Background

Advanced coronary artery disease (CAD) is currently a leading cause of morbidity and mortality in the western world and the most common indication for heart transplantation. Even after successful transplantation, allograft vasculopathy affects as many as 60% of cardiac grafts within one year [1] and is the principal cause of late graft loss. While their natural histories differ, CAD and allograft vasculopathy share certain features of their pathogenesis and histopathology. Prominent among these features is the recruitment and retention of peripheral blood monocytes in the vascular wall, an event that is thought to trigger the formation of vascular lesions.

Monocyte-derived macrophages play a central role in the pathogenesis of both native atherosclerosis and allograft vasculopathy [2, 3]. Macrophages are an integral cellular component of the atherosclerotic plaque where they function to sequester lipids, giving rise to "foam cells". The release of extracellular matrix-degrading proteases by these cells, combined with their pro-apoptotic effect on adjacent vascular smooth muscle cells [4], are thought to destabilise the plaque, progressively leading to rupture. Furthermore, atherosclerotic plaque macrophages promote local coagulation by releasing prothrombotic mediators such as tissue factor. Similarly, though their precise role remains poorly defined, macrophages abound in the neointimal lesions associated with allograft vasculopathy and formation of these lesions is defective in macrophage-deficient mice [3]. Moreover, treatments that have proven effective in reducing neointimal lesion formation also reduced the macrophage burden of the lesion [5, 6]. It seems likely, therefore, that macrophage turnover in the vascular wall may influence the rate of progression of both native atherosclerosis and allograft vasculopathy.

Statins are highly efficacious in controlling hyperlipidaemia and reducing the risk of acute coronary events and cardiovascular death [7]. The biological activity of statins stems from their chemical structure, which resembles that of mevalonic acid. Statins suppress de novo cholesterol biosynthesis by inhibiting HMG-CoA reductase, the rate-limiting enzyme of the mevalonate pathway [8]. Amazingly, statin therapy is well tolerated with few major adverse effects, usually attributable to metabolic interactions with other drugs [9].

Early animal studies suggested that, in addition to its anti-atherosclerotic effect, statin treatment might also attenuate the development of allograft vasculopathy [10, 11]. The first evidence for an association of post-transplant statin treatment with reduced incidence and progression of allograft vasculopathy in human cardiac allograft recipients came from a prospective study by Kobashigawa et al [12]. This finding was subsequently corroborated by others [13] and spurred interest in characterising the vasculoprotective effects of statins [14, 15].

Statins are now known to have multiple effects on native cellular components of the vascular wall as well as on monocytes / macrophages [16]. Given the involvement of macrophages in CAD and allograft vasculopathy, one plausible mechanism through which statins exert their vasculoprotective actions could be the induction of macrophage apoptosis. When grown in vitro monocytes differentiate into macrophages, a phenotypic transition heralded by down-regulation of the IL-1β response to lipopolysaccharide (LPS) [17, 18]. Using this simple model, we explored the hypothesis that mevastatin treatment arrests monocyte-to-macrophage differentiation and, instead, steers these cells into apoptosis.

Methods

All aqueous solutions were prepared using endotoxin-free water from a MilliQ Biocel purification unit (Millipore, Bedford MA) and filter-sterilised. Reagents were obtained from Sigma (St Louis MO) unless otherwise indicated.

Subjects and materials

Peripheral blood mononuclear cells (PBMCs) obtained by venipuncture from six individuals were used in this study. All subjects were healthy volunteers recruited using procedures and documentation approved by the Cambridge Local Ethics Committee.

Mevastatin was reconstituted to 4 mM in dimethyl sulfoxide (DMSO), stored at 4°C in the dark and added to cell cultures at a final concentration of 10 μM. Mevalonolactone (ICN Biochemicals, Cleveland OH), a membrane-permeable internal ester of mevalonate that is hydrolysed by cytoplasmic esterases, was reconstituted to 40 mM in DMSO, stored at 4°C in the dark and used at a final concentration of 100 μM. Lipopolysaccharide (LPS; from E coli strain 055:B5) was dissolved in phosphate-buffered saline (PBS) and used at 5 μg/mL final concentration. Recombinant human gamma-interferon (γIFN; Peprotech, London UK) was reconstituted in PBS and used at 50 ng/mL (1000 U/mL) final concentration.

