Published ahead of print on December 10, 2004, doi:10.1164/rccm.200406-729OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200406-729OC
A Role for Hydroxy-Methylglutaryl Coenzyme A Reductase in Pulmonary Inflammation and Host DefenseDepartment of Medicine, National Jewish Medical and Research Center; and Department of Medicine, University of Colorado Health Sciences Center, Denver, Colorado Correspondence and requests for reprints should be addressed to Michael B. Fessler, M.D., Department of Medicine, National Jewish Medical and Research Center, 1400 Jackson Street, D403 Neustadt, Denver, CO 80206. E-mail: michael.fessler{at}uchsc.edu
Rationale: A growing literature indicates that hydroxy-methylglutaryl coenzyme A reductase inhibitors (statins) modulate proinflammatory cellular signaling and functions. No studies to date, however, have addressed whether statins modulate pulmonary inflammation triggered by aerogenic stimuli or whether they affect host defense. Objectives: To test whether lovastatin modulates LPS-induced pulmonary inflammation and antibacterial host defense. Methods: To address these questions, and to confirm any effect of statins as dependent on inhibition of hydroxy-methylglutaryl coenzyme A reductase, we treated C57Bl/6 mice with three oral doses of 10 mg/kg lovastatin (or vehicle) and three intraperitoneal doses of 10 mg/kg mevalonic acid (or saline), and then exposed them to the following: (1) aerosolized LPS, (2) intratracheal keratinocyte-derived chemokine (KC), or (3) intratracheal Klebsiella pneumoniae. Measurements and main results: LPS- and KC-induced airspace neutrophils were reduced by lovastatin, an effect that was blocked by mevalonic acid cotreatment. Lovastatin was also associated with reduced parenchymal myeloperoxidase and microvascular permeability, and altered airspace and serum cytokines after LPS. Native pulmonary clearance of K. pneumoniae was inhibited by lovastatin and extrapulmonary dissemination was enhanced, both reversibly with mevalonic acid. Ex vivo studies of neutrophils isolated from lovastatin-treated mice confirmed inhibitory effects on Rac activation, actin polymerization, chemotaxis, and bacterial killing. Conclusion: Lovastatin attenuates pulmonary inflammation induced by aerosolized LPS and impairs host defense.
Key Words: acute respiratory distress syndrome • lipopolysaccharide • pneumonia • rho • statins Hydroxymethyl-glutaryl (HMG) coenzyme A (CoA) reductase inhibitors (statins) exert pleiotropic effects on cellular signaling and cellular functions involved in inflammation. Statin-sensitive signaling molecules include Rho guanosine triphosphatases (GTPases), mitogen-activated protein kinases, and Akt (1–3); statin-sensitive cellular functions include adhesion, chemotaxis, and release of superoxide anion (O2–) and cytokines (4–7). Although recent reports indicate that some statins may inhibit leukocyte function antigen-1 through direct binding (8), it is generally believed that the majority of the observed antiinflammatory effects of statins stem from their inhibiting the prenylation of signaling proteins (e.g., Rho GTPases) (9). In support of the notion that statins inhibit protein prenylation by depleting the cellular pool of isoprenoids (e.g., geranylgeranyl-pyrophosphate) downstream of mevalonic acid (the product of HMG CoA reductase), reports document that various effects of statins on cellular signaling and functions are blocked by coincubation of cells with mevalonic acid or isoprenoids (5, 10, 11). To date, most of the mechanistic reports on the antiinflammatory effects of statins have arisen from in vitro models (2, 3, 5, 10, 12). Although there are reports of statins inhibiting inflammation in in vivo model systems, including cutaneous inflammation (13), peritonitis (14), ischemia-reperfusion (15), and endotoxemic sepsis (6), only one of these (13), to our knowledge, has sought to confirm that the observed effects were caused by inhibition of the mevalonic acid pathway. Moreover, despite recent cries for a clinical trial of statins in sepsis (16), no reports, with the exception of two recent retrospective clinical studies (17, 18), have rigorously tested whether the observed antiinflammatory effects of statins might carry with them untoward influences on host defense. The present investigation tested whether lovastatin exerts effects on parallel murine models of pulmonary inflammation and infection: (1) aerosolized LPS and (2) Gram-negative bacterial pneumonia. Moreover, we isolated the primary immune cell from these animals central to the pathogenesis of both models, the neutrophil (PMN), for ex vivo mechanistic studies. This article reports that lovastatin reduces LPS-induced influx of PMNs into the airspaces by a mevalonate-dependent mechanism. Lovastatin also reduces lung parenchymal PMN burden, attenuates measures of microvascular injury, and modulates cytokine release into the airspace and serum. We trace an important element of the operative mechanism to a direct inhibitory effect of lovastatin on PMN chemotaxis, and to underlying inhibition of the Rho GTPase Rac and actin cytoskeletal remodeling in the PMN. Last, we report that lovastatin impairs native compartmentalization and clearance of murine pulmonary bacterial infection, and that this effect at least partly reflects direct inhibition of the bactericidal capacity of the murine PMN. In sum, our report documents a previously underappreciated in vivo cellular target of statins, the PMN, and carries with it implications for the use of statins in the critically ill patient. Some of the results of these studies have been previously reported in the form of an abstract (19).
