate-dependent enzymes inhibited by MTX and MTX polyglutamates (-Gn). FA, folic acid; DHF, dihydrofolate; THF, tetrahydrofolate; AICAR, aminoimidazole carboxamide ribotide monophosphate; IMP, inosine monophosphate; dUMP, deoxyuridine; dTMP, deoxythymidine; 5-FU, 5-fluorouracil. Enzymes: (I) dihydrofolate reductase (EC 1.5.1.3); (II) folylpolyglutamate synthetase (III) thymidylate synthetase (EC 2.1.1.6) (IV) 10-formyl tetrahydrofolate synthetase; 5,10-methenyltetrahydrofolate cyclohydrolase; 5,10-methylene tetrahydrofolatedehydrogenase (EC 6.3.4.3; EC 3.5.4.9; EC 1.5.1.5., respectively); (V) AICAR-transformylase (EC 2.1.2.3). Modified from Baggott et al. 1993; Cronstein et al., 1993.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
2. Materials and methods
2.1. Animals Female rats from the Dark Agouti (DA) strain with an approximate body weight of 165 g were used in the study. ¨ They were obtained from Mollegaards Breeding Center, Denmark and kept in sawdust-covered cages with food and water ad libitum. The temperature was thermostatically maintained at 1228C, the relative humidity was 50%, and the light was on in 12 h periods from 6 a.m. to 6 p.m. The Animal Ethics Committee of Lund, Sweden approved the experiments. The rats were conditioned at least 1 week before immunisation.
2.2. Immunisation; induction of AIA; administration of substances The rats were sensitised intradermally at the tail root with 1 mg methylated bovine serum albumin (mBSA) dissolved in 50 ml saline and emulsified in 50 ml Freund’s complete adjuvant. Ten days later the rats were challenged with a unilateral intraarticular injection of 50 mg mBSA (1 mg / ml, dissolved in saline). Substances were administered orally (p.o.), by gavage, during a brief enflurane anaesthesia, or intraperitoneally (i.p.).
2.3. Evaluation The arthritis development was followed during 10 days after challenge by quantifying the joint swelling (knee diameters) with a caliper. The local changes in the joint have previously been extensively characterised by others and us (Verschure et al. 1989; Carpenter et al., 1994; Andersson et al., 1998a,b; 1999). There is a marked inflammatory activity with high energy metabolism, influx of inflammatory cells and initially a high vascular porosity. There is also a transient elevation in cellular proliferation rate and the formation of an inflammatory pannus tissue. Finally, there is a progressive degradation of the joint cartilage. Correlation analysis has shown that the presently used parameter, the joint swelling, is mainly a reflection of the pannus mass (Andersson et al., 1998a). The joint swelling has a low variance and is the only non-radiological method for continuous measurement of arthritis intensity in single animals. It is also used for monitoring of arthritis development in clinical practice. We thus regard it as a simple and robust way to get an overall estimate of arthritis development. Unless stated otherwise, the area under the curve (AUC) was calculated for each animal and these values were used in the statistical analysis. The most abundant adverse effects of MTX in human use are gastrointestinal disturbances such as anorexia, nausea and vomiting (Bannwarth et al., 1994). We thus tried to estimate MTX-induced side effects by measuring weight loss. However, this parameter exhibited a relatively large
335
variation, and since the arthritis in itself tends to reduce body weight the data were sometimes difficult to interpret. Statistical analysis was performed by analysis of variance followed by the Fisher protected least significant difference (PLSD). All calculations were performed on an Apple computer using the STATVIEW 4.0 (Berkeley, CA, USA) software. In all experiments each group contained six animals and each experiment contained a vehicletreated control group. Unless stated otherwise, all animals survived to the end of the experiment. In each figure, all presented data are from a separate experiment. All values are given as mean6SEM.
2.4. Experiments with cell primary cultures — cytokine production Three days after the induction of arthritis, pannus tissue was obtained from the left knee joint under sterile conditions and placed in PBS containing antibiotics. Cells were isolated from the pannus tissue, which was cut into small pieces and digested at 378C for 3 h with collagenase. The primary culture contained a mixture of cell types; the majority were fibroblasts, macrophages, dendritic cells, granulocytes and a few lymphocytes (Verschure et al., 1989). After washing, cells were seeded at a concentration of 10 6 cells / ml in 48-well plates in RPMI 1640 medium containing foetal bovine serum 10% and antibiotics. Cells were treated with MTX (final concentration 0.1, 1, 10 or 100 mM) and after 30 min incubation medium was added with (final concentration 0.1 mg / ml) or without lipopolysaccharides (LPS). After incubation at 378C for 16 h, the medium was collected and the supernatant (after centrifugation) was frozen at 2208C for later analysis of cytokines. In RA the local production of several cytokines like interleukin-1b (IL-1b), tumor necrosis factor-a (TNFa)and interleukin-6 (IL-6). Synthesis of IL-6 is generally believed to be induced by TNFa and IL-1b (Feldmann et al., 1996). In AIA joints the tissue levels of IL-1b seems not to be elevated during the first days of inflammation (Andersson et al., in press) and we have not detected any release of this cytokine in the primary cultures. In the joints there is, however, an early peak of TNFa (Andersson et al., in press). Thus, we measured the production of TNFa and IL-6. Analysis was performed using the BioSource Cytoscreen immunoassay kit (BioSource, Camarillo, CA, USA).
2.5. Substances MTX and folinic acid (Leucovorin ) were purchased from Wyeth Lerdele (Wayne, NJ, USA). mBSA, Freund’s complete adjuvant, aminopterin, 5-FU, SAM, LPS and folic acid from Sigma (St. Louis, MO, USA). R-(2)-N6(2-Phenylisopropyl)adenosine (R-PIA), xanthine amine congener (XAC), 8-p-sulphophenyltheophyllamine (8-
336
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
SPT), DMPX, 8-cyclopentyl-1,3-dipropylxanthine (CPX) from Research Biochemicals (Natick, MA, USA). Enflurane (Efrane ) from Abbot (Campoverde di Aprilia, Italy).
3. Results
3.1. Anti-arthritic effect of antifolates MTX was given orally p.o., once weekly, in the doses 2, 3 or 4 mg / kg / week. Treatment was carried out Day 7 before challenge, on the day of challenge and on Day 7 after challenge. A clear dose–response relationship was seen and a statistical significant reduction of joint swelling was reached for the two highest doses (Fig. 2). The changes in body weight were similar in all groups: from the beginning to the end of the experiment the AIA1MTX 2 mg / kg / week group lost 462 g; AIA13 mg / kg / week, 1065 g; AIA14 mg / kg / week, 862 g and AIA1vehicle, 861 g. This experiment has been repeated several times and there has been some variation in the response; the effect of the 2 mg / kg / week has sometimes, and 3 mg / kg / week dose has not always, reached statistical significance (data not shown). We have also performed experiments with daily administration of MTX, starting on the day
before challenge. It was then seen that the anti-arthritic effect of 4 mg / kg / week was comparable to that of 0.3 mg / kg / day with an about 30% reduction in AUC (data not shown). Also another anti-folate, aminopterin (4-aminofolic acid) inhibited the arthritis. This compound is structurally similar to MTX, subject to polyglutamation and was successfully used to treat RA in the 1950s (Baggott et al., 1993). In our experiments aminopterin seemed more potent than MTX; when it was given in the dose 0.3 mg / kg / day, treatment starting on the day before challenge arthritis was totally abolished on Day 5 after onset. However, also the side effects, measured as reduced body weight, were more pronounced and the experiment had to be terminated prematurely because of this.
