Downregulation of monocytic differentiation via modulation of CD147 by 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors

CD147 is an activation induced glycoprotein that promotes the secretion and activation of matrix metalloproteinases (MMPs) and is upregulated during the differentiation of macrophages. Interestingly, some of the molecular functions of CD147 rely on its glycosylation status: the highly glycosylated forms of CD147 induce MMPs whereas the lowly glycosylated forms inhibit MMP activation. Statins are hydroxy-methylglutaryl coenzyme A reductase inhibitors that block the synthesis of mevalonate, thereby inhibiting all mevalonate-dependent pathways, including isoprenylation, N-glycosylation and cholesterol synthesis. In this study, we investigated the role of statins in the inhibition of macrophage differentiation and the associated process of MMP secretion through modulation of CD147. We observed that differentiation of the human monocytic cell line THP-1 to a macrophage phenotype led to upregulation of CD147 and CD14 and that this effect was inhibited by statins. At the molecular level, statins altered CD147 expression, structure and function by inhibiting isoprenylation and N-glycosylation. In addition, statins induced a shift of CD147 from its highly glycosylated form to its lowly glycosylated form. This shift in N-glycosylation status was accompanied by a decrease in the production and functional activity of MMP-2 and MMP-9. In conclusion, these findings describe a novel molecular mechanism of immune regulation by statins, making them interesting candidates for autoimmune disease therapy.


Cell culture
All experiments were performed in the human monocytic cell line THP-1 [8,11] (a kind gift from Dr Angelika Bierhaus, University of Heidelberg, Germany). Cells were cultured in RPMI 1640 medium supplemented with 5% FCS (Gibco BRL, Gaithersburg, MD), 1% penicillin/ streptomycin and 2% L-glutamine (Gibco) at 37˚C in a humidified atmosphere of 5% CO 2 . For induction of THP-1 differentiation, cells (2-3 x 10 6 ) were seeded in the presence of 200 nM phorbol-12-myristate-13 acetate (PMA; Merck) and incubated for 24 h [11]. After incubation, non-attached cells were removed by aspiration, and the adherent cells were washed three times with medium. Undifferentiated THP-1 cells (seeded and incubated without PMA) were used as a control.

Cytofluorometric analysis
The expression patterns of CD147 and CD14 were determined by flow cytometry. THP-1 cells (10 5 per well) were washed with PBS (PAA Laboratories, Pasching, Austria) staining buffer containing 2% FCS and incubated with their respective antibodies for 45 min. After being washed three times with staining buffer, cells were analyzed in a FACScan 1 instrument (Becton Dickinson) using FlowJo 1 software (Treestar, Ashland, OR).

Cell permeabilization assay
THP-1 cells were incubated with statins or controls and differentiated with PMA to upregulate CD147. Cells were simultaneously permeabilized and fixed in cell permeabilization buffer containing paraformaldehyde (BD Cytofix/Cytoperm 1 , Becton Dickinson). Flow cytometry was used to monitor the intracellular expression of CD147.

CD147 immunoblotting
THP-1 cells were preincubated with statins or controls and differentiated with PMA for 24 h. Cells were then lysed in 1 ml RIPA buffer supplemented with sodium orthovandate (Sigma) and a complete protease inhibitor cocktail (Roche, Indianapolis, IN) for 45 min with intermittent shaking on ice. The proteins were separated on a 10% SDS-PAGE gel and transferred to a PVDF membrane using a Transblot instrument (Biorad, Hercules, CA) at 20 V for 1 h. The immunoblot was blocked for 2 h in 5% lipid free milk and incubated with an anti-CD147 antibody (Immunotools) for 1 h, followed by a 30 min wash. The primary antibody was conjugated to goat anti-mouse secondary antibody-horseradish peroxidase (HRP; Becton Dickinson) and incubated for 1 h. Finally, the blot was developed using an ECL development system (Amersham, Buckinghamshire, UK) or an enhanced ECL system (Pierce, Rockford, IL). Images were analyzed using ImageJ software (NIH, Bethesda, MD).

