Single doses of p38 MAP kinase inhibitors or prednisolone affect CRP and IL-6 in patients with active Rheumatoid Arthritis (RA) ()
1. INTRODUCTION
Rheumatoid arthritis (RA) is a chronic inflammatory disease characterised by over-production of inflammatory cytokines (e.g. TNFa, IL-6, IL-1) [1] which result in increased levels of acute phase proteins (e.g. C-Reactive Protein (CRP)) in the blood [2]. CRP is a well-validated surrogate marker of disease activity and correlates with functional outcome in RA [2,3]. CRP has a half-life measured in days [4] and is not subject to circadian rhythms. These factors reduce the variability in measuring CRP and contribute to its success as a surrogate marker. Treatments that affect CRP could reasonably be expected to translate into clinical efficacy in RA [3]. In fact, CRP is reduced significantly by a single dose of infliximab in RA patients with maximum inhibition observed 3 days postdose [5]. Transcription of CRP is regulated by pro-inflammatory cytokines [6] including IL-6 which may be derived directly from the inflamed synovium [4,7] and therefore a reduction in IL-6 should precede a decrease in CRP. However, given the relatively short half-life of IL-6 [4] and the known circadian variability of this cytokine [8], IL-6 is more variable and difficult to correlate with outcome measures. Thus, observing an early effect on IL-6 and a subsequent effect on CRP in the same subjects with RA could provide confidence that a novel therapy has the potential to deliver clinical efficacy.
p38 mitogen-activated protein kinase (MAPK) is a serine/threonine-directed protein kinase which phosphorylates a number of intracellular signal transduction molecules that are involved in regulating the biosynthesis of inflammatory cytokines [9]. Consequently, inhibitors of this kinase, were expected to be effective in diseases where the pro-inflammatory mechanisms are cytokinedependent such as rheumatoid arthritis. It is now evident that clinical studies with p38 MAPK inhibitors have generally been disappointing in RA. Recently, RA data for pamapimod (Roche) [10] and VX-702 (Vertex) [11] have failed to show efficacy in RA patients. There is a growing body of evidence that some initial activity in RA patients may be observed but that for reasons unknown this effect may not be sustained [12]. Losmapimod [13] and dilmapimod (potent and selective p38 MAPK inhibitors) were safe and well-tolerated in proof of concept studies in RA but did not provide significant benefit in this disease (manuscript in preparation). Recently, however, BMS-582949 has been reported to provide efficacy in RA patients [14], and this is attributed to its apparent dual activity which inhibits p38 MAPK activity and activation [15]. Thus, there remains a need to understand this mechanism and the single-dose response studies described here add to the body of knowledge in this significant pathway.
Prior to investing in large proof of concept studies, p38 MAP kinase inhibitors have been tested in smaller proof of mechanism studies. Healthy subjects have been challenged systemically with bacterial lipopolysaccharide (LPS) and inhibition of resultant inflammatory cytokines was observed in the presence of p38 mitogen activated protein kinase (MAPK) inhibitors [16]. In addition, VX-702, another p38 MAPK inhibitor, was dosed for 5 days in patients with acute coronary syndrome undergoing percutaneous coronary intervention (PCI) and showed a detectable reduction in CRP 24 hours post-dose [17,18]. These results indicate the potential of p38 MAPK inhibitors to deliver an anti-inflammatory effect in an acute inflammatory setting. This apparent disconnect between the early experimental medicine studies and the RA PoC studies may be due to a number of factors. There may be differences in the mechanisms driving an acute inflammatory response and those driving chronic inflammation in RA [19]. Furthermore, there is a history of p38 MAPK inhibitors and a risk of hepatotoxicity which is likely to have limited administration of sufficiently high doses for extended periods [20]. There is also a recent discussion in the literature regarding transient effects of p38 MAPK inhibitors in RA which tend to wane with duration of dosing [12]. Further understanding of these observations requires ongoing investigation into the anti-inflammatory response in RA and the role of p38 MAPK.
