Effects of p38 Mitogen-Activated Protein Kinase Inhibition on Vascular and Systemic Inflammation in Patients With Atherosclerosis

Study design

This was a phase II, randomized, double-blind, placebo-controlled study conducted at 4 sites in the United Kingdom (Cambridge University Hospitals National Health Service Foundation Trust; Bart's and The London School of Medicine and Dentistry; University of Oxford; and King's Health Partners, London). The protocol was approved by Oxfordshire Research Ethics Committee and registered with ClinicalTrials.gov (NCT00633022). The study complied with the Declaration of Helsinki and written informed consent was obtained from all participants.

Study population

Patients age 50 to 80 years with a history of atherosclerosis (clinically stable, at least 6 months after myocardial infarction, transient ischemic attack/stroke, or symptomatic peripheral vascular disease) and with a body mass index between 19 and 35 kg/m² were eligible. Patients were on stable statin therapy. Prior to enrolment, eligible patients underwent a screening PET/CT to determine whether they had sufficient vascular inflammation for study entry, defined as an average arterial FDG whole vessel tissue-to-background ratio (TBR) of ≥1.6 (11) in either of the carotids or the ascending aorta. Patients with New York Heart Association functional class II to IV heart failure, atrial fibrillation, hepatic or renal disease, poorly controlled type II diabetes, insulin-dependent diabetics, and those with chronic inflammatory conditions and malignancy were excluded.

Intervention

Patients were randomized (1:1:1) to receive oral losmapimod 7.5 mg twice daily (higher dose [HD]), losmapimod 7.5 mg once daily (lower dose [LD]), or placebo for 84 days. Losmapimod and placebo tablets were indistinguishable and study personnel and patients were blinded to treatment allocation until the trial was complete. Study medication was manufactured by GlaxoSmithKline.

Vascular and fat PET/CT imaging

Vascular PET/CT imaging (19,20) was performed at study entry (pre-dose) and repeated at day 84. Vessels were identified and sectioned into 5-mm contiguous “segments.” Regions of interest were drawn around the arterial wall (in the axial plane) for every segment of the coregistered PET/CT images. The maximum standard uptake value (SUV) of FDG in each segment was recorded and normalized to background blood FDG activity, yielding a TBR. For each patient, the artery with the highest average maximum TBR at baseline was designated the “index vessel” (either carotid artery or aorta) and was used for further analyses. The index vessels were evaluated using a “whole vessel” approach and an “active segment” approach (Fig. 1). With the whole vessel approach, all segments composing the index vessel were analyzed, regardless of whether or not active inflammation was present at baseline. In the active segment approach, noninflamed segments were excluded, similar to previous work in which the effect of simvastatin was assessed only in locations with increased FDG uptake at baseline (14,21). Prior FDG-PET imaging studies with pathological correlations demonstrate that a TBR value <1.6 is associated with <5% inflammation within the atheroma (11,22–24). Therefore, segments with TBR ≥1.6 were defined as active. Measurements of visceral and subcutaneous fat were performed as previously described (25). See the supplementary methods in the Online Appendix for additional details.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Analysis Methods Used To Quantify Vascular Inflammation in the Index Vessel

A series of contiguous segments from the index vessel are illustrated. Atherosclerotic plaques in the vessel wall are indicated in yellow. The value within each segment represents the maximum tissue-to-background ratio (TBR) for that segment. Active segments were pre-defined as segments with TBR ≥1.6.

Laboratory assessments

Blood samples were collected pre-dose on days 1, 7, 14, 28, 56, and 84 and 2 weeks post-cessation of drug for measurement of hsCRP and pre-dose on days 1, 28 or 42, and 84 for measurement of inflammatory biomarkers. All analyses were conducted centrally using standard laboratory methods.

Safety assessments

Adverse events, safety laboratory parameters, hemodynamic variables, and electrocardiograms were assessed throughout the study.

Statistical methods

A sample size of 30 patients per group provided 90% power to detect a 15% difference in change from baseline in TBR across all segments within the index vessel with a 5% level of significance (14).

The safety population included patients who received at least 1 dose of the investigational product. Change from baseline analyses for pharmacodynamics and FDG-PET/CT included patients with both baseline and post-baseline values.

We used 3 approaches to quantify vascular inflammation (Fig. 1): 1) the pre-specified whole vessel primary endpoint of change in the average maximum TBR for all segments within the index vessel (19), with a complementary post hoc analysis displaying the group distribution of TBR using a frequency histogram; 2) an analysis of change from baseline in average maximum TBR for active segments within the index vessel; and 3) an analysis of the probability of a segment being active within the index vessel.

Change from baseline in average maximum TBR was analyzed using analysis of covariance, fitting treatment as fixed effect, and including baseline value as a covariate. Point estimates and corresponding 95% confidence intervals (CI) were constructed for the relevant comparisons of interest.

TBR data were plotted to show the distribution from all segments from all index vessels within each treatment group (at pre-dose and post-dose). The Kolmogorov-Smirnov statistic was applied to these data to measure the effect of treatment on TBR distribution. The difference between losmapimod and placebo was calculated and tested using a nonparametric permutation test at the patient level. Baseline correction is not feasible with this analysis approach.

The number of active segments and the total number of segments were included in logistic regression analyses to model the probability of a segment being active. For baseline correction within each group, a model was fitted with terms for treatment and day. For placebo and baseline correction, a model was fitted with treatment term and including the baseline proportion of active segments as covariate. In the baseline correction within each group model, the generalized estimating equation method was used to adjust for the fact that multiple data points (day 1 and day 84) were from the same vessel (patient). Point estimates and corresponding 95% CI were constructed to establish the odds ratio for the relevant comparisons of interest.