3.2. Does folate supplementation affect the response to MTX? In several experiments we investigated if supplementation with folic acid / folinic acid could inhibit the effect of MTX. In two experiments MTX and folate was given in equal amounts (MTX–folate ratio 1:1). In the first MTX was given p.o. once weekly in the dose 4 mg / kg. In supplemented groups folic acid or folinic acid were given 0.67 mg / kg p.o. daily 6 days a week. Supplementation was
Fig. 2. Effect of MTX on joint swelling in AIA. MTX was administered p.o. once weekly in the doses 2, 3 or 4 mg / kg / week. **, P,0.01; ***, P,0.001 for AUC compared to vehicle.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
not given on the same day as MTX since, theoretically, it would be advantageous if the drug would not have to compete with the folate for transport across the cell membrane and subsequent polyglutamation. In this experiment no significant inhibition of the anti-arthritic effect was seen and there was no difference between folic acid and folinic acid in this respect (Fig. 3a). The weight reduction was mild in all groups and no attenuation of this parameter could be detected in the supplemented groups. Until the end of the experiment vehicle-treated animals lost 0.760.8 g; MTX-treated, 5.162.2; MTX1folate, 4.663.0 g; MTX1folinic acid, 5.662.3 g. In the second experiment MTX and folic acid were given daily in the dose 0.3 mg / kg p.o. Again there was no significant attenuation of the anti-arthritic effect in the supplemented group (data not shown). The folate did, however, abolish the MTX-induced reduction in body weight: vehicle-treated animals, 1.861.6 g; MTX-treated, 29.162.6 g (P,0.0001); MTX1folic acid, 3.761.1 (n.s. compared to vehicle). In the second group of experiments the MTX–folate ratio was 1:6. MTX was given 4 mg / kg once weekly. Supplementation with either folic acid or folinic acid was given with 4 mg / kg / day 6 days a week; the day without supplementation was either the same day as MTX was administered or, in separate groups, the day after. In this dosage folic acid totally inhibited both the effects of MTX on joint swelling (Fig. 3b) and weight reduction (data not shown), independent of the days of administration. For folinic acid, the same result was seen in the group when supplementation was given on the same day as MTX. In the group where folinic acid was given instead on the day after MTX there was a large variation and the AUC for joint diameter values in this group neither differed significantly from the MTX group or from the vehicle group (data not shown). We also performed one experiment with the MTX– folate ratio of 1:10. As expected, this dose of folate abolished all effects of MTX (data not shown).
3.3. Are adenosine receptor agonists anti-arthritic? If MTX is anti-arthritic secondary to adenosine release then a similar effect would be seen following administration of an adenosine receptor agonist. As mentioned, it has previously been shown that A 2 receptor agonists have anti-inflammatory properties. We now tested R-PIA, an agonist with some selectivity for the A 1 receptor. It was administered i.p., 250 mg / kg, twice daily from the onset of arthritis and this significantly reduced the joint swelling (Fig. 4).
3.4. Is the MTX-effect due to release of adenosine? Whether different adenosine antagonists could reduce the anti-arthritic effect of MTX was tested in two experiments. In the first, MTX (4 mg / kg / week) was combined
337
with the unselective (A 1 .A 2 , according to the supplier) adenosine receptor antagonist 8-SPT given i.p. in the dose 30 mg / kg / twice daily. Unexpectedly, 8-SPT markedly enhanced the effect of MTX (Fig. 5). Here two animals died on Day 7 and AUC for the whole experiment could thus not be calculated. When 8-SPT was given in the same dose to AIA animals not receiving MTX, no anti-arthritic effect was detected (data not shown). In the second experiment the same dose of MTX was combined with either the adenosine A 2 antagonist DMPX; 3 mg / kg / day p.o. or the A 1 antagonist CPX; 1.5 mg / kg / day p.o. These doses are reported to be effective in vivo (Seale et al., 1988). Similarly to 8-SPT, both these compounds potentiated the anti-arthritic effect of MTX and after 4 days no joint swelling was detected in the groups with combination treatment, the experiment was then ended (Fig. 6). Finally, the A 1 antagonist XAC was tested. XAC (1.7 mg / kg / day) was given i.p. twice daily together with MTX (2 mg / kg / week p.o.) but had no modulating effect (data not shown).
3.5. Is the MTX-effect due to inhibition of thymidylate synthase? The specific thymidylate synthase inhibitor 5-FU was given p.o. at doses 0.3, 3 or 30 mg / kg / day. The highest dose of the compound markedly reduced body weight (vehicle-treated: 23.661.3 g; 5-FU: 224.663.3 g, P, 0.0001) but despite this there was no effect on joint swelling (data not shown).
3.6. Is the MTX-effect due to inhibition of polyamine synthesis? MTX could inhibit polyamine synthesis by reducing the levels of available SAM. We therefore tested if we could inhibit the effect of MTX by supplementing with SAM (10 mg / kg / day i.p.). There was, at some days, a slight reduction of the anti-arthritic effect in SAM treated animals but the difference for AUC did not reach statistical significance (Fig. 7).
3.7. Can MTX inhibit cytokine production in synoviocytes in vitro? In the synoviocyte primary cultures there was a spontaneous production of IL-6 (concentration in medium around 1400 pg / ml), and to a lesser extent, TNFa (Fig. 8). Addition of LPS enhanced the TNFa but not the IL-6 concentration in the medium, which shows that there is no obvious correlation between the formation of the two cytokines in this system. MTX in concentrations up to 100 mM affected neither the spontaneous nor the LPS-stimulated production of either cytokine.
338
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
Fig. 3. Effect of supplementation with folic acid or folinic acid to AIA rats treated with MTX 4 mg / kg / week p.o. **, P,0.01 for AUC compared to vehicle. (a) Folic acid and folinic acid given 0.67 mg / kg / day p.o. 6 days a week. Supplementation was not given on the same day as MTX. *, P,0.05 for AUC compared to vehicle. (b) Effect of folic acid given 4 mg / kg / day p.o. 6 days a week. In one group supplementation was not given on the same day as MTX. In one group (marked with x) supplementation was not given the day after MTX.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
339
Fig. 4. Effect of R-PIA 250 mg / kg twice daily on joint swelling in AIA. ***, P,0.001 for AUC compared to vehicle.
Fig. 5. Effect of MTX (4 mg / kg / week) with or without 8-SPT (30 mg / kg / twice daily) on joint swelling in AIA. *, P,0.05 for AUC compared to vehicle.
340
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
Fig. 6. Effect of MTX (4 mg / kg / week) with or without DMPX (3 mg / kg / day) or CPX (1.5 mg / kg / day) on joint swelling in AIA. **, P,0.01 for AUC compared to vehicle.
4. Discussion
4.1. Verisimilitude of the model Hypotheses regarding the mode of action for MTX have so far been mainly derived from observations made in experiments in vitro or in non-arthritic in vivo models. We wanted to investigate if any of them could be confirmed in a more complex disease model. An essential question then is whether MTX acts on the same target in humans as in the rats. Two characteristics of the response indicate that this could the case; firstly, both in the model and in RA patients, MTX is effective after weekly dosing. This suggests that the compound is entrapped intracellularly due to polyglutamation and that the target may be inhibited by MTX–polyglutamates (Baggott et al., 1993). Secondly, the sensitivity to folate supplementation is similar in the model and in humans: in the rats supplementation with folate in a 1:1 ratio (MTX–folic acid) did not significantly reduce the anti-arthritic effect. When this ratio was increased to 1:6 or higher the effect was lost. This is in agreement with results from clinical trials where patients medicated with MTX have been supplemented with folic acid in ratios of up to 1: 2.85 without any marked inhibition of the anti-arthritic efficacy (Morgan et al., 1994). However, supplementation with higher doses seems to diminish the clinical effect (Stenger et al., 1992; Morgan et al., 1994). Some differ-
ences between rats and humans should also be kept in mind: the effective dose is about ten times higher in rats (possibly due to a higher metabolic rate), and the time for onset for the anti-arthritic effect of MTX is shorter.
4.2. Folate antagonism — DHFR The result from the folate supplementation experiments, and the anti-arthritic effect of aminopterin, suggests that the mechanism of action is inhibition of a folate dependent process. The relative insensitivity for supplementation suggests, however, that only a partial inhibition is sufficient for the clinical response and / or that MTX has a higher affinity for the target than folate. Folinic acid is fully reduced and one-carbon substituted folate and, if MTX’s main target is DHFR, it is thus for theoretical reasons a more potent antagonist to MTX since it will bypass the enzyme (Morgan et al., 1994 and references therein). In our experiments we could not, however, see any consistent difference in potency between the two analogues.
