Intro
Inflammation occurs following a challenge of host tissue by microorganisms, trauma, or other injuries ( Gilroy and Lawrence, 2008 ). Following this challenge, pro-inflammatory signals are released to trigger key events such as recruitment of neutrophils from the blood to the site of the injury to clear microorganisms and cell debris through rapid phagocytosis and to release antimicrobial factors ( Kolaczkowska and Kubes, 2013 ). Throughout this process, neutrophils undergo apoptosis and monocytes are recruited to clear the site of infection via lymphatic uptake ( Ginhoux and Jung, 2014 ). This inflammatory stage bridges innate and adaptive immune responses, and leads to the recruitment of B- and T-cells ( Lanier, 2013 ). All of these events result in the resolution of inflammation and restoration of tissue homeostasis ( Serhan et al., 2008 ). Disruption of this process can lead to chronic inflammation, which is observed in autoimmune diseases ( Tabas and Glass, 2013 ), and is thought to play a role in the development and tumor progression in cancer ( Landskron et al., 2014 ). The onset and peak of inflammation is an active process controlled by several factors including cytokines and chemokines and recently, the resolution of inflammation has also been shown to be an active, rather than a passive, process ( Buckley et al., 2013 ). For an in depth review of resolution, readers are encouraged to refer to the excellent reviews by Dr. Charles Serhan and colleagues ( Serhan, 2014 , Serhan et al., 2015a ). Briefly, resolution of inflammation involves the cessation of polymorphonuclear leukocytes (PMNL) influx through the upregulation of PMNL apoptosis, restoration of normal cytokine gradients, and the clearance of cell debris by macrophages ( Serhan, 2014 , Serhan et al., 2015b , Basil and Levy, 2016 ).
One class of molecules that has been gaining prominence in relation to inflammation is the lipid mediators (LM) derived from polyunsaturated fatty acids (PUFA). PUFA are appreciated for their beneficial actions in the immune ( Hubler and Kennedy, 2016 ), neural ( Crupi et al., 2013 , Bazinet and Layé, 2014 , Hashimoto et al., 2014 , Song et al., 2016 ) and cardiovascular systems ( Dessì et al., 2013 , Lorente-Cebrián et al., 2013 , Nicholson et al., 2013 , Colussi et al., 2014 ). Most notably, PUFA such as arachidonic acid (AA), eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) are precursors for the biosynthesis of a variety of LM ( Tessaro et al., 2015 ). AA is an ω-6 PUFA that is generated from dietary linoleic acid through the successive actions of Δ6-desaturase, elongase and Δ5-desaturase ( Astudillo et al., 2012 ) or from cellular phospholipids via cytosolic phospholipase 2 (cLPA 2 ) ( Serini et al., 2009 ). EPA and DHA are ω-3 PUFA that can be generated in the body from α-linoleic acid via elongase and Δ5-desaturase ( Neff et al., 2011 ), but the conversion rate is minute ( Goyens et al., 2005 ), necessitating dietary intake of EPA and DHA. The major dietary source of EPA and DHA are marine animal oils ( Tur et al., 2012 ), though recently algae oil has been identified as a vegetarian source of DHA ( Lane et al., 2014 ). EPA and DHA have been shown to have direct anti-inflammatory properties through the inhibition of NLRP3 pathway ( Zhou et al., 2011 , Calder, 2013 , Yan et al., 2013 ), and are transported to the site of inflammation via lipoproteins coinciding with edema generation ( Kasuga et al., 2008 ). Omega-3 dietary supplementation has been investigated for the treatment of cardiovascular disease ( Rizos et al., 2012 ), metabolic syndrome ( Lorente-Cebrián et al., 2013 ), and depression ( Lorente-Cebrián et al., 2013 ), among others. Interestingly, Omega-3 dietary supplementation has been investigated for the treatment of periodontitis in a variety of preclinical and clinical models ( El-Sharkawy et al., 2010 , Deore et al., 2014 , Chee et al., 2016 ). More importantly, EPA and DHA are important substrates to produce pro-resolving LM, such as resolvins and protectins ( Hong and Lu, 2013 , Buckley et al., 2014 , Spite et al., 2014 ), whose functions will be discussed in detail in the following sections. LM