Materials Endotoxin-free reagents and plastics were used in all experiments. Lovastatin, mevalonolactone, and Evan's blue were purchased from Sigma (St. Louis, MO), p21-binding domain-glutathione-S-transferase (GST) agarose and mouse anti-Rac antibody from Upstate Biotechnology (Lake Placid, NY), and rabbit anti-Cdc42 antibody from Santa Cruz Biotechnology (Santa Cruz, CA). Escherichia coli 0111:B4 LPS was purchased from List Biological Laboratories (Campbell, CA), recombinant keratinocyte-derived chemokine (KC) and macrophage inflammatory protein-2 (MIP-2) from R&D Systems (Minneapolis, MN), and Klebsiella pneumoniae 43816, serotype 2, from American Type Culture Collection (Rockville, MD).
Animals
Animal Treatments
Neutrophil Functional Assays
Statistical Analysis Additional details, including the preparation of lovastatin and mevalonic acid, BAL and serum collection, MPO assay, Evan's blue assay, quantitative culture of K. pneumoniae, and in vitro rho GTPase activation assay, are included in the online supplement.
Lovastatin Reduces LPS-induced Recruitment of PMNs into the Airspaces Although most of the literature to date describing the effects of statins on inflammation biology has arisen from in vitro systems, a small number of reports have suggested that statins attenuate acute inflammation in in vivo models (6, 13–15). Of these reports, only one (13) has used mevalonate "rescue" to confirm mechanistically that the observed in vivo antiinflammatory effects were caused by inhibition of the mevalonic acid pathway, rather than by other described pharmacologic effects of statins (8). We queried whether pretreatment with lovastatin would attenuate PMN accumulation in the airspaces in a well-characterized murine model of pulmonary inflammation, aerosolized LPS (21). To address this question, female C57Bl/6 mice aged 6 to 8 weeks were dually treated with the following: (1) 10 mg/kg lovastatin (or vehicle) orally and (2) 10 mg/kg intraperitoneal mevalonic acid (or sterile saline) 20, 12, and 0.5 hours preceding exposure to aerosolized LPS (300 µg/ml, 20 minutes), as previously described (13). Subgroups of mice were then killed 4, 8, 24, and 48 hours after LPS exposure, and BALF total leukocyte and differential cell counts were performed. As depicted in Figure 1A, lovastatin was associated with a statistically significant reduction in BAL leukocyte counts at all time points. Moreover, this effect was specific for PMN influx, because the PMN percentage of BAL leukocytes was also reduced by lovastatin (Figure 1A), whereas the total number of airspace mononuclear cells was unaffected (data not shown). Similar findings were encountered with three preexposure doses of 1.25 mg/kg lovastatin (data not shown). The effect of lovastatin on both BAL total leukocyte count and PMN differential was blocked by cotreatment with mevalonic acid (Figure 1B), confirming the mechanism to be dependent on inhibition of HMG CoA reductase. Of note, isolated mevalonic acid supplementation was not associated with a significant change in LPS-induced BAL cell counts compared with control mice (Figure 1B), suggesting that the native murine innate immune response to inhaled LPS is not limited by mevalonic acid levels. No change in baseline (i.e., nonstimulated) BAL total white blood cell (WBC) count or differential was noted with lovastatin or mevalonate treatment (data not shown).
