The differential statin effect on cytokine production of monocytes or macrophages is mediated by differential geranylgeranylation-dependent Rac1 activation

Abstract

Monocytes and macrophages contribute to pathogenesis of various inflammatory diseases, including auto-inflammatory diseases, cancer, sepsis, or atherosclerosis. They do so by production of cytokines, the central regulators of inflammation. Isoprenylation of small G-proteins is involved in regulation of production of some cytokines. Statins possibly affect isoprenylation-dependent cytokine production of monocytes and macrophages differentially. Thus, we compared statin-dependent cytokine production of lipopolysaccharide (LPS)-stimulated freshly isolated human monocytes and macrophages derived from monocytes by overnight differentiation. Stimulated monocytes readily produced tumor necrosis factor-α, interleukin-6, and interleukin-1β. Statins did not alter cytokine production of LPS-stimulated monocytes. In contrast, monocyte-derived macrophages prepared in the absence of statin lost the capacity to produce cytokines, whereas macrophages prepared in the presence of statin still produced cytokines. The cells expressed indistinguishable nuclear factor-kB activity, suggesting involvement of separate, statin-dependent regulation pathways. The presence of statin was necessary during the differentiation phase of the macrophages, indicating that retainment-of-function rather than costimulation was involved. Reconstitution with mevalonic acid, farnesyl pyrophosphate, or geranylgeranyl pyrophosphate blocked the retainment effect, whereas reconstitution of cholesterol synthesis by squalene did not. Inhibition of geranylgeranylation by GGTI-298, but not inhibition of farnesylation or cholesterol synthesis, mimicked the retainment effect of the statin. Inhibition of Rac1 activation by the Rac1/TIAM1-inhibitor NSC23766 or by Rac1-siRNA (small interfering RNA) blocked the retainment effect. Consistent with this finding, macrophages differentiated in the presence of statin expressed enhanced Rac1-GTP-levels. In line with the above hypothesis that monocytes and macrophages are differentially regulated by statins, the CD14/CD16-, merTK-, CX₃CR1-, or CD163-expression (M2-macrophage-related) correlated inversely to the cytokine production. Thus, monocytes and macrophages display differential Rac1-geranylgeranylation-dependent functional capacities, that is, statins sway monocytes and macrophages differentially.

Fig. 1. Statins retain the cytokine production of monocyte-derived macrophages in the differentiation phase, but do not affect cytokine production of freshly isolated monocytes.

a Experimental design. On day 1, mononuclear cells (MNC) were prepared from heparinized whole blood. Monocytes (Mo) were isolated from MNC using CD14-antibodies linked to magnetic beads. Mo (left part of the figure; marked in yellow) were incubated in VLE-RPMI-1640, containing 10% fetal calf serum (FCS), 1% l-glutamine, and 1% antibiotics, without or with statin (10 µg/ml; M or S, respectively; blue letters) and without or with LPS (100 ng/ml; N or L, respectively; red letters), added immediately after the isolation of the Mo. On day 2, the supernatants and/or the cells were harvested and stored for analyses (SN-IC_Mo). On the other hand, macrophages (Mac; right part of the figure; marked in yellow) were derived from the Mo by incubating Mo without (M) or with statin (S) for 24 h (Diff, differentiation phase; gray box), but without stimulus. On day 2, medium (N) or medium with LPS (L) was added as the stimulus and the cultures were incubated for further 24 h (Stim, stimulation phase; gray box). Thereafter (on day 3), the supernatants and/or the cells of the Mac were harvested and stored for analyses (SN-IC_Mac). b A summary of multiple experiments shows that fluvastatin does not affect the cytokine production of the Mo, but retains the cytokine production of the Mac. Mo and Mac were isolated and prepared as described in Fig. 1a. The cytokine data of the four controls (Mo, Mo + Stat, Mac, and Mac + Stat; blue, orange, black, and red columns, respectively; all LPS-stimulated) always present in the numerous experiments performed in the study (IL-1, n = 70; IL-6, n = 48) were normalized (%) to the respective highest cytokine level of these four controls. The mean, the SD, and the significance of these data were calculated in SPSS (Levene’s test, Welch’s ANOVA, and Games–Howell post hoc analysis). The asterisks above the columns reflect the significance of “Mac” vs. “Mo”, “Mo + Stat,” or “Mac + Stat”, respectively; other comparisons are indicated by the lines (***p < 0.001; **p < 0.01; *p < 0.05; ns, not significant). c The retainment effect of the statin is initiated during the differentiation phase. Freshly isolated Mo (100,000 cells/cm²) were incubated in 24-well plates (Nunc) on day 1 and LPS (cond 1) or LPS and fluvastatin (cond 2) were added. Supernatants were harvested on day 2. In order to produce Mac cultures, the Mo were incubated on day 1 in the absence (cond 3) or presence (cond 4) of fluvastatin. On day 2, LPS was added to both cultures (blue letters). In parallel cultures to cond 4 (i.e., statin-pretreated Mac), the statin was removed by a washing step on day 2, before LPS was added (light gray columns; wash + L; cond 5). On the other hand, in parallel cultures to cond 3 (i.e., medium-pretreated Mac), LPS and statin were added simultaneously on day 2 (dark gray columns; L + S; cond 6). Supernatants were harvested after further 24 h. The cytokine concentration was determined in ELISA. Four additional experiments showed similar results. Statistics and color code as in Fig. 1b.