are generated via two major biosynthetic pathways, specifically, cyclooxygenase and lipoxygenase pathways. These pathways have been widely described in systemic diseases for over 30 years ( Gwebu, 1979 , Samuelson and Paoletti, 1982 , Pace-Asciak and Granström, 1983 ), and have been investigated in conjunction with oral diseases ( El Attar and Lin, 1983 , Porteder et al., 1984 , Mendieta et al., 1985 , Williams et al., 1988 , Offenbacher et al., 1989 ). Interestingly, many inflammatory systemic diseases are associated with increased incidence of oral diseases. For example, there is strong evidence of a link between periodontitis and atherosclerosis and other cardiovascular diseases ( Dietrich et al., 2013 , Loos et al., 2016 ), and between periodontitis and diabetes ( Taylor et al., 1996 , Hasturk and Kantarci, 2016 ). It has been suggested that some LM complicit in oral diseases are potential risk factor markers of other systemic diseases ( Bäck et al., 2007 ). This review seeks to provide an overview of the current state of the art regarding biosynthesis and functions of various LM. We also hope to provide clinical providers and researchers in the oral health field with a detailed account regarding the roles of pro-inflammatory and pro-resolving LM in both systemic and oral diseases. Additionally, we conclude by addressing some major questions, including the potential effects of gender differences on the progression of chronic inflammatory diseases, and future research directions regarding diagnostic and therapeutic potential of LM in oral health.
Lipid
The ubiquity of LM in physiological and pathological processes has led to an increased interest in identifying LM in various body fluids. Particularly, as they represent a simple and noninvasive approach for research and diagnostic purposes.
Serum levels of various LM have been investigated to study whether they correlated with certain diseases. High serum LTB 4 levels have been shown to be correlated with an elevated risk of acute coronary syndrome in adults ( He et al., 2014 ). Additionally, high LTB 4 levels may predispose children to tonsillar hypertrophy ( Alexopoulos et al., 2015 ). Further, serum levels of Cys-LTs are elevated during anaphylaxis and have been suggested to be a reliable marker of this condition ( Nassiri et al., 2016 ). Finally, elevated serum levels of LTB 4 and Cys-LTs are correlated with higher levels of circulating neutrophils in children with sleep-disordered breathing ( Shen et al., 2014 ).
Increased PGE 2 serum levels indicate overexpression of aromatase ( Subbaramaiah et al., 2012 ), which is linked to hormone-related breast cancer ( Subbaramaiah et al., 2008 ). Additionally, PGE 2 serum levels, along with 8-Isoprostane, C-reactive protein, and amyloid A, have been suggested as inflammatory markers for antiphospholipid syndrome ( Sciascia et al., 2012 ). Similarly, serum thromboxane levels are an indicator of platelet inhibition by aspirin, and can thus be used for epidemiological studies on the effects of aspirin ( Reny et al., 2012 , Zantek et al., 2014 ). However, measurement methods for serum thromboxane need to be standardized to produce consistent results across studies ( Brun et al., 2016 ).
In contrast to the pro-inflammatory leukotrienes and prostaglandins, decreased lipoxin serum levels are typically observed in several pathological conditions. For example, lowered LXA 4 serum levels were found in patients with sepsis ( Tsai et al., 2013b ) and in wheezy infants ( Eke Gungor et al., 2014 ). Serum LXA 4 levels were also found to have an inverse relationship to the risk of metabolic syndrome ( Yu et al., 2015 ). Of note to dental professionals, periodontitis patients might have elevated LXA 4 serum levels compared with healthy controls ( Doğan et al., 2015 ).
RvD1 and RvE1 serum levels have been investigated as markers for the efficacy of diclofenac in acute pancreatitis treatment, with increased resolvin and lipoxin serum levels observed in the drug receiving group ( Zhao et al., 2014 ). Additionally, higher RvD1 serum levels have been described as a potential biomarker for Familial Mediterranean Fever ( Taylan et al., 2015 ).
