Resolvin D1 prevents TNF-α-mediated disruption of salivary epithelial formation

Abstract

Sjögren's syndrome is a chronic autoimmune disorder characterized by inflammation of salivary glands resulting in impaired secretory function. Our present studies indicate that chronic exposure of salivary epithelium to TNF-α and/or IFN-γ alters tight junction integrity, leading to secretory dysfunction. Resolvins of the D-series (RvDs) are endogenous lipid mediators derived from DHA that regulate excessive inflammatory responses leading to resolution and tissue homeostasis. In this study, we addressed the hypothesis that activation of the RvD1 receptor ALX/FPR2 in salivary epithelium prevents and/or resolves the TNF-α-mediated disruption of acinar organization and enhances monolayer formation. Our results indicate that 1) the RvD1 receptor ALX/FPR2 is present in fresh, isolated cells from mouse salivary glands and in cell lines of salivary origin; and 2) the agonist RvD1 (100 ng/ml) abolished tight junction and cytoskeletal disruption caused by TNF-α and enhanced cell migration and polarity in salivary epithelium. These effects were blocked by the ALX/FPR2 antagonist butyloxycarbonyl-Phe-Leu-Phe-Leu-Phe. The ALX/FPR2 receptor signals via modulation of the phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathways since, in our study, blocking PI3K activation with LY294002, a potent and selective PI3K inhibitor, prevented RvD1-induced cell migration. Furthermore, Akt gene silencing with the corresponding siRNA almost completely blocked the ability of Par-C10 cells to migrate. Our findings suggest that RvD1 receptor activation promotes resolution of inflammation and tissue repair in salivary epithelium, which may have relevance in the restoration of salivary gland dysfunction associated with Sjögren's syndrome.

proper salivary gland function is critical for oral health. Sjögren's syndrome is an autoimmune chronic inflammatory disease that causes significant loss of salivary gland function leading to xerostomia (14). The ensuing salivary gland hypofunction decreases the quality of life for these patients (57). To date, several approaches have been used to treat inflammation-mediated hyposalivation in humans, but these reagents have had limited efficacy (9, 41, 55). Therefore, new approaches to modulate inflammatory responses in salivary glands and restore salivary tissue integrity in Sjögren's syndrome are necessary.

Recent studies demonstrate that human and animal cells convert ω-3 polyunsaturated fatty acids into resolvins (Rvs), which are novel, highly potent, short-lived, anti-inflammatory agents that control the duration and magnitude of inflammation in models of complex diseases (27, 29, 58–60, 62). Resolvin D1 (RvD1, 7S, 8R, 17S-trihydroxy docosahexaenoic acid [DHA]) is produced in resolving exudates in vivo and is a product of transcellular biosynthesis with human leukocytes and endothelial or epithelial cells (i.e., leukocytes can take up 17-HDHA from epithelial cells or endothelial cells and convert it to RvD1) (3). Aspirin, an anti-inflammatory drug that acetylates cycloxygenase-2 (COX-2) and changes COX-2 function from a cyclooxygenase to a lipoxygenase, generates the aspirin-triggered (AT) form AT-RvD1, which appears to be more biologically stable than RvD1 (62, 63). Resolution pathways initiated by RvD1 includes binding to the high-affinity G protein-coupled receptors (GPCRs), ALX, and GPR32 (40). Upon binding their ligands, activation of these receptors decreases proinflammatory cytokine (TNF-α)-induced IL-1β transcripts in microglial cells (29). RvD1 also enhances the phagocytic and clearance functions of human macrophages and rapidly stops polymorphonuclear neutrophils (PMN) migration in microfluidic chambers, indicating its potent actions at the single-cell level (37). In murine peritonitis in vivo, RvD1 and AT-RvD1 proved equipotent, limiting PMN infiltration in a dose-dependent fashion (62).

RvD1's potent anti-inflammatory actions in many tissues prompted us to investigate its impact on salivary epithelial integrity under physiological and inflammatory conditions using the polarized rat parotid cell line Par-C10 (54). Our results indicate that RvD1 treatment in Par-C10 cells prevents TNF-α-mediated disruption of salivary epithelial formation and enhances cell migration and cell polarity via phosphatidylinositol 3-kinase (PI3K)/Akt signaling. These studies support new anti-inflammatory and proresolving properties of RvD1 in damaged salivary epithelium.

MATERIALS AND METHODS

Experimental animals.

Female C57BL/6 mice at 16 wk of age were anesthetized with 80–100 mg/kg ketamine + 10 mg/kg xylazine. Mice were euthanized by abdominal exsanguination, and submandibular glands were removed for preparation of dispersed cell aggregates or frozen in isopentane in liquid nitrogen for preparation of frozen sections. All animal usage, anesthesia, and surgery were conducted under the strict guidelines and approval of the State University of New York at Buffalo Institutional Animal Care and Use Committee.

Preparation of dispersed cell aggregates from mouse submandibular gland.

Dispersed cell aggregates from submandibular glands (SMGs) of C57BL/6 mice were prepared as described previously (68). Protocols conformed to Institutional Animal Care and Use guidelines of the State University of New York at Buffalo. Briefly, mice were anesthetized with pentobarbital sodium injection (125 mg/kg body wt), and SMGs were removed. Glands were finely minced in dispersion medium consisting of DMEM-Ham's F12 (1:1) (Hyclone, Logan, UT) and 0.2 mM CaCl₂, 1% (wt/vol) BSA, 50 U/ml collagenase (Worthington Biochemical, Freehold, NJ), and 400 U/ml hyaluronidase at 37°C for 40 min with aeration (95% air-5% CO₂). Cell aggregates in dispersion medium were suspended by pipetting at 20, 30, and 40 min. The dispersed cell aggregates were washed with enzyme-free assay buffer (in mM: 120 NaCl, 4 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 1 CaCl₂, 10 glucose, 15 HEPES, pH 7.4) containing 1% (wt/vol) BSA and filtered through a nylon mesh. Cells were cultured in DMEM-Ham's F12 (1:1) containing 2.5% (vol/vol) FBS (GIBCO, Gaithersburg, MD) and the following supplements: 0.1 μM retinoic acid, 80 ng/ml epidermal growth factor, 2 nM triiodothyronine, 5 mM glutamine, 0.4 μg/ml hydrocortisone, 5 μg/ml insulin, 5 μg/ml transferrin, 5 ng/ml sodium selenite, 50 μg/ml gentamicin, and 8.4 ng/ml cholera toxin for 6 h. Cells were centrifuged at room temperature at 1,000 rpm for 30 s, the supernatant was removed, and pellet was lysed for Western blot analysis.