In the early development of two p38 MAPK inhibitors (losmapimod and dilmapimod, previously known as GW856553 and SB-681323 respectively) the novel study design described here was tested in order to show antiinflammatory effects and to indicate an appropriate dose range. These studies were conducted in patients with RA in order to overcome the shortcomings of testing these molecules in an acute inflammatory model. Prednisolone, which has proven efficacy in RA was included as a positive control [21-23]. These clinical studies provide additional information about p38 MAPK inhibitors in RA and describe a new study design which may be used in the early clinical development of other anti-inflammatory mechanisms in this disease.
2. PATIENTS AND METHODS
2.1. Ethical Conduct of the Studies
These studies were reviewed and approved by national, regional, or investigational centre ethics committees or institutional review boards as required in the relevant countries.
These studies were conducted in accordance with Good Clinical Practice and all applicable regulatory requirements, and, the guiding principles of the Declaration of Helsinki. Written informed consent was obtained from each subject prior to the performance of any studyspecific procedures. Electronic case report forms (eCRFs) were provided for each subject’s data to be recorded. Study A (GSK protocol number: RA3103730) can be found on ClinicalTrials.gov under the study identifier NCT00256919 and the study B (GSK protocol number RA1104046) identifier is NCT00134693.
2.2. Patients
Male and female subjects aged ³18 years, with a diagnosis of RA according to revised 1987 American College of Rheumatology criteria were included. Subjects had serum CRP ³ 10 mg/L and ³ 3/66 swollen or ³ 3/68 tender/painful joints and were receiving stable weekly doses of methotrexate (2.5 mg to 25 mg). Other antirheumatic therapy, including disease modifying antirheumatic drugs [DMARDs], non-steroidal anti-inflammatory drugs [NSAIDs], cyclo-oxygenase-2 (COX-2) inhibitors and, in Study A prednisolone (≤10 mg/day), was permitted providing the doses had been stable for 4 weeks prior to the study. Patients receiving biological anti-rheumatic therapies were excluded. Any history of liver disease or elevated liver transaminases (>1.5 times the upper limit of normal) were also exclusion criteria.
2.3. Study Design
The two studies were multi-centre, double blind, double dummy (study B only), randomized, placebo controlled and parallel group in design.
In Study A, eligible subjects were randomized to receive single doses of losmapimod (7.5, 20, and 60 mg) or placebo. In Study B, subjects were randomized to receive single doses of dilmapimod (7.5, 15 and 25 mg), or prednisolone (10, 20 and 50 mg) or placebo. Prednisolone was over-encapsulated in order to maintain the blinding. Dissolution testing showed no significant effect of the over-encapsulation against the compendial specification for prednisolone tablets. Subjects received the study treatment on a day between the two weekly doses of methotrexate and at least 1 hour after a light breakfast and two hours before the next meal.
Screening took place within 10 days prior to the first dose and the 120-h post-dose visit was used as a followup visit.
2.4. Endpoints
The primary endpoint was the single-dose response for serum CRP at 72 h post-dose (using placebo as the zero dose). Secondary endpoints were the single-dose response for serum CRP at 24 and 48 h and IL-6 singledose response (measured at 1, 3, 24 and 72 h post-dose).
CRP was measured using a high sensitivity assay and IL-6 using an ELISA (Quest diagnostics). Other exploratory biomarkers were assayed either at Pathway Diagnostics or at GSK R&D, Stevenage as previously described [24].
Other endpoints included estimations of previously described serum biomarkers of inflammation A-SAA, BCL/CXCL13, MPIF-1, TNFSFR14, calprotectin, sIL- 6R, and TNFα [24] measured at 1, 3, 24 and 72 hrs post dose. In addition, blood cytokine messenger ribonucleic acid (mRNA) levels were measured at 45, 90 and 180 minutes after dosing. These mRNAs were previously shown to be reduced by p38 MAPK inhibition (IL-6, TNF-a, IL-1b, COX2, IL-8, MMP9, iNOS manuscript in preparation), modulated by prednisolone (FKBP5, GILZ, IFI30, IL-1R2, DDIT4, TXNIP and ZNF145 manuscript in preparation) or modulated by both (DUSP1).