4.3. Mechanism of action — adenosine release? We then continued to study which folate dependent mechanism(s) that could be involved. According to the ‘adenosine-hypothesis’ the main effect is due to inhibition
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
341
Fig. 7. Effect of MTX (4 mg / kg / week) with or without combination with SAM (10 mg / kg / day) on joint swelling in AIA. *, P,0.05; **, P,0.01 compared to vehicle.
of AICAR-transformylase, which will cause release of adenosine with anti-arthritic properties. The A 2 antagonist DMPX inhibited the effect of MTX on carrageenan-induced leukocyte influx indicating involvement of this receptor (Cronstein et al., 1993). A 1 agonists have, however, also been reported to have anti-inflammatory prop-
erties in some models (Schrier et al., 1990; Lesch et al., 1991). Here we found that R-PIA, which has some selectivity for the A 1 receptor significantly, reduced the arthritis. It should here be kept in mind that selectivity for most compounds is a relative concept, which is only valid when the substance is given in a defined concentration,
Fig. 8. Production of TNFa in unstimulated or LPS-stimulated synoviocyte primary cultures with different concentrations of MTX added.
342
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
usually in an in vitro experiment. When the dose is increased the selectivity is usually lost. Our result thus indicate that adenosine has anti-arthritic properties. Furthermore they render it possible that this effect could be mediated by both A 1 and A 2 receptors but more detailed studies are required to confirm this. Next, we tried to establish if it was possible to inhibit the MTX-effect with adenosine antagonists, similar to what was seen in the air-pouch model. We tested this hypothesis with four compounds; the unselective antagonist 8-SPT, the A 1 antagonists CPX and XAC, and the A 2 antagonist DMPX. As stated above, we do not know the real degree of selectivity in this in vivo situation and our approach was therefore to test a range of compounds with various antagonistic properties. Contrary to what was expected, three of the compounds, (8-SPT, DMPX and CPX) markedly enhanced the anti-arthritic effect whereas one (XAC) showed no modulation. The MTX-enhancing effect is interesting but cannot, at present, be fully explained. Since 8-SPT had no effect of its own, one interpretation would be that MTX elevates the extracellular adenosine concentration but that the mediator has a proinflammatory role. This is not in accordance with the observed anti-arthritic effect of R-PIA or with other reported findings regarding A 2 agonists. This discrepancy could be an indication that some of our results were attributed to unspecific effects of the compounds, and this should be taken into account when their effect on MTX is interpreted. However, some of them, such as 8-SPT, have been used to prove other hypotheses where adenosine has been proposed to be the mediating agent such as the ischaemic preconditioning of the heart (Liu et al., 1991). Similarly, as mentioned, the effect of DMPX was one of bases for the ‘adenosine hypothesis’. The dosing of DMPX (3 mg / kg / day p.o.) in our experiment was not markedly different from what was used in the carrageenan–air pouch system (IC 50 50.2 mg / kg given locally in a 4-h experiment). We are thus not able to define the role of endogenous adenosine for the development of arthritis. It can be concluded, however, that we cannot find supportive evidence for the ‘adenosine hypothesis’ in this model. It is also possible that AICAR-transformylase is the main target for the anti-folates but that the effect is not dependent on adenosine release.
is also in line with our earlier finding that there was no reduction in cellular proliferation rate in pannus tissue from AIA rats treated with MTX (Andersson et al., 1998b).
4.5. Mechanism of action-inhibition of polyamine synthesis? Polyamines are reported to be important in cell mediated immune reactions (Nesher and Moore, 1990). It has been proposed that MTX would reduce the available amounts of SAM and thus subsequently impair polyamine synthesis. In agreement with this, it has been reported that polyamine levels are markedly decreased in peripheral blood lymphocytes from patients taking MTX (Nesher et al., 1997) and that addition of SAM in vitro can prevent this effect (Nesher et al., 1996). Supplementation with SAM did not markedly reduce the effect of MTX in our experiment; they were thus not supportive for the ‘polyamine-hypothesis’. It remains possible, however, that an inhibitory effect would have been seen after a higher dose of SAM. In the final part of the study we investigated whether local cytokine production could be reduced by MTX. It has been reported that in joints of arthritic animals, as in RA-patients, MTX reduces the levels of cytokines such as interleukin-1b (IL-1b), arachidonic acid derivatives and also leukocyte influx (Cronstein, 1996 and references therein). We could not find any indication that this effect is due to a primary inhibition of the production in the synoviocytes. The mechanism is thus more likely located upstream of this process. In summary, the present study shows that MTX has a marked anti-arthritic effect in the AIA rat. This response has several features in common with the effect in human RA. The primary target seems to be a folate dependent enzyme. We could not obtain any data, which would support that the effect of MTX was due to: adenosine release, inhibition of polyamine synthesis or inhibition of thymidylate synthase. Local cytokine production in the joint does not seem to be the primary target either. MTX’s mode of action is thus still enigmatic but our results indicate that there is the potential possibility to solve this question with further studies using a combined biochemical and in vivo pharmacological approach.
4.4. Mechanism of action-inhibition of TS? One important mechanism for the anti-proliferative effects of MTX is inhibition of TS. In the present experiments the specific TS inhibitor 5-FU did not ameliorate joint swelling despite that it was given in a dose that elicited significant side-effects. TS inhibition and subsequent effects on cell proliferation seems thus to be of minor, if any, importance for the anti-arthritic effect. This is in agreement with what is believed to be mode of action for the anti-rheumatic effect of MTX (see Introduction). It
References ¨ ¨ Andersson, S.E., Lexmuller, K., Ekstrom, G.M., 1998a. Physiological characterization of the mBSA antigen induced arthritis in the rat. I. Vascular leakiness and pannus growth. J. Rheumatol. 25, 1772–1777. ¨ ¨ Andersson, S.E., Johansson, A., Lexmuller, K., Ekstrom, G.M., 1998b. Physiological characterization of the mBSA antigen induced arthritis in the rat. II. Joint blood flow, glucose metabolism, and cell proliferation. J. Rheumatol. 25, 1778–1784. ¨ ¨ Andersson, S.E., Lexmuller, K., Alving, K., Ekstrom, G.M., 1999. Periarticular tissue levels of cytokine- and endothelin-1-like immuno-
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343 reactivity during the course of antigen-induced arthritis in the rat. Infl. Res. 48, 491–496. ´ G.S., Koopman, W.J., Baggott, J.E., Morgan, S.L., Ha, T.-S., Alarcon, Krumdieck, C.L., 1993. Antifolates in rheumatoid arthritis: a hypothetical mechanism of action. Clin. Exp. Rheumatol. 11 (suppl. 8), S101–5. Bannwarth, B., Labat, L., Moride, Y., Schaeverbeke, T., 1994. Methotrexate in rheumatoid arthritis. An update. Drugs 47, 25–50. Carpenter, T.A., Everett, J.R., Hall, L.D., Harper, G.P., Hodgson, R.J., James, M.F., Watson, P.J., 1994. High-resolution magnetic resonance imaging of arthritic pathology in the rat knee. Skeletal. Radiol. 23, 429–437. Cronstein, B.N., 1996. Molecular therapeutics. Methotrexate and its mechanism of action. Arthr. Rheum. 39, 1951–1960. Cronstein, B.N., Naime, D., Ostad, E., 1993. The anti-inflammatory mechanism of methotrexate. Increased adenosine release at inflamed sites diminishes leukocyte accumulation in an in vivo model of inflammation. J. Clin. Invest. 92, 2675–2692. ¨ Ekstrom, G., Roempke, K., Szajewski, K., Andersson, S.E., 1997. Comparison of drug effects in collagen-induced arthritis (CIA) and antigen-induced arthritis (AIA). Ann. Int. Med. 148, 89. Feldmann, M., Brennan, F.M., Maini, R.N., 1996. Role of cytokines in rheumatoid arthritis. Annu. Rev. Immunol. 14, 397–440. Galivan, J., Rehder, M.-C., Kerwar, J., 1986. Methotrexate in adjuvant arthritis. In: Chemistry and Biology of Pteridines, Walter de Gruyter, Berlin. Lesch, E.M., Ferin, M.A., Wright, C.D., Schrier, D.J., 1991. The effects of (R)-N-(1-methyl-2-phenyethyl) adenosine (L-PIA), a standard A 1 selective adenosine agonist on acute rat models of inflammation and neutrophil function. Agents Actions 34, 25–27. Liu, G.S., Thornton, J., Van Winkle, D.M., Stanley, A.W.H., Olsson, R.A., Downey, J.M., 1991. Protection against infarction afforded by preconditioning is mediated by A 1 adenosine receptors in rabbit heart. Circulation 84, 350–356. Morgan, S.L., Baggott, J.E., Vaughn, W.H., Austin, J.S., Veitch, B.S., Lee, ´ G.S., 1994. SuppleJ.Y., Koopman, W.J., Krumdieck, C.L., Alarcon, mentation with folic acid during methotrexate therapy for rheumatoid arthritis. A double-blind, placebo-controlled trial. Ann. Intern. Med. 121, 833–841.