Par-C10 cell monolayer culture.

The rat parotid cell line (Par-C10) was derived from freshly isolated rat parotid gland acinar cells by transformation with simian virus 40 (SV40), Par-C10 cells exhibit similar morphological, biochemical, and functional characteristics of native acinar cells (54). Par-C10 cells (5×10⁵; passages 30–60) were plated on plastic for signaling studies or on permeable supports (diameter 1.2 cm, pore size 0.4 μm; Becton Dickinson, Franklin Lakes, NJ) for structural studies. The cultures were grown to confluence in DMEM-Ham's F12 (1:1) with supplements as described previously (5). Cells were cultured at 37°C in a humidified atmosphere of 95% air-5% CO₂ and used for assays at specified time points.

Par-C10 acinar-like spheres culture.

Fifty microliters of growth-factor-reduced (GFR)-Matrigel (8 mg/ml; 2:1 GFR-Matrigel: DMEM-Ham's F12 [1:1] medium; Becton Dickinson Labware) was allowed to solidify in a 37°C incubator for 1 h in 8-well chambers mounted on No. 1.5 German borosilicate coverglasses (Nalge Nunc International, Naperville, IL). Then, Par-C10 cells (1×10⁴ cells/well; passages 30–60) were plated on the GFR-Matrigel in DMEM-Ham's F12 (1:1) medium with supplements, as defined previously (7). Differentiated 3D cultures of Par-C10 acinar-like spheres were used for assays after incubation at 37°C with 95% air and 5% CO₂ at specified time points.

RvD1 and TNF-α treatment of Par-C10 cells.

Par-C10 cells were treated with RvD1 (100 ng/ml; Cayman Chemical, Ann-Arbor, MI) and/or TNF-α (100 ng/ml; Becton Dickinson Pharmingen, San Diego, CA) at plating and incubated for 24 h and 60 h at 37°C. For signaling studies, cells were stimulated with RvD1 (100 ng/ml; Cayman) for various time periods, lysed, and subjected to Western blot analysis for Akt phosphorylation.

Intracellular free Ca²⁺ concentration measurements.

The intracellular free Ca²⁺ concentration ([Ca²⁺]_i) was quantified in single Par-C10 cells grown on coverslips and preloaded with fura-2, a Ca²⁺-sensitive fluorescent dye, using a MetaFluor dual wave length fluorescence imaging system (Molecular Devices, Sunnyvale, CA) as described previously (32). For fura-2 preloading, Par-C10 cells on glass coverslips were incubated in assay buffer [in mM: 120 NaCl, 4 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 10 glucose, 15 HEPES, pH 7.4, and 0.1% (wt/vol) BSA] containing 4 μM fura-2-acetoxymethylester (fura-2-AM; Molecular Probes, Eugene, OR) for 45 min at 25°C, washed, and further incubated for 20 min in calcium containing cell culture medium at 25°C in a 5% CO₂ atmosphere. The cells were positioned on the stage of a fluorescence microscope and stimulated with carbachol (100 μM) or UTP (100 μM) at room temperature, as described in figure legends. Fura-2 AM fluorescence images at 340 nm and 380 nm excitation wavelengths were captured with a video camera. Average whole-cell ratio values from single cells were determined (see figure legends).

Inhibition studies.

Par-C10 cells grown to 80% confluence were cultured in serum-free DMEM overnight before the addition of inhibitors or agonists. In some experiments, Par-C10 cells were preincubated for 30 min in serum-free DMEM with or without the selective PI3K inhibitor LY294002 (10 μM; Cell Signaling Technology, Danvers, MA). Then, medium was replaced with DMEM without LY294002 to prevent cell apoptosis (34, 49, 70). In another set of experiments, cells were treated with or without TNF-α (100 ng/ml; Becton Dickinson Pharmingen) and then immediately incubated for 30 min with or without the ALX/FPR2 receptor antagonist Boc-2 (38) (butyloxycarbonyl-Phe-Leu-Phe-Leu-Phe; 10 μM; Phoenix Pharmaceuticals, Burlingame, CA), followed by treatment with or without RvD1 (100 ng/ml; Cayman) and used for cell chemotaxis and chemokinetic assays.

siRNA-mediated suppression of AKT expression.

Par-C10 cells were transfected in reduced-serum medium (Opti-MEM; Invitrogen) with 100 nM of double-stranded small interference Akt RNA (i.e., Akt siRNA; Cell Signaling Technology) to suppress Akt expression as described previously (6), whereas 100 nM nonspecific siRNA (Cell Signaling Technology) was used as a negative control. Transfection of cells with Akt siRNA was carried out using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions at 1:2.5 (vol/vol) siRNA:Lipofectamine. All siRNAs were dissolved in RNase-free buffer as described in manufacturer's protocol. After 24 h, cells were treated in the absence or presence RvD1 (100 ng/ml) in serum-free DMEM for 6 h and lysed in 200 μl of 2× Laemmli buffer for Western blot analysis or plated for cell chemotaxis and chemokinetic assays.

5-Bromo-2′-deoxyuridine cell proliferation assay.

Par-C10 cell proliferation was measured by a luminometric immunoassay based on 5-bromo-2′-deoxyuridine (BrdU) incorporation into DNA using a cell proliferation enzyme-linked immunosorbent BrdU kit (EMD Chemicals, Gibbstown, NJ). Briefly, cells were plated in DMEM-Ham's F12 (1:1) containing 2.5% FBS (Invitrogen, Carlsbad CA), with growth supplements at a density of 1×10⁴ cells/well on 96-well plates, and then incubated for 24 h. Then, the growth media was replaced with serum-free media for an additional 12 h followed by the addition of RvD1 (100 ng/ml), TNF-α (100 ng/ml), or 5% FBS. For cotreatment studies, RvD1 was added 30 min before TNF-α. Subsequently, 10 μl of BrdU was added to each well for labeling de novo synthesized DNA during 24 h. Then, cells were fixed, denatured, and incubated with anti-BrdU antibody. Cell proliferation rate was determined according to the manufacturer's protocol.

Chemokinetic assay.

This assay was performed with the Cell Motility Hit kit (Cellomics, Pittsburgh, PA) following the manufacturer's protocol. Cell suspensions (∼500 cells) were added to a collagen-coated 96-well microplate containing a lawn of microscopic blue fluorescent beads. After 24 h of incubation at 37°C with the indicated reagents, cells were washed and fixed, and phagokinetic tracks were visualized with a Zeiss Axio Imager motorized fluorescence microscope at ×40 magnification, and images were analyzed using the Axio Vision 4.8 software (Carl Zeiss Jena, Germany).