Lovastatin Reduces LPS-induced Parenchymal PMN Burden and Microvascular Injury Airspace PMNs may represent a small minority of extravasated PMNs in the inflamed lung and do not invariably track in parallel with parenchymal PMN burden (20). By contrast, interstitial PMNs are more intimately associated with clinically relevant measures of acute lung injury, such as increased microvascular permeability and decreased lung compliance (27, 28). It has been reported that statins decrease parenchymal MPO and microvascular injury in a model of pulmonary ischemia-reperfusion (15), but no reports have examined whether statins modulate these measures in response to an aerogenic inflammatory stimulus. We thus sought to evaluate the effect of lovastatin on LPS-induced parenchymal PMN burden, as quantified by a tissue MPO assay we have previously described (20). As depicted in Figure 2A, lovastatin was associated with a statistically significant reduction in parenchymal MPO levels at both 8 and 24 hours after LPS exposure. In parallel, LPS-induced BALF protein (Figure 2B) (29) and parenchymal Evan's blue dye (Figure 2C) were both reduced with lovastatin, indicating reduced microvascular injury (30). Baseline (i.e., nonexposed) BALF protein was not altered by lovastatin treatment (data not shown).
Lovastatin Modulates LPS-induced BALF Cytokines Several reports indicate that statins modulate cytokine and chemokine production elicited by LPS and other agonists (6, 12, 31), and some suggest that statins may attenuate PMN tissue infiltration and attendant inflammation by reducing end-organ chemokine levels (13). Inhaled LPS induces the production of airspace cytokines (e.g., tumor necrosis factor-  [TNF-]) and chemokines (e.g., KC, MIP) that contribute to the pathogenesis of acute lung injury (32). Hence, we investigated the effect of lovastatin pretreatment on LPS-induced BALF cytokine and chemokine levels. As depicted in Figure 3A, lovastatin was associated with no significant change in airspace TNF- but was associated with a further increase in LPS-induced BALF MIP-2 and KC (8-hour time point). These findings suggest that the reduction in airspace PMNs associated with lovastatin is not a simple consequence of reduced end-organ chemokine levels. To address the possibility that lovastatin may independently alter chemokine levels in the serum compartment, and thus induce more complex effects on the effective alveolar-serum chemokine gradient driving PMN migration into the airspaces, we measured post-LPS serum chemokines. Lovastatin was associated with a statistically significant increase in post-LPS serum MIP-2 comparable to that observed in BALF, and with no significant change in post-LPS serum KC (Figure 3B). TNF- was not detected in any of the postinhaled LPS serum specimens (data not shown).
Lovastatin Inhibits PMN Chemotaxis Ex Vivo and In Vivo Because the reduction in airspace PMNs observed with lovastatin did not appear to be a simple consequence of decreased airspace chemoattractants, we queried whether it might reflect inhibition of PMN chemotaxis. Inhibition of in vitro leukocyte chemotaxis by statins has been reported (5); however, previous reports have neither confirmed inhibition in the intact animal nor sought to identify an underlying mechanism. To address these questions, we isolated morphologically mature bone marrow PMNs from vehicle- and lovastatin-treated mice and evaluated their chemotaxis to MIP-2 using videomicroscopy in a Zigmond chamber, as we have previously reported (21, 25). In contrast to some other assays for chemotaxis, such as the modified Boyden chamber, this method applies the McCutcheon index (26) to discriminate between the migratory behavior of cells (i.e., MI  0.6 designates chemotactic cells, MI < 0.6 designates nondirectionally migrating [NDM] cells) and their migratory velocity. As depicted in Figure 4, lovastatin was associated with a statistically significant reduction in mean and peak migratory rate of both chemotactic and NDM PMNs. However, lovastatin did not significantly alter the percentage of either chemotactic (Figure 4C) or immobile (i.e., no uropod release) PMNs (data not shown). This finding suggests that lovastatin does not prevent polarization of PMNs in a chemotactic gradient or uropod release but likely inhibits the efficiency of the cytoskeletal remodeling underlying cellular locomotion. Of note, these results are in contrast to what we have previously reported for p38 inhibition in the migrating human PMN, which is associated with a selective reduction in the percentage of chemotactic PMNs without any change in the migratory velocity of either chemotactic or NDM cells (21).
