|Home | About | Journals | Submit | Contact Us | Français|
Unresolved inflammation and tissue destruction are underlying mechanisms of periodontitis, which is linked to dysregulated polymorphonuclear neutrophil (PMN) functions. Lipoxin A4 (LXA4) is a specialized proresolving lipid mediator (SPM) that dampens excessive inflammation, promotes resolution, and protects from leukocyte-mediated tissue damage. Human periodontal ligament stem cells (hPDLSCs) represent key players during tissue regeneration and may contribute to resolution of inflammation; thus, they may represent a promising tool in regenerative dentistry. In the present study, we investigated the actions of hPDLSCs on PMN apoptosis and antimicrobial functions, and determined the impact of LXA4 on hPDLSCs. hPDLSCs significantly reduced apoptosis and stimulated microbicidal activity of human PMNs, via both cell-cell interactions and paracrine mechanisms. Lipid mediator metabololipidomics analysis demonstrated that hPDLSCs biosynthesize SPMs, including resolvin D1, D2, D5, and D6; protectin D1; maresins; and LXB4; as well as prostaglandins D2, E2, and F2α. LXA4 significantly enhanced proliferation, migration, and wound healing capacity of hPDLSCs through the activation of its cognate receptor ALX/FPR2, expressed on hPDLSCs. Together, these results demonstrate that hPDLSCs modulate PMN functions, and provide the first evidence that stem cells generate SPM and that the LXA4-ALX/FPR2 axis regulates regenerative functions of hPDLSCs by a novel receptor-mediated mechanism.
These findings uncovered unappreciated features of stem cells from the periodontal ligament, supporting the notion that these cells may act as master regulators of pathophysiological events through the release of mediators that promote the resolution of inflammation and bacterial killing. The study also demonstrated that it is possible to modulate important functions of periodontal stem cells using lipoxin A4, a potent endogenous stop signal of inflammation. Thus, this study revealed an unappreciated anti-inflammatory proregenerative circuit that may be exploited to combat periodontal pathologies using resident stem cells. Moreover, the data may represent a more general template to explain the immunomodulatory functions of stem cells.
Acute inflammation is a protective host response, which, if timely and self-limited, leads to elimination of pathogens and to homeostasis restoration [1, 2]. In oral tissues, excessive or unresolved pathogen-induced inflammation causes tissue injury and periodontitis (PD) [3, 4]. Hence, while PD etiology is bacterial, a failure to resolve inflammation is a pathophysiological determinant in PD progression . PD is associated with cardiovascular disease ; therefore, the need for efficacious treatment of periodontal inflammation is urgent.
Resolution of acute inflammation is an active and highly coordinated response, involving endogenous proresolving mediators, biosynthesized locally, that control leukocyte trafficking and restore homeostasis [7, 8]. Among them, specialized proresolving lipid mediators (SPMs), derived from polyunsaturated fatty acids (PUFAs), including lipoxins (LXs), resolvins (Rvs), protectins, and maresins, perform multipronged actions that improve the outcome of inflammation-related pathologies in experimental models and clinical trials [8–11]. LXA4 and its isomer 15-epi-LXA4 are biosynthesized from arachidonic acid (AA) by transcellular mechanisms during inflammation resolution [12, 13]. They control polymorphonuclear neutrophil (PMN) infiltration into inflamed tissues and attract monocytes/macrophages, enhancing efferocytosis and bacterial phagocytosis, hallmarks of resolution . These actions are mediated by the activation of a specific G-protein-coupled receptor (GPCR) termed ALX/FPR2 [15, 16]. LXA4 is found in the gingival crevicular fluid and in peripheral blood of patients with PD [17, 18]. Moreover, SPMs combat excessive inflammation and tissue destruction in experimental PD . Along these lines, LXA4, 15-epi-LXA4, and related synthetic stable analogs limit PMN infiltration and activation triggered by microorganisms in vitro [17, 19] and in an animal model of ligature plus infection-induced PD, improving the clinical outcome . In addition, LXA4 significantly reduces PD-related bone loss , supporting the relationship between nonresolving inflammation and tissue destruction in PD, and providing a rational approach to stimulate wound healing and tissue regeneration. Nevertheless, it is unclear whether the protective actions of LXs in PD include, in addition to anti-PMN-driven periodontal damage, the stimulation of proregenerative activities of cells in the periodontium.