Chemotaxis assay.

Chemotaxis assays were performed with Culturex BME-coated 8-μm pore size polycarbonate membrane (Corning) as described previously (10). In brief, cells were harvested by trypsinization, washed, resuspended in 100 μl of serum-free DMEM (5×10⁴ cells), and placed in the upper chamber of the transwells. The lower chamber was filled with 500 μl of serum-free DMEM supplemented with varying concentrations of RvD1, as indicated (see figure legends). The cells were allowed to transmigrate for 24 h at 37°C. Cells migrating to the lower side of the membrane were dissociated and stained with calcein AM (Trevigen, Gaithersburg, MD). The number of cells was quantitated in a Bio-Tek Epoch spectrophotometer at 485 nm excitation and 520 nm emission. Analysis was performed using the Gen 5 software, on three transwells per condition, and experiments were repeated at least three times. For Akt silencing experiments, chemotaxis assay was performed using the QCM cell invasion assay (Millipore, Bedford, MA) according to the manufacturer's protocol.

Fluorescence microscopy analysis.

Par-C10 grown on GFR-Matrigel or on permeable supports or mouse SMG frozen sections (10 μm) were fixed in 4% paraformaldehyde for 10 min at room temperature, incubated with 0.1% Triton X-100 in PBS for 5 min, and washed three times with PBS. The cells or sections were then incubated with 5% goat serum containing 10 μM digitonin for 2 h at room temperature and washed three times with PBS. The cells or sections were incubated overnight at 4°C with rabbit anti-ZO-1 (Invitrogen) or rabbit anti-ALX/FPR2 (Alomone Labs, Jerusalem, Israel) antibody at 1:500 dilution in 5% goat serum containing 10 μM digitonin. The next day, cells or sections were washed three times for 5 min with PBS (acinar-like spheres, were warmed to room temperature for 20 min before wash). Cells or sections were incubated for 45 min with Alexa Fluor 488-conjugated goat anti-rabbit (1:1,000 dilution in 5% goat serum containing 10 μM digitonin) and washed three times with PBS. Cells or sections were stained for 5 min with 1:10,000 dilution in PBS of Hoechst nuclear stain (Sigma; final concentration 0.05 pg/ml) and 1:500 dilution in PBS of phalloidin F-actin stain (Sigma; final concentration 13.2 pM). Images for Par-C10 cells grown on GFR-Matrigel were obtained using a Carl Zeiss 510 confocal microscope. Images for Par-C10 cells grown on permeable supports or sections were obtained using a Zeiss AxioImager fluorescence microscope. All images were analyzed using the AxioVision software (version 4.8). The ZO-1 integrated density values were obtained from defined tight junction (TJ) areas in Par-C10 cell monolayers and analyzed using ImageJ NIH software.

Measurement of transepithelial resistance in Par-C10 cell monolayers.

Changes in transepithelial resistance (TER) were measured as a function of time using an epithelial volt-ohmmeter (EVOM; World Precision Instruments, New Haven, CT) with miniature dual chopstick electrodes. After subtraction of bare filter resistance (120 Ω), tissue resistance values in Ω were multiplied by effective membrane area (π) (d²)/4 = (3.14) (1.20 cm)²/4 = 1.13 cm². Therefore, TER is expressed as Ω·cm².

Western blot analysis.

All of the cells were cultured at 37°C in a humidified atmosphere of 95% air-5% CO₂ and lysed in 200 μl of 2× Laemmli buffer [120 mM Tris·HCl, pH 6.8, 10% (vol/vol) glycerol, 2% (wt/vol) SDS, 1 mM DTT, and 0.002% (wt/vol) bromophenol blue], sonicated for 5 s with a VibraCell sonifier (microtip; output level 3; duty cycle 50%; Sonics and Materials, Dansbury, CT), and boiled for 5 min. Cell lysates were subjected to 7.5–12% (wt/vol) SDS-PAGE on minigels and transferred to nitrocellulose membranes. Membranes were blocked for 1 h with 5% (wt/vol) nonfat dry milk in Tris-buffered saline [0.137 M NaCl, 0.025 M Tris (hydroxymethyl)-aminomethane, pH 7.4] containing 0.1% (vol/vol) Tween-20 (TBST) and immunoblotted overnight with primary antibodies at 4°C in TBST containing 3% (wt/vol) BSA. The following antibodies were utilized: rabbit anti-ZO-1 that recognizes amino acids 334–634 of human ZO-1 (1:1,000 dilution; Invitrogen); rabbit phospho-Akt that recognizes endogenous phospho-Akt when phosphorylated at Ser473 (1:500 dilution; Cell Signaling Technology); rabbit anti-ALX/FPR2 (1:500 dilution; Alomone) that recognizes a sequence corresponding to amino acid residues 184–196 of human FPR2. After incubation with the primary antibodies, membranes were washed three times each for 15 min with TBST and incubated with peroxidase-linked goat anti-rabbit IgG antibody (1:2,000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA), at room temperature for 1 h. The membranes were washed three times for 15 min each with TBST, treated with chemiluminescence detection reagent (Pierce Biotechnology, Rockford, IL), and protein bands were visualized on X-ray film. Quantification of the bands was performed using a computer-driven scanner and Quantity One software (Bio-Rad, Hercules, CA). For signal normalization, membranes were treated with stripping buffer [0.1 M glycine, pH 2.9, and 0.02% (wt/vol) sodium azide] and reprobed with rabbit pan-Akt (1:1,000 dilution; Cell Signaling Technology) that detects endogenous levels of total Akt regardless of posttranslational modifications such as phosphorylation and acetylation.

Statistical analysis.

Data are means ± SE of results from three or more determinations. Data were analyzed by one-way ANOVA followed by pairwise post hoc Tukey's t-test where P < 0.05 represents significant differences between experimental groups.

RESULTS

RvD1 treatment enhances acinar formation.