PBMC purification and handling

Blood was processed immediately after phlebotomy. PBMCs were purified by centrifugation over Percoll (AP Biotech, Piscataway NJ), spin-washed once with cold modified Hanks' Balanced Salt Solution (containing 5 mM Na4EDTA, 1% BSA) and once with cold PBS at 250 × g for 5 min, resuspended in PBS and counted. The monocytic (CD33+hi) content of the PBMC population and the extent of neutrophil contamination were quantified by immunofluorescent staining of 5 × 105 cells with an anti-human CD33-FITC MAb (clone WM53; Serotec, Oxford UK) and analysis by flow cytometry. All PBMC preparations used in this study were essentially neutrophil-free (neutrophil content <0.5%).

The remaining PBMCs were resuspended in RPMI 1640 medium and recounted. Aliquots of this suspension, calculated to contain 2.5 × 105 CD33+hi cells each, were seeded in individual wells of a 24-well tissue culture plate. Each well contained 1 mL of pre-warmed complete growth medium (RPMI 1640 supplemented with 10% FCS, 2 mM L-glutamine, 50 U/mL penicillin and 50 μg/mL streptomycin). Cell culture was performed at 37°C / 5% CO2 in a humidified incubator.

Analysis of apoptotic monocytes

Apoptotic cells were identified by double supravital staining with recombinant FITC-conjugated annexin V and propidium iodide (PI) [19], using the TACS Apoptosis Detection kit (R&D Systems, Minneapolis MN) according to the manufacturer's instructions. All cultures were performed and assayed in triplicate. Flow cytometric analysis was performed immediately after supravital staining. For each sample, 2,500 events were obtained in the monocyte region and gated onto a fluorescence dot plot, where the fraction of total annexin V-positive cells was determined. Common settings for forward / side scatter, fluorescence gains, colour compensation and fluorescence threshold were used throughout the study. These were determined by assaying unstained and single-stained cells (with PI or annexin V alone).

Measurement of cytokine release

The concentrations of IL-1β and IL-1Ra in cell culture supernatants were determined by ELISA using human Cytoset™ kits (BioSource, Camarillo CA) according to the manufacturer's instructions. Cell culture supernatants were depleted of residual cells and cell debris by centrifugation, collected in microtubes and stored at -20°C prior to the assay. Each supernatant was thawed once and both cytokines were assayed in parallel. All standards, samples and controls were assayed in duplicate.

Statistical analysis

Data were compiled and analysed using MicroSoft Excel™ (MicroSoft, Redmont WA). Descriptive statistics (mean; standard deviation; standard error of mean) were calculated for each treatment group. For both the apoptosis and cytokine release studies, group means were compared using the appropriate version of Student's unpaired t-test, as determined by the F-test for equality of variances. Test results are reported as two-tailed p values, where p < 0.05 was considered statistically significant. Summary data are reported as mean +/- SEM.

Results

Mevastatin-treated differentiating human monocytes undergo delayed apoptosis

Prolonged mevastatin treatment was associated with massive apoptosis of cultured human peripheral blood monocytes, but not lymphocytes (figure 1). Incubation with mevastatin for 24 h did not compromise monocyte viability when compared to untreated or vehicle-treated cell cultures. However, a 48-h treatment with mevastatin resulted in the appearance of a large apoptotic monocyte population, as defined by light scatter criteria (granular cells of diminishing size) and annexin V-positive staining. At this time-point, 27.5 +/-1.8% of mevastatin-treated monocytes were undergoing apoptosis compared with 8.8 +/-1.6% of vehicle-treated monocytes (p = 0.001; figure 2).