Fig. 2. The isoprenoid pathway, rather than cholesterol synthesis, is involved in the retainment effect.

a Schematic overview of the cholesterol and isoprenoid pathways, as well as the used inhibitors and reconstituting compounds. Red letters, the used inhibitors. Blue letters, the compounds used for reconstitution/spiking. Green letters, the respective pathway. The broken lines indicate omission of steps. FPP, farnesyl pyrophosphate; FTI, farnesyl transferase inhibitor; GGPP, geranylgeranyl pyrophosphate; GGTI, geranylgeranyl transferase inhibitor; Mev, mevalonic acid; Stat, statin; Zara, zaragozic acid A. b Mevalonic acid, but not squalene, reverses the retainment effect. Mac (50,000 cells/cm²; 24-well plate) were incubated as outlined in Fig. 1a. The cells were preincubated during day 1 with medium (M; black column), statin (S; red column), statin plus mevalonic acid (S Mev; light gray column; 10 µM; Sigma-Aldrich) or statin plus squalene (S Squa; dark gray column; 20 µM; Sigma-Aldrich). On day 2, LPS was added to all cultures and the supernatants were harvested 24 h later. Two experiments with similar results were performed. Data analysis and color code as in Fig. 1c (“M” vs. “S”, “S Mev” or “S Squal”, respectively). c FPP, like mevalonic acid, reverses the retainment effect. Experimental design as in Fig. 2b, except for the use of farnesyl pyrophosphate (S FPP, 20 µM; Echelon Biosciences, Mobitec, Göttingen, Germany). Two experiments with similar results were performed. Data analysis and color code as in Fig. 1c (“M” vs. “S”, “S Mev” or “S FPP”, respectively). d GGPP blocks the retainment effect in a concentration-dependent fashion. Experimental design as in Fig. 2b, except the use of 5, 10, 20 or 40 µM geranylgeranyl pyrophosphate (GGPP; Echelon). Five additional experiments with similar results were performed. Data analysis and color code as in Fig. 1c (“M” vs. “S” or “S GGPP”, respectively). All comparisons of “S” vs. “S GGPP” were <0.001. e GGPP interferes with Mac but not with Mo. Mo and Mac (50,000/cm²; 24-well plate) were prepared and incubated as described in Fig. 1a. Statin and GGPP were added on day 1 to Mo and Mac. To Mo LPS was also added on day 1 and the cells were incubated for 24 h. To Mac LPS was added only on day 2 and the cells were then incubated for further 24 h. Color code and statistics (No GGPP vs. With GGPP) as described in Fig. 1c. N, no statin; S, statin (20 µg/ml); C, no GGPP; G, GGPP (5 µM). f Like statin, GGTI, but not FTI or Zara, retains the cytokine production of Mac. Mac (100,000 cells/cm²; 24-well plate) were prepared in the presence of medium (M) or medium containing fluvastatin (S), geranylgeranyl transferase inhibitor 298 (GGTI; 8 µM), farnesyl transferase inhibitor 277 (FTI; 8 µM) or the cholesterol synthesis inhibitor zaragozic acid A (Zara; 20 µM; Sigma-Aldrich), respectively, for 24 h. On day 2 LPS was added, supernatants were harvested on day 3 and cytokines were measured in ELISA. Two experiments with similar results were performed. Data analysis and color code as in Fig. 1c (“M” vs. “S”, “GGTI”, “FTI” or “Zara”, respectively). Please note the axis break.