Periodontal ligament (PDL), the soft connective tissue involved in teeth stability, homoeostasis, and repair [21, 22], contains a niche of multipotent stem cells [23, 24] that retain tissue regenerative capacity in adulthood [21, 22]. Human periodontal ligament stem cells (hPDLSCs) possess self-renewal and multilineage (osteo-, chondro-, odonto-, adipo-, and neurogenic) differentiation capability [24, 25]. Like bone marrow mesenchymal stem cells (BM-MSCs), hPDLSCs express mesenchymal antigens as well as cementoblastic/osteoblastic markers . Because of these properties, hPDLSCs have been used in preclinical studies to assess their periodontal healing capability [26, 27]. Along these lines, the potential therapeutic applications of MSCs in regenerative medicine have been broadened by their unique immunomodulatory properties [28, 29] via cell-cell contact and/or paracrine mechanisms [30, 31]. Similarly to BM-MSCs, hPDLSCs, expanded ex vivo, possess paracrine immunosuppressive capability by inhibiting peripheral blood mononuclear cell (PBMC) proliferation . Whether they also form and release proresolution mediators remains to be determined.
In the present study, we provide evidence that hPDLSCs regulate PMN functions related to inflammation resolution and bacterial killing, and demonstrate that they biosynthesize SPMs. Moreover, we show that LXA4 stimulates receptor-dependent functions of hPDLSCs relevant to tissue healing. Together, these results contribute to the elucidation of novel pathways that can be exploited for stem cell-based treatment of PD.
The protocol and informed consent to use human periodontal ligament biopsy specimens were approved by the medical ethics committee at the Medical School, “G. D’Annunzio” University of Chieti-Pescara, Italy (no. 266/17.04.14). PDL biopsy specimens were obtained from healthy subjects (ages 20–35 years) without systemic and oral diseases who signed an informed consent form before specimen collection.
hPDLSCs were isolated according to Trubiani et al. . PDL tissues were incubated with MSC basal medium supplemented with a specific bullet kit (Lonza Group, Walkersville, MD, https://www.lonza.com) at 37°C in 5% carbon dioxide (CO2) for 20 days, during which cells spontaneously migrated from the explants. Stemness surface molecules and pluripotency-associated markers were evaluated, as previously described . The hPDLSC phenotype was determined by flow cytometry using the following antibodies: fluorescein isothiocyanate (FITC)-conjugated CD13, CD14/FITC, phycoerythrin (PE)-conjugated CD29, CD31/FITC, peridinin chlorophyll protein (PerCP)-cyanine (Cy) 5.5-conjugated CD34, CD44/FITC, CD45/FITC, CD73/PE, CD90/FITC, CD105/PE, CD106/FITC, allophycocyanin (APC)-conjugated CD117, CD133 /PE, CD144/FITC, CD146/PE, CD166/FITC, CD326/PerCP-Cy5.5, Alexa488-conjugated human leukocyte antigen (HLA)-ABC, HLA DR/PE (BD Bioscience, Franklin Lakes, NJ, https://www.bdbiosciences.com). Pluripotent embryonic markers NANOG/PE, octamer-binding transcription factor 4 (Oct4)/PE, stage-specific embryonic antigen (SSEA4)/FITC, and (sex determining region Y)-box 2 (Sox2)/FITC were purchased from BD Bioscience. Cells were analyzed with a FACSCalibur flow cytometer (BD Bioscience), using the Cellquest (BD Bioscience) and FlowJo (Ashland, OR, http://www.flowjo.com) software. Unstained cells were the negative control.
To assess their colony-forming unit (CFU) capability, hPDLSCs were seeded at 1 × 103 cells per 10 ml onto 100-mm culture dishes in growth medium (Lonza Group) kept at 37°C in 5% CO2. After 14 days, cells were fixed with cold methanol, stained with 0.5% crystal violet, washed with distilled water, and dried. Aggregates of 50 or more cells were scored as 1 colony forming unit-fibroblast.
The formation of a mineralized bone matrix was monitored for 21 days using human MSC (hMSC) osteogenic differentiation medium (Lonza Group), as previously described . Osteogenic differentiation was determined by staining mineralized deposits with Alizarin Red S (Sigma-Aldrich, Milan, Italy, https://www.sigmaaldrich.com) and by assessing the expression of osteogenic markers (i.e., alkaline phosphatase [ALP], runt-related transcription factor 2 [RUNX2], and osteopontin [OPN]) by real-time polymerase chain reaction (PCR).
To drive hPDLSC chondrogenic differentiation, cells were grown in three-dimensional (3D) cultures incubated for 21 days with chondro-inductive medium (CIM) made of high-glucose Dulbecco’s modified Eagle’s medium (DMEM) containing 1 μM dexamethasone, 1 μM ascorbate-2-phosphate (Sigma-Aldrich), 1% sodium pyruvate (Sigma-Aldrich), 10% insulin-transferrin-selenium 100× concentration (Thermo Fisher Scientific, Waltham, MA, https://www.thermofisher.com), and 10 ng/ml transforming growth factor (TGF)-β3 (PeproTech, London, U.K., https://www.peprotech.com). After 3 weeks, cells were fixed, cut into 3-μm sections and stained with 1% (wt/vol) Alcian blue 8GX (Sigma-Aldrich) for histological examination. In parallel, hPDLSC monolayers were cultured in 6-well plates for 14–21 days with CIM to assess chondrogenic-related gene expression (see below) and chondrogenic matrix deposition using Alcian blue. Adipogenesis was induced using the hMSC adipogenic differentiation medium (Lonza Group) for 4 weeks. Lipid-drop accumulation and triglyceride content were respectively evaluated by Red Oil staining and the AdipoRed assay (Lonza Group) . All experiments were carried out with hPDLSCs at passages 3–7.