Figure 1
figure 1

Mevastatin-treated differentiating monocytes are particularly susceptible to apoptosis. After 24 h of statin treatment monocytes still constituted a single, viable cell population (panel A, black arrow). By 48 h this viable monocytic population had decreased in size and a second population of smaller, annexin V-positive cells had appeared (B). By 72 h the majority of monocytes were undergoing apoptosis (C). Conversely, peripheral blood lymphocytes (white arrow) remained largely viable throughout the observation period. Results shown here are representative of 18 experiments (three per individual).

Full size image
Figure 2
figure 2

Kinetics of apoptosis in mevastatin-treated human peripheral blood monocytes. Cultures were treated with vehicle (0.25% DMSO); 10 μM mevastatin; or mevastatin plus 100 μM mevalonolactone. Error bars represent SEM. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Full size image

Monocyte viability was severely compromised after 72 h of mevastatin treatment, when 53.0 +/- 3.9% of monocytes were undergoing apoptosis compared to 12.9 +/- 1.3% of vehicle-treated monocytes (p = 0.0006; figure 2). Dose-response experiments showed that mevastatin triggered apoptosis at concentrations >1 μM (EC50 ~ 2.5 μM at 72 h; data not shown) and that this effect was already maximal at 5 μM. Consistent with 48-h time-point observations, flow cytometric analysis revealed relatively few annexin V single-positive monocytes (<18% of apoptotic population), indicating rapid progression to apoptosis. In contrast to the lymphocyte population, the size of which did not vary significantly over time, viable monocyte numbers had plummeted after 72 h of treatment with mevastatin (figure 1C).

Supplementation of mevastatin-treated cultures with mevalonolactone attenuated monocyte apoptosis by 83.1% (range 72.6–99.3%; p = 0.01) and 77.4% (range 56.4–90.9%; p = 0.01) at 48 h and 72 h, respectively (figure 2).

Mevastatin treatment arrests monocyte differentiation in vitro

Interleukin-1β release after LPS stimulation is suppressed early during monocyte-to-macrophage differentiation in vitro and in vivo. Continuous mevastatin treatment of non-stimulated human PBMCs for 48 h did not affect constitutive IL-1β output (47.7 +/- 13.6 vs 45.1 +/- 9.0 pg/mL in vehicle-treated cultures). However, when LPS was added at 24 h mevastatin-treated cells released approximately 40-fold more IL-1β compared to vehicle-treated cells (15,559.1 +/- 5,518 vs 389.2 +/- 163 pg/mL; p = 0.05), a level comparable to that obtained with freshly harvested, undifferentiated monocytes (figure 3A).

Figure 3
figure 3

Mevastatin arrests the functional maturation of cultured human peripheral blood monocytes. LPS was added to PBMC cultures either at the onset (undifferentiated control), or after 24 h, of cell culture and supernatants were harvested 24 h later in each case. Pre-treatment with γIFN was only applicable to differentiating cells. Mevastatin treatment preserved monocyte responsiveness to LPS in terms of IL-1β release (A) but suppressed IL-1Ra release (B). *, p < 0.05; **, p < 0.01 (brackets indicate groups being compared).

Full size image

Gamma-interferon is known to potentiate induction of IL-1β by LPS [20]. In this study, incubation of vehicle-treated monocytes with γIFN for 6 h prior to addition of LPS increased IL-1β output approximately 4-fold. Under the same conditions, mevastatin-treated monocytes released approximately 11-fold more IL-1β than vehicle-treated cells (18,423.4 +/- 4,152 vs 1,649.7 +/- 338.4 pg/mL; p = 0.03), although the enhancing effect of γIFN was no longer apparent.

Inclusion of mevalonolactone in the growth medium completely prevented the mevastatin-induced blockade of monocyte differentiation, both in the presence and in the absence of γIFN (p = 0.02 and 0.05, respectively; figure 3A). In fact, mevalonolactone supplementation further suppressed IL-1β release below that of vehicle-treated monocytes. When LPS was used alone, mevalonolactone-treated monocytes released 62% less IL-1β than vehicle-treated cells (147.7 +/- 38.3 vs 389.2 +/- 163.0 pg/mL; p = 0.25). When LPS was used in combination with γIFN, mevalonolactone suppressed IL-1β release by 81.5%, compared to vehicle treatment (305.2 +/- 65.3 vs 1,649.7 +/- 338.4 pg/mL; p = 0.03).

