Pathophysiological regulation of lung function by the free fatty acid receptor FFA4

See allHide authors and affiliations

Science Translational Medicine  19 Aug 2020:
Vol. 12, Issue 557, eaaw9009
DOI: 10.1126/scitranslmed.aaw9009

Breathing easy

Asthma and chronic obstructive pulmonary disease result in airway swelling that hinders breathing. Prihandoko et al. found that intranasal delivery of a free fatty acid receptor 4 (FFA4) agonist reduced airway resistance in mouse models of acute and chronic ozone pollution–mediated inflammation as well as house dust mite and cigarette smoke–induced inflammation. This airway relaxation was mediated in part by the release of prostaglandin PGE2. FFA4 agonism also mediated smooth muscle relaxation of ex vivo human airway samples, suggesting FFA4 as a potential respiratory disease target in humans.


Increased prevalence of inflammatory airway diseases including asthma and chronic obstructive pulmonary disease (COPD) together with inadequate disease control by current frontline treatments means that there is a need to define therapeutic targets for these conditions. Here, we investigate a member of the G protein–coupled receptor family, FFA4, that responds to free circulating fatty acids including dietary omega-3 fatty acids found in fish oils. We show that FFA4, although usually associated with metabolic responses linked with food intake, is expressed in the lung where it is coupled to Gq/11 signaling. Activation of FFA4 by drug-like agonists produced relaxation of murine airway smooth muscle mediated at least in part by the release of the prostaglandin E2 (PGE2) that subsequently acts on EP2 prostanoid receptors. In normal mice, activation of FFA4 resulted in a decrease in lung resistance. In acute and chronic ozone models of pollution-mediated inflammation and house dust mite and cigarette smoke–induced inflammatory disease, FFA4 agonists acted to reduce airway resistance, a response that was absent in mice lacking expression of FFA4. The expression profile of FFA4 in human lung was similar to that observed in mice, and the response to FFA4/FFA1 agonists similarly mediated human airway smooth muscle relaxation ex vivo. Our study provides evidence that pharmacological targeting of lung FFA4, and possibly combined activation of FFA4 and FFA1, has in vivo efficacy and might have therapeutic value in the treatment of bronchoconstriction associated with inflammatory airway diseases such as asthma and COPD.


Asthma and chronic obstructive pulmonary disease (COPD) are chronic respiratory inflammatory diseases that affect more than 500 million people worldwide, causing substantial morbidity and exacting a considerable cost to health services (13). Despite the effectiveness of current therapies, the heterogeneity of these conditions means that ~45% of asthmatics remain uncontrolled, and in COPD, the combination of corticosteroid, long-acting muscarinic antagonists and β-adrenoceptor agonists inadequately control symptoms and exacerbations in many patients (14). We have addressed the urgent need to understand paradigms in lung physiology that might reveal therapeutic strategies by focusing on the role and potential clinical value of a member of the G protein–coupled receptor (GPCR) superfamily, FFA4 (previously known as GPR120). This GPCR responds to a range of free circulating long-chain (>C12) fatty acids (LCFAs) that include the diet-derived essential atty acids linoleic and α-linolenic as well as omega-3 polyunsaturated fatty acids prevalent in fish oils (57). Most of the previous studies on FFA4 have focused on the regulation of physiological responses associated with food intake. Expressed on entero-endocrine cells of the gut (8, 9), various cell types in pancreatic islets (10, 11), white adipose tissue (12), and immune cells, particularly macrophages (13), FFA4 has been implicated in the regulation of glucose homeostasis (11) and inflammation, particularly in connection with adipose tissue in animals fed a high-fat diet (14). These data, together with the emergence of a number of small “drug-like” ligands to FFA4 (6), have fueled interest from both academia and industry in targeting FFA4 for the treatment of metabolic disease including type 2 diabetes and obesity (7).

It has, however, recently become clear that FFA4 is abundantly expressed in lung epithelial cells (15), and a recent study suggested that omega-3 fatty acids acting via FFA4 receptors might have some beneficial effects on airway epithelial repair after naphthalene-induced airway injury (16). Other than this study, the role that FFA4 might play in lung physiology is unknown. Here, we used a combination of pharmacological agents, genetically engineered mice, and both ex vivo and in vivo techniques to determine that FFA4 receptors expressed in mouse and human airways mediate airway smooth muscle (ASM) relaxation in a manner that is relevant in both normal physiology and inflammatory airway disease.