To determine whether RvD1 stimulates the rate of acinar formation under physiological or inflammatory conditions, Par-C10 cells were grown on GFR-Matrigel, and lumen formation rates were investigated. As shown in Fig. 1, A–C, untreated Par-C10 at 24 h cells have not formed a lumen yet; however, at 60 h (Fig. 1, D–F) they formed polarized acinar-like spheres with discernible lumens and apically localized ZO-1. These features were enhanced (e.g., intense ZO-1 staining and larger-sized lumen) in Par-C10 cells treated with RvD1 (100 ng/ml) (Fig. 1, G–I). In contrast, cells treated with the proinflammatory cytokine TNF-α (100 ng/ml) exhibited no lumen formation and ZO-1 staining was shown in the center of the collapsed acinar sphere (Fig. 1, J–L). Strikingly, cells cotreated with RvD1 and TNF-α recovered lumen formation and ZO-1 organization similar to untreated or RvD1-treated cells at 60 h (Fig. 1, M–O). As summarized in Fig. 1P, quantification of lumen size indicates that cells treated with RvD1 alone formed larger-sized lumens compared with untreated cells. Furthermore, cells treated with TNF-α did not form a lumen; however, lumen formation was restored when cells were treated with RvD1+ TNF-α. A demonstrative view of the later effect can be seen in the supplemental material (Video 1; supplemental material can be found with the online version of this article). These results indicate that RvD1 not only blocks the disruption of acinar formation caused by TNF-α but also enhances cell polarity in Par-C10 acinar-like spheres.

Fig. 1.

Resolvin D1 (RvD1) treatment enhances acinar formation in Par-C10 cells grown on growth-factor-reduced (GFR)-Matrigel. Par-C10 cells grown on GFR-Matrigel in 8-well chambers as described in materials and methodsand incubated in the absence (A–C, 24 h and D–F, 60 h) or presence of RvD1 (100 ng/ ml; G–I, 60 h), TNF-α (100 ng/ml; J–L, 60 h), or RvD1 and TNF-α (100 ng/ml; M–O, 60 h) added at plating. Acinar spheres were subjected to immunofluorescence by using goat anti-rabbit anti-ZO-1 (A, D, G, J, and M; green) followed by Hoechst nuclear stain (B, E, H, K, and N; blue). The xy cross section images were obtained and analyzed using a Carl Zeiss 510 confocal microscope. Lumen sizes were quantified and expressed as means ± SE of results from 3 or more experiments (P), where *P <0.05 indicates significant differences from control cells.

Since RvD1 treatment enhanced polarized acinar formation on cells grown on GFR-Matrigel, we investigated whether cell polarity was also affected by RvD1 treatment in the absence of an extracellular matrix. Therefore, Par-C10 cells were grown on permeable supports, and TJ organization was investigated. As shown in Fig. 2, A, G, M, and S at 24 h, Par-C10 cells started to attach to the permeable supports and formed little islands of monolayers with no signs of cell polarization (e.g., absence of ZO-1 staining). However, at 60 h, islands of Par-C10 cells were bigger, and cell monolayers untreated or treated with RvD1 began to show signs of cell polarity (e.g., intense ZO-1 staining in an organized fashion) (Fig. 2, B, H, and T), in particular, cells treated with RvD1 alone (Fig. 2I) exhibited higher ZO-1 organization (i.e., a more intense staining and a more organized punctate pattern) than untreated cells (Fig. 2C, arrowheads). TNF-α-treated cells (Fig. 2O), display a lower intensity staining for ZO-1 and a less organized pattern than untreated or RvD1-treated cells; however, these effects were prevented when cells were coincubated with RvD1 and TNF-α (Fig. 2U). Furthermore, a pixel quantification analysis of the TJ areas from confocal images indicate that RvD1-treated cells display a significantly higher ZO-1 fluorescence intensity compared with untreated cells, while TNF-α-treated cells display a significantly lower pixel density for ZO-1 compared with RvD1 + TNF-α treatment group (Fig. 2Y), although RvD1-treated cells exhibited a stronger ZO-1 fluorescence intensity in both monolayers and acinar-like spheres. Because cell density was lower in TNF-α-treated monolayers than cells cotreated with RvD1 and TNF-α or treated with RvD1 alone, we controlled for cell density by equal protein loading in a Western blot analyses of Par-C10 cell monolayers. Our results indicated that ZO-1 protein expression was not altered by RvD1 treatment (Fig. 2 Z–Z1), suggesting that RvD1 did not provoke a significant upregulation of the ZO-1 protein but was able to enhance the organization of this protein within the epithelial cell junction (e.g., apically located ZO-1 to the regions of the junctional complex in Par-C10 cell monolayers) in salivary epithelium. Interestingly, RvD1 was able to block a downregulation of ZO-1 expression caused by TNF-α (Fig. 2, Z–Z1).Together, these results indicate that RvD1 treatment enhances cell polarity and blocks TNF-α-induced disruption of TJ integrity independent of an extracellular matrix.

Fig. 2.

RvD1 treatment enhances monolayer formation in Par-C10 cells grown on permeable supports. Par-C10 cells were cultured on permeable supports, as described in materials and methodsand incubated in the absence (A–F) or presence of RvD1 (100 ng/ ml; G–L), TNF-α (100 ng/ml; M–R), or RvD1 and TNF-α (100 ng/ml; S–X) added at plating for 24 h and 60 h. Monolayers were subjected to immunofluorescence using goat anti-rabbit anti-ZO-1 (C, I, O, and U; green) followed by phalloidin staining (D, J, P, and V; red) and Hoechst nuclear stain (E, K, Q, and W; blue). The xy cross sections images were obtained and analyzed using a Zeiss AxioImager fluorescence microscope. ZO-1 fluorescence intensity of defined tight junction (TJ) areas was quantified using ImageJ NIH software and expressed as means ± SE of results from 3 or more experiments (Fig. 2Y), where *P or **P < 0.05 indicates significant differences from control cells. ZO-1 expression was detected by Western blot analysis as described in materials and methods (Z). Data represent the means ± SE of results from 3 experiments (Z1), where *P < 0.05 indicates significant differences from untreated and RvD1-treated cells. Arrowheads indicate ZO-1 expression. Arrows indicate actin stress fibers.

Because TJ disorganization occurs concomitantly with activation of the cytoskeleton machinery (1, 65), F-actin organization was studied. As shown in Fig. 2D, F-actin staining in untreated Par-C10 cell monolayers were densely arrayed in perijunctional rings typical of polarized epithelium. However, cells treated with RvD1 alone, showed thick filamentous structures (arrow) that resembled actin stress fibers, indicative of actively migrating cells (Fig. 2J). Par-C10 cells cultured in the presence of TNF-α displayed thin actin filaments without particular orientations and a pattern of diffuse distribution for F-actin (Fig. 2P). The effect of TNF-α on F-actin was attenuated in cells treated with RvD1 and TNF-α (Fig. 2V). These results indicate that RvD1 is likely to induce migration in Par-C10 cells and reverses the cytoskeleton disruption caused by TNF-α.