To confirm our ex vivo chemotaxis results in the intact animal in isolation from either LPS or TNF- , we tested the effect of lovastatin on an in vivo model of chemotaxis to the lung, as we have previously reported (20). The chemokine KC is a potent and selective chemoattractant for murine PMNs that triggers little of the inflammatory cascade (20). KC was administered intratracheally to mice treated with lovastatin (or vehicle) and mevalonate (or saline), and BAL leukocyte counts were performed 4 hours later. As shown in Figure 5, lovastatin was associated with a significant reduction in KC-induced BAL PMNs, and this effect was blocked by cotreatment with mevalonic acid. BAL analysis demonstrated no measureable release of MIP-2 or TNF-, and levels of KC decreased rapidly after administration (data not shown). These results suggest that our ex vivo chemotaxis results are biologically significant in the intact animal and that the lovastatin-associated reduction in LPS-induced accumulation of airspace PMNs may reflect a selective inhibition of PMN chemotaxis, independent of other factors such as pulmonary TNF- production.
Lovastatin Inhibits LPS-induced Rac Activation in the Murine PMN Chemotaxis is dependent on a complex spatiotemporal activation and polarized segregation of Rho GTPases in the migrating cell (33). Such activation is believed to permit the orchestrated remodeling of the actin cytoskeleton responsible for cellular locomotion. Among the Rho GTPases, Rac2 appears necessary for chemotaxis, because this function is markedly impaired in PMNs from Rac2-null mice (34). Statins are reported to modulate Rho GTPase activation, presumably by inhibiting the downstream prenylation of these proteins, which permits their translocation to the plasma membrane (9). To investigate whether the observed inhibitory effects on PMN chemotaxis might reflect inhibition of Rho GTPases, we isolated mature bone marrow PMNs from untreated and lovastatin-treated mice and tested them in vitro for LPS-induced Rac and Cdc42 activation using a commercially available p21-binding domain-GST agarose pull-down assay (35). Consistent with a previous report in THP-1 cells (3), we found that lovastatin treatment was associated with a blockade of LPS-induced Rac activation (Figure 6). Under identical conditions, no activation of Cdc42 was detected (data not shown).
Lovastatin Modulates Actin Polymerization in the Murine PMN Cellular locomotion relies on efficient remodeling of the actin cytoskeleton. Several reports indicate that statins inhibit actin polymerization in nonhematopoetic cells (36, 37). By contrast, p38 inhibitors, which do not retard PMN migratory rate (21), as observed for statins in the murine PMN (Figure 4), also do not inhibit actin polymerization in the PMN (22). To address the possibility that lovastatin modulates actin remodeling in the PMN, we isolated mature bone marrow PMNs from untreated and lovastatin-treated mice and quantified their relative actin polymerization in response to LPS and TNF- in vitro using previously reported techniques (22). As shown in Figure 7, lovastatin was associated with a statistically significant reduction in both LPS- and TNF-–induced actin polymerization. These studies provide a tentative mechanism for the reduced migratory rate observed for statin-treated PMNs (Figure 4).
Lovastatin Impairs Pulmonary Antibacterial Host Defense through an Effect on the PMN Although the PMN is central to the pathogenesis of acute inflammation, it also plays an integral role in host defense to acute infection. Moreover, several cytoskeletal-based PMN functions (e.g., phagocytosis, granule release, O2– generation) are central to the PMNs microbicidal capacity. There have been no clinical reports, to our knowledge, of effects of statins on infection. Nevertheless, given the observed inhibitory effects of lovastatin on recruitment of PMNs to the lung (Figures 1 and 5), and on PMN cytoskeletal remodeling (Figure 7), we queried whether they may modulate the host response to pulmonary bacterial infection. To address this possibility in a parallel fashion to our gram-negative LPS model of pulmonary inflammation, we tested the effect of lovastatin on a murine model of gram-negative bacterial pneumonia. Untreated and lovastatin-treated mice were inoculated intratracheally with 2 x 103 cfu of K. pneumoniae (24), and lung homogenate cfu were quantitated 24 and 48 hours later. As shown in Figure 8A, compared with control animals, lungs from lovastatin-treated mice had a statistically significant, approximately threefold higher cfu count of K. pneumoniae 24 hours after intratracheal inoculation, and an approximately 1.6-log higher count at 48 hours. The effect on pulmonary bacterial growth was completely reversed by cotreatment with mevalonic acid. To determine whether lovastatin has any effect on systemic bacterial dissemination, we cultured splenic homogenates 24 hours after intratracheal inoculation of K. pneumoniae. Compared with controls, significantly higher K. pneumoniae cfu were detected in the spleens of lovastatin-treated mice (Figure 8B). Near-complete reversal of this phenomenon was noted with mevalonate cotreatment, thus confirming that statins promote decompartmentalization/dissemination of bacteria from the murine lung by a mechanism that is dependent on inhibition of HMG CoA reductase.
