Anti-inflammatory potential of simvastatin and amfenac in ARPE-19 cells; insights in preventing re-detachment and proliferative vitreoretinopathy after rhegmatogenous retinal detachment surgery

All experiments were conducted using a commercial human RPE cell line [ARPE-19; American Type Culture Collection (ATCC)]. Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM/F-12 1:1; Life Technologies, Paisley, UK) with an antibiotic mixture of penicillin (100 U/ml) and streptomycin (100 μg/ml; Life Technologies, Grand Island, NY, USA), L-glutamine (2 mM, Life Technologies, Paisley, UK), and fetal bovine serum (FBS 10%, GE Healthcare Life Sciences, South Logan, UT, USA). Experiments were performed in serum-free DMEM/F-12 medium with L-glutamine and antibiotics. For thiazole blue/3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), lactate dehydrogenase (LDH), interleukin (IL)-6, IL-8, monocytic chemoattractant protein (MCP-1), vascular endothelial growth factor (VEGF), and pigment epithelium-derived factor (PEDF) measurements, cells were seeded onto 12-well plate (Corning Incorporated Costar, Kennebunk, ME, USA) at 200 000 cells per well, for Nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and prostaglandin E2 (PGE2) measurements onto 6-well plate at 400 000 cells per well, and for reactive oxygen species (ROS) detection onto 96-well plate at 15 000 cells per well. Cells were cultured until confluency for three days in a humidified incubator at 5% CO₂ and + 37 °C. Before exposures, cells were washed once with serum-free medium and then pre-treated with 5 μM simvastatin (Sigma Aldrich, USA) and/or 0.1 μM amfenac (Cayman Chemical, Michigan, USA) for 24 h in the incubator at 5% CO₂, + 37 °C. Dimethyl sulfoxide (DMSO, 0.15%; Sigma Aldrich, St. Louis, MO, USA) was used as a diluent control for amfenac and simvastatin. Thereafter, recombinant human IL-1α (10 pg/ml; R&D Systems, Minneapolis, MN, USA) was added to cell cultures for additional 24 h excluding NF-κB and PGE2 measurement or ROS detection for which 4 h or 1 h incubations were applied, respectively. For ROS measurement, hydroquinone at 125 µM concentration was used as a positive control [28]. The release of LDH and cytokines (IL-6, IL-8, and MCP-1) were determined also after an exposure of cells to recombinant human IL-1β (5 pg/ml; Gibco, Carlsbad, CA, USA) for 24 h at 5% CO₂, + 37 °C.

After treatments, cells were photographed under an Axio Vert A1 Zeiss microscope (Jena, Germany), after which medium samples were collected into microtubes (Sarstedt, Numbrecht, Germany), centrifuged (381×g, 10 min, + 4 °C), and stored at − 20 °C until analyses. Cells were purified from medium by adding Dulbecco’s phosphate-buffered saline (DPBS; Life Technologies, Paisley, UK) and scraped into microtubes with 200 µl DPBS, centrifuged (16 090×g, 10 min) and stored at − 80 °C. After thawing, cells were lysed by adding 60 μl of 1X lysis buffer (Cell lysis buffer 10X; Cell Signaling Technology, Leiden, Netherlands) and incubated for 5 min on ice. Thereafter, cell lysates were sonicated three times á 10 s and centrifuged at 16 090×g and + 4 °C for 10 min. Protein levels of the supernatants were determined using a method originated from the Bradford analysis, as described previously [29, 30].

The cell membrane integrity was studied using the commercial CytoTox96 Non-Radioactive Cytotoxicity Assay (Promega, Madison, WI, USA) just after the sample collection avoiding the loss LDH enzyme activity due to freezing. Medium samples were collected as described above, and 50 μl per well was added onto a 96-well plate (Greiner Bio-One GmbH, Frickenhausen, Germany). Briefly, 50 μl of substrate mixture was added into each well and incubated with medium samples at room temperature for 30 min in darkness. Thereafter, stop solution (50 μl) was added, and absorbances were measured using a spectrophotometer (BioTek, ELx808; Instruments Inc., Winooski, VT, USA) at the 490 nm wavelength.

Mitochondrial activity of the cells was monitored using the MTT (Sigma-Aldrich, St. Louis, MO, USA) assay. After exposures, 500 μl medium was left in each well of the cell culture plates, MTT salt (10 mg/ml) in DPBS was added, and cells were incubated for 3 h at 5% CO₂, + 37 °C. Thereafter, medium and MTT salt solution were replaced with 1000 μl of DMSO (Fischer Scientific, Leics, UK) and incubated for 20 min at room temperature. Thereafter, 200 μl of DMSO was transferred into wells of 96-well plate (Greiner Bio-One GmbH, Frickenhausen, Germany), and absorbance values were determined using the BioTek ELx808 spectrophotometer at the wavelength of 560 nm.