GAPDH mRNA, cyclophilin mRNA and β-Actin mRNA were used in the statistical analysis to control for variability between samples caused by confounding technical factors.
2.5. Dose Rationale
Doses of 15 - 20 mg/day of oral prednisolone are administered to control a flare of RA activity but this is rapidly tapered to 5 - 10 mg/day within the first few weeks of dosing. In addition doses of 60 mg/day have been administered for short periods to RA patients with significant benefit [25]. Therefore, a dose range of 10 - 50 mg oral prednisolone was considered reasonable for this single dose study.
The IC50 for inhibition of TNFa production by losmapimod in a whole blood assay of LPS-induced TNFa production, was 25 nM (9.5 ng/ml) [13]. Dilmapimod had a similar IC50 in the same assay (not shown).
Single doses of losmapimod administered to healthy subjects resulted in 30% (7 mg), 48% (20 mg) and 59% (60 mg) weighted mean reduction in ex vivo whole blood LPS-induced TNF-a production. Single doses of dilmapimod (7.5 mg and 25 mg) were administered to patients with COPD and shown to be active in an ex vivo whole blood assay of LPS-stimulated production of TNFa (33.4% weighted mean inhibition by 7.5 mg and 40% inhibition by 25 mg) [26].
All dosing in the studies described here was carried out between two weekly doses of methotrexate. This was based on a previous study which measured CRP levels in RA patients during weekly or daily administration of methotrexate [26,27]. This indicated that CRP levels remained stable with daily dosing of methotrexate, but, in subjects given weekly doses, CRP levels decreased after dosing and increased again prior to the subsequent dose [27]. Therefore, in the present studies, it was considered advantageous to dose subjects when CRP levels were at their lowest i.e. between two weekly doses of methotrexate. Theoretically, this would increase the probability of detecting a difference between active and placebo.
2.6. Pharmacokinetics
Plasma samples collected in study A were assayed for losmapimod using a validated analytical method based on protein precipitation followed by HPLC/MS/MS analysis. This method was validated over the range 0.2 - 200 ng/ml and the lower limit of quantification was 0.2 ng/ml using a 50 ml sample of human plasma. At all validation sample concentrations examined for losmapimod, the bias was less than 15%.
Pharmacokinetics samples were not collected in study B; however the PK characteristics of dilmapimod were studied in a separate study (not shown).
2.7. Statistical Methods
It was estimated that 12 evaluable subjects in each group were required to provide 90% power to detect a dose-response decrease in log CRP based on a test of significance at the two-sided 5% level [28]. For each dose and time separately, point estimates and corresponding 95% confidence intervals were constructed and back transformed and all pairwise treatment differences between doses of p38 MAPK inhibitors and placebo presented accordingly.
Principal components analysis was used to identify trends in the mRNA transcripts. A second principle component analysis was used as a covariate in a repeated measures analysis of mRNA transcripts to adjust for any technical factors.
The maximum plasma concentrations (Cmax) of losmapimod, the first time the plasma concentration reached maximum (tmax), and the area under the plasma concentration-time profiles by time t (t hours post dose) (AUC 0 - t) were calculated using a non-compartmental analysis method. The weighted mean concentration by time t was defined as AUC 0 - t divided by t.
The effect of systemic exposure of losmapimod on IL-6 reduction was studied by exploring the relationship between IL-6 ratio to baseline at time t and the weighted mean concentration by time t using a nonlinear mixed effects model that considered both the population and subject level effects.
3. RESULTS
3.1. Patient Population
The baseline and demographic characteristics of the subjects are shown in Table 1. Study A involved 51 and Study B involved 77 RA subjects receiving methotrexate. In study A approximately 50% in each group concomitantly received low dose glucocorticoids: prednisolone, prednisone or methylprednisone (Table 1). No subject in study B was receiving glucocorticoids. Baseline rheumatological assessments and serum levels of CRP and IL-6 were consistent with active rheumatoid arthritis (Table 1).