Mevastatin suppresses LPS-induced IL-1Ra release from differentiating monocytes

The interleukin-1 receptor antagonist is an important endogenous regulator of IL-1. IL-1Ra inducibility by LPS is also progressively down-regulated during monocyte-to-macrophage differentiation [21]. In contrast to our observations with IL-1β, mevastatin treatment of non-stimulated PBMCs down-regulated constitutive IL-1Ra output at 48 h by approximately 40% (p = 0.24; figure 3B); interestingly, the magnitude of this reduction is similar to the rate of monocyte apoptosis seen at that same time-point. Following LPS stimulation, mevastatin-treated monocytes again released less IL-1Ra than vehicle-treated cells (p = 0.29). This consistent observation reached statistical significance after combined γIFN / LPS stimulation, when mevastatin-treated monocytes released approximately 50% less IL-1Ra than vehicle-treated cells (3,836.2 +/- 841.6 vs 7,286.9 +/- 928.8 pg/mL; p = 0.03). Crucially, mevalonolactone apparently failed to prevent the suppression of IL-1Ra release by mevastatin (figure 3B).

Discussion

Fibroproliferative vascular disease, thought to arise as a result of endothelial stress or trauma, is characterised by a complex pathology that culminates in a spectrum of manifestations, depending on the nature of the triggering event. Thus, native atherosclerosis is thought to occur mainly as a result of chronic oxidative stress, while mechanical endothelial trauma initiates restenosis after angioplasty or stenting. Lastly, allograft vasculopathy is thought to arise chiefly as a result of endothelial dysfunction precipitated by both immune and non-immune injury. Monocyte recruitment and retention in the vascular wall is a unifying feature of these diverse outcomes and correlates strongly with the timing and severity of vascular lesion formation [22–25].

Peripheral blood monocytes represent a heterogeneous population of immature cells that differentiate upon exiting the circulation. The tissue microenvironment plays a crucial role in initiating and guiding monocyte differentiation: arrest on a solid matrix is generally sufficient to initiate this process, while the presence of particular cytokines and growth factors functions to steer it towards a variety of phenotypes [26–28]. This process can be faithfully reproduced in vitro where, in the absence of exogenous growth factors, approximately 80% of peripheral blood monocytes spontaneously develop into macrophages [29]. One of the earliest functional indices of monocyte-to-macrophage differentiation, both in vitro and in vivo, is a markedly attenuated IL-1β release in response to LPS [17, 18].

Here we have demonstrated that mevastatin, an inhibitor of HMG-CoA reductase, abolishes the profound down-regulation of IL-1β release associated with monocyte differentiation; and that this effect coincides temporally with the induction of monocyte-specific apoptosis in human PBMC cultures. While monocyte differentiation proceeds spontaneously in vitro, lymphocytes differentiate only in response to signals via the T/B cell receptor complexes, which are lacking in pure PBMC cultures from single individuals; notably, these cells remained viable throughout the observation period. Therefore, unlike previous observations with murine J744 myeloid cells [30], the susceptibility of primary monocytes to statin-induced apoptosis appears to be tightly linked to the process of cell differentiation.

The precise molecular mechanisms underlying these findings are currently under investigation; depletion of downstream products of the mevalonate pathway such as the phosphorylated geranylgeranyl and farnesyl isoprenoids, which would compromise G-protein function, is likely to be involved. Shortage of such intermediates may be exacerbated in the face of increased utilization of the mevalonate pathway that occurs during monocyte differentiation [31], explaining the failure of mevalonolactone supplementation to completely prevent apoptosis in our experimental set-up. Conversely, mevalonolactone appeared to accelerate the suppression of LPS-induced IL-1β release in differentiating monocytes, hinting that the mevalonate pathway regulates this process both dynamically and bi-directionally. Hence this metabolic pathway may represent a promising therapeutic target for diseases involving dysregulated macrophage function.