Similar to serum, urine levels of various LM have been used to investigate multiple diseases. High levels of urinary leukotriene have been observed in pediatric patients with wheezing and recurrent asthma ( Morales et al., 2016 ). High levels of urinary leukotrienes have also been detected in adult individuals with systemic mastocytosis ( Lueke et al., 2016 ), as well as in aspirin-intolerant asthmatic individuals ( Yamaguchi et al., 2015 , Hagan et al., 2016 ). Urinary leukotriene levels are directly correlated to the severity of arterial occlusive disease following transluminal angioplasty ( Maga et al., 2016 ). Additionally, they are inversely correlated with renal function in Type 2 diabetes ( Rafnsson and Bäck, 2013 ). Therefore, they have been suggested as prognostic markers for these diseases.
While levels of total unmodified prostaglandin in urine have been correlated with only a few diseases (e.g., cardiovascular disease ( Raatz et al., 2012 ) and anaphylaxis ( Lieberman, 2013 )), levels of prostaglandin metabolites in urine have attracted more scrutiny. High levels of the PGD 2 metabolite tetranor PGDM in urine, have been correlated with Duchenne muscular dystrophy, systemic mastocytosis, rheumatoid arthritis, colitis, and colon cancer ( Nakagawa et al., 2013b , Nakagawa et al., 2013a , Iwanaga et al., 2014 , Cho et al., 2015 ). High levels of urinary PGE 2 metabolite levels appear to be associated with colorectal adenoma ( Shrubsole et al., 2012 , Davenport et al., 2015 ), pancreatic cancer ( Zhao et al., 2015 ), and breast cancer in postmenopausal women ( Cui et al., 2014 ). Additionally, elevated levels of urinary thromboxane have been suggested as a predictor of atherthrombotic risk ( Neath et al., 2013 ).
In contrast to the pro-inflammatory LM, there is a dearth of studies examining urinary levels of pro-resolving LM. Decreased urine levels of lipoxins and resolvins have been hypothesized to be related to inflammatory bowel disease ( Das, 2016 ), while elevated urinary LXA 4 /creatinine ratio has been observed in systematic lupus erythematosus ( Abdou et al., 2015 ).
Lung fluids are an obvious target for analysis of LM involved in inflammatory lung diseases. Asthmatic patients had elevated sputum levels of Cys-LTs and PGE 2 compared to healthy subjects ( Papadaki et al., 2013 ), with smoking asthmatics exhibiting even higher sputum levels of these LM ( Kontogianni et al., 2013 ). COPD patients, on the other hand, only exhibited elevated PGE 2 sputum levels, but no differences in Cys-LT levels, compared to healthy controls ( Drozdovszky et al., 2013 ). An analysis of sputum from adult cystic fibrosis patients found elevated levels of LTB 4 and PGE 2 ( Yang et al., 2012 ). Pediatric cystic fibrosis patients, however, did not have elevated LTB 4 sputum levels, but the ratio of LXA 4 to LTB 4 was depressed ( Ringholz et al., 2014 ). Accurate measurements of LTB 4 using UPLC-MS/MS require a refinement of sputum processing due to the instability of LTB 4 in the presence of standard preparation reagents ( Jian et al., 2013 ). In exhaled breath condensate (EBC), similar findings with regards to higher leukotriene levels were observed as well. For example, smoking asthmatic patients displayed elevated EBC Cys-LT levels ( Celik et al., 2013 ). Cys-LT EBC levels are correlated with the severity of asthma ( Kazani et al., 2013 ). Cys-LT levels in exhaled breath condensate were found to be elevated in children with asthma and allergic rhinitis ( Wan et al., 2013 ). 8-iso-PGE 2 levels in exhaled breath condensate were found to be elevated in aspirin-hypersensitive asthmatics ( Mastalerz et al., 2015 ). Lipoxin/Leukotriene imbalance was also detected in exhaled breath condensate of severe refractory asthma patients ( Sedlák et al., 2014 ). Exhaled breath condensate has recently emerged as a simple and accurate medium to aid the diagnosis and evaluation of inflammatory severity in asthmatic patients ( Thomas et al., 2013 ).