343
Nesher, G., Moore, T.L., 1990. The in vitro effects of methotrexate on peripheral blood mononuclear cells. Modulation by methyl donors and spermidine. Arthr. Rheum. 33, 954–959. Nesher, G., Osborn, T.G., Moore, T.L., 1996. In vitro effects of methotrexate on polyamine levels in lymphocytes from rheumatoid arthritis patients. Clin. Exp. Rheumatol. 14, 395–399. Nesher, G., Osborn, T.G., Moore, T.L., 1997. Effect of treatment with methotrexate, hydroxychloroquine and prednisone on lymphocyte polyamine levels in rheumatoid arthritis: Correlation with the clinical response and rheumatoid factor synthesis. Clin. Exp. Rheumatol. 15, 343–347. Noaves, G.S., Mello, S.B.V., Laurdino, I.M.L., Cossermelli, W., 1996. Low dose methotrexate decreases intraarticular prostaglandin and interleukin 1 levels in antigen-induced arthritis in rabbits. J. Rheumatol. 23, 2092–2097. Rau, R., Schleusser, B., Herborn, G., Krager, T., 1997. Longterm treatment of destructive rheumatoid arthritis with methotrexate. J. Rheumatol. 24, 1881–1889. ´ G.S., Morgan, S.L., 1995. Adverse effects in Sandoval, D.M., Alarcon, methotrexate-treated rheumatoid arthritis patients. Br. J. Rheumatol. 34 (suppl. 2), 49–56. Schrier, D.J., Lesch, M.E., Wright, C.D., Gilbertsen, R.B., 1990. The anti-inflammatory effects of adenosine receptor agonists on the carrageenan-induced pleural inflammatory response in rats. J. Immunol. 145, 1874–1879. Seale, T.W., Abla, K.A., Shamim, M.T., Carney, J.M., Daly, J.W., 1988. 3,7-Dimethyl-1-propargylxanthine: a potent and selective in vivo antagonist of adenosine analogs. Life Sci. 43, 1671–1684. Stenger, A.A.M.E., Houtman, P.M., Bruyn, G.A.W., 1992. Does folate supplementation make sense in patients with rheumatoid arthritis treated with methotrexate? Ann. Rheum. Dis. 51, 1019–1020. Suzuki, Y., Nakagawa, M., Masuda, C., Ide, M., Uehara, R., Ischikawa, Y., Mizushima, Y., 1997. Short term low dose methotrexate ameliorates abnormal bone metabolism and bone loss in adjuvant induced arthritis. J. Rheumatol. 24, 1890–1895. Verschure, P.J., Van Noorden, C.J.F., Dijkstra, C.D., 1989. Macrophages and dendritic cells during the early stages of antigen-induced arthritis in rats: Immunohistochemical analysis of cryostat sections of whole knee joint. Scand. J. Immunol. 29, 371–381.
Anti-arthritic effect of methotrexate: is it really mediated by adenosine? ˚ ¨ ¨ Sven E. Andersson*, Lars-Hakan Johansson, Kristina Lexmuller, Gunilla M. Ekstrom Department of Immunopharmacology, Preclinical R& D, Astra Draco AB, Lund, Sweden Received 23 June 1999; received in revised form 23 September 1999; accepted 30 September 1999
Abstract The mechanism of action for the anti-arthritic effect of methotrexate (MTX) was investigated in rats with antigen-induced arthritis (AIA). Arthritis intensity was quantified as area under the curve (AUC) for the joint swelling. The response to MTX was in several respects similar to what is seen in the clinic. The drug reduced the AUC in a dose-dependent manner after oral weekly (2–4 mg / kg / week) or daily (0.3 mg / kg / day) dosing. This effect was not affected by supplementation with an equal dose of folate. The model thus seemed suitable for this type of study. Supplementation with folate in excess abolished the effect of MTX. A structurally similar antifolate, aminopterin, also reduced the arthritis. The effect thus seemed to be due to folate antagonism although a complete inhibition of dihydrofolate reductase (DHFR) might not be essential. Hence, it could be that the main target is a process downstream of DHFR. It has been proposed that inhibition of AICAR-transformylase induce the release of adenosine with anti-inflammatory properties. Here the adenosine antagonist R-PIA reduced the arthritis but when MTX was combined with adenosine antagonists no attenuation of the anti-arthritic effect was seen. On the contrary, three adenosine agonists (8-p-sulphophenyltheophyllamine 30 mg / kg i.p. twice daily; 3,7-dimethyl-1-propargylxanthine, p.o. 3 mg / kg / day and 8-cyclopentyl-1,3-dipropylxanthine, 1.5 mg / kg / day p.o.) potentiated MTX. The specific thymidylate synthase inhibitor 5-fluourouracil (0.3–3.0 mg / kg / day) had no anti-arthritic effect. Neither did our data support the hypotheses that syntheses of polyamines or cytokines were primary targets. It is thus possible that the mechanism of action is inhibition of a process downstream of DHFR but the release of adenosine seems not to be important. 2000 Published by Elsevier Science B.V. All rights reserved. Keywords: Adenosine; IL-6; Methotrexate; Polyamines; Arthritis; TNFa
1. Introduction Methotrexate (MTX) has become the most widely used second line drug for the treatment of rheumatoid arthritis (RA). It is relatively well tolerated and has a well documented clinical efficacy, although its ability to slow down radiological progression of the disease is uncertain (Bannwarth et al., 1994; Rau et al., 1997). It thus seems likely that knowledge of the mechanism of action for the drug would provide important information on the pathophysiology of the disease. This would possibly also enable the development of new anti-rheumatic drugs. At present, however, MTX’s main targets are largely unknown. MTX is a potent inhibitor of dihydrofolate reductase (DHFR) which subsequently inhibits de novo purine and
*Corresponding author. Department of Internal Medicine, Med. Mott., Lund University Hospital. S-221 85 Lund, Sweden. Tel.: 146-46-171000. E-mail address: [email protected] (S.E. Andersson)
pyrimidine synthesis. Based on these properties it was developed as a cytostatic agent. There is, however, much evidence indicating that the anti-arthritic effects of MTX are due more to an anti-inflammatory rather than antiproliferative or immunomodulatory effect (Bannwarth et al., 1994; Cronstein, 1996). Firstly, the dosing differs from that used for treatment of oncological disease in which MTX is administered in pulses of 20–250 mg / kg; for RA treatment it is administered in weekly doses of 0.1–0.3 mg / kg (Cronstein et al., 1993). Secondly, there are reports that RA patients can be supplemented with folate in doses as high as 27.5 mg / week without compromising the drugs efficacy (Morgan et al., 1994). Although there is no consensus over this view (Stenger et al., 1992) such data suggest that complete inhibition of DHFR is not essential in rheumatological use. Interestingly it has been reported that only antifolates with a high structural similarity to MTX have anti-arthritic effect (Galivan et al., 1986; Baggott et al., 1993). It has therefore been proposed that inhibition of other folate dependent enzymes, located downstream of DHFR, such as thymidylate synthase (TS)
0928-0987 / 00 / $ – see front matter 2000 Published by Elsevier Science B.V. All rights reserved. PII: S0928-0987( 99 )00073-1
334
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