RvD1 treatment enhances monolayer formation in Par-C10 cells.

Given that both F-actin and ZO-1 are potential regulators of TJ permeability and they are thus integral to defining the characteristics of columnar epithelial barrier function (45, 69), we assessed whether treatment with RvD1 with or without TNF-α caused changes in TER during Par-C10 cell monolayer formation. At 60 h, Par-C10 cells treated with RvD1 alone displayed the highest TER values, in particular, a significant difference was observed compared with untreated cells. Thus RvD1-treated group reflects the most polarized phenotype of all treatment groups (Fig. 3, A and B). However, at 5 days of plating (i.e., when cells reach confluence) the TER values were similar in untreated and RvD1-treated cells (Fig. 3C). Par-C10 cells treated with TNF-α (100 ng/ml) alone showed the lowest TER values throughout the entire period of monolayer formation (Fig. 3A). These values were significantly higher in cells incubated with a combination of RvD1 and TNF-α or with RvD1 alone at 60 h (Fig. 3B). The TER values in TNF-α-treatment group were lower than RvD1 alone and RvD1 and TNF-α treatment groups at 5 days (Fig. 3C). These results indicate that exposure of salivary gland epithelium to RvD1 accelerates monolayer formation and enhances barrier function in salivary epithelium.

Fig. 3.

RvD1 increases transepithelial resistance (TER) in untreated and TNF-α-treated Par-C10 cell monolayers. A: Par-C10 cells were cultured on permeable supports as described in materials and methodsand exposed to TNF-α (100 ng/ml) and/or RvD1 (100 ng/ml) at plating. TER was measured at different times during monolayer formation until confluence was reached. Following subtraction of medium resistance (120 Ω), tissue resistance was multiplied by the effective membrane area (1.13 cm²) and expressed as means ± SE of results from 3 or more experiments. B: results at 60 h are shown in a bar graph, where *P < 0.05 indicates significant differences from untreated cells, **P < 0.05 indicates significant differences between TNF-α and TNF-α + RvD1-treated cells, and ***P < 0.05 indicates significant differences between TNF-α and RvD1-treated cells. C: results at 5 days are shown in a bar graph, where *P < 0.05 indicates significant differences from untreated cells, **P < 0.05 indicates significant differences between TNF-α and RvD1-treated cells, and n.s. indicates nonsignificant differences.

RvD1 enhances migration but not proliferation in Par-C10 cells.

Because RvD1 alone was able to improve salivary epithelial barrier properties, we evaluated the role of RvD1 on other functions related to acinar formation including cell migration and proliferation. Migration was evaluated by chemotaxis and chemokinetic studies. For chemotaxis, Par-C10 single cells were seeded in the top compartment of a modified Boyden chamber, while varying concentrations of RvD1 were added to the bottom of the chamber. As shown in Fig. 4A, RvD1 induced a dose-dependent increase in the migration of Par-C10 cells compared with untreated cells. This migratory response was maximal at an RvD1 concentration of 100 ng/ml, (Fig. 4A), indicating that this ligand acts as a potent chemoattractant in Par-C10 cells. To determine whether RvD1 and/or TNF-α treatment affect cell chemokinesis, Par-C10 cells were plated on a lawn of microscopic fluorescent beads in collagen-coated 96-well plates. As shown in Fig. 4B, Par-C10 cells treated with RvD1 (100 ng/ml) or with FBS (10%) induced long tracks in a 24-h period indicating high migratory activity. However, Par-C10 cells treated with TNF-α (100 ng/ml) displayed small and round phagokinetic tracks typical of restricted migration; Par-C10 cells treated with a combination of RvD1 and TNF-α displayed longer phagokinetic tracks compared with TNF-α-treated cells, indicating restoration of migratory activity when RvD1 is present. Thus, the RvD1 treatment induced chemotaxis and chemokinesis and blocked the confined migratory response caused by TNF-α. To determine whether the ALX/FPR2 receptor was involved in RvD1-mediated effects, we incubated Par-C10 cells with TNF-α followed by the addition of the ALX/FPR2 receptor antagonist Boc-2 for 30 min before RvD1 treatment. As shown in Fig. 4C, the ALX/FPR2 receptor antagonist Boc-2 (10 μM) blocked the ability of RvD1 to restore migration in TNF-α-treated cells. Boc-2 also abolished cell migration in RvD1-treated cells compared with cells treated in the absence of the antagonist (Fig. 4C). To further analyze whether ALX/FPR2 was involved in RvD1-induced cell migration, we incubated Par-C10 cells with Boc-2 for 30 min before RvD1 treatment and performed a chemotaxis assay. As shown in Fig. 4D, Boc-2 significantly blocked RvD1-induced chemotaxis of Par-C10 cells compared with cells treated in the absence of the antagonist. These data indicate that RvD1-mediated migratory responses are likely due to activation of the ALX/FPR2 receptor.

Fig. 4.

RvD1 promotes organized cell migration in Par-C10 cells. A: chemotaxis of Par-C10 cells. Cells (5×10⁴) were seeded into the upper chamber of transwells. Lower chambers contained serum-free medium with or without RvD1 (10–100 ng/ml). Cell migration was evaluated 24 h after RvD1 stimulation and is expressed as the number of cells that moved across the transwell membranes in response to RvD1 compared with untreated controls. B: chemokinetic movement of Par-C10 cells. Cells were plated on a lawn of microscopic fluorescent beads, treated with TNF-α (100 ng/ml) and/or RvD1 (100 ng/ml), serum FBS-containing (10%) or serum-free growth medium for 24 h and visualized using a Zeiss AxioObserver fluorescence microscopy system. Phagokinetic tracks (black) produced as the cells (white dots) moved across the lawn of beads indicate the magnitude of cell movement. The lines within the images are due to breaks in image continuity. Due to the highly homogenous background the stitching software utilized had difficulty aligning the images. C: effect of Boc-2 on chemokinetic movement in Par-C10. Cells were treated with or without TNF-α (100 ng/ml) and subsequently treated with or without the ALX/FPR2 receptor antagonist Boc-2 (10 μM) for 30 min, and then cells were incubated with or without RvD1 (100 ng/ml) and chemokinetic movement was analyzed as described in Fig. 4B. D: effect of Boc-2 on Par-C10 cell chemotaxis. Cells were treated with or without the ALX/FPR2 receptor antagonist Boc-2 (10 μM) for 30 min, and then cells were incubated with or without RvD1 (100 ng/ml), and chemotaxis was analyzed as described in Fig. 4A. *P < 0.05 indicates significant differences from RvD1-treated cells. E: cell proliferation of Par-C10 cells. Cells were incubated in the presence or absence of RvD1 and/or TNF-α for 24 h and bromodeoxyuridine (BrdU) incorporation assay was performed as described in materials and methods. Results from a representative of 3 experiments are shown.