PMNs were isolated from healthy volunteers (25–40 years old), who denied taking medication at least for 5 days, using dextran-histopaque sedimentation . Isolated PMNs were, on average, 93% viable, as determined by flow cytometric analysis of annexin V (annV) and propidium iodide (PI) staining (annexin V-FITC Apoptosis Detection Kit; eBioscience, San Diego, CA, https://www.ebioscience.com). For coculture experiments, PMNs were incubated with hPDLSCs at different ratios (1:1, 5:1, 10:1) for 18 hours at 37°C and 5% CO2. In some experiments, PMNs were cultured for 18 hours with hPDLSC-conditioned medium (CdM). In others, PMNs were exposed to human interleukin (IL)-8 (100 ng/ml) (PeproTech) for 30 minutes at room temperature (RT) before coculture with hPDLSCs. PMNs maintained in complete medium were used as controls.
To obtain CdM, hPDLSCs (passages 3–6) were seeded in 6-well plates at 3 × 105 cells per well in 2 ml of complete DMEM. Cells were cultured for 24, 48, and 72 hours at 37°C and 5% CO2. At each time point, CdM was collected, centrifuged to remove cellular debris, and snap frozen at −80°C until use.
Apoptosis of PMNs with or without hPDLSCs or hPDLSC-CdM was evaluated by flow cytometry using annexinV and PI (eBioscience) double staining. Approximately 104 cellular events were acquired for each sample.
The reference laboratory Pseudomonas aeruginosa strain PAO1 (ATCC 15692, ATCC, Manassas, VA, https://www.atcc.org) was grown (at 37°C overnight) on Muller-Hinton (MH) agar plates. At 24 hours before experiments were conducted, a single colony was picked up from the agar plate, inoculated in 5 ml of MH broth, and grown overnight under continuous agitation (300 rpm, 37°C). The following day, confluent PA01 was diluted (1:10) with fresh MH medium and grown for ~2 hours in a shaking incubator (300 rpm, 37°C) to midlog phase (optical density of 550 nm [OD550] = 0.45 ± 0.05), corresponding to ~2 × 108 CFU/ml). Bacteria (2 ml) were opsonized (30 minutes, 37°C) with 10% AB human serum (Sigma-Aldrich), adjusted to an OD550 of ~0.45 with Dulbecco’s phosphate buffered saline (PBS), and further diluted (1:5) with Roswell Park Memorial Institute (RPMI) medium containing 20% AB human serum. PMNs cocultured with hPDLSCs (10:1) for 18 hours were counted, washed twice with PBS, and suspended in antibiotic-free RPMI at a density of 106 cells per milliliter. Opsonized PA01 suspension (~3 × 107 CFU, 500 μl) was added to 500 µl of human PMNs (~6 × 105) and the mixtures were maintained in constant motion on a rotating wheel at 37°C. PA01 kept alone served as a control of maximal bacterial growth. After 90 minutes, aliquots of each mixture were removed, serially diluted with sterile PBS, and plated on MH agar to enumerate the viable CFUs. The PMN killing index was calculated as follows: [(CFU in the absence of PMN − CFU in the presence of PMN)/CFU in the absence of PMN × 100].
The concentration of epidermal growth factor (EGF), IL-10, IL-6, IL-8, and vascular endothelial growth factor (VEGF) in 24-hour cultured hPDLSC-CdM was determined using a Milliplex human cytokines/chemokines kit (EMD Millipore, Billerica, MA, https://www.emdmillipore.com) according to the manufacturer’s instructions. Briefly, 25 µl of medium was added to 25 µl of assay buffer, followed by 25 µl of antibody-coated magnetic beads incubated for 2 hours at RT during shaking. Plates were washed twice with buffer and incubated (1 hour, RT) with 25 µl of biotinylated antibodies followed by streptavidin-PE (30 minutes, RT). Plates were analyzed using a Luminex 100/200 platform (Luminex, Austin, TX, https://www.luminexcorp.com) equipped with the xPONENT 3.1 software. Standard curves for each analyte were generated using reference proteins. Analyte concentrations were determined by a five-parameter logistic curve. Analyses were carried out in duplicate and all incubation steps were performed in the dark.