or 5-aminoimidazole-4-carboxamide ribotide transformylase (AICAR-transformylase) are of greater importance (Baggott et al., 1993; Sandoval et al., 1995). One interesting hypothesis is that MTX exerts its anti-inflammatory effect by elevating the extracellular levels of adenosine (Cronstein et al., 1993). The line of evidence states that MTX is taken up by cells and converted to long-lived polyglutamates. AICAR-transformylase is potently inhibited by the MTX polyglutamates and subsequently the intracellular AICAR concentration will rise. High AICAR levels will inhibit adenosine degradation and hence lead to an accumulation of that mediator in the extracellular fluid (Baggott et al., 1993). Adenosine has been reported to exhibit a number of anti-inflammatory effects. Furthermore, MTX raised the adenosine and AICAR contents and inhibited the carrageenan-induced leukocyte influx in a murine air-pouch system. 3,7-Dimethyl-1-propargylxantine (DMPX), an adenosine A 2 antagonist, and a adenosine degrading enzyme reversed the MTX effect but an A 1 antagonist was ineffective (Cronstein et al., 1993). For TS, a degree of inhibition has been demonstrated in lymphocytes from MTX-treated patients. However, 5fluorouracil (5-FU), a specific inhibitor of TS is a much weaker immunosuppressant than MTX in animal models of transplantation (see: Baggott et al., 1993). Another hypothesis is that MTX interferes with polyamine synthesis (Nesher and Moore, 1990, Nesher et al., 1997). Polyamines are important in immune-mediated cellular reactions. The synthesis of polyamines is depen-
dent on availability for precursors such as methionine, which is required for the synthesis of S-adensosylmethionine (SAM). The regeneration of methionine from homocysteine is, in turn, folate dependent and will thus be inhibited by MTX (Nesher and Moore, 1990). Finally, local cytokine release in the joints is important for the propagation of the disease and it could be that MTX inhibits this production. The relevant parts of the folate metabolism are presented in Fig. 1. Mode of action studies for the anti-arthritic effect of MTX have mostly been performed in vitro or in nonarthritic in vivo models. There are only few reported attempts to confirm the findings in more complex disease models. This is somewhat surprising since there is an abundance of animal arthritis models and MTX seems to be effective in most of them: The drug ameliorates ¨ et al., collagen-induced arthritis (CIA) in rats (Ekstrom ¨ unpublished). In the same dose 1997) and mice (Ekstrom, interval (weekly dosing) as in the present study it is also reported to be efficient in rats with adjuvant arthritis (Suzuki et al., 1997), and in rabbits with antigen-induced arthritis (Noaves et al., 1996). In preliminary experiments we found that the compound also inhibited the development of antigen-induced arthritis (AIA). The general effect suggests that in these models the drug act on the same target. We wanted to investigate some of the hypotheses on the mechanism of action for MTX in an arthritis model. We chose AIA since this model has a high reproducibility and permits quantifications of the response with a continuous parameter (joint diameter).
Fig. 1. Some folate-dependent enzymes inhibited by MTX and MTX polyglutamates (-Gn). FA, folic acid; DHF, dihydrofolate; THF, tetrahydrofolate; AICAR, aminoimidazole carboxamide ribotide monophosphate; IMP, inosine monophosphate; dUMP, deoxyuridine; dTMP, deoxythymidine; 5-FU, 5-fluorouracil. Enzymes: (I) dihydrofolate reductase (EC 1.5.1.3); (II) folylpolyglutamate synthetase (III) thymidylate synthetase (EC 2.1.1.6) (IV) 10-formyl tetrahydrofolate synthetase; 5,10-methenyltetrahydrofolate cyclohydrolase; 5,10-methylene tetrahydrofolatedehydrogenase (EC 6.3.4.3; EC 3.5.4.9; EC 1.5.1.5., respectively); (V) AICAR-transformylase (EC 2.1.2.3). Modified from Baggott et al. 1993; Cronstein et al., 1993.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
2. Materials and methods
2.1. Animals Female rats from the Dark Agouti (DA) strain with an approximate body weight of 165 g were used in the study. ¨ They were obtained from Mollegaards Breeding Center, Denmark and kept in sawdust-covered cages with food and water ad libitum. The temperature was thermostatically maintained at 1228C, the relative humidity was 50%, and the light was on in 12 h periods from 6 a.m. to 6 p.m. The Animal Ethics Committee of Lund, Sweden approved the experiments. The rats were conditioned at least 1 week before immunisation.
2.2. Immunisation; induction of AIA; administration of substances The rats were sensitised intradermally at the tail root with 1 mg methylated bovine serum albumin (mBSA) dissolved in 50 ml saline and emulsified in 50 ml Freund’s complete adjuvant. Ten days later the rats were challenged with a unilateral intraarticular injection of 50 mg mBSA (1 mg / ml, dissolved in saline). Substances were administered orally (p.o.), by gavage, during a brief enflurane anaesthesia, or intraperitoneally (i.p.).
2.3. Evaluation The arthritis development was followed during 10 days after challenge by quantifying the joint swelling (knee diameters) with a caliper. The local changes in the joint have previously been extensively characterised by others and us (Verschure et al. 1989; Carpenter et al., 1994; Andersson et al., 1998a,b; 1999). There is a marked inflammatory activity with high energy metabolism, influx of inflammatory cells and initially a high vascular porosity. There is also a transient elevation in cellular proliferation rate and the formation of an inflammatory pannus tissue. Finally, there is a progressive degradation of the joint cartilage. Correlation analysis has shown that the presently used parameter, the joint swelling, is mainly a reflection of the pannus mass (Andersson et al., 1998a). The joint swelling has a low variance and is the only non-radiological method for continuous measurement of arthritis intensity in single animals. It is also used for monitoring of arthritis development in clinical practice. We thus regard it as a simple and robust way to get an overall estimate of arthritis development. Unless stated otherwise, the area under the curve (AUC) was calculated for each animal and these values were used in the statistical analysis. The most abundant adverse effects of MTX in human use are gastrointestinal disturbances such as anorexia, nausea and vomiting (Bannwarth et al., 1994). We thus tried to estimate MTX-induced side effects by measuring weight loss. However, this parameter exhibited a relatively large
335
variation, and since the arthritis in itself tends to reduce body weight the data were sometimes difficult to interpret. Statistical analysis was performed by analysis of variance followed by the Fisher protected least significant difference (PLSD). All calculations were performed on an Apple computer using the STATVIEW 4.0 (Berkeley, CA, USA) software. In all experiments each group contained six animals and each experiment contained a vehicletreated control group. Unless stated otherwise, all animals survived to the end of the experiment. In each figure, all presented data are from a separate experiment. All values are given as mean6SEM.