Gelatin zymography
Cellular MMP production was measured in THP-1 cell supernatants. To assess MMP secretion, supernatants were collected after incubation with statins or controls, and MMP activity was determined by SDS-polyacrylamide gel zymography. Samples were centrifuged to remove cellular debris, and supernatants were collected and stored at -20˚C. Ten μl of supernatant was mixed with 10 μl SDS loading buffer (Invitrogen, Carlsbad, CA) and loaded onto a 10% polyacrylamide gel containing 0.1% gelatin (Sigma). Positive controls for MMP-2 and -9 (R&D Systems, Minneapolis, MN) were added to each gel. After electrophoresis at 125 V for 150 min, the gel was renatured in a renaturating buffer (Invitrogen) containing 25% Triton X-100 for 30 min. After equilibration in developing buffer (Invitrogen) for 30 min, fresh developing buffer was added, and the gelatin containing gel was allowed to develop overnight at 37˚C. The gelatin gels were stained with 0.5% Coomassie blue (Sigma) and destained in a buffer consisting of 10% acetic acid, 50% methanol and 40% distilled water for 30 min to visualize the zymogen bands produced by MMP digestion. An image of each gel was scanned after drying. Images were analyzed using ImageJ software (NIH).

Membrane biotinylation
To assess the proportion of CD147 molecules that reached the cell surface of THP-1 cells, 5 x 10 6 cells were seeded in a 40 ml dish, differentiated with PMA and treated with 10 μM of tunicamycin or the indicated statin. Cells were then washed three times with cold PBS and slowly agitated with 1 mg/ml biotin (Sulfo-NHS-Biotin; Pierce) for 30 min on ice. Cells were then lysed using lysis buffer containing 10% SDS, 0.4 M phenylmethylsulfonyl fluoride and 0.1 M benzamidine. Streptavidin agarose beads were washed in lysis buffer, mixed with the cell lysate and thoroughly mixed with the cell lysate. The biotinylated proteins were pulled down with beads and eluted in SDS Laemmli buffer for immunoblotting as described above.

Statistics
Where applicable, Student's T-test was performed for statistical analysis. A p-value of < 0.05 was accepted to be significant.

Results
In this study, we examined the effect of statins on the structure and function of CD147 in the monocytic cell line model THP-1, which can be differentiated in vitro by treatment with PMA [11,34]. We used two approaches to further elucidate the pathways responsible for these effects: (i) rescue experiments (add-on approach) in which the processes inhibited by statins (i.e., isoprenylation and N-linked glycosylation involving mevalonate-derivative products) were selectively restored at key steps; and (ii) inhibition experiments, where the statin-induced effects were mimicked using known inhibitors of isoprenylation and N-linked glycosylation (Fig 1).
To this end, in all experimental setups, the following scenarios were investigated: 1. non-differentiated THP-1 cells were compared to differentiated cells; 2. the effects of statins on differentiated THP-1 cells were assessed; 3. rescue experiments were performed to reverse statin effects at different key steps of the cholesterol biosynthesis pathway; 4. the effects of statins were compared to those of inhibitors of processes downstream of HMG-CoA reductase.
Statin treatment alters the morphology of PMA-differentiated THP-1 cells In culture, naïve THP-1 cells are round and non-adherent (Fig 2A). Differentiation of THP-1 cells with PMA resulted in a change in morphology, with cells becoming flat, elongated, amoeboid and adherent ( Fig 2B). In contrast, treatment with various statins (pravastatin, atorvastatin, and fluvastatin) followed by differentiation with PMA resulted in a cellular morphology similar to that of undifferentiated THP-1 cells; pravastatin treatment also resulted in the formation of bunched clusters of cells (Fig 2C-2E). Next, rescue experiments were performed to reverse the effects of statins. Treatment with mevalonate, the product of HMG-CoA reductase, reversed the effect of fluvastatin and resulted in a differentiated cellular morphology ( Fig 2F). Addition of downstream intermediates (FPP, GGPP, and dolichol) also reversed the effects of fluvastatin (although to decreasing degrees with dolichol exhibiting the weakest effects) and induced a differentiated cellular morphology (Fig 2G-2I). Tunicamycin, an inhibitor of N-glycosylation, prevented differentiation of THP-1 cells in a similar manner to statins ( Fig 2J). FTI-277 and GGTI-298, which inhibit isoprenylation, were partially able to prevent differentiation of THP-1 cells (Fig 2K and 2L).
Statin treatment inhibits expression of CD147 and CD14 on PMAdifferentiated THP-1 cells Flow cytometric analysis revealed that the surface expression of CD147 was upregulated in PMA-differentiated THP-1 cells (Fig 3A). Treatment of cells with pravastatin, atorvastatin or fluvastatin resulted in CD147 downregulation ( Fig 3B). Rescue experiments revealed that treatment with dolichol ( Fig 3C) or FPP (Fig 3D) rescued the expression of CD147, whereas GGPP treatment only partially rescued CD147 expression ( Fig 3E). In the inhibitor experiments, tunicamycin treatment induced the most potent inhibition of CD147 expression, followed by FTI-277 and GGTI-298 (Fig 3F), indicating that farnesylation may be the dominant pathway regulating the expression of CD147.
Expression of the monocytic differentiation antigen CD14 is regulated in a manner similar to that of CD147; we therefore monitored the expression of CD14 along with CD147. PMAmediated differentiation was accompanied by increased CD14 expression (Fig 4A) that was inhibited by treatment with any of the three statins ( Fig 4B). Rescue experiments confirmed that dolichol ( Fig 4C) and FPP (Fig 4D) almost completely rescued the differentiation of THP-1 cells, whereas GGPP failed to do so (Fig 4E). In the inhibition experiments, the inhibitors tested had similar effects on CD14 and CD147 expression (Fig 4F).
To assess if the downregulation of CD147 cell surface expression is caused by a decrease in overall protein levels or in membrane translocation, we next measured the expression of CD147 in permeabilized cells (which represents the total cellular level of CD147) and compared it to CD147 cell surface expression in non-permeabilized cells. These results demonstrate that in PMA-differentiated THP-1 cells, most of the CD147 molecules were expressed on the cell surface ( Fig 5A), but that fluvastatin induced intracellular retention of CD147 ( Fig  5B) that was reversed upon FPP treatment (Fig 5C). This effect of fluvastatin was mimicked by tunicamycin treatment (Fig 5D).