Although these findings might conceivably reflect diminished recruitment of PMNs to the murine airspace, as observed with aerosolized LPS (Figure 1), they might also be explained by lovastatin-induced impairment of PMN microbicidal capacity. To address the latter possibility, we performed an in vitro killing assay of K. pneumoniae by PMNs. K. pneumoniae was quantitatively cultured in vitro, and titers of viable bacteria at 2, 6, and 10 hours after inoculation were compared with bacterial titers derived from cocultures of K. pneumoniae with mature bone marrow PMNs from the following: (1) vehicle-treated mice, (2) lovastatin-treated mice, and (3) lovastatin-/mevalonate-treated mice. As depicted in Figure 8C, coculture with PMNs from both untreated and lovastatin-/mevalonate-treated mice induced an approximately 50% reduction in viable bacteria, whereas no significant reduction in titer was observed in the coculture with PMNs from lovastatin-treated mice. Of note, PMNs from lovastatin-treated mice had no detectable decrease in O2– release elicited by coincubation with K. pneumoniae compared with untreated PMNs (data not shown), suggesting that impaired oxidant release does not account for the observed blockade by statins of PMN in vitro bacterial killing capacity. Moreover, as previously reported for coincubations of K. pneumoniae serotype 2 with the PMN (38), the bacteria did not induce detectable MPO (primary granule constituent) release during coculture out to 4 hours (data not shown). These findings confirm that lovastatin exerts a direct effect on the murine PMN, markedly impairing its antibacterial killing capacity by a mechanism that is dependent on inhibition of HMG CoA reductase. Moreover, they suggest that, within the context of the present model system, this phenomenon does not reflect effects on either O2– or primary granule release.
A growing number of reports indicate that statins exert assorted inhibitory effects on inflammatory cell signaling and functions in vitro (2, 3, 5, 10, 12) and in vivo (6, 13–15). It is believed that the majority of these effects stem most proximally from statin-mediated depletion of cellular isoprenoids (e.g., geranylgeranyl-pyrophosphate) and resultant inhibition of the prenylation of signaling proteins (e.g., Rho GTPases), which normally permits their association with membranes. Perhaps because of an historic focus on the effects of statins on endothelium, the literature on statin intervention in in vivo models of inflammation has, to date, largely centered on vascular insults and has excluded any investigations into effects of statins on the second of the two faces of innate immunity, host defense. The lung is a unique vascular and externally exposed organ that is subject both to PMN-mediated inflammation and PMN-opposed infection. The present study reports, for the first time, that statins attenuate acute inflammation arising from an airway-introduced stimulus, LPS. We demonstrate this effect to be dependent on statin inhibition of the mevalonic acid pathway and suggest that it reflects a direct effect of statins on PMN chemotaxis, likely arising from inhibition of the Rho GTPase Rac, and actin cytoskeletal remodeling. In contrast to p38 inhibitors, which selectively impair chemotactic directionality of PMNs without affecting migratory velocity (21), we provide evidence that statins inhibit migratory velocity without impairing chemoattractant tropism. Moreover, we report, for the first time, that statins inhibit in vitro PMN bacterial killing capacity and, in parallel, impair native compartmentalization and clearance of murine pulmonary infection. Together, these findings highlight the significance of a previously underappreciated in vivo cellular target of statins, the PMN. The observed effects of statins on both pulmonary inflammation and host defense are unified by our demonstration of the PMN as an important in vivo cellular target of statins. The direct effects of statins on the murine PMN were confirmed not only by inhibition of in vivo chemotaxis to the relatively selective PMN chemoattractant, KC (Figure 5), but by inhibition of ex vivo chemotaxis to MIP-2, as well as inhibition of Rac activation, actin remodeling, and bacterial killing capacity of isolated PMNs (Figures 4, 6, 7, and 8C). We furthermore propose that Rac inhibition (Figure 5) may be the unifying, proximal cause. In this light, PMNs from Rac2-null mice have similarly been reported to have impaired killing capacity, as well as impaired actin remodeling, chemotaxis, and O2– production (34). Although it is interesting that we did not find any effect of lovastatin on whole-bacteria–triggered O2– release by murine PMNs (data not shown), neither Rac2-null PMNs nor statin-exposed PMNs invariably demonstrate impaired O2– release to all agonists tested (7, 39). Moreover, we speculate that the effects of statins on activation of particular Rho GTPases may conceivably be cell- and/or agonist-dependent. For example, although most reports suggest that statins inhibit Rho GTPases, including