The DNA binding of NF-κB was measured from cell lysate samples with 15 μg of total protein in each sample using the commercial Active Motif TransAM NF-κB p65 (Carlsbad, CA, USA) assay. PGE2 was measured from cell lysates as pg/ml values using commercial Prostaglandin E2 Express Elisa kit (Cayman chemical, USA). Enzyme-linked immunosorbent assay (ELISA) analyses for cytokines IL-6, IL-8 and MCP-1 were performed from medium samples. IL-6 and IL-8 levels were detected using the BD OtpEIA™ Human ELISA sets (BD Biosciences, San Diego, CA, USA) and MCP-1 using the Human CCL2 (MCP-1) ELISA set (Thermo Fisher Scientific, San Diego, CA, USA). VEGF and PEDF were determined from medium samples using DuoSet Human VEGF and Human Serpin F1/PEDF DuoSet ELISA kits (both R&D Systems; Minneapolis, MN, USA). ELISA measurements were performed according to the manufacturers’ instructions and absorbance values were measured using a Bio-Rad Model 550 spectrophotometer (Bio-Rad Laboratories Inc., Hercules, CA, USA) at the wavelength of 450 nm with the correction wavelength of 655 nm.

After treatments, cells were washed once with serum-free medium. Thereafter, cells were exposed to the membrane permeable 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA; 5 μM) probe (Molecular probes, Life Technologies, Eugene, USA) for 1 h at 5% CO₂, + 37 °C. The probe is cleaved by ROS to 2′,7′-dichlorofluorescein (DCF) resulting in fluorescence emission. Hydroquinone (125 µM; Sigma-Aldrich, Saint Louis, MO, USA) was used as a positive control for ROS production [28]. After the probe incubation, cells were washed two times with 100 μl DPBS, 100 μl DPBS was added, and the fluorescence was measured using a fluorometer (BioTek Cytation3; Winooski, VT, USA) at the wavelengths of 488 nm and 528 nm (excitation/emission).

IL-1 cytokines are critical contributors to inflammatory and degenerative diseases in the retina [9]. The selected IL-1α concentration has been associated with mild inflammation in the eye, such as in corneal epithelial erosion patients [31, 32]. Simvastatin concentration is based on our previous study where it reduced inflammation in RPE cell cultures upon exposure to lipopolysaccharide (LPS) in ARPE-19 cells [20], and amfenac concentration was selected based on literature [25, 33] and preliminary results (Supplementary Figs. 1, 2). IL-1α did not affect the cell membrane integrity or metabolic activity of ARPE-19 cells when compared to cells exposed only to DMSO, which was used as a solvent for simvastatin and amfenac (Fig. 1). The used DMSO volume was not harmful for the cells (Fig. 1). The addition of simvastatin to IL-1α-treated cells neither caused changes in cell viability (Fig. 1). Instead, amfenac reduced LDH release and did not change metabolic activity of the cells when compared to cells exposed only to IL-1α and DMSO (Fig. 1). Together, simvastatin and amfenac increased metabolic activity in RPE cells without changing the cell membrane integrity when compared to IL-1α and DMSO -treated cells (Fig. 1). Microscopic analysis supported good condition of IL-1α-treated cells upon exposure either to simvastatin, amfenac, or both (Fig. 2).

Since inflammation contributes to the development of PVR after unsuccessful recovery from RRD surgery [3], we wanted to measure the effect of simvastatin and amfenac on the activity of NF-κB that regulates the production of many pro-inflammatory cytokines. IL-1α exposure did not significantly increase NF-κB levels when compared to cells exposed to DMSO (Fig. 3a). We have previously shown that DMSO does not alter the NF-κB activity in comparison to untreated cells [28]. Simvastatin alone or together with amfenac significantly reduced the DNA binding of NF-κB p65 whereas amfenac had no effect (Fig. 3a). Since prostaglandin and NF-κB regulate each other and NSAIDs are known COX-inhibitors [34‐36], we measured also PGE2 levels. The used IL-1α concentration did not change PGE2 levels in ARPE-19 cells when compared to DMSO-treated control cells (Fig. 3b). Instead, simvastatin alone reduced PGE2 levels in comparison to IL-1α but not when combined with amfenac (Fig. 3b). Amfenac alone did not reduce intracellular PGE levels. These data suggest that simvastatin has potential to regulate NF-κB activation and PGE2 levels in human RPE cells upon IL-1α exposure.