Because RvD1 enhanced the rate of Par-C10 cell migration, its effects on cell proliferation were also evaluated. To address this question, Par-C10 cells were cultured in the presence or absence of RvD1 and/or TNF-α for 24 h and BrdU incorporation was monitored. As shown in Fig. 4E, no significant differences were observed between control and RvD1- and/or TNF-α-treated cell monolayers after 24 h, indicating that neither of these components differentially affected cell proliferation in Par-C10 cells.

Par-C10 cells express RvD1 receptor.

RvD1 binds the G protein-coupled receptor (GPCR) ALX/FPR2 to activate specific signaling pathways that resolve inflammation (40). To evaluate whether the ALX/FPR2 was expressed in Par-C10 cell monolayers, Western blot analysis was performed as described in materials and methods. Results illustrate a protein band with an apparent molecular weight of 44 kDa, indicating that the ALX/FPR2 receptor is expressed in Par-C10 cells. This receptor was also detected in cell lysates from fresh isolated cells from mouse submandibular gland (Fig. 5A). ALX/FPR2 was also detected in Par-C10 cell monolayers as well as mouse submandibular glands (SMG) by immunohistochemistry. In Par-C10 cells the ALX/FPR2 receptor was visualized both in the membrane and within the cytoplasmic compartment, while in mouse, SMG was visualized on the basolateral side of acinar cells (Fig. 5B, white arrowheads). In the presence of isotype control antibody, there was minimal diffuse staining in Par-C10 cell monolayers and in mouse SMG (data not shown).

Fig. 5.

RvD1 receptor is expressed in salivary epithelium. A: lysates were prepared from Par-C10 cell monolayers and submandibular gland (SMG) cells freshly isolated from C57BL/6 mice and expression of ALX/FPR2 receptor was detected by Western blot analysis. Results from a representative of 3 experiments are shown. B: cell monolayers and frozen sections of mouse submandibular glands were fixed and expression of ALX/FPR2 was detected in SMG (visualized on the basolateral side of acinar cells, arrowheads) using immunofluorescence microscopy with rabbit anti-ALX/FPR2 (green). Images were obtained and analyzed using a Carl Zeiss 510 confocal microscope. The left image is rabbit anti-ALX/FPR2 stain (green); the center image is Hoechst nuclear stain (blue), and the right image is the merged image (green/blue). Par-C10 cells were stimulated with carbachol (C; 100 μM), UTP (D; 100 μM), or RvD1 (E; 100 ng/ml), and changes in the Fura-2 fluorescence ratio recorded at 340 nm and 380 nm were monitored, as described in materials and methods. Results from a representative of 3 experiments are shown for each agonist. F: representative Western blots of Par-C10 cells serum starved for 18 h and treated with or without RvD1 (100 ng/ml) for the indicated times. Protein extracts were subjected to SDS-PAGE and immunoblotted for either p-Akt or pan-Akt. G: data are expressed as fold increases in Akt phosphorylation induced by RvD1, compared with untreated control and represent the means ± SE of results from 3 or more experiments.

To determine whether the ALX/FPR2 receptor is functional in Par-C10 cells, we studied whether RvD1 was able to induce intracellular calcium release. As shown in Fig. 5E, RvD1 was not able to stimulate intracellular Ca²⁺ mobilization in Par-C10 cells, although progressive increases in intracellular Ca²⁺ mobilization were observed in Par-C10 cells stimulated with the muscarinic agonist carbachol (100 μM) (Fig. 5C) or the purinergic agonist UTP (100 μM) (Fig. 5D), both agonists known to elicit calcium responses in Par-C10 cells (5). These results indicate that RvD1 signaling in Par-C10 cells does not activate the classic GPCR second messenger system. Since RvD1 induced Par-C10 cell migration (Fig. 5), and previous studies indicated that Akt regulates migratory responses in salivary cells (30, 50), therefore, we investigated whether RvD1 was able to induce Akt phosphorylation. As shown in Fig. 5F, Akt, a downstream target of PI3K exhibited a time-dependent RvD1-mediated phosphorylation that reached maximal values at 60 min. However, Akt phosphorylation in Par-C10 cells was minimal in the absence of RvD1 (i.e., at 0 min time point) (Fig. 5F). These data indicate that the ALX/FPR2 receptor is expressed in salivary epithelium, and that RvD1 is able to elicit cell responses through a PI3K component of this GPCR signaling pathway.

RvD1-mediated cell migration depends on PI3K signaling pathway.

Because RvD1 was able to phosphorylate Akt (Fig. 5F), and this molecule is linked to the PI3K signaling pathway (13, 39, 48, 53, 72), we determined whether RvD1-mediated Par-C10 cell migration was dependent on PI3K and Akt activation. As shown in Fig. 6A, the selective PI3K inhibitor LY294002 (10 μM) significantly decreased the ability of RvD1-treated Par-C10 cells to migrate compared with cells treated in the absence of the inhibitor (see Fig. 4B). To further analyze whether PI3K activity was involved in RvD1-induced cell migration, we performed a chemotaxis assay in the presence or absence of LY294002. As shown in Fig. 6B, LY294002 (10 μM) almost completely abolished RvD1-induced migration of Par-C10 cells. Thus, these data indicate the presence of a PI3K-dependent migratory response activated by RvD1 in Par-C10 cells. To conclusively determine whether Akt was the downstream molecule responsible for RvD1-induced cell migration, Par-C10 cells were transfected with siRNAs to inhibit expression of the endogenous Akt. As shown in Fig. 7A, inhibition of Akt expression with the corresponding siRNAs blocked RvD1-stimulated cell migration, whereas nonspecific siRNA or Lipofectamine treatment alone did not affect RvD1-stimulated Par-C10 cell migration. To confirm that cell migration was affected, chemotaxis assay was also performed. As shown in Fig. 7B, cells transfected with Akt siRNA lost their ability to migrate in response to RvD1, whereas cells transfected with nonspecific silencing RNA or mock-transfected cells did not lose their ability to migrate in response to RvD1. The extent of suppression of Akt mRNA expression by the siRNAs is shown in Fig. 7C. These results indicate that Akt is the primary signaling molecule responsible for RvD1-mediated cell migration in Par-C10 cells.