LXA 4 sputum levels in pediatric severe asthmatics were found to be depressed, along with reduced ALX/FPR2 expression in induced sputum cells ( Gagliardo et al., 2016 ). Conversely, LXA 4 levels in exhaled breath condensate were found to be elevated in asthmatic children with exercise-induced bronchoconstriction ( Tahan et al., 2016 ). Such investigations are eminently useful for understanding the pathologies of different respiratory disorders. As for resolvins, there is currently a lack of rigorous analysis of the levels of resolvins in sputum and exhaled breath condensate. More research into this area would contribute greatly to the understanding of the development and resolution of various respiratory conditions.
Breast milk is extremely important for developing innate immunity and healthy intestinal homeostasis in babies ( Cederlund et al., 2013 , Jakaitis and Denning, 2014 ). Elevated PUFA levels in breast milk decrease during the first month after birth, but lipoxin and resolvin levels remain stable, indicating an important role in neonatal immunity ( Weiss et al., 2013 ). LM profile in breast milk could be associated with conditions that affect neonatal health. For example, maternal smoking can negatively affect the LM profile in breast milk ( Szlagatys-Sidorkiewicz et al., 2013 ). Furthermore, mastitis results in elevated LTB 4 and decreased lipoxin and resolvin levels ( Arnardottir et al., 2015 )
LM profiling of synovial fluid from rheumatoid arthritis reveals the presence of both pro-inflammatory and pro-resolving LM ( Giera et al., 2012 ). Elevated levels of PGD 2 and PGE 2 in synovial fluid from arthritic patients have been linked to dendritic cells in the synovial fluid ( Moghaddami et al., 2013 ). Of note for dental clinicians, synovial fluid from painful, dysfunctional temporomandibular joints (TMJ) was found to contain elevated levels of LTB 4 and PGE 2 ( Quinn and Bazan, 1990 ). LTB4 synovial fluids were found to be prognostic markers for the success of arthrocentesis of the TMJ, with higher LTB4 levels indicating unsuccessful treatment ( Kaneyama et al., 2007 ). Similarly, glucosamine-chondroitin sulfate treatment for TMJ derangement results in lower PGE2 synovial fluid levels ( Damlar et al., 2015 ).
Gingival crevicular fluid (GCF) is an important periodontal diagnostic marker ( Gupta, 2013 ). LTB 4 levels in GCF are higher in chronic periodontitis patients ( Emingil et al., 2001 , Pradeep et al., 2007 ). Similarly, PGE 2 levels in GCF are a strong marker of gingivitis and periodontitis ( Heasman et al., 1998 , Camargo et al., 2015a ), and are correlated with the severity of periodontal disease ( Kumar et al., 2013 ). GCF from orthodontic patients displaying gingival inflammation who received a chlorhexidine/thymol-containing dental varnish administered exhibited a reduction in LTB 4 and PGE 2 levels ( Sköld et al., 1998 , Yucel-Lindberg et al., 1999 ).
GCF levels of pro-resolving LM are generally decreased in periodontal patients. For example, LXA 4 levels were found to be significantly lower in periodontal patients compared to healthy controls, combined with a negative correlation between LXA 4 levels and clinical attachment loss ( Tarannum and Faizuddin, 2016 ). GCF LXA 4 levels were similarly found to be lowest in smokers with aggressive periodontitis ( Heasman et al., 1998 ). The ration of pro-resolving and pro-inflammatory LM has been suggested as a better marker for aggressive periodontitis ( Elabdeen et al., 2013 ).
Saliva diagnostics are an intuitive and comprehensive approach for oral diseases ( Zhang et al., 2015 ), and they have recently emerged as a method for determining general health status without the invasiveness of traditional methods ( Punyadeera and Slowey, 2013 , Schafer et al., 2014 ). Metabolomic and bioinformatic approaches are bringing these applications even closer ( Ai et al., 2012 , Zhang et al., 2012 ). The development of higher sensitivity assays should accelerate the adoption of salivary diagnostics over serum ( Khaitan et al., 2015 ). In addition to the traditional inflammatory biomarkers, salivary levels of LM can also be used for diagnostic and prognostic purposes, though they lag behind other body fluids. For example, elevated salivary LTB 4 and PGE 2 are correlated with arterial stiffness ( Labat et al., 2013 ). Aspirin-intolerant asthma patients exhibit elevated Cys-LT levels in saliva ( Gaber et al., 2008 , Ono et al., 2011 ). Salivary prostaglandin levels have long been suggested to be indicators of major depression ( Ohishi et al., 1988 , Nishino et al., 1989 ).