hPDLSCs were cultured in 150-mm plates with 9 ml of phenol-red free medium until 80% confluence. After treatment, medium and cells were harvested, using a scraper in 15-ml tubes, snap frozen in an acetone dry ice bath, and rapidly stored at −80°C. Two volumes of ice-cold methanol containing 500 pg of deuterium (d)-labeled d4-LTB4, d5-LXA4, d5-RvD2, and d4-PGE2 (Cayman Chemical, Ann Arbor, MI, https://www.caymanchem.com) were then added and samples stored at 4°C for 45 minutes to allow for protein precipitation. Supernatants were collected after centrifugation (3,000 rpm, 10 minutes, 4°C) and brought to less that 10% of methanol content using nitrogen. All samples for liquid chromatography-tandem mass spectrometry (LC-MS-MS) analysis were extracted using C18 solid-phase extraction (SPE) columns (Waters, Milford, MA, https://www.waters.com), equilibrated with 6 ml of methanol and 12 ml of H2O. Sample supernatants were acidified (pH ~3.5) and immediately loaded onto the SPE column. Samples were washed with 12 ml of hexane, eluted with 12 ml of methyl formate, taken to dryness, and suspended in phase (1:1 methanol to water) for LC-MS-MS analysis. The LC-MS-MS system (QTrap6500; ABSciex, Framingham, MA, https://sciex.com) equipped with an Agilent HP1100 binary pump (Agilent Technologies, Santa Clara, CA, https://www.chem.agilent.com) and an automatic injector was setup and operated, as described by Colas et al. . To monitor and quantify the levels of the various lipid mediators (LMs), a multiple reaction monitoring method was developed with signature ion fragments for each molecule, also as described by Colas et al. .
Surface and intracellular expression of ALX/FPR2 was evaluated in hPDLSCs from different donors using a monoclonal anti-ALX/FPR2 primary antibody (Aldevron, Fargo, ND, https://www.aldevron.com). Briefly, 5 × 105 cells per sample were fixed with 4% paraformaldehyde (PFA), washed, and incubated with buffer containing primary antibody for 60 minutes at 4°C. After removing the excess of primary antibody, cells were incubated with an anti-mouse FITC-conjugated secondary antibody. For intracellular staining, cells were permeabilized with FACS lysing and Perm2 solutions (BD Bioscience) before staining. Secondary antibody-matched controls were used to assess unspecific fluorescence.
hPDLSCs grown on glass coverslips were fixed with 4% PFA, washed, and processed for ALX/FPR2 immunofluorescence. Cells were permeabilized with Triton X-100, neutralized with 5% skimmed milk in PBS for 30 minutes. Incubation with primary mouse anti-ALX/FPR2 antibody (Aldevron) (1:100) was followed by exposure to Alexa Fluor 488 green fluorescence conjugated goat anti-mouse (Molecular Probes; Thermo Fisher Scientific, Eugene, OR, https://www.thermofisher.com). For actin cytoskeleton and nuclei counter-staining, cells were incubated with AlexaFluor 594 conjugated phalloidin (Thermo Fisher Scientific), and TO-PRO (Thermo Fisher Scientific). Samples were analyzed using a Zeiss LSM510 META confocal system (Zeiss, Jena, Germany, www.zeiss.com) connected to an inverted Zeiss Axiovert 200 microscope equipped with a Plan Neofluor oil-immersion objective (40×/1.3 NA). Images were collected using an argon laser beam with excitation lines at 488 nm and a helium-neon source (543 nm and 665 nm).
hPDLSC total RNA was isolated with the RNeasy Plus Universal Mini Kit (Qiagen, Valencia, CA, https://www.qiagen.com) according to the manufacturer’s instructions. Quantitative PCR for osteogenic markers and ALX/FPR2 was performed with an ABI PRISM 7900 HT Sequence Detection System (Applied Biosystems, Foster City, CA, https://www.appliedbiosystems.com). TaqMan gene expression assays (ALP, Hs01029144_m1, RUNX2, Hs00231692_m1, OPN, Hs00959010_m1, fatty acid binding protein 4 [FABP4], Hs01086177_m1, peroxisome proliferator-activated receptor γ [PPARγ], Hs01115513_m1, and ALX/FPR2 Hs00265954_m1 from Applied Biosystems; and ACAN, Hs.PT.56a.742783 and COMP, Hs.PT.58.2694031 from Integrated DNA Technologies, San Jose, CA, https://www.idtdna.com) were used according to standard protocols. Beta-2 microglobulin (B2M, Hs99999907_m1; Applied Biosystems) was used for template normalization. Duplicates were set up for each sample and the mRNA expression was analyzed by the comparative 2-Ct relative quantification method .