Fig. 6.

The selective PI3K inhibitor LY294002 blocks RvD1-mediated Par-C10 cell migration. A: Par-C10 cells were incubated with or without LY294002 (10 μM) for 30 min and chemotaxis assay was performed as described in Fig. 4. *P < 0.05 indicates significant differences from LY294002 + RvD1-treated cells. Par-C10 cells treated (C) with or (B) without LY294002 were incubated with RvD1 (100 ng/ml) in serum-free growth medium for 24 h and chemokinetic movement of cells was performed as described in Fig. 4. Results from a representative of 3 experiments are shown. As stated for Fig. 4, the lines within the images are due to breaks in image continuity. Due to the highly homogenous background the stitching software utilized had difficulty aligning the images.

Fig. 7.

RvD1-mediated cell migration in Par-C10 cells is decreased by Akt siRNA. A: Par-C10 cells were transfected with small interfering Akt RNA (100 nM) using Lipofectamine RNAiMAX, and chemotaxis assay was performed as described in Fig. 4. Cells transfected with nonspecific siRNA or treated with Lipofectamine alone served as controls. Par-C10 cells transfected with siRNAs or treated with RNAiMAX were incubated with RvD1 (100 ng/ml) in serum-free growth medium for 24 h, and chemokinetic movement of cells was performed and (B) quantitated as described in materials and methods. Results from a representative of 3 experiments are shown. C: total Akt expression was detected by Western blot analysis as described in materials and methods. Data represent the means ± SE of results from 3 experiments, where siRNA suppressed 55% of total protein expression. Results from a representative experiment are shown at the top of the figure.

DISCUSSION

Previous studies indicate that RvD1 exhibits potent anti-inflammatory and proresolving properties (19, 35, 40, 63, 67, 71), these responses include the inhibition of PMN migration and shortening the resolution phase of acute inflammation (8); RvD1 also reduced the expression of inflammatory signaling molecules and PMN transmigration across choroid retinal endothelial cell barriers (67). Furthermore, in peritonitis and ischemia-reperfusion injury, RvD1 attenuated PMN migration and tissue injury (37), and more recently RvD1 was shown to reduce conjunctival goblet cell secretion stimulated by leukotriene, which is an important component of ocular allergy and early dry eye (15). Here, we demonstrate that RvD1 not only blocks inflammatory responses caused by the proinflammatory cytokine TNF-α, but also is able to restore and enhance tissue architecture in Par-C10 cells, suggesting a new anti-inflammatory and proresolving role for RvD1 in salivary epithelium.

RvD1 effects on enhancing and restoring tissue integrity are shared by other lipid mediators, for instance, topical application of RvE1 (a member of the Rv family derived from EPA) in rabbit periodontitis improved regeneration of periodontal bone and prevented the progression of tissue destruction (27); RvE1 also decreased inflammation and enhanced corneal epithelial integrity in a dry eye mouse model (42). Other studies indicate that lipoxin A₄ (a lipid mediator derived from arachidonic acid) and neuroprotectin D1 (PD1, a lipid mediator derived from DHA) decreased proinflammatory chemokine production and accelerated corneal re-epithelialization in mouse eyes (23) and more recent findings demonstrate that lipoxin A₄ significantly increased TJ expression and barrier function in human bronchial epithelial cells (24). Thus, resolvins, lipoxins, and protectins, not only terminate inflammatory events but also allow tissue repair, supporting the notion that the resolution of inflammation is an active process.

Previous studies using a β-arrestin-based ligand receptor system indicated that RvD1 selectively activates two separate GPCRs, such as ALX/FPR2 and GPR32 (40). ALX/FPR2 has been identified in human (20), mouse (64), and rat (11) tissues. Our results show that Par-C10 cells as well as submandibular cells express the ALX/FPR2 receptor (Fig. 5, A–D). Previous studies have also indicated that the ALX/FPR2 receptor is expressed in other tissues such as in human intestinal and pulmonary epithelia (11, 12, 20, 64). Par-C10 cells did not express the GPR32 receptor (data not shown), which is not surprising, considering that GPR32 receptor is a pseudogene in the rat and the murine counterparts of human GPR32 remain unknown (25). Studies to characterize ALX/FPR2 receptor activity in Par-C10 cells indicated that RvD1 does not directly evoke Ca²⁺ mobilization (Fig. 5E), consistent with previous studies in phagocytes, indicating the absence of RvD1-induced classical GPCR signaling responses such as intracellular calcium mobilization or generation of cAMP (40). However, RvD1 induced phosphorylation of the serine/threonine protein kinase Akt at the serine residue 473 (Fig. 5F) and suggest the ability of RvD1 to signal in Par-C10 likely through the ALX/FPR2 receptor. These results are consistent with a previous study indicating that Akt phosphorylation is enhanced in adipose tissue and in the vasculature of mice treated with RvD1 (28). GPCRs, can readily activate an epitope-tagged form of Akt kinase in COS-7 epithelial cells (47). The signal-transducing molecules, βγ complexes and Gα_q and Gα_i subunits, generated upon GPCR activation can effectively promote Akt activation in a PI3K-dependent manner (47). Since the ALX/FPR2 receptor is likely coupled to Gα_i/o proteins (40), it is possible that activation of Akt by this receptor occurs through Gα_i/o proteins; however, future studies are necessary to confirm this notion. Receptor-mediated activation of the PI3K pathway occurs through the recruitment of the p85 regulatory subunit of PI3K via its Src-homology 2 domains to phophotyrosine residues located within the receptor (18). After being recruited to the plasma membrane, the p110 catalytic subunit of PI3K phosphorylates phosphatidylinositol-4,5-bisphosphate (PIP₂) and generates PIP₃ (44). The resulting PIP₃ serves to recruit phospholipid-binding domain-containing proteins to the plasma membrane (31). In particular, Akt and phosphoinositide-dependent kinase 1 (PDK1) are recruited to the membrane via their plekstrin homology (PH) domains (51).The direct homodimerization of the two PH domains between Akt and PDK1 might also mediate protein proximity and subsequently phosphorylate Thr308 in Akt, which stabilizes the activation loop in an active conformation and renders Ser473 phosphorylation by the rapamycin-insensitive mTORC2, resulting in full activation of Akt kinase (2). mTORC2, therefore, is the elusive PDK2 molecule for Akt/PKB (2). mTORC2 has been shown to function as an important regulator of the cytoskeleton through its stimulation of F-actin stress fibers, paxillin, RhoA, Rac1, Cdc42, and protein kinase C-α (PKCα) (56).