2.4. Experiments with cell primary cultures — cytokine production Three days after the induction of arthritis, pannus tissue was obtained from the left knee joint under sterile conditions and placed in PBS containing antibiotics. Cells were isolated from the pannus tissue, which was cut into small pieces and digested at 378C for 3 h with collagenase. The primary culture contained a mixture of cell types; the majority were fibroblasts, macrophages, dendritic cells, granulocytes and a few lymphocytes (Verschure et al., 1989). After washing, cells were seeded at a concentration of 10 6 cells / ml in 48-well plates in RPMI 1640 medium containing foetal bovine serum 10% and antibiotics. Cells were treated with MTX (final concentration 0.1, 1, 10 or 100 mM) and after 30 min incubation medium was added with (final concentration 0.1 mg / ml) or without lipopolysaccharides (LPS). After incubation at 378C for 16 h, the medium was collected and the supernatant (after centrifugation) was frozen at 2208C for later analysis of cytokines. In RA the local production of several cytokines like interleukin-1b (IL-1b), tumor necrosis factor-a (TNFa)and interleukin-6 (IL-6). Synthesis of IL-6 is generally believed to be induced by TNFa and IL-1b (Feldmann et al., 1996). In AIA joints the tissue levels of IL-1b seems not to be elevated during the first days of inflammation (Andersson et al., in press) and we have not detected any release of this cytokine in the primary cultures. In the joints there is, however, an early peak of TNFa (Andersson et al., in press). Thus, we measured the production of TNFa and IL-6. Analysis was performed using the BioSource Cytoscreen immunoassay kit (BioSource, Camarillo, CA, USA).
2.5. Substances MTX and folinic acid (Leucovorin ) were purchased from Wyeth Lerdele (Wayne, NJ, USA). mBSA, Freund’s complete adjuvant, aminopterin, 5-FU, SAM, LPS and folic acid from Sigma (St. Louis, MO, USA). R-(2)-N6(2-Phenylisopropyl)adenosine (R-PIA), xanthine amine congener (XAC), 8-p-sulphophenyltheophyllamine (8-
336
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
SPT), DMPX, 8-cyclopentyl-1,3-dipropylxanthine (CPX) from Research Biochemicals (Natick, MA, USA). Enflurane (Efrane ) from Abbot (Campoverde di Aprilia, Italy).
3. Results
3.1. Anti-arthritic effect of antifolates MTX was given orally p.o., once weekly, in the doses 2, 3 or 4 mg / kg / week. Treatment was carried out Day 7 before challenge, on the day of challenge and on Day 7 after challenge. A clear dose–response relationship was seen and a statistical significant reduction of joint swelling was reached for the two highest doses (Fig. 2). The changes in body weight were similar in all groups: from the beginning to the end of the experiment the AIA1MTX 2 mg / kg / week group lost 462 g; AIA13 mg / kg / week, 1065 g; AIA14 mg / kg / week, 862 g and AIA1vehicle, 861 g. This experiment has been repeated several times and there has been some variation in the response; the effect of the 2 mg / kg / week has sometimes, and 3 mg / kg / week dose has not always, reached statistical significance (data not shown). We have also performed experiments with daily administration of MTX, starting on the day
before challenge. It was then seen that the anti-arthritic effect of 4 mg / kg / week was comparable to that of 0.3 mg / kg / day with an about 30% reduction in AUC (data not shown). Also another anti-folate, aminopterin (4-aminofolic acid) inhibited the arthritis. This compound is structurally similar to MTX, subject to polyglutamation and was successfully used to treat RA in the 1950s (Baggott et al., 1993). In our experiments aminopterin seemed more potent than MTX; when it was given in the dose 0.3 mg / kg / day, treatment starting on the day before challenge arthritis was totally abolished on Day 5 after onset. However, also the side effects, measured as reduced body weight, were more pronounced and the experiment had to be terminated prematurely because of this.
3.2. Does folate supplementation affect the response to MTX? In several experiments we investigated if supplementation with folic acid / folinic acid could inhibit the effect of MTX. In two experiments MTX and folate was given in equal amounts (MTX–folate ratio 1:1). In the first MTX was given p.o. once weekly in the dose 4 mg / kg. In supplemented groups folic acid or folinic acid were given 0.67 mg / kg p.o. daily 6 days a week. Supplementation was
Fig. 2. Effect of MTX on joint swelling in AIA. MTX was administered p.o. once weekly in the doses 2, 3 or 4 mg / kg / week. **, P,0.01; ***, P,0.001 for AUC compared to vehicle.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
not given on the same day as MTX since, theoretically, it would be advantageous if the drug would not have to compete with the folate for transport across the cell membrane and subsequent polyglutamation. In this experiment no significant inhibition of the anti-arthritic effect was seen and there was no difference between folic acid and folinic acid in this respect (Fig. 3a). The weight reduction was mild in all groups and no attenuation of this parameter could be detected in the supplemented groups. Until the end of the experiment vehicle-treated animals lost 0.760.8 g; MTX-treated, 5.162.2; MTX1folate, 4.663.0 g; MTX1folinic acid, 5.662.3 g. In the second experiment MTX and folic acid were given daily in the dose 0.3 mg / kg p.o. Again there was no significant attenuation of the anti-arthritic effect in the supplemented group (data not shown). The folate did, however, abolish the MTX-induced reduction in body weight: vehicle-treated animals, 1.861.6 g; MTX-treated, 29.162.6 g (P,0.0001); MTX1folic acid, 3.761.1 (n.s. compared to vehicle). In the second group of experiments the MTX–folate ratio was 1:6. MTX was given 4 mg / kg once weekly. Supplementation with either folic acid or folinic acid was given with 4 mg / kg / day 6 days a week; the day without supplementation was either the same day as MTX was administered or, in separate groups, the day after. In this dosage folic acid totally inhibited both the effects of MTX on joint swelling (Fig. 3b) and weight reduction (data not shown), independent of the days of administration. For folinic acid, the same result was seen in the group when supplementation was given on the same day as MTX. In the group where folinic acid was given instead on the day after MTX there was a large variation and the AUC for joint diameter values in this group neither differed significantly from the MTX group or from the vehicle group (data not shown). We also performed one experiment with the MTX– folate ratio of 1:10. As expected, this dose of folate abolished all effects of MTX (data not shown).
3.3. Are adenosine receptor agonists anti-arthritic? If MTX is anti-arthritic secondary to adenosine release then a similar effect would be seen following administration of an adenosine receptor agonist. As mentioned, it has previously been shown that A 2 receptor agonists have anti-inflammatory properties. We now tested R-PIA, an agonist with some selectivity for the A 1 receptor. It was administered i.p., 250 mg / kg, twice daily from the onset of arthritis and this significantly reduced the joint swelling (Fig. 4).
3.4. Is the MTX-effect due to release of adenosine? Whether different adenosine antagonists could reduce the anti-arthritic effect of MTX was tested in two experiments. In the first, MTX (4 mg / kg / week) was combined
337
with the unselective (A 1 .A 2 , according to the supplier) adenosine receptor antagonist 8-SPT given i.p. in the dose 30 mg / kg / twice daily. Unexpectedly, 8-SPT markedly enhanced the effect of MTX (Fig. 5). Here two animals died on Day 7 and AUC for the whole experiment could thus not be calculated. When 8-SPT was given in the same dose to AIA animals not receiving MTX, no anti-arthritic effect was detected (data not shown). In the second experiment the same dose of MTX was combined with either the adenosine A 2 antagonist DMPX; 3 mg / kg / day p.o. or the A 1 antagonist CPX; 1.5 mg / kg / day p.o. These doses are reported to be effective in vivo (Seale et al., 1988). Similarly to 8-SPT, both these compounds potentiated the anti-arthritic effect of MTX and after 4 days no joint swelling was detected in the groups with combination treatment, the experiment was then ended (Fig. 6). Finally, the A 1 antagonist XAC was tested. XAC (1.7 mg / kg / day) was given i.p. twice daily together with MTX (2 mg / kg / week p.o.) but had no modulating effect (data not shown).
3.5. Is the MTX-effect due to inhibition of thymidylate synthase? The specific thymidylate synthase inhibitor 5-FU was given p.o. at doses 0.3, 3 or 30 mg / kg / day. The highest dose of the compound markedly reduced body weight (vehicle-treated: 23.661.3 g; 5-FU: 224.663.3 g, P, 0.0001) but despite this there was no effect on joint swelling (data not shown).