Expression of FFA4 in mouse airways

Using both reverse transcription polymerase chain reaction (RT-PCR) (Fig. 1A) and quantitative RT-PCR (qRT-PCR) (Fig. 1B), we confirmed the presence of FFA4 mRNA in the mouse lung to a degree substantially higher than seen for a second receptor for LCFAs, FFA1 (Fig. 1, A and B). We confirmed the lack of expression of FFA4 mRNA in equivalent samples from a genetically engineered mouse strain where the first exon of the FFA4 gene was replaced with the coding sequence of β-galactosidase [FFA4-KO(βgal) mice] (Fig. 1, A and B). To investigate the tissue distribution of FFA4 further, we stained for β-galactosidase in FFA4-KO(βgal) mice as this acts as a surrogate marker for FFA4-expressing cells. Using this approach, we established that FFA4 is expressed primarily in cells of the airway epithelium (Fig. 1C). These data are supported by studies using an in-house generated mouse FFA4-selective antiserum (17), which confirmed that FFA4 expression was primarily associated with the epithelium of the mid and lower airways, with lower expression associated with ASM, as defined by the presence of α-actin (Fig. 1D). Costaining with an antiserum to club cell specific protein 10 (CC10) demonstrated that club cells were at least one cell type expressing FFA4 in the airway epithelium (Fig. 1E), a result consistent with a recent mRNA analysis that identified enrichment of FFA4 transcripts in club cells (18).

Fig. 1 FFA4 receptors are expressed in murine lung and localize primarily to the epithelial layer.

(A) FFA1, FFA4, and tubulin (control) transcripts were identified in lung from wild-type and FFA4-KO(βgal) mice using RT-PCR. Shown is a representative gel from three independent experiments. Expected sizes of PCR products are as follows: FFA4, 800 base pairs (bp), tubulin, 750 bp, and tubulin, 530 bp. (B) qRT-PCR of lung tissue samples from wild-type and FFA4-KO(βgal) mice using the GAPDH housekeeping gene as a control. Data are means ± SD of n = 4 [FFA4-KO(βgal)] and n = 10 (WT animals). (C) Representative confocal image of lung tissue sections obtained from FFA4-KO(βgal) mice and stained for β-galactosidase activity on 5-Bromo-4-Chloro-3-Indolyl β-D-Galactopyranoside substrate (XGAL) as a surrogate for FFA4 expression. (D) Representative images of lung tissue sections obtained from wild-type (left) and FFA4-KO(βgal) (right) mice costained with an in-house generated mouse FFA4-specific antiserum (green), α-actin antibody to stain ASM (red), and DAPI to identify nuclei (blue). (E) Immunofluorescence costaining of mouse FFA4 and club cell marker (CC10). Images are representative of at least three independent experiments.

FFA4 mediates ASM relaxation and bronchodilation

To probe the activity of FFA4, we used a well-characterized agonist, TUG-891 (1923), that has activity at both FFA1 and FFA4 (Fig. 2A). In addition, we used the recently generated agonist TUG-1197 (24) that we confirmed here selectively activates FFA4 but not FFA1 (Fig. 2B). FFA4 has previously been shown to signal in a bimodal fashion, initiating signal transduction via heterotrimeric G proteins as well as through a mechanism operating via the recruitment of arrestin adaptor proteins (17, 1921). Our previous studies have established that the recruitment of arrestin3 in response to TUG-891 is nearly totally dependent on receptor phosphorylation and mediates processes such as receptor internalization (17, 20). Here, we found that TUG-1197 acted as an agonist at FFA4 in both G protein–dependent signaling, as measured by Gq/11-mediated calcium mobilization (Fig. 2B) and activation of extracellular signal–regulated kinase 1/2, (ERK1/2; Fig. 2, C and D). In addition, TUG-1197 was also an agonist at promoting arrestin3 recruitment to FFA4 (Fig. 2E), receptor phosphorylation (Fig. 2F), and receptor internalization (Fig. 2G). Hence, the signaling properties of TUG-1197 indicated that this ligand is a selective agonist at FFA4, initiating signaling through canonical G protein and arrestin pathways. TUG-1197 is therefore a useful tool ligand for probing the physiological activity of FFA4. Moreover, although able to activate FFA1, TUG-891 was also able to mimic the effects of TUG-1197 at FFA4 (Fig. 2, C to G).

Fig. 2 Pharmacology of FFA4 agonist ligands in cell-based assays and ex vivo murine lung tissue preparations.