Biomarker data were analyzed by analysis of covariance fitting terms for regimen, day, and interaction of day and regimen as fixed effects; patient as a random effect; and baseline biomarker at day 1 as a covariate.

Statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, North Carolina). No multiplicity adjustment was made and p values <0.05 were considered statistically significant.

Vascular PET/CT imaging

All Segments

Figure 3 illustrates the patchy nature of vascular inflammation in atherosclerosis. The magnitude of reduction in average maximum TBR of all segments within the index vessel was larger in the losmapimod groups, but there was no significant difference when compared with placebo (Table 2). There was, however, a significant leftward shift in TBR distribution from baseline to day 84 in the losmapimod groups (HD vs. placebo: p = 0.007; LD vs. placebo: p = 0.031), with no significant change in the placebo group (Fig. 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Typical FDG-PET/CT Images

(A) From left to right, sagittal CT, PET, and combined CT and PET images from a patient at pre-dose (baseline) are shown. (B) Matching images from the same patient after 84 days of treatment with losmapimod HD. The arrows highlight the descending aorta, demonstrating heterogeneous atherosclerotic fluorine ¹⁸F-fluorodeoxyglucose (FDG) uptake, which is lowered post-dose. Other abbreviations as in Figure 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Frequency Histograms of All Segments From the Index Vessel

Frequency histograms of all segments from the index vessel histograms show tissue-to-background ratio (TBR) data from all patients in the placebo group (A) at pre-dose baseline (black dashed line) and after 84 days (black solid line), in the losmapimod LD group (B) at pre-dose baseline (green dashed line) and after 84 days (green solid line), and in the losmapimod HD group (C) at pre-dose baseline (pink dashed line) and after 84 days (pink solid line).

View this table:

• View inline
• View popup

Table 2

Change from Baseline in FDG Uptake in the Index Vessel

Active Segments

When only active segments were considered, there was a statistically significant reduction in average maximum TBR in these inflamed areas for both HD and LD losmapimod compared with placebo (p = 0.0125 and p = 0.0194, respectively) (Table 2).

The odds of having an active segment on day 84 was significantly lower than on day 1 for the HD group (p < 0.001) and the LD group (p = 0.001), without significant change in the placebo group (p = 0.736) (Table 2, Fig. 5). After adjusting for baseline and correcting for placebo, the odds of having an active segment on day 84 were significantly lower in the HD group than in the placebo group (p = 0.002), but no significant difference was observed for the LD group versus the placebo group (p = 0.429). This result was maintained irrespective of the TBR cutoff used to define significant inflammation (Online Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Change in Probability of a Segment Being Active

This figure depicts the odds ratio (OR) and associated 95% confidence intervals for having an inflamed arterial segment (TBR ≥1.6) in the index vessel on day 84 versus baseline in the placebo, losmapimod LD, and losmapimod HD groups. The p values above each bar denote the comparison within each group on day 84 compared with baseline (based on logistic regression model with baseline correction within each group); p values above the horizontal parentheses denote the comparison for LD or HD groups versus placebo, after baseline adjustment (based on logistic regression model with both baseline and placebo correction). The numbers underneath the x-axis show the mean (SD) of proportion of active segments at baseline and day 84 for each group. Abbreviations as in Figures 1 and 2.

Biomarkers

There was a statistically significant decrease from baseline in average hsCRP over the 84-day treatment period in the HD group compared with the placebo group, and there was a nonsignificant trend for a decrease in the LD group (Table 3,Fig. 6). There was a rebound in hsCRP above baseline levels 2 weeks after cessation of losmapimod treatment (Fig. 6). In the HD group, compared with the placebo group, statistically significant reductions from baseline were also observed at day 84 for interleukin 8, matrix metalloproteinase 9-neutrophil gelatinase associated lipocalin dimer, and monocyte chemotactic protein-1. No reductions were observed for other biomarkers (Table 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Percentage Change From Baseline in hsCRP

Percentage change from baseline in high-sensitivity C-reactive protein (hsCRP) percentage change (95% confidence intervals) from baseline in hsCRP in the placebo group (black dashed line), losmapimod LD group (green solid line), and losmapimod HD group (pink solid line). Day 7 data includes a subset of the total population (HD: n = 15; LD: n = 13; placebo: n = 14). Abbreviations as in Figure 2.

Visceral and subcutaneous fat imaging

At baseline, there were no significant differences between groups in SUV for FDG for either subcutaneous or visceral fat (Table 4). At day 84, there was a significant reduction from baseline in the HD (but not LD) group in maximum SUV for visceral fat (p = 0.002) that remained statistically significant when compared with placebo (p = 0.018) (Table 4). There were no changes from baseline in maximum SUV for subcutaneous fat in any of the groups. There were no changes in glucose, adiponectin, or insulin levels or body mass index over the course of the study (data not shown).

Safety

Losmapimod was well-tolerated in this study. There were no clinically meaningful changes in laboratory parameters, vital signs, or electrocardiograms over time in any of the groups. Adverse events were reported by a similar proportion of patients in each group (Online Table 3).

Study limitations

A limitation of our exploratory study is that a stable patient group was enrolled with well-controlled risk factors, low levels of systemic inflammation, and background statin therapy, likely making any apparent change more difficult to determine. It is possible that the effect would have been greater in patients with higher inflammatory burden. We also excluded patients with chronic disease without active vessel inflammation in whom plaques may be particularly quiescent; therefore, the effect of losmapimod in these especially stable patients is unknown. Finally, we accept the exploratory nature of our approach including the imaging technique and analyses we used to determine vascular effect. The 10% change in FDG-PET/CT observed in a previous statin study (15) suggests that the additional changes reported herein could have clinical relevance.