Monocyte turnover in solid tissue is determined by the balance between recruitment and clearance. Overproduction of monocyte chemoattractant protein (MCP-1) by native vascular cells and macrophages is a major mechanism promoting macrophage accumulation in atherosclerotic [32, 33] and neointimal [34, 35] lesions. Statin treatment was recently shown to down-regulate MCP-1 production in vitro and in vivo [36]; and to suppress the release of various other atherothrombotic mediators by plaque macrophages [15, 36–40]. However, in most cases, the potential contribution of cell death to these findings was not concurrently investigated. Here we show that HMG-CoA reductase inhibition suppressed the release of IL-1Ra from LPS-stimulated monocytes; but that this suppression was largely accounted for by apoptotic cell death. Unlike IL-1β, the release of which occurs via stimulation of the P2X7 receptor at the level of the cell surface [41] and is thus refractory to apoptosis, secretion of IL-1Ra follows the classical pathway, which is effectively shut down during apoptosis [42]. While suppression of endothelial MCP-1 synthesis might genuinely contribute to the benefits of statin therapy, we suggest that induction of monocyte apoptosis may underlie many downstream statin actions, including the reduction in plaque macrophage counts and inhibition of in situ MMP and TF production [37–40] (figure 4). It is noteworthy that, since osteoclasts are also developmentally derived from peripheral blood monocytes, this hypothesis fits well with the recently proposed role of statins as modulators of bone formation [43].

Figure 4
figure 4

Statin-mediated monocyte apoptosis in the context of atherosclerosis. During formation of the atherosclerotic plaque monocytes are recruited and retained in the vascular wall, where they differentiate into macrophages. Macrophages incorporate lipids giving rise to foam cells, which exhibit a dysregulated phenotype possibly due to oxidative stress. Mature macrophages and foam cells are thought to promote plaque rupture by eroding the fibromuscular cap. This is likely achieved through release of matrix-degrading proteases and also via FasL-mediated cytotoxic effects on adjacent vascular smooth muscle cells. Statins modulate both primary vascular release of MCP-1 and monocyte chemotaxis. In situ monocyte apoptosis might also contribute to their downstream effects on plaque stability.

Full size image

Interestingly, in vivo studies so far do not lend support to our hypothesis. Libby and colleagues have postulated that foam cell apoptosis might perpetuate vascular lipid accumulation, contributing to atherosclerotic plaque growth [2]. Furthermore, in a recent, very elegant study these authors concluded that apoptosis is not involved in the reduction of plaque macrophage counts by cerivastatin and, instead, attributed these effects to suppression of cell growth [38]. The data presented here clearly challenge this conclusion and we feel that rejecting apoptosis as a possible mechanism of statin vasculoprotection may be premature on several counts. Firstly, it is well known that apoptotic cells are rapidly cleared in vivo and, in their paper, Aikawa and colleagues acknowledged that very few apoptotic cells were actually observed in both the control and statin-treated groups [38]. In view of this, it is unfortunate that these authors did not assess the role of apoptosis in their in vitro studies of monocyte differentiation. Moreover, several reports have cautioned that the TUNEL assay, used by Aikawa et al, may not allow accurate quantification of apoptotic cells [44, 45]. It might be argued that cerivastatin doses investigated in the aforementioned study (0.01–0.05 μM) reflect therapeutic plasma levels of that drug, which may not be pro-apoptotic. However, in support of our own findings, a recent study by Kaneider et al [46] showed significant monocyte pro-apoptotic activity of low-dose (0.01 μM) cerivastatin, in terms of caspase-3 activation. It should be noted that cerivastatin is relatively hydrophilic and a more potent inhibitor of HMG-CoA reductase than mevastatin; thus our findings with mevastatin at relatively high doses may still be therapeutically relevant. Further in vitro studies with clinically used statins are currently in progress to address this issue.

Conclusions

HMG-CoA reductase inhibition in human peripheral blood monocytes appears to maintain them in a functionally immature state and to render them susceptible to apoptosis. We suggest that this may represent one mechanism whereby statins decrease macrophage load and the release of prothrombotic mediators in vascular lesions, potentially leading to their prevention or regression.