(A and B) Calcium concentration-response curves in Flp-In T-REx 293 cells expressing mouse FFA1 or FFA4 and stimulated with (A) TUG-891 or (B) TUG-1197. (C) ERK1/2 concentration-response curves measured using Homogeneous Time Resolved Fluorescence FRET (HTRF) of CHO Flp-In cells expressing mouse FFA4 after 5 min stimulation with TUG-891 or TUG-1197. (D) Kinetics of TUG-891– and TUG-1197–mediated ERK1/2 responses at maximal concentration (1 μM) of each ligand. HTRF ratio corresponds to the transfer of energy from the donor-conjugated antibody to acceptor-conjugated antibody and is directly proportional to the amounts of ERK1/2 phosphorylation (E) Arrestin3 recruitment to FFA4 concentration response curves of Flp-In T-REx 293 cells expressing recombinant mouse FFA4 and stimulated with varying concentrations of TUG-891 or TUG-1197. (F) FFA4 phosphorylation in response to TUG-891 and TUG-1197 in Western blots of lysates prepared from Flp-In T-REx 293 expressing mouse FFA4 and probed with an anti-phospho antiserum to pThr347/pSer350 on mouse FFA4. Parallel immunoblotting to detect GAPDH (glyceraldehyde-3-phosphate dehydrogenase) acted as a loading control. (G) Internalization of enhanced yellow fluorescent protein (eYFP)–tagged mouse FFA4 expressed in Flp-In T-REx 293 cells after 30-min application of TUG-891 or TUG-1197. (H to J) Influx of intracellular calcium evoked by TUG-1197 in precision cut lung slices from (H) wild-type mice, (I) FFA4 KO(βgal) mice, and (J) wild-type mice in the presence of the Gq inhibitor FR900359. The white boxes indicate the area of interest where the fluorescence changes are illustrated in the graph to the right of the images. The FURA-2 ratio represents changes in 340 of 380 nm emission from the FURA2-AM calcium indicator. Arrows indicate the time point at which the FFA4 agonist was added. The experiments shown are representative of at least three independent experiments. Data in (A) to (E) are the means ± SEM.

In precision-cut lung slices (PCLS) derived from wild-type mice, TUG-1197 induced a rapid increase in intracellular calcium (Fig. 2H). In contrast, no substantial calcium response to TUG-1197 was observed in PCLS derived from FFA4-KO(βgal) mice (Fig. 2I). Furthermore, the calcium response to TUG-1197 in PCLS from wild-type mice was markedly inhibited by the Gq/11 blocker, FR900359 (Fig. 2J) (25). From these data, we concluded that FFA4 in the lung is functionally active and couples to the Gq/11/phospholipase C/inositol phosphate signal transduction cascade and that the activity of our tool compounds TUG-891 and TUG-1197 in driving a calcium response in lung tissue is mainly via FFA4.

As FFA4 mediated a strong calcium response in PCLS, we examined the possibility that FFA4 might promote ASM contraction. Application of the bronchoconstrictors carbachol or serotonin resulted in a narrowing of airways in PCLS (fig. S1, A to C), responses that act here as a positive control. In contrast, the application of TUG-891 or TUG-1197 at concentrations that resulted in robust calcium responses in PCLS (for example, 50 μM; Fig. 2H) did not result in any apparent change in the diameter of airways in PCLS (fig. S1, B and C). These data indicated that FFA4 did not mediate ASM contraction. Next, we tested the possibility that FFA4 agonists could mediate ASM relaxation in precontracted airways. Our data revealed that both TUG-891 and TUG-1197 acted in a concentration-dependent manner to increase the diameter of airways precontracted with carbachol (Fig. 3, A to D), with a potency [-log10 of the half maximal response (pEC50)] of 4.60 ± 0.10 and 4.77 ± 0.15, respectively.

Fig. 3 Activation of FFA4 leads to airway relaxation in precontracted murine airways.

(A and B) Representative images of the concentration-dependent relaxation responses to (A) TUG-891 and (B) TUG-1197 in precision cut lung slices (PCLS) precontracted with carbachol (CCh). (C and D) Quantification of the data in (A) and (B). Data presented are means ± SEM (n = 5). (E) Representative images of the effect of the FFA4 antagonist AH7614 on the TUG-1197–mediated relaxation response. (F) Quantification of the experiment in (E) from n = 6 mice per group. (G) Representative images of the TUG-1197–mediated relaxation responses in PCLS from wild-type and FFA4-KO(βgal) mice. (H) Quantification of the experiment in (G) from n = 6 mice. Data in (F) and (H) are means ± SEM *P < 0.05 and **P < 0.01 as determined by ANOVA with Bonferroni post hoc test. n.s., not significant.