Statins induce lowly glycosylated forms of CD147
CD147 exists in two different glycosylated forms: a highly glycosylated, higher molecular weight (HG) form and a lowly glycosylated, lower molecular weight (LG) form. HG CD147 forms homo-oligomers and induces MMP secretion and activation, whereas LG CD147 inhibits MMP secretion. Experimentally induced changes in CD147 N-glycosylation are illustrated in Fig 6A and 6B. Treatment of THP-1 cells with fluvastatin induced LG CD147 expression in a dose-dependent manner. Rescue with mevalonate, squalene, or cholesterol was able to reduce the fluvastatin-induced increase in expression of LG CD147. Inhibition of isoprenylation by FTI-277, GGTI-298 or FPT1 (a combined inhibitor of farnesylation and geranylgeranylation) increased LG CD147. Consistently, treatment with AP-9, a specific peptide antagonist of CD147, increased LG CD147. Treatment of THP-1 cells with tunicamycin affected the levels of both LG and HG CD147: the expression of LG CD147 (33 kDa) and HG CD147 (51 kDa) was greatly diminished. The molecular weight of the form that appeared upon tunicamycin treatment (27 kDa) is consistent with that of the non-glycosylated core protein. The small amount of 51 kDa HG CD147 remaining was likely synthesized before tunicamycin treatment (Fig 6A).

Statins impair translocation of CD147 to the cell surface
To confirm the results from the flow cytometry experiments comparing permeabilized and non-permeabilized THP-1 cells and to quantify intracellular retention, surface biotinylation was performed to measure the expression of CD147 on the cell surface. Cell surface levels of HG-CD147 were markedly downregulated after treatment with any of the three statins. The control substance tunicamycin decreased the translocation of CD147 to the cell surface ( Fig  7A and 7C) to a greater degree than the statins. Consistently, the biotinylated (cell surface) form of the core protein could not be detected by western blot; both HG-CD147 and LG-CD147 were detected in the cell lysate (Fig 7B and 7C). Statins downregulate matrix metalloproteinase activity MMP-9 exists in several forms: a pro-form (92 kDa), an active form (82 kDa), a heterodimer (135 kDa) and a homodimer (260 kDa). Gelatin zymography is able to measure the activities of all of these forms of MMP-9. PMA-induced differentiation of THP-1 cells promoted the activation of MMP-9, as indicated by the emergence of the active form of MMP-9. Treatment with fluvastatin resulted in a decrease in the levels of active MMP-9, whereas pravastatin and atorvastatin failed to do so. To rescue the effects of fluvastatin, we added mevalonate, FPP, GGPP, or dolichol. Both mevalonate and dolichol were able to increase the levels of activated MMP-9, whereas FPP and GGPP were not. Inhibition experiments revealed that tunicamycin, FTI-277 and GGTI-298 inhibited the expression of active MMP-9 (Fig 8A and 8C).
MMP-2 exists in two forms, a pro-form (72 kDa) and an active form (62 kDa). PMA differentiation upregulated the levels of active MMP-2 in THP-1 cells. Fluvastatin reduced active MMP-2 levels, whereas pravastatin and atorvastatin did not. Dolichol and mevalonate treatment rescued active MMP-2. Tunicamycin and GGTI-298 treatment mimicked the effects of fluvastatin (Fig 8B and 8C).