3.6. Is the MTX-effect due to inhibition of polyamine synthesis? MTX could inhibit polyamine synthesis by reducing the levels of available SAM. We therefore tested if we could inhibit the effect of MTX by supplementing with SAM (10 mg / kg / day i.p.). There was, at some days, a slight reduction of the anti-arthritic effect in SAM treated animals but the difference for AUC did not reach statistical significance (Fig. 7).
3.7. Can MTX inhibit cytokine production in synoviocytes in vitro? In the synoviocyte primary cultures there was a spontaneous production of IL-6 (concentration in medium around 1400 pg / ml), and to a lesser extent, TNFa (Fig. 8). Addition of LPS enhanced the TNFa but not the IL-6 concentration in the medium, which shows that there is no obvious correlation between the formation of the two cytokines in this system. MTX in concentrations up to 100 mM affected neither the spontaneous nor the LPS-stimulated production of either cytokine.
338
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
Fig. 3. Effect of supplementation with folic acid or folinic acid to AIA rats treated with MTX 4 mg / kg / week p.o. **, P,0.01 for AUC compared to vehicle. (a) Folic acid and folinic acid given 0.67 mg / kg / day p.o. 6 days a week. Supplementation was not given on the same day as MTX. *, P,0.05 for AUC compared to vehicle. (b) Effect of folic acid given 4 mg / kg / day p.o. 6 days a week. In one group supplementation was not given on the same day as MTX. In one group (marked with x) supplementation was not given the day after MTX.
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
339
Fig. 4. Effect of R-PIA 250 mg / kg twice daily on joint swelling in AIA. ***, P,0.001 for AUC compared to vehicle.
Fig. 5. Effect of MTX (4 mg / kg / week) with or without 8-SPT (30 mg / kg / twice daily) on joint swelling in AIA. *, P,0.05 for AUC compared to vehicle.
340
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
Fig. 6. Effect of MTX (4 mg / kg / week) with or without DMPX (3 mg / kg / day) or CPX (1.5 mg / kg / day) on joint swelling in AIA. **, P,0.01 for AUC compared to vehicle.
4. Discussion
4.1. Verisimilitude of the model Hypotheses regarding the mode of action for MTX have so far been mainly derived from observations made in experiments in vitro or in non-arthritic in vivo models. We wanted to investigate if any of them could be confirmed in a more complex disease model. An essential question then is whether MTX acts on the same target in humans as in the rats. Two characteristics of the response indicate that this could the case; firstly, both in the model and in RA patients, MTX is effective after weekly dosing. This suggests that the compound is entrapped intracellularly due to polyglutamation and that the target may be inhibited by MTX–polyglutamates (Baggott et al., 1993). Secondly, the sensitivity to folate supplementation is similar in the model and in humans: in the rats supplementation with folate in a 1:1 ratio (MTX–folic acid) did not significantly reduce the anti-arthritic effect. When this ratio was increased to 1:6 or higher the effect was lost. This is in agreement with results from clinical trials where patients medicated with MTX have been supplemented with folic acid in ratios of up to 1: 2.85 without any marked inhibition of the anti-arthritic efficacy (Morgan et al., 1994). However, supplementation with higher doses seems to diminish the clinical effect (Stenger et al., 1992; Morgan et al., 1994). Some differ-
ences between rats and humans should also be kept in mind: the effective dose is about ten times higher in rats (possibly due to a higher metabolic rate), and the time for onset for the anti-arthritic effect of MTX is shorter.
4.2. Folate antagonism — DHFR The result from the folate supplementation experiments, and the anti-arthritic effect of aminopterin, suggests that the mechanism of action is inhibition of a folate dependent process. The relative insensitivity for supplementation suggests, however, that only a partial inhibition is sufficient for the clinical response and / or that MTX has a higher affinity for the target than folate. Folinic acid is fully reduced and one-carbon substituted folate and, if MTX’s main target is DHFR, it is thus for theoretical reasons a more potent antagonist to MTX since it will bypass the enzyme (Morgan et al., 1994 and references therein). In our experiments we could not, however, see any consistent difference in potency between the two analogues.
4.3. Mechanism of action — adenosine release? We then continued to study which folate dependent mechanism(s) that could be involved. According to the ‘adenosine-hypothesis’ the main effect is due to inhibition
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
341
Fig. 7. Effect of MTX (4 mg / kg / week) with or without combination with SAM (10 mg / kg / day) on joint swelling in AIA. *, P,0.05; **, P,0.01 compared to vehicle.
of AICAR-transformylase, which will cause release of adenosine with anti-arthritic properties. The A 2 antagonist DMPX inhibited the effect of MTX on carrageenan-induced leukocyte influx indicating involvement of this receptor (Cronstein et al., 1993). A 1 agonists have, however, also been reported to have anti-inflammatory prop-
erties in some models (Schrier et al., 1990; Lesch et al., 1991). Here we found that R-PIA, which has some selectivity for the A 1 receptor significantly, reduced the arthritis. It should here be kept in mind that selectivity for most compounds is a relative concept, which is only valid when the substance is given in a defined concentration,
Fig. 8. Production of TNFa in unstimulated or LPS-stimulated synoviocyte primary cultures with different concentrations of MTX added.
342
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343
usually in an in vitro experiment. When the dose is increased the selectivity is usually lost. Our result thus indicate that adenosine has anti-arthritic properties. Furthermore they render it possible that this effect could be mediated by both A 1 and A 2 receptors but more detailed studies are required to confirm this. Next, we tried to establish if it was possible to inhibit the MTX-effect with adenosine antagonists, similar to what was seen in the air-pouch model. We tested this hypothesis with four compounds; the unselective antagonist 8-SPT, the A 1 antagonists CPX and XAC, and the A 2 antagonist DMPX. As stated above, we do not know the real degree of selectivity in this in vivo situation and our approach was therefore to test a range of compounds with various antagonistic properties. Contrary to what was expected, three of the compounds, (8-SPT, DMPX and CPX) markedly enhanced the anti-arthritic effect whereas one (XAC) showed no modulation. The MTX-enhancing effect is interesting but cannot, at present, be fully explained. Since 8-SPT had no effect of its own, one interpretation would be that MTX elevates the extracellular adenosine concentration but that the mediator has a proinflammatory role. This is not in accordance with the observed anti-arthritic effect of R-PIA or with other reported findings regarding A 2 agonists. This discrepancy could be an indication that some of our results were attributed to unspecific effects of the compounds, and this should be taken into account when their effect on MTX is interpreted. However, some of them, such as 8-SPT, have been used to prove other hypotheses where adenosine has been proposed to be the mediating agent such as the ischaemic preconditioning of the heart (Liu et al., 1991). Similarly, as mentioned, the effect of DMPX was one of bases for the ‘adenosine hypothesis’. The dosing of DMPX (3 mg / kg / day p.o.) in our experiment was not markedly different from what was used in the carrageenan–air pouch system (IC 50 50.2 mg / kg given locally in a 4-h experiment). We are thus not able to define the role of endogenous adenosine for the development of arthritis. It can be concluded, however, that we cannot find supportive evidence for the ‘adenosine hypothesis’ in this model. It is also possible that AICAR-transformylase is the main target for the anti-folates but that the effect is not dependent on adenosine release.
is also in line with our earlier finding that there was no reduction in cellular proliferation rate in pannus tissue from AIA rats treated with MTX (Andersson et al., 1998b).