Rac in THP-1 cells (3) and RhoA in endothelial cells (40), it has also been reported that statins induce Rac activation in endothelial cells (12, 41). Although the mechanism underlying impaired bacterial killing remains uncertain from the data presented, and is the subject for future studies, we speculate that it may reflect impaired cytoskeletal-based, antimicrobial PMN functions, such as phagocytosis (42) or perhaps phagosome-granule fusion (43). Granule constituents play an important role in pulmonary host defense against K. pneumoniae (44). A strength of our functional approach to studying PMN bacterial killing as an endpoint, however, is that killing is the ultimate summation of PMN antimicrobial functions. Even phagocytosis is not commensurate with killing, because intracellular Klebsiella is well described to resist killing in the PMN (45). In any event, within the context of our model system, impaired killing does not appear to reflect a statin-mediated effect on either O2– or primary granule release (data not shown). Of note, although statins have been shown to modulate multiple signaling cascades, it appears that they may share a complex and unique, if not intimate, relationship with the prototypical stimulus of innate immunity, LPS, through a common intersection at the mevalonic acid pathway. It has been reported that LPS upregulates hepatocyte HMG CoA reductase abundance and activity in mice, yet inhibits the downstream sterol branch of the mevalonic acid pathway, thereby likely increasing the cellular pool of isoprenoids (46). Furthermore, patients with the Asp299Gly polymorphism of TLR4, the LPS receptor, have an enhanced clinical response to statin therapy (47), and statins have been reported to decrease cell surface expression of the CD14 component of the LPS receptor complex (48). Finally, LPS is also reported to inhibit cellular cholesterol efflux through interactions with the liver X receptor (LXR) pathway (49), perhaps explaining how it induces foam cell formation in macrophages (50), whereas LPS itself is reported to bind to serum lipoproteins (51) and to undergo cellular uptake through the Cla-1 scavenger receptor system (52). Although the full significance of these assorted findings is unclear, they perhaps suggest that cellular inflammatory phenotype may be determined by interactions between LPS signaling and statins on a cellular isoprenoid or cholesterol "fingerprint." They also raise the intriguing possibility that at least some of the well-described cellular responses to LPS may be indirect consequences of altered cellular cholesterol or isoprenoid metabolism. Limitations of the present study must be noted. First, although the present study focuses on the PMN, we do not dispute the probability of additional, independent effects of statins on other cell types in our mouse model. PMN microvascular transmigration is a complex, multistep process that involves intimate interactions between the PMN and endothelial cell (11, 53). In this light, others have reported that statins reduce thrombin-induced endothelial barrier dysfunction (12) and endothelial expression of P-selectin (11) and intercellular adhesion molecule-1 (ICAM-1) (40). It has also been reported that statins reduce monocytic surface expression of CD11a, CD18, and VLA4 (53). Similarly, important effects of statins on the macrophage have been reported (42, 54). Although airspace cytokines/chemokines were not reduced by lovastatin in the present study (Figure 3A), we do not dispute the probability that statins may modulate the function of tissue-resident macrophages and circulating monocytes in our system. Second, although reversal of statin-mediated lung protection with mevalonic acid rescue does confirm that the operative mechanism depends on inhibition of HMG CoA reductase, it does not discriminate between effects on the three major alternate downstream forks of the mevalonic acid pathway: (1) protein geranylgeranylation, (2) protein farnesylation, and (3) cellular cholesterol synthesis. Among these three, even cellular cholesterol levels have been described to modulate signaling (55, 56). Third, although the p21-binding domain-GST agarose pull-down and the analogous rhotekin-GST pull-down have been used extensively in the statin literature (3, 12, 41), they measure GTP-Rac and GTP-Cdc42, and GTP-Rho, respectively, regardless of their subcellular localization to cytosol or membrane. Statins do not directly inhibit GTP-loading but rather are believed to inhibit membrane localization of Rho GTPases by inhibiting their prenylation. Although it is generally accepted that GTP-loaded ("activated") Rho GTPases translocate to the cell membrane (57, 58), the effect of membrane exclusion of Rho GTPases on their measureable GTP-loading, and on downstream activation of various Rho GTPase effectors, remains unclear. The complexity of the relationship between Rho GTPase prenylation and activation is suggested by the findings that prenylation not only permits membrane translocation of Rho GTPases but also