Fig. 3. The blockade of the statin-mediated retainment effect by GGPP is reversed by GGTI.

a GGTI reverses the GGPP effect. Monocytes and macrophages (50,000 cells/cm²; 24-well plate) were cultured and stimulated as described in Fig. 1a. The cytokine production of the four controls is presented in the left part of the figure. To parallel cultures of the statin-treated Mac (red column), GGPP at various concentrations, in the absence (gray columns) or presence (rose columns) of GGTI (8 µM), was added on day 1. To all macrophage cultures, LPS was added on day 2 and the supernatants were harvested 24 h later. The cytokine levels were measured in ELISA. One additional experiment with similar results was performed. Data analysis (“Stat + GGPP” vs. “Stat + GGPP + GGTI”; other comparisons are indicated by the lines) and color code (except rose and gray, compare above) as in Fig. 1c. b The blockade of the retainment by GGPP and its reversal by GGTI is also detectable at the RNA-level. Monocytes and macrophages were prepared as described in Fig. 3a in 25 cm² culture flasks (Falcon, Corning GmbH, Kaiserslautern, Germany; 100,000 cells/cm²; Stat, 10 µg/ml; GGPP, 5 µM; GGTI, 8 µM). IL-1 and IL-6-RNA was measured using the “iScript protocol”. Cytokine levels were normalized to GAPDH. The highest normalized value of each experiment was determined 100% and the mean ± SD of three experiments calculated. Data analysis was performed in SPSS (ANOVA and LSD post hoc; “Mac” vs. “Mo”, “Mo + Stat”, “Mac + Stat”, “Mac + Stat + GGPP” or “Mac + Stat + GGPP + GGTI”, respectively; other comparisons are indicated by the lines; compare Fig. 1c). Color code as in Fig. 3a.

Fig. 4. Rac1 is involved in the retainment of the cytokine production.