Discussion
In this study we used PMA-differentiated THP-1 cells as an in vitro model of monocytic differentiation [8,11,36]. We are aware that the THP-1 model is artificial and does not necessarily reflect the behavior of natural monocytes and macrophages in human diseases which limits Statins downregulate CD147 in monocytes the interpretability of the data. Another limitation of this study is that-as often seen in in vitro models-the concentrations of compounds are higher than in vivo. Concentrations of the statins used in our model exceed maximum concentrations (c max) of respective statins after oral Statins downregulate CD147 in monocytes intake of highest approved dose (80 mg per day) up to 100-fold. Another limitation of our study is its observational character.
Differentiated THP-1 cells acquire an adherent phenotype, increased surface expression of both the macrophage-specific differentiation antigen CD14 and the highly glycosylated form of CD147 and activation of MMP-9 and MMP-2.
Statins inhibit all of the elements of this macrophage differentiation process, retaining THP-1 cells in a dormant state. Possible molecular mechanisms described so far include:  To elucidate the mevalonate-dependent molecular mechanisms underlying the anti-monocytic differentiation properties of statins, two approaches were applied. First, rescue experiments were performed to reverse statin effects at several key steps of the cholesterol biosynthesis pathway; second, the effects of statins were compared to those of known inhibitors of intermediates downstream of HMG-CoA reductase.
Most of the statin effects we observed were reversed by FPP and dolichol and were mimicked by tunicamycin; in contrast, direct inhibitors of farnesylation and geranylgeranylation were not as effective as the statins, indicating that the N-glycosylation pathway, rather than the isoprenylation pathway, is the predominant regulator of CD147.

Role of monocytes in inflammation
During inflammation and infection, monocytes migrate from peripheral compartments to target organs where they mature and differentiate into tissue macrophages [41]. This maturation process results in the release of several inflammatory cytokines [42] and other factors, such as MMPs [43,44].

CD147 as a key player in monocytic differentiation
CD147 is one of the major factors reported to influence the monocytic differentiation process [45]. The pivotal role of CD147 in monocyte differentiation has been confirmed by the fact that AP-9, a specific peptide antagonist of CD147, prevents monocyte differentiation [46]. Furthermore, it has been demonstrated that AP-9 significantly inhibits the secretion and activation of MMP-2 and MMP-9 by THP-1 cells, thus emphasizing the role of CD147 in regulating MMPs [11]. CD147 interacts with adhesion markers such as CD29 (β1-integrin) and CD98 (large neutral amino acid transporter 1) [47] and participates in integrin-mediated adhesion, cellular differentiation and apoptosis [48]. Inhibition of CD147 has also been shown to impair the translocation of monocarboxylate transporter (MCT)-1 to the cell surface, disrupting the cellular lactate acid shuttle; this leads to intracellular acidification and metabolic starvation. Both of these effects lead to inhibition of monocytic differentiation [49].

Statins as potential immunomodulators
Statins are HMG-CoA reductase inhibitors that have been shown to promote beneficial effects in autoimmune conditions [50]. Although the original function of statins was to lower plasma low density lipoprotein cholesterol, cholesterol independent effects have also been demonstrated [51]. In monocytes, statins have been shown to inhibit proinflammatory responses [52], abort the functional differentiation of monocytes and inhibit MMP secretion and activation [53]. However, the mechanisms by which they regulate these processes have not yet been completely elucidated.