their binding to Rho guanosine diphosphate (GDP) dissociation inhibitor (59). Although neither membrane translocation nor prenylation of Rac2 are necessary or sufficient for PMN nicotinamide adenine diphosphate (NADPH) oxidase activity (60, 61), prenylation of rac is nevertheless required for nucleotide exchange in a cell-free system (61). Our observation that statins inhibit GTP-loading of rac (Figure 6) is consistent with this finding. Although our data are derived from an animal model only distantly related to clinical acute lung injury, we nevertheless speculate that they may have implications for the critically ill patient. First, our data suggest that statins may hold promise in the prophylaxis of acute lung injury in patients deemed "at risk" (e.g., sepsis, massive transfusion), or perhaps even in the treatment of early acute lung injury. Although our statin-dosing protocol exceeds that used on a weight basis in patients, the current standard for clinical statin dosing was originally developed for subacute lowering of serum cholesterol over weeks, and recent reports in the cardiology literature have demonstrated improved outcomes in acute coronary syndromes with high-dose statin therapy (e.g., atorvastatin 80 mg/day). The present report and others (62) indicate that statins exert antiinflammatory effects, even in normocholesterolemic animals. Nevertheless, hypercholesterolemia has been reported to induce PMN-mediated inflammation in animal models (63, 64), and lipid-rich parenteral alimentation has been reported to aggravate clinical acute lung injury (65). It is therefore interesting to speculate whether at least some of the effects of statins on acute inflammation may be dependent on the cholesterol status of patients. Finally, the clinical significance of the impairment of antibacterial host defense observed in the present study remains unclear. To our knowledge, no reports exist of adverse effects of statins in clinical infection. Even if the phenomenon observed in our murine model does translate to patients, the possibility remains that the effect is subclinical, particularly in the antibiotic age. In this light, one retrospective clinical study has reported that statins are independently associated with a reduction in mortality in patients with bacteremia (17). A second group has reported reduced incidence of severe sepsis in statin-treated patients who develop localized infection. Because patients with documented infection are generally treated with antibiotics, we speculate that the prophylactic effects of statins on development of sepsis syndrome (6, 18) and acute lung injury may outweigh impaired PMN microbicidal capacity, and possibly even increased tendency to bacterial dissemination, at least in some clinical settings. In closing, we present the first report that statins modulate both LPS-induced pulmonary inflammation and pulmonary bacterial infection and connect these observations with direct effects of statins on the circulating PMN. Future studies need to dissect further the mechanisms of the observed statin effect, and perhaps investigate whether more specific inhibitors of the mevalonic acid pathway (e.g., geranylgeranyl transferase inhibitors) may exert more discriminatory effects on the inflammatory versus host defense effects of statins on the PMN.
The authors acknowledge the technical assistance of Michelle Palic and Travis Walker.
Supported by the American Heart Association (0275035N) and the National Institutes of Health (HL 068743, HL 67179, HL 061407-05). This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: M.B.F. has received a research grant from Pfizer, proprieter of atorvastatin. S.K.Y. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; S.J. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.G.L. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.G.A. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; J.A.N. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; G.S.W. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 9, 2004; accepted in final form December 8, 2004
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
p = not significant by Student's t test compared with non–pre-treated animals).
p < 0.01 compared with untreated cells by Student's t test). (C) The effect of lovastatin on the population composition of chemotactic versus NDM PMNs per high-powered field (hpf) is depicted as mean percentage ± SEM from four consecutive experiments. There is no significant change in the percentage of chemotactic PMNs with lovastatin treatment. Open bars = untreated; solid bars = lovastatin.
p = 0.023, compared with untreated animals by Student's t test). (C) A total of 5 x 105 cfu/ml of K. pneumoniae was cultured alone (light gray) or cocultured with mature bone marrow PMNs from mice treated with oral vehicle (medium gray), lovastatin (dark gray), and lovastatin/mevalonate (black). Bacterial titers were then quantified at the indicated time points. The data shown represents mean cfu ± SEM for three consecutive experiments (*p < 0.05; 