a Statin-pretreated Mac (“low-GGPP Mac”) express enhanced Rac1-activation. Macrophages (100,000/cm²) were prepared by incubation (75 cm² culture flask; 24 h) in the absence (M) or presence of statin (S; 10 µg/ml), statin and GGPP (SG; G, 5 µM) or statin, GGPP, and GGTI (SGT; T, 8 µM). The cells were harvested and lysed in pull-down lysis buffer. Pull-down for Rac1-GTP was performed with PAK-PBD beads, the beads were washed and used in Western blot. Control samples for GAPDH and total Rac1 (Rac1) were directly taken from the lysate and applied to the Western blot. The numbers at the right indicate the molecular weight (kDa) taken from a molecular weight standard. Four additional experiments with similar results were performed. b Rac1- and PI3K-inhibitors block the retainment effect during the differentiation phase. Monocytes and macrophages (50,000 cells/cm²; 24-well plate) were incubated as described in Fig. 1a. Parallel to macrophages treated with statin (A; red letter above the columns), macrophages were treated with statin and a Rac1/TIAM1-inhibitor (A + Rac1 Inhib; NSC23766, Tocris, Bio-Techne GmbH, Wiesbaden, Germany) or a PI3K-inhibitor (A + PI3K Inhib; LY294002, Tocris), respectively, both at 10 µM, under the following protocols: “1”, statin and inhibitors were present during day 1 and 2; “1^w”, statin and inhibitors were present during day 1 and removed by washing on day 2; “2”, statin was added on day 1 and the inhibitors were added on day 2. To all macrophage cultures LPS was added on day 2. IL-1 and IL-6 were measured in ELISA. One additional experiment with similar results was performed. Data analysis (“Mac + Stat” vs. “Mac + Stat + Inhibitors”; other comparisons are indicated by the lines) and color code as in Fig. 1c. c Rac1-siRNA blocks the retainment in macrophages. Monocytes and macrophages were cultured as described in Fig. 1a in 6-well plates (100,000 cells/cm²; 100 ng/ml LPS), in the absence or presence of statin (10 µg/ml). The control siRNA (Ctrl; 88 nM) or the Rac1-siRNA (both are presented as gray columns; both are “Silencer®select” siRNAs, Ambion) were also added on day 1 in the depicted concentrations to the Mac. LPS was added on day 2 to all Mac and the supernatants were harvested on day 3. IL-1 was measured in ELISA. Two experiments with similar results were performed. Data analysis (“Ctrl” vs. “Rac1-siRNA”) and color code as in Fig. 1c.

Fig. 5. microRNAs 146a, 146b, and 155 are down-regulated in statin-treated Mac.

a microRNA array. Total RNA was isolated from macrophages differentiated in the absence or presence of statin (25 cm² flasks, 100,000 cells/cm²) with the “RNeasy Plus Mini Kit”. Deep sequencing was performed by the “Core Unit DNA”, Universität Leipzig, using the “TruSeq™Small RNA sample prepkit v2” (illumina, San Diego, USA). The mean of the normalized data of macrophages pretreated without statin and macrophages pretreated with statin was calculated and data of samples with a mean >100 counts (192 samples; compare Supplementary Table 2) were included into the analysis and blotted against each other. The orange line indicates unchanged expression. The red dots mark three selected miRs, which were down-regulated in statin-pretreated macrophages, as compared to Mac prepared in the absence of statin (compare Supplementary Table 2). A second array showed a similar result. b Blockade of miR-146a and miR-155 reverses the hypo-responsiveness in macrophages only to some degree. Macrophages were incubated as described in Fig. 1a (6-well plate; 100,000 cells/cm²). The respective “miRCURY LNA™” anti-miR (Exiqon, Qiagen, Vedbaek, Denmark) were prepared in “Lipofectamine RNAiMax” and 250 µl of this solution was added to 2750 µl of culture medium in the absence (−) or presence (+) of statin. After 24 h LPS was added. After further 24 h, the supernatants were harvested and analyzed in ELISA. Four experiments with similar results were performed. Data analysis and color code as in Fig. 1c (“Ctrl” vs. “anti-miR”).

Fig. 6. The retainment effect is paralleled by reciprocal expression of macrophage-related surface markers.