The RvD1 effects on cell migration were dose dependent and were mediated through the ALX/FPR2 receptor (Fig. 4, C and D), PI3K (Fig. 6), and Akt signaling (Fig. 7). Specifically in our study, the ALX/FPR2 antagonist Boc-2 blocked RvD1 from restoring TNF-α-mediated disruption of cell migration (Fig. 4C) as well as RvD1-mediated effects on cell migration (Fig. 4, C and D). Previous studies indicated that Boc-2 blocked increases in cell proliferation induced by lipoxin A₄ in breast tumor cell lines (38). Boc-2 also blocked the ability of 15-epi-lipoxin A₄ to reduce PMN adherence to IL-1β-stimulated endothelial cell (46). Boc-2 has shown antagonist activity at both FPR1 and FPR2, and hence is considered a nonselective antagonist (52); therefore, future studies will be necessary to confirm whether ALX/FPR2 receptor alone is involved in RvD1-mediated cell migration.

Regarding the ALX/FPR2 downstream signaling molecules, blocking PI3K activation with LY294002, a potent and selective PI3K inhibitor, also prevented RvD1-induced cell migration. Furthermore, Akt gene silencing almost completely blocked the ability of Par-C10 cells to migrate. Activation of PI3K activity alone is sufficient to remodel actin filaments to increase cell migration through the activation of Akt in chicken embryo cells (53). Akt also has been shown to be involved in the migratory responses of a variety of cell types (13, 17, 72), including salivary glands (26). Thus, it is logical to suggest that the ALX/FPR2 receptor and its downstream signaling molecules PI3K and Akt are important for RvD1-mediated migration in Par-C10 cells likely by activating actin stress fiber formation (Fig. 8).

Fig. 8.

Proposed signaling mechanisms mediated by RvD1 in Par-C10 cells. RvD1 binds to the ALX/FPR2 receptor and activates phosphatidylinositol 3-kinase (PI3K) and Akt signaling enhancing cell polarity (i.e., ZO-1 apical localization) and migration (i.e., actin stress fiber formation) and possibly blocking TNF-α signaling. Broken arrow indicates that mTORC2 is likely activated by PI3K. P indicates phosphorylated residue.

Regarding the proliferation rates between Par-C10 cells treated with TNF-α and RvD1, we did not observe significant differences; furthermore combination of the two ligands did not affect the rate of proliferation in Par-C10 cells plated on plastic (4E). Although cell density may vary depending on the extracellular matrix on which cells are plated (22) the effects of RvD1 versus TNF-α on cell proliferation are likely to be similar to cells plated on permeable supports or Matrigel. Nonetheless, these results indicate that TNF-α induced cell proliferation in Par-C10 cells at the same degree as RvD1, suggesting that TNF-α may cause proliferation but not organized migration and differentiation in Par-C10 cells.

RvD1 enhanced ZO-1 organization and increased TER in Par-C10 cells (Figs. 2 and 3), suggesting that RvD1 enhances cell polarity and epithelial integrity (Fig. 8). Both TJ and paracellular ion channels may contribute to TER (66), and, therefore, it is likely that RvD1 affects both parameters. Understanding whether RvD1 affects the unit composition of paracellular-tight junction channels and the overall molecular architecture of the TJ is challenging and remains to be determined. Previous studies indicated that lipoxin A₄ significantly increased ZO-1, claudin-1, and occludin protein expression at the plasma membrane of confluent human bronchial epithelial cells (24). Furthermore, lipoxin A₄ stimulated monolayer formation rates of the human airways epithelial cell monolayer as measured by TER (24). However, in our studies, RvD1 did not have an effect on ZO-1 expression, but it was able to block the TNF-α-mediated ZO-1 downregulation (Fig. 2, Z–Z1). The TNF-α effects on ZO-1 expression are only observed when this cytokine is added at plating but not after cells achieve confluence (5). These studies indicate that lipoxin A₄ and RvD1 may share similar effects on epithelial integrity and cell polarity; moreover, both ligands activate the ALX/FPR2 receptor (40), suggesting a common signaling pathway and a similar effect on migration and tissue repair for lipoxin A₄ and RvD1.

Resolution of inflammation is an active process with many control points, regulated by a unique class of chemical mediators that are anti-inflammatory and proresolving (3, 4). Our previous studies indicate that chronic exposure of Par-C10 cells to inflammatory cytokines, such as TNF-α and/or IFN-γ, alters TJ integrity leading to secretory dysfunction (5, 7). RvD1 enhanced lumen formation and assembly of TJ and blocked inflammatory signals mediated by TNF-α (Figs. 1–3). Since the proinflammatory cytokine TNF-α is involved in the pathogenesis of Sjögren's syndrome (21), the effects of RvD1 on blocking this cytokine could be attributable to its anti-inflammatory role. Furthermore, activation of the PI3K/Akt pathway observed in Par-C10 cells (Figs. 5–7) suggests that RvD1 may also activate prosurvival pathways since in salivary glands, the PI3K pathway has been shown to protect salivary glands from programmed cell death (43) and might explain the RvD1-mediated blocking of TNF-α, although, this mechanism was not further investigated. The RvD1 effects may be also due to activation of nonnuclear pathways such as mTORC2, which seems to promote the organization of the actin cytoskeleton and cell migration (16, 33). These proposed signaling pathways are shown in Fig. 8.

In summary, this study reveals for the first time that RvD1 effects in salivary epithelium enhances 1) organized cell migration during the first 60 h of monolayer formation, 2) enables cell polarity after 60 h, and 3) reduces the actions of inflammatory signaling molecules induced by TNF-α. Our studies provide a new insight into RvD1-mediated tissue repair besides its anti-inflammatory effects similar to other resolvins effects seen in conditions, such as colitis (4), cornea inflammation (42, 73), and periodontitis (27). These studies may lead to better therapeutic strategies for minimizing autoimmune-associated inflammation of salivary gland that contributes to secretory dysfunction in Sjögren's syndrome.

GRANTS

This work was supported by National Institute of Dental and Craniofacial Research Grant R21-DE19721-01A1.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: O.O., S.C., A.M., M.E.D., and O.J.B. performed experiments; O.O., A.M., and O.J.B. analyzed data; O.O., A.M., and O.J.B. prepared figures; O.O., M.E.D., and O.J.B. edited and revised manuscript; O.O., S.C., A.M., M.E.D., and O.J.B. approved final version of manuscript; O.J.B. conception and design of research; O.J.B. interpreted results of experiments; O.O. drafted manuscript.