In the case of FFA4-mediated ASM relaxation, the maximal effective concentration of either TUG-891 or TUG-1197 returned PCLS airways to 70 to 80% of the diameter observed before carbachol pretreatment (Fig. 3, A to D). However, the FFA4-selective antagonist AH7614 (26) significantly reduced the response to both TUG-891 and TUG-1197 (P < 0.05) (Fig. 3, E and F, and fig S2, A and B). Similarly, ASM relaxation to TUG-1197 was significantly diminished (P < 0.05) in PCLS derived from FFA4-KO(βgal) mice (Fig. 3, G and H). In both antagonist experiments and in FFA4-KO(βgal) mice, there was a small response to TUG-1197 and TUG-891 that may represent some off-target effect of these ligands. However, further support for a broncho-relaxation response from FFA4 activation was evident from our studies using a chemically distinct, and previously described, selective FFA4 agonist compound A (14). This agent showed similar FFA4 agonist properties (pEC50 and Emax) to TUG-1197 and TUG-891 in in vitro calcium mobilization assays (pEC50 values of 6.76, 6.77, and 6.00 for TUG-891, TUG-1197, and compound A, respectively) (fig. S3A). We found that compound A mimicked the effects of TUG-891 and TUG-1197 in reducing ASM contraction in PCLS (fig. S3, A to C). Last, the relaxation response to FFA4 agonists was not restricted to airways pretreated with carbachol as airways precontracted with serotonin (5-hydroxytryptamine) also similarly relaxed in response to TUG-1197 (fig. S4, A and B).

FFA4 mediates bronchodilation in normal lung physiology and disease

The data from the above ex vivo experiments led to the prediction that FFA4 agonists would increase the diameter of airways and thereby decrease airway resistance in vivo. We tested this possibility in anesthetized mice whose airway resistance was elevated via administration of nebulized acetylcholine (27). Under these conditions, the response to acetylcholine alone in FFA4-KO(βgal) was comparable to that observed in wild-type mice (fig. 4A). We tested the effect of activation of FFA4 on the acetylcholine response by coadministration of nebulized TUG-891 or TUG-1197 with acetylcholine. Under these conditions, the acetylcholine-mediated increase in resistance was significantly (P < 0.01) attenuated (Fig. 4A and fig. S5). There was no substantial decrease in resistance after TUG-1197 treatment in FFA4-KO(βgal) mice (Fig. 4A).

Fig. 4 FFA4 agonism reduces airway resistance in healthy mice and produces broncho-relaxant, anti-inflammatory, and prophylactic effects in an ozone model of inflammatory lung disease.

(A) Broncho-relaxant effect of TUG-1197 (0.364 mg/ml) on acetylcholine-induced (185 mg/ml) airway resistance in healthy wild-type and FFA4-KO(βgal) mice. (B) Broncho-relaxant effect of TUG-1197 (0.364 mg/ml) on acetylcholine-induced (185 mg/ml) airway resistance in wild-type and FFA4-KO(βgal) mice that had been exposed to ozone (3 ppm, 3 hours) 24 hours before the lung resistance experiments. (C to J) Anti-inflammatory effect of repeated administration of TUG-891 in a 3-week ozone exposure model. Mice were exposed to either control air or 3 ppm ozone for 3 hours, twice a week for 3 weeks. Before each exposure, mice were treated with either vehicle or TUG-891 (0.036 mg/ml) by intranasal administration. Cells in BAL fluid were obtained and enumerated for total cell counts (C) and analyzed using flow cytometry to identify (D) alveolar macrophages, (E) monocyte-derived macrophages (MDM), and (F) neutrophil populations. Lung tissue (left lobe) was isolated, and RT-qPCR analysis was performed to identify changes in the expression of (G) KC/CXCL1 and (H) TNFα genes. (I) Prophylactic and (J) broncho-relaxant effect of TUG-891 in 3-week ozone model where TUG-891 (0.036 mg/ml) was administered 1 hour before each ozone exposure (I) and at the end of the ozone exposure (J) during acetylcholine-induced (50 mg/ml) lung resistance measurements. Data are means ± SEM of n = 4 to 7 animals. *P < 0.05, **P < 0.01, ***P < 0.001, Bonferroni multiple comparison test.

We next tested the effects of FFA4 activation within the context of an acute model of ozone pollution–induced inflammation. Under these conditions, we observed minimal hyperresponsiveness induced by exposure to ozone, but nevertheless, administration of TUG-1197 was again seen to decrease airway resistance (P < 0.05) in wild-type mice (Fig. 4B). A reduction in airway resistance to TUG-1197 was again absent in FFA4-KO(βgal) mice (Fig. 4B). We further tested the effects of FFA4 activation in a model of chronic (3 weeks) ozone exposure. This model demonstrated a profound inflammatory response as indicated by an increase in immune cells in broncho-alveolar lavage (BAL) fluid that included alveolar macrophages, monocyte-derived macrophages, and neutrophils, as well as up-regulation of inflammatory mediators KC/CXCL1 and tumor necrosis factor–α (TNFα) (Fig. 4, C to H). This inflammatory response was accompanied by significant (P < 0.05) airway hyperresponsiveness (Fig. 4, I and J). Under these conditions, administration of TUG-891 1 hour before each ozone exposure over the 3-week period significantly (P < 0.01) decreased airway hyperresponsiveness (Fig. 4I) and airway inflammation as indicated by a reduction of immune cells in BAL fluid (Fig. 4, C to F) and reduced inflammatory markers (Fig. 4, G and H), respectively.

The