4.5. Mechanism of action-inhibition of polyamine synthesis? Polyamines are reported to be important in cell mediated immune reactions (Nesher and Moore, 1990). It has been proposed that MTX would reduce the available amounts of SAM and thus subsequently impair polyamine synthesis. In agreement with this, it has been reported that polyamine levels are markedly decreased in peripheral blood lymphocytes from patients taking MTX (Nesher et al., 1997) and that addition of SAM in vitro can prevent this effect (Nesher et al., 1996). Supplementation with SAM did not markedly reduce the effect of MTX in our experiment; they were thus not supportive for the ‘polyamine-hypothesis’. It remains possible, however, that an inhibitory effect would have been seen after a higher dose of SAM. In the final part of the study we investigated whether local cytokine production could be reduced by MTX. It has been reported that in joints of arthritic animals, as in RA-patients, MTX reduces the levels of cytokines such as interleukin-1b (IL-1b), arachidonic acid derivatives and also leukocyte influx (Cronstein, 1996 and references therein). We could not find any indication that this effect is due to a primary inhibition of the production in the synoviocytes. The mechanism is thus more likely located upstream of this process. In summary, the present study shows that MTX has a marked anti-arthritic effect in the AIA rat. This response has several features in common with the effect in human RA. The primary target seems to be a folate dependent enzyme. We could not obtain any data, which would support that the effect of MTX was due to: adenosine release, inhibition of polyamine synthesis or inhibition of thymidylate synthase. Local cytokine production in the joint does not seem to be the primary target either. MTX’s mode of action is thus still enigmatic but our results indicate that there is the potential possibility to solve this question with further studies using a combined biochemical and in vivo pharmacological approach.
4.4. Mechanism of action-inhibition of TS? One important mechanism for the anti-proliferative effects of MTX is inhibition of TS. In the present experiments the specific TS inhibitor 5-FU did not ameliorate joint swelling despite that it was given in a dose that elicited significant side-effects. TS inhibition and subsequent effects on cell proliferation seems thus to be of minor, if any, importance for the anti-arthritic effect. This is in agreement with what is believed to be mode of action for the anti-rheumatic effect of MTX (see Introduction). It
References ¨ ¨ Andersson, S.E., Lexmuller, K., Ekstrom, G.M., 1998a. Physiological characterization of the mBSA antigen induced arthritis in the rat. I. Vascular leakiness and pannus growth. J. Rheumatol. 25, 1772–1777. ¨ ¨ Andersson, S.E., Johansson, A., Lexmuller, K., Ekstrom, G.M., 1998b. Physiological characterization of the mBSA antigen induced arthritis in the rat. II. Joint blood flow, glucose metabolism, and cell proliferation. J. Rheumatol. 25, 1778–1784. ¨ ¨ Andersson, S.E., Lexmuller, K., Alving, K., Ekstrom, G.M., 1999. Periarticular tissue levels of cytokine- and endothelin-1-like immuno-
S.E. Andersson et al. / European Journal of Pharmaceutical Sciences 9 (2000) 333 – 343 reactivity during the course of antigen-induced arthritis in the rat. Infl. Res. 48, 491–496. ´ G.S., Koopman, W.J., Baggott, J.E., Morgan, S.L., Ha, T.-S., Alarcon, Krumdieck, C.L., 1993. Antifolates in rheumatoid arthritis: a hypothetical mechanism of action. Clin. Exp. Rheumatol. 11 (suppl. 8), S101–5. Bannwarth, B., Labat, L., Moride, Y., Schaeverbeke, T., 1994. Methotrexate in rheumatoid arthritis. An update. Drugs 47, 25–50. Carpenter, T.A., Everett, J.R., Hall, L.D., Harper, G.P., Hodgson, R.J., James, M.F., Watson, P.J., 1994. High-resolution magnetic resonance imaging of arthritic pathology in the rat knee. Skeletal. Radiol. 23, 429–437. Cronstein, B.N., 1996. Molecular therapeutics. Methotrexate and its mechanism of action. Arthr. Rheum. 39, 1951–1960. Cronstein, B.N., Naime, D., Ostad, E., 1993. The anti-inflammatory mechanism of methotrexate. Increased adenosine release at inflamed sites diminishes leukocyte accumulation in an in vivo model of inflammation. J. Clin. Invest. 92, 2675–2692. ¨ Ekstrom, G., Roempke, K., Szajewski, K., Andersson, S.E., 1997. Comparison of drug effects in collagen-induced arthritis (CIA) and antigen-induced arthritis (AIA). Ann. Int. Med. 148, 89. Feldmann, M., Brennan, F.M., Maini, R.N., 1996. Role of cytokines in rheumatoid arthritis. Annu. Rev. Immunol. 14, 397–440. Galivan, J., Rehder, M.-C., Kerwar, J., 1986. Methotrexate in adjuvant arthritis. In: Chemistry and Biology of Pteridines, Walter de Gruyter, Berlin. Lesch, E.M., Ferin, M.A., Wright, C.D., Schrier, D.J., 1991. The effects of (R)-N-(1-methyl-2-phenyethyl) adenosine (L-PIA), a standard A 1 selective adenosine agonist on acute rat models of inflammation and neutrophil function. Agents Actions 34, 25–27. Liu, G.S., Thornton, J., Van Winkle, D.M., Stanley, A.W.H., Olsson, R.A., Downey, J.M., 1991. Protection against infarction afforded by preconditioning is mediated by A 1 adenosine receptors in rabbit heart. Circulation 84, 350–356. Morgan, S.L., Baggott, J.E., Vaughn, W.H., Austin, J.S., Veitch, B.S., Lee, ´ G.S., 1994. SuppleJ.Y., Koopman, W.J., Krumdieck, C.L., Alarcon, mentation with folic acid during methotrexate therapy for rheumatoid arthritis. A double-blind, placebo-controlled trial. Ann. Intern. Med. 121, 833–841.
343
Nesher, G., Moore, T.L., 1990. The in vitro effects of methotrexate on peripheral blood mononuclear cells. Modulation by methyl donors and spermidine. Arthr. Rheum. 33, 954–959. Nesher, G., Osborn, T.G., Moore, T.L., 1996. In vitro effects of methotrexate on polyamine levels in lymphocytes from rheumatoid arthritis patients. Clin. Exp. Rheumatol. 14, 395–399. Nesher, G., Osborn, T.G., Moore, T.L., 1997. Effect of treatment with methotrexate, hydroxychloroquine and prednisone on lymphocyte polyamine levels in rheumatoid arthritis: Correlation with the clinical response and rheumatoid factor synthesis. Clin. Exp. Rheumatol. 15, 343–347. Noaves, G.S., Mello, S.B.V., Laurdino, I.M.L., Cossermelli, W., 1996. Low dose methotrexate decreases intraarticular prostaglandin and interleukin 1 levels in antigen-induced arthritis in rabbits. J. Rheumatol. 23, 2092–2097. Rau, R., Schleusser, B., Herborn, G., Krager, T., 1997. Longterm treatment of destructive rheumatoid arthritis with methotrexate. J. Rheumatol. 24, 1881–1889. ´ G.S., Morgan, S.L., 1995. Adverse effects in Sandoval, D.M., Alarcon, methotrexate-treated rheumatoid arthritis patients. Br. J. Rheumatol. 34 (suppl. 2), 49–56. Schrier, D.J., Lesch, M.E., Wright, C.D., Gilbertsen, R.B., 1990. The anti-inflammatory effects of adenosine receptor agonists on the carrageenan-induced pleural inflammatory response in rats. J. Immunol. 145, 1874–1879. Seale, T.W., Abla, K.A., Shamim, M.T., Carney, J.M., Daly, J.W., 1988. 3,7-Dimethyl-1-propargylxanthine: a potent and selective in vivo antagonist of adenosine analogs. Life Sci. 43, 1671–1684. Stenger, A.A.M.E., Houtman, P.M., Bruyn, G.A.W., 1992. Does folate supplementation make sense in patients with rheumatoid arthritis treated with methotrexate? Ann. Rheum. Dis. 51, 1019–1020. Suzuki, Y., Nakagawa, M., Masuda, C., Ide, M., Uehara, R., Ischikawa, Y., Mizushima, Y., 1997. Short term low dose methotrexate ameliorates abnormal bone metabolism and bone loss in adjuvant induced arthritis. J. Rheumatol. 24, 1890–1895. Verschure, P.J., Van Noorden, C.J.F., Dijkstra, C.D., 1989. Macrophages and dendritic cells during the early stages of antigen-induced arthritis in rats: Immunohistochemical analysis of cryostat sections of whole knee joint. Scand. J. Immunol. 29, 371–381.