a The enhanced CD14/CD16-expression in macrophages prepared without statin is not observed in macrophages prepared in the presence of statin. The surface marker expression of LPS-stimulated Mo and Mac prepared as described in Fig. 1a was analyzed by flow cytometry. For this purpose the cells (25 cm²; 100,000/cm²) were harvested, centrifuged (300 × g; 10 min), resuspended in 100 µl PBS and transferred into conical 96-well plates (Greiner). The plates were centrifuged (400 × g; 3 min) and incubated in PBS containing Zombie Aqua™ (BioLegend, San Diego, USA) for 15 min, followed by centrifugation and resuspension in PBS containing 1% BSA, 0.1% sodium azide and 1 mM EDTA (FACS-buffer). In order to avoid unspecific binding of the antibodies, the cells were incubated with 10% FcR-blocking reagent (Miltenyi) for 15 min at 4 °C in FACS-buffer. Antibodies against CD14 or CD16 (compare Supplementary Table 1) were added (15 min; 4 °C; in the dark). Analysis was performed in a LSR-Fortessa™, using the “FlowJo LLC” software (Ashland, OR, USA). Aggregated cells were excluded by FSC-H- and FSC-A-scatter and dead cells were excluded by gating Zombie Aqua™-negative cells (compare “gating strategy” in Supplementary Fig. 5A–D). The dashed circles indicate the same position of CD14⁺/CD16⁺-cells in each graph. Five experiments with similar results were performed. The numbers in the gates reflect the respective percentages. B The CD163- and CX₃CR1-expression is upregulated in untreated macrophages, but not in statin-treated macrophages. Mo and Mac were prepared as described in Fig. 1a in 25 cm² flasks (57,353 cells/cm²). After the respective incubation, the cultures were gently scraped, the cells centrifuged (300 × g; 10 min), the supernatants harvested and the cell pellets resuspended twice in 1 ml MACS-buffer (PBS, 2 mM EDTA, 2% FCS; 4 °C). FcR-blocking reagent (2%; Miltenyi) was added for 10 min. Antibodies against CD86, CCR2/CD192, CX₃CR1 or CD163, or the respective isotype controls (compare Supplementary Table 1; thin gray line in the figure) were added. After 20 min of incubation in the dark, the cells were washed with MACS-buffer and analyzed using a LSR-Fortessa™ (BD Biosciences). Aggregated cells were identified in the FSC-H (forward scatter-high) FSC-A (forward scatter-area) window and excluded from the analysis (compare “gating strategy” in Supplementary Fig. 5A–D). The monocyte region was then determined and gated based on the FSC and SSC (side scatter) parameters. Dead cells were excluded by gating of cells, which were not stained for 7-AAD (7-aminoactinomycin D; BD Biosciences). Visualization and analysis were performed using the “FlowJo LLC” software and the expression of the respective marker (normalized to ratio) was presented. A representative experiment out of seven is shown. The numbers in the lower right corner reflect the MFI (geometric mean) of “isotype control”, “Mo”, “Mo + Stat”, “Mac,” and “Mac + Stat”, respectively, taken from “FlowJo LLC”.

Fig. 7. Hypothesis, summary, conclusion, and outlook.

a Hypothesis and Summary—Monocytes and macrophages respond differentially to statin. Hypothesis: From our and other authors previous data we hypothesized that in Mo and Mac GGPP (blue/italic letters) might be present at different levels. Summary: This hypothesis appears to be correct, since in the present manuscript we show that statin does not significantly change the level of inflammatory cytokines in the monocytes. This was paralleled by a lack of changes in surface marker expression in the monocytes. However, we observed profound statin effects in overnight-differentiated macrophages. Green letters, high levels; orange letters, low levels. b Conclusion and Outlook—The statin effects on macrophages are geranylgeranylation-dependent. Conclusion: In the absence of statin (No Stat), high GGPP is present in the overnight-differentiated macrophages, which results in low Rac1-activation and low IL-1-levels. On the other hand, in the presence of statin (Stat) during the overnight differentiation, low GGPP is present, which results in enhanced Rac1- and subsequent high IL-1-levels. This was proven by external GGPP and GGTI: a) addition of GGPP (GGPP_a; blue letters; blue arrow) to the Mac treated with statin reversed statin’s effect. b) On the other hand, addition of GGTI (GGTI_b; blue letters; blue arrow) to Mac prepared without statin blocked the geranylgeranylation, resulting in high Rac1 and subsequent high IL-1. c) Addition of a Rac1-inhibitor (Rac-I_c; blue letters; blue arrow) to Mac prepared in the presence of statin may result in low Rac1-activity and, subsequently, in low IL-1-production. The open arrowheads indicate that the regulation pathway resulting in Rac1-activation is not yet defined. However, Rac1 may be regulated by some GEF, such as TIAM1, which may be determined in the future. Outlook: The retainment of function(s) in statin-treated Mac may be of importance in various diseases related to inflammatory processes, such as auto-inflammatory diseases, sepsis, cancer or atherosclerosis.