3.2. Pharmacokinetics of Losmapimod
Overall, 37 subjects provided a total of 268 samples for pharmacokinetic analysis of losmapimod in study A. The exploratory analysis of the losmapimod plasma concentration data obtained from these patients showed approximately dose proportional increase in exposure for doses of 7.5 mg and 20 mg. There was relatively less than proportional increase in the plasma concentration following the 60 mg dose. The elimination phase appeared slightly longer following the 60 mg dosing of losmapimod compared to the other two dose levels. The geometric mean observed maximum plasma concentration (Cmax) was 42 ng/mL (95% CI: 33; 55), 75 ng/mL (95% CI: 56; 102), and 126 ng/mL (95% CI: 101; 156) for 7.5 mg, 20 mg and 60 mg group, respectively. The Cmax was reached approximately 3 hours post dose. The area under the plasma concentration-time curve (AUC) up to 24 hours post dose was 333 ng·h/mL (95% CI: 281; 393), 714 ng·h/mL (95% CI: 573; 891) and 1377 ng·h/mL (95% CI: 1164; 1630) for 7.5 mg, 20 mg and 60 mg dose groups, respectively.
The pharmacokinetic data for dilmapimod was not collected in study B, but was measured in a previous study where dilmapimod was administered at 7.5 mg/day (2.5 mg in the morning and 5 mg in the evening) for 15 days to nine patients with RA on a background of stable doses of methotrexate. In this study the geometric mean Cmax (after morning dose of 5 mg) was 33 ng/ml (95% CI: 21; 51 ng/ml) with a Tmax of 0.75 hours and geometric mean AUC (0 - 24) was 125 ng·h/ml (95% CI: 82; 188).
3.3. CRP and IL-6
Prednisolone decreased circulating levels of serum CRP and baseline-adjusted ratios to placebo (Figure 1(c) and (f)). Single doses of prednisolone (10, 20 and 50 mg) produced a dose-dependent decrease in CRP at 48 hours only although the trend did not achieve statistical significance (slope = −0.01, p = 0.070) (Figure 3(c)). The percentage decrease in CRP compared to placebo at 48 hours was 41% (95% CI: 11%; 62%) for the 10 mg dose; 33% (95% CI: 0%; 55%) for the 20 mg dose and 39% (95% CI: 10%; 59%) for the 50 mg dose.
Prednisolone decreased circulating levels of serum IL-6 (Figure 2(c)) and baseline-adjusted ratios to placebo (Figure 2(f)). There was a statistically significant single dose response relationship between prednisolone and IL-6 at 3 hours (Figure 3(f)) (largely driven by the highest dose) and 24 hours post-dose. The percentage decrease in IL-6 compared to placebo at 3 hours was 40% (95% CI: −27%; 72%) for the 10mg dose; 11% (95% CI: −78%; 56%) for the 20 mg dose and 53% (95% CI: 12%; 75%) for the 50 mg dose, the percentage decrease in IL-6 compared to placebo at 24 hours was −6% (95% CI: −164%; 57%) for the 10 mg dose; 11% (95% CI: −166%; 57%) for the 20 mg dose and 56% (95% CI: −4%; 82%) for the 50 mg dose.
Losmapimod (7.5, 20, and 60 mg) showed no effect on CRP at any time point either compared with baseline or placebo (Figures 1(a), 1(d) and 3(a)).
All doses of losmapimod produced a statistically significant reduction in serum IL-6 at 3 hours post-dose although the differences were similar across the doses (Figures 2(a) and (d)). The percentage decrease in IL-6 compared to placebo at 3 hours was 55% (95% CI: 18%; 76%) for the 7.5 mg dose; 53% (95% CI: 12%; 75%) for the 20 mg dose and 59% (95% CI: 25%; 77%) for the 60 mg dose (Figure 3(d)).
Dilmapimod had no effect on CRP (Figures 1(b) and (e)) or IL-6 (Figures 2(b) and (e) and Figure 3(e)).
3.4. Relationship between Plasma Levels of Losmapimod and Serum IL-6
Figure 4 shows the relationship of IL-6 ratio to baseline versus the weighted mean concentration of losmapimod for each dose group (Figures 4(a)-(c)) and all dose groups including placebo (Figure 4(d)). Overall,