hPDLSCs (2 × 103 cells per well) were seeded in 96-well microplates with complete DMEM overnight at 37°C and 5% CO2. The following day, cells were starved overnight with DMEM plus 1% fetal bovine serum (FBS), gently washed, and exposed to 0.001 to 1.0 nM LXA4 (Calbiochem, San Diego, CA, https://www.merckmillipore.com/) or vehicle (0.01% ethanol maximum) diluted with 1% FBS-containing DMEM. Cells were enumerated after 24, 48, and 72 hours using trypan blue exclusion and data were expressed as proliferation index calculated as ratio between LXA4-and vehicle-treated hPDLSCs. In some experiments, hPDLSCs were exposed to the ALX/FPR2 antagonist WRW4 (Calbiochem) (10 µM, 30 minutes) before LXA4, and the cellular proliferation index was calculated as ratio between LXA4- and WRW4-LXA4 treated cells. Each condition was tested at least in triplicate and experiments were repeated three times using hPDLSCs from different donors.
hPDLSCs (5 × 104 cells per well) were grown in 6-well plates in complete medium until 80% to 90% confluence and then starved overnight with DMEM plus 1% FBS. Wounds (~580 µm wide) were created by scraping the monolayers with a 200-µl pipette tip. Cell debris was removed by vigorous rinsing with PBS to avoid random reattachment. Thereafter, serum-free medium containing increasing concentrations of LXA4 (0.01 to 1 nM) or vehicle (0.01% ethanol, maximum) were added to each well. hPDLSCs maintained in DMEM plus 10% FBS served as a control of maximal healing capacity. Wounds were monitored at 0, 3, 6, 24, and 48 hours using an optical microscope equipped with a digital camera (Canon, Tokyo, Japan, https://www.canon.com). Each well was photographed at three locations along the original wound boundary and wound width was quantified in three fields using ImageJ software (U.S. National Institutes of Health, Bethesda, MD, https://imagej.nih.gov/ij). To examine the involvement of the ALX/FPR2 receptor in LXA4-regulated migration, wounded cells were exposed to the ALX/FPR2 antagonist WRW4 (10 µM, 30 minutes) before LXA4, and wound closure was monitored at 0, 3, 6, 24, and 48 hours. Duplicates for each condition were examined.
Data were expressed as mean ± SEM. Differences between groups were determined using the Student’s t test. Differences were considered significant when p < .05.
hPDLSCs from PDL biopsy specimens consisted of cells with a homogeneous fibroblast-like morphology (Fig. 1A), reminiscent of other MSCs (e.g., bone marrow)  and capable of generating fibroblast colonies from single cells after 14 days of culture (Fig. 1B). hPDLSCs expressed specific MSC antigens, such as CD13, CD29, CD44, CD73, CD90, CD105, CD117, CD146, and CD166. Furthermore, these cells expressed the histocompatibility antigen HLA-ABC as well as markers of pluripotency (NANOG, Oct4, SSEA4, and Sox2). They stained negative for the hematopoietic markers CD14, CD31, CD34, CD45, and HLA-DR as well as for surface vascular endothelial-cadherin (CD144) and epithelial cell adhesion molecule (CD326) (Fig. 1C). Thus, hPDLSCs met the minimal criteria for being classified as MSCs .
To evaluate hPDLSC differentiation, we induced osteogenic; chondrogenic, and adipogenic commitment. Under osteo-inductive conditions for 3 weeks, hPDLSCs produced mineralized extracellular matrix that was stained with Alizarin Red S and quantified (Fig. 1D, upper panel). Consistent with this, ALP and RUNX2 mRNA accumulated at early differentiation stages (7–14 days) to progressively decrease at 21 days when OPN was upregulated (Fig. 1D, lower panel).
After 21 days in chondrogenic conditions, the deposition of chondrogenic-like matrix was revealed by Alcian blue staining (Fig. 1E, upper panel), whereas real-time PCR showed a higher expression of chondrogenic-related genes as aggrecan (ACAN) and cartilage oligomeric matrix protein (COMP) compared with undifferentiated hPDLSCs (p < .01) (Fig. 1E, lower panel).
When hPDLSCs were subjected to adipogenic differentiation for 4 weeks, Oil Red O-positive lipid-droplet accumulation within cells was observed accompanied by an increase in triglyceride content compared with control cells (p < .0001) (Fig. 1F, upper panel). Consistent with this, hPDLSCs under adipogenic culturing conditions expressed high levels of adipogenic genes as FABP4 and PPARγ compared with control cells (Fig. 1F, lower panel).
To evaluate the potential pathophysiological relevance of hPDLSCs in PD, we examined their impact on selected functions of PMNs, key regulators of the periodontal homeostasis [2, 40, 41]. To this end, PMNs were cocultured with hPDLSCs at different ratios for 18 hours and their survival and bactericidal activity were evaluated. hPDLSCs protected PMNs from spontaneous apoptosis (p < .0001), with the 1:1 ratio being most efficient (~36% inhibition) (Fig. 2A). Also, the percentage of early apoptotic PMNs (annV+/PI−) was markedly reduced (from approxiately ~90% to 50%; p = .0002) by hPDLSCs, while the percentage of viable cells (annV−/PI−) increased (from ~7% to 37%; p = .0007).
Since IL-8 drives PMN recruitment at inflammatory sites , we ran hPDLSC/PMN coincubations in the presence of IL-8 to mimic an inflammatory environment. Figure 2C shows that hPDLSCs also protected IL-8-activated PMNs from apoptosis, compared with PMNs treated with IL-8 alone.
To investigate whether hPDLSCs, in addition to sustaining PMN viability, maintained their bactericidal activity, we exposed PMNs from cocultures with hPDLSCs to the P. aeruginosa strain PAO1. PMNs derived from hPDLSC cocultures (1:10) showed higher bactericidal activity compared with PMNs kept alone (p = .0197) (Fig. 2D), as indicated by the lower CFU number in the presence of PMNs isolated from cocultures (7.36 log10 ± 0.16 vs. 7.53 log10 ± 0.16 CFU/ml; p = .0197). Collectively, these results provide the first evidence that hPDLSCs promote survival and antibacterial activities of PMNs.
To assess whether hPDLSCs antagonized PMN apoptosis by releasing survival factors, we incubated PMNs (18 hours, 37°C) with CdM collected from hPDLSCs cultured for 24, 48, and 72 hours. As shown in Figure 2E, hPDLSC-CdM from all 3 time points significantly inhibited PMN apoptosis (p = .0003; p < .0001; p < .0001, respectively). Notably, 72-hour hPDLSC-CdM exhibited greater protective effect (26.61 ± 1.77 annV+/PI− PMNs) compared with 48 hours (44.79 ± 2.71 annV+/PI− PMNs) or 24 hours (61.33 ± 2.85 annV+/PI− PMNs) CdM (Fig. 2E). Remarkably, the antiapoptotic effect of 72-hour CdM was more potent than that observed with hPDLSCs (46.50 ± 2.64 annV+/PI− PMNs, 1:1 ratio) (Fig. 2A). To gain insights into this finding, we carried out multiplex cytokine analysis of hPDLSC-CdM and consistently observed that hPDLSC supernatants contain IL-6 and IL-8, which are known to inhibit PMN death [41, 42], as well as IL-10, VEGF, and EGF (Table 1). These results indicate that hPDLSCs can regulate PMN functions in a paracrine manner, producing a protective milieu that can influence resolution of inflammation and bacterial clearance.
We asked whether SPM biosynthesis was a component of the immunomodulatory properties of hPDLSCs, because SPMs promote inflammation resolution and are protective in experimental PD . To address this point, we performed LC-MS-MS-based LM metabololipidomics. hPDLSCs generated PUFA-derived LM from both lipoxygenase (LOX) and cyclooxygenase (COX) pathways (Fig. 3A, ,3B;3B; Table 2). In these incubations, we identified mediators from the docosahexaenoic acid (DHA)-derived bioactive metabolome including the D-series resolvins (RvD1, RvD2, RvD5, and RvD6), protectin D1 (PD1), and maresins (MaR1 and 7S,14S-diHDHA). These mediators were identified using characteristic retention times and MS-MS fragmentation in accordance with published criteria , as illustrated for MaR1, RvD1, and LXB4 (Fig. 3B). In these hPDLSC incubations, we also identified mediators from the AA bioactive metabolome, including the LOX products LXB4 and 15-epi-LXA4, as well as the COX products PGD2, PGE2, and PGF2α. Moreover, we identified RvE2 and RvE3 from the eicosapentaenoic acid (EPA) metabolome (Fig. 3A, ,3B;3B; Table 2). These findings demonstrate that explanted hPDLSCs produce SPMs, suggesting that they may contribute to the resolution of periodontal inflammation in vivo.
Next, we sought to determine whether proresolving mediators and receptors influenced hPDLSC activities. We focused on the ALX/FPR2-LXA4 axis, which is known to improve the outcome of experimental PD . hPDLSCs express ALX/FPR2 at the mRNA and protein levels, as confirmed by real-time PCR, and flow cytometry and confocal microscopy, respectively (supplemental online Fig. 1). Furthermore, ALX/FPR2 expression was not altered by cryopreservation (data not shown).
Once ALX/FPR2 expression was demonstrated in these cells, we examined LXA4 bioactions. To this end, we determined whether exposure to LXA4 for 24 hours influenced the lipid mediator profile of hPDLSCs. LXA4 increased D series Rvs as well as PGD2, while it decreased PD1 and PGE2 (Table 2). Although these changes did not reach statistical significance, because of interdonor variability in these primary hPDLSC cultures, these results suggest that LXA4 may stimulate the production of SPMs by hPDLSCs, likely triggering a proresolution feed-forward loop in the periodontium.
Because hPDLSCs modulate PMN activities and release mediators of inflammation and resolution, their self-renewal and migration capabilities are fundamental to achieve a pathophysiological impact. Therefore, we determined whether these functions could be modulated by LXA4. hPDLSCs exposed to increasing concentrations of LXA4 (0.001–1 nM) for 24, 48, and 72 hours displayed a concentration- and time-dependent increment in proliferation, with a maximum at 24 hours (Fig. 4A). This effect was suppressed by WRW4, an ALX/FPR2 antagonist peptide  (Fig. 4B). Likewise, LXA4 concentrations as low as 0.01–1 nM significantly enhanced hPDLSC directional migration in vitro, accelerating wound closure (Fig. 5A, ,5B).5B). At 3 and 6 hours after wounding, the percentage of filled wound in LXA4-treated cells was 3.5-fold higher compared with vehicle-exposed cells, and comparable to that observed in cells with 10% FBS, our control for maximal wound-healing capacity (Fig. 5A, ,5B).5B). LXA4 shortened the time needed to reduce wound width by 50% (W50) from 22 to 14 hours (~50%) (Fig. 5C). This effect was blunted by WRW4 (Fig. 5A–5C), which, instead, did not influence W50 in cells exposed to FBS (Fig. 5D).
PD is characterized by persistent inflammation and it is associated with local and systemic morbidities [3, 6] Therefore, knowledge of mechanisms and therapeutic approaches for this pathology is relevant for human health. In this study, we focused on stem cells isolated from the periodontal ligament. These cells have been added recently to the growing list of stem cells  and are likely to represent a promising tool to combat PD and related systemic consequences. hPDLSCs exhibit fibroblast-like morphology and colony-forming efficiency, and express embryonic and mesenchymal, but not hematopoietic or endothelial, markers (Fig. 1). They also express SSEA4, an embryonic marker, and the transcription factors NANOG, Oct4, and Sox2, associated with pluripotency (Fig. 1). With appropriate media, hPDLSCs differentiated into osteogenic, chondrogenic, and adipogenic lineages (Fig. 1). They can also produce cementum and PDL Sharpe’s fibers [23, 45]. These properties, combined with their easy accessibility, make hPDLSCs excellent candidates for periodontal regenerative medicine.
It is widely accepted that stem cells exert immunoregulatory functions, which, in some clinical settings, represent a primary therapeutic mechanism . In PD, the regulation of the host immune response is paramount for resolution ; PMNs are key players because they contribute to the removal of oral pathogens and tissue rebuilding . Consistent with this scenario, PMN key activities, for instance chemotaxis and bacterial phagocytosis, are impaired in patients with periodontitis, thus contributing to tissue damage [40, 41].
Here, we show for the first time that hPDLSCs sustain PMN survival and bactericidal activity (Fig. 2). Cell-cell contact was not strictly necessary for these effects to take place, indicating paracrine immunoregulatory functions of hPDLSCs. Indeed, they release cytokines and growth factors (i.e., IL-8, VEGF, and IL-6) that promote PMN recruitment and survival [49, 50] (Table 1). They also release IL-10 (Table 1), which limits the proinflammatory activities of PMNs  and improve wound repair during PD . This is consistent with a proresolutive, antimicrobial profile of the hPDLSC secretoma and prompted us to evaluate the capability of these cells to produce SPMs, pivotal regulators of inflammation resolution .
Using an established LM metabololipidomic approach , we obtained the first evidence that hPDLSCs produce and release a broad spectrum of proresolutive mediators, including DHA-derived D-series resolvins, protectin D1, and maresin 1; EPA-derived resolvin E2 and E3; and AA-derived LXB4 (Fig. 3). These mediators carry immunoregulatory/proresolutive bioactions by regulating granulocyte trafficking, stimulating efferocytosis and microbial clearance [8, 54]. In particular, D- and E-series resolvins regulate PMN chemotaxis and increase their capability to phagocytose and kill bacteria through specific cell-surface receptors, namely ALX/FPR2 and chemerin receptor 23 (ChemR23), widely expressed on PMNs [37, 54, 55]. Furthermore, these SPMs prevent bone resorption in experimental periodontitis . hPDLSCs also produce large amounts of PGE2, which regulates bone formation and resorption . Hence, hPDLSCs generate LMs that orchestrate inflammation resolution, immunomodulation, and bone metabolism.
Although it will be interesting to evaluate SPM production by stem cells from other sources, the present results provide new elements for the understanding of stem cell biology. It has been reported that stem cells exert a number of paracrine functions related to immunomodulation and control of the inflammation [28, 29]. Thus, the timely release of lipid mediators of inflammation and resolution may be regarded as a pivotal mechanism underlying these functions. Evidence that supports this hypothesis is accumulating. For instance, MSCs from umbilical cords reduce mast cell degranulation and alleviate murine atopic dermatitis through PGE2 release . PGE2 is also key for the attenuation of graft-versus-host disease by BM-MSCs  and for the anti-inflammatory effects of adipose-derived MSCs on chondrocytes and synoviocytes from osteoarthritis patients . Although PGE2 was the most abundant eicosanoid generated by hPDLSCs under routine culture conditions (Table 1), the observation, never reported before, to our knowledge, that these cells can generate SPM from DHA and EPA, which act at subnanomolar to nanomolar concentrations, suggest that: (a) the relative abundance of bioactive lipid mediators generated by stem cells could be modulated by the dietary intake of precursor fatty acids and might change in pathological conditions, and (b) the overall impact of stem cells on immuno-inflammatory responses may be dictated by a network of lipid mediators, which act in concert with peptides (Table 1).
On the other hand, the features of hPDLSCs uncovered in the present study propose these cells as key players within the context of PD, as their recruitment, proliferation, and release of proresolutive mediators can influence the outcome of PD. Here, we provide the first evidence that these processes can be modulated by an LXA4-driven, receptor-mediated mechanism as hPDLSCs express ALX/FPR2 (supplemental online Fig. 1). In addition to LXA4, this receptor recognizes RvD1 and annexin-A1, both carrying anti-inflammatory, proresolution functions . Consistent with earlier reports showing expression of this GPCR in human BM-MSCs and its role in stem cell adhesion, migration, and homing for tissue repair [60, 61], the activation of ALX/FPR2 by LXA4 stimulated hPDLSC proliferation and migration in vitro (Figs. 4, ,5).5). The effect of LXA4 on wound healing was significant already at 3 hours, ruling out the involvement of cell proliferation, at least at the earlier time points. These effects are of considerable relevance as hPDLSCs can differentiate into osteoblasts and regenerate PDL ; therefore, their recruitment and expansion within a diseased periodontium may accelerate tissue healing. Along these lines, LXA4 levels are reduced in crevicular fluids and plasma from patients with PD [17, 18], thus potentially limiting endogenous hPDLSC-dependent tissue-repair mechanisms. Whether LXA4 can trigger differentiation pathways in hPDLSCs, similar to what occurs with embryonic stem cells exposed to PD1 , remains to be determined.
LXA4 also changed the LM profile of hPDLSCs (Table 2), although it did not have appreciable impact on PMN survival. Cells exposed to LXA4, showed an increase in D-series Rvs, which control leukocyte trafficking and have antimicrobial properties in vivo . LXA4 also changed the AA metabolome: It increased PGD2 while decreasing PGE2 formation (Table 2). This is consistent with a proresolution profile . Therefore, it appears that LXA4 can sustain a receptor-mediated proresolution feed-forward loop that involves hPDLSCs in the periodontium.
The main finding of this study is that stem cells isolated from the periodontal ligament can release SPM and other mediators that carry relevant immunomodulatory, proresolution, and prohealing properties. This hPDLSC characteristic can be harnessed for therapeutic purposes in PD, although it is likely to represent a more general mechanism by which stem cells from different sources can influence the outcome of inflammatory disorders. We also demonstrate that the ALX/FPR2 receptor is crucial to promote the recruitment and expansion of hPDLSCs, thus representing a novel molecular target to enhance stem cell-mediated proresolving circuits. Collectively, these results highlight the involvement of SPM and their receptors in stem cell biology, providing useful knowledge for innovative stem cell therapeutics.
We thank the dentists of the Department of Medical Science, Oral and Biotechnology of the University G. d’Annunzio, Chieti, Italy, and all the people who donated the periodontal ligament and blood. We also thank Ilaria Merciaro and Alessia Lamolinara for technical support. This work was partially supported by a grant from the Cari-Chieti Foundation (Italy) to the StemTeCh Group and grants from the Italian Ministry of Education, University and Research to M.R. (PRIN 2010YK7Z5K_002), O.T. (PRIN 20102ZLNJ5), and S.M. (FIRB RBAP1047J_006). J.D., R.C., and C.N.S. were supported by NIH GM-095467 to C.N.S.
E.C. and A.R.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; O.T.: financial support, collection and assembly of data, data analysis and interpretation, final approval of manuscript; F.D.: provision of patients, collection and assembly of data, data analysis and interpretation, final approval of manuscript; M.M.: collection and assembly of data, data analysis and interpretation, final approval of manuscript; S.M.: financial support, data analysis and interpretation, final approval of manuscript; R.A.C. and J.D.: collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; C.N.S.: data analysis and interpretation, manuscript writing, final approval of manuscript; M.R.: conception and design, financial support, manuscript writing, final approval of manuscript.
C.N.S. is a compensated Corbus scientific advisory board member; has compensated research funding from the NIH; and is an uncompensated inventor on patents assigned to Brigham and Women’s Hospital and Partners HealthCare on the composition of matter, uses, and clinical development of anti-inflammatory and pro-resolving lipid mediators. These include lipoxins and resolvins and related compounds that are licensed for clinical development. The resolvins are licensed to Resolvyx Pharmaceutical. The other authors indicated no potential conflicts of interest.