Saliva levels of various LM are more widely used for studying oral diseases than for systemic diseases. For example, patients with radiotherapy-caused oral mucositis displayed higher saliva levels of PGD 2 , PGE 2 , and PGF 2 ( Grunberg et al., 2013 ). Similarly, a different study found elevated levels of LTA 4 hydrolase in oral mucositis patients ( Jehmlich et al., 2015 ). Elevated salivary levels of PGE2 were found to be correlated with gingivitis ( Syndergaard et al., 2014 , Gümüş et al., 2016 ) and periodontitis ( Sánchez et al., 2013b ). In chronic periodontitis patients, salivary LTB 4 levels were to found to correlate with the severity of alveolar bone loss ( Sánchez et al., 2013a ). Ratios of pro-resolving LM to pro-inflammatory LM that were found to be correlated with aggressive periodontitis in GCF were not found to be similarly correlated in saliva ( Elabdeen et al., 2013 ). No studies to date have examined salivary levels of pro-resolving LM, such as lipoxins and resolvins, for diagnostic studies of oral disease.
One interesting aspect of saliva diagnostics is the potential to explain the etiology of various diseases. Some diseases display marked gender differences, notably SS, which affects older females primarily. Differences in LM levels have been observed between males and females in multiple diseases, such as diabetes ( Tessaro et al., 2015 ), COPD ( Balgoma et al., 2016 ), Alzheimer’s ( Pomponi et al., 2011 ), and metabolic syndrome ( Yu et al., 2015 ).
To test this hypothesis in saliva, we investigated differences in salivary LM levels between healthy male and female donors. For this previously unpublished study, we recruited patients and students at the University of Utah School of Dentistry student clinics, who were between 18 and 40 years old and had no history or visible signs of gingivitis or periodontitis (11 females and 19 males). All human specimen usage was conducted under the strict guidelines and approval of the University of Utah Health Sciences Institutional Review Board, and informed consent was obtained for each patient (IRB approval: IRB_00081821 on October 29 th 2015). Donor demographic information is shown in Table 1 . Unstimulated whole saliva was collected by having subjects spit into a sterile 50 mL conical tube for five minutes after a brief mouth rinse with water, then transported to the laboratory on ice. We opted for the spitting method over cotton-based methods due to the potential of interference with immunoassay results ( Shirtcliff et al., 2001 ). Samples were centrifuged at 5000 × g for 10 minutes at 4°C to pellet out debris, and the supernatant was collected and broken up via trituration with a 19G hypodermic needle to reduce its viscosity. Saliva was consecutively filtered through a 40 μm and a 0.2 μm filter to remove debris and bacteria, respectively. Enzyme-linked immunosorbent assays (ELISA) were performed immediately for the following LM: LTB 4 , PGD 2 , PGE 2 , PGF2α, TXB 2 , RvD1, and RvD2 (Cayman Chemical, Ann Arbor, MI). Measured LM concentrations (mean ± standard deviation) are shown in Table 2 .
Using this method, we show that we can detect salivary levels of LTB 4 , PGD 2 , PGE 2 , PGF 2α , TXB 2 , RvD 1 , and RvD 2 . We observed significantly lower PGF 2α in female saliva as compared to male saliva. As indicated above, PGF 2α plays a role in many female functions such as reproduction and menstrual cycle ( Wilks et al., 1973 , Downie et al., 1974 ); therefore, we could speculate that females demand and use this lipid mediator for more functions than males, thereby levels in saliva are decreased. Additionally, we observed that saliva levels of RvD2 from male subjects were significantly lower than female subjects. This finding raises interesting questions regarding the etiology of some inflammatory diseases that skew females, such as SS. One hypothesis that needs further investigation is that higher levels of pro-resolving LM, such as RvD2, might indicate the presence of inflammatory assaults that the body is attempting to resolve over time, leading to persistent chronic inflammation. More research needs to be conducted to confirm such a hypothesis, including comparing saliva LM levels of SS patients to healthy subjects, and investigating whether gender-related differences that are observed in healthy subjects disappear in SS patients.