Since the regular systemic use of statins prior to surgery has been shown to reduce the need of re-surgery for retinal detachment, and topical NSAIDs reduce macular edema after cataract surgery [2, 37], we wanted to study whether simvastatin and/or amfenac have direct anti-inflammatory effects on RPE cells. We have previously shown that 0.5% DMSO reduces the release of cytokines (IL-6, IL-8, and MCP-1) when compared to untreated control cells [28]. DMSO has anti-inflammatory features e.g. by reducing cytokine production as well as the levels of prostaglandin and mitogen-activated protein kinases (MAPKs) [38]. In the presence of IL-1α, 0.15% DMSO reduced IL-8 levels (Fig. 4). The used IL-1α concentration alone did not induce IL-6 but increased IL-8 and MCP-1 production by ARPE-19 cells (Fig. 4). Neither simvastatin nor amfenac had effect on IL-6 levels when compared to cells treated by IL-1α without additional exposure. Simvastatin alone reduced the release of IL-8 and MCP-1, whereas amfenac with or without simvastatin reduced IL-8 but had no effect on MCP-1. Collectively, these data suggest that simvastatin and amfenac have potential to reduce the release of chemoattractant proteins from human RPE cells, both may reduce active inflammation by reducing the chemotaxis of neutrophils [39], and simvastatin can also reduce the attraction of monocytes to the retina [40].

Since IL-1β production is an important contributor in early retinal detachment [10, 41], we tested the anti-inflammatory capacity of simvastatin and amfenac also upon IL-1β exposure of RPE cells. In the presence of IL-1β, DMSO reduced the release of IL-6, IL-8, and MCP-1 cytokines when compared to untreated cells (Fig. 5). IL-1β induced the production of IL-6 and MCP-1 (Fig. 5). The pre-treatment of cells with simvastatin alone or together with amfenac significantly reduced the release of all three cytokines (Fig. 5). Amfenac alone significantly reduced the MCP-1 secretion but had no effect on the release of IL-6 or IL-8 when compared to IL-1β-treated RPE cells. Together simvastatin and amfenac compromised cell viability seen as significantly increased LDH release but separately both were well tolerated (Fig. 6). However, visible damage was not present in cell images (Fig. 7). Taken together, our present data allude that simvastatin could reduce the levels of chemokines that recruit neutrophils and monocytes to the inflamed tissue [39, 40, 42], whereas amfenac has slightly narrower capacity to prevent the release of chemokines. It is noteworthy that simvastatin also reduced IL-6 levels, which is critical factor for the EMT process in RPE cells as well as for the PVR formation [13, 14, 17].

ROS production increased during induced retinal detachment in a rat model, whereas its reduction improved the photoreceptor viability [43]. On the other hand, low level of ROS is produced as well in normal conditions but its excessive amount is detrimental e.g. induces inflammation [44], for which we took the effect of simvastatin and amfenac on the ROS production into consideration, as well. IL-1α alone or together with simvastatin or both simvastatin and amfenac, had no effect on the ROS production (Fig. 8a). Instead, amfenac increased intracellular ROS levels upon IL-1α-induced inflammation (Fig. 8a). In non-inflammatory conditions, simvastatin and amfenac alone and together increased ROS production in ARPE-19 cells (Fig. 8b). Collectively, amfenac and simvastatin showed properties to increase ROS levels in non-inflammatory and in induced inflammatory conditions in ARPE-19 cells although the production of pro-inflammatory cytokines was reduced (Fig. 4).

Since topical NSAID administration has been shown to reduce the need of anti-VEGF treatment for macular edema after cataract surgery [27], we tested whether simvastatin and/or amfenac has direct effects on the VEGF production of RPE cells under inflammatory conditions [45]. In addition, we measured the levels of PEDF, an anti-angiogenic factor that regulates VEGF [46, 47]. IL-1α (10 pg/ml) increased the VEGF release from RPE cells when compared to those exposed solely to the solvent control DMSO (Fig. 9a). Simvastatin or amfenac alone had no effect on the VEGF release from RPE cells after IL-1α exposure, whereas their combination significantly increased VEGF secretion. Neither IL-1α, simvastatin, nor amfenac changed PEDF levels (Fig. 9b). The VEGF/ PEDF ratio increased in IL-1α-treated ARPE-19 cells with the combined application of simvastatin and amfenac when compared to cells exposed only to IL-1α (Fig. 9c). Collectively, simvastatin together with amfenac had potential to increase VEGF production without any change in the PEDF release from RPE cells upon inflammatory conditions.