Statins and CD147
We hypothesized that inhibition of monocytic differentiation by statins is mediated through CD147 as a major player in the monocyte differentiation process [11]. In this study, statins inhibited the differentiation of THP-1 cells, which was evidenced by decreases in adherence. Statin treatment also altered the morphology of THP-1 cells, thus indicating that statins interfere with cellular differentiation.
Because of the known effects of statins on farnesylation and dolichol, we hypothesized that statins influence the molecular structure of CD147. The extracellular domain of CD147 has three putative N-linked glycosylation sites [18] that could potentially be inhibited by statin treatment. CD147 exists in two different glycosylation states: a HG (highly glycosylated) and a LG (lowly glycosylated) form. MMP induction is promoted by HG CD147 via homo-oligomerization, whereas LG CD147 acts as an inhibitor of MMP-2 and MMP-9, probably through its affinity for caveolin [9,19]. Thus, the glycosylation status of CD147 acts as a molecular switch for the activation of MMPs [8,54]. Our results show that fluvastatin treatment increases the expression of LG CD147, confirming the role of CD147 in mediating statin-associated MMP inhibition.
Cell permeabilization studies with statin-treated THP-1 cells revealed that the statininduced inhibition of CD147 expression was more pronounced at the cell surface compared to the intracellular compartment. We therefore hypothesized that statins regulate the expression of CD147 through post-translational mechanisms, e.g., isoprenylation, N-glycosylation or intracellular retention of immature pro-forms of CD147, rather than changes at the genomic level.
From these observations, we surmised that both isoprenylation and N-glycosylation contribute to the expression and activity of CD147 and that both of these mevalonate-dependent pathways are inhibited by treatment with statins. One possible mechanism for statin-induced effects on N-linked glycosylation is through dolichol, which acts as a carbohydrate donor during the N-glycosylation of membrane targeted proteins; dolichol production is regulated by FPP, which is downstream of mevalonate [55].
Two distinct inhibitors of isoprenylation (FTI-277, a farnesylation inhibitor, and GGTI-298, a geranylgeranylation inhibitor) only partially inhibited CD147 surface expression and MMP activation, indicating that isoprenylation plays only a minor role in these pathways. However, the same isoprenylation inhibitors did induce the expression of LG forms of CD147. We speculate that this discrepancy is due to a potential effect of statins on caveolin-1 itself (mediated by isoprenylation) because upregulation of caveolin-1 also promotes LG CD147 forms and subsequently decreases self-association of CD147 on the cell surface [18]. Thus, statins act on isoprenylation pathways to impair trafficking of CD147 to the cell surface. Tunicamycin specifically inhibits the N-glycosylation of newly synthesized proteins [56]. In our experiments, tunicamycin drastically reduced de novo N-glycosylation of CD147, resulting in greatly increased expression of the non-glycosylated core protein. Taken together, these data suggest that statins regulate CD147 on multiple levels, particularly through isoprenylation and N-glycosylation.
Most transmembrane proteins are not as sensitive to changes in glycosylation status as CD147. The requirement for high levels of glycosylation for the MMP stimulating activity of CD147 [18] is unusual but not unique. Like CD147, insulin-like growth factor (IGF)-1 receptor is a membrane-targeted molecule that requires N-glycosylation for its proper function; reduced glycosylation activity causes proreceptor retention within the endoplasmic reticulum [32]. Consistent with our findings for CD147, two other groups have described downregulation of IGF-1 receptor by statin treatment and demonstrated that this inhibitory effect is via the isoprenylation and N-glycosylation pathways, which act synergistically to promote IGF-1 receptor activity [32,57].
Further analysis of statins' effects on other key players surrounding CD147, such as CD29, CD98, caveolin-1 and cyclophilins, will increase the understanding of the mechanisms underlying the effects of these potent pleiotropic agents.
We postulate that statins produce their anti-inflammatory effects on monocytes and macrophages by influencing CD147, making them interesting candidates for therapeutic strategies against autoimmune disorders such as MS or RA [58,59]. However, formal proof of their clinical efficacy is still pending. In contrast, statins are already widely used for primary prevention of atherosclerosis [29]. Other CD147-mediated properties of statins, such as their effects on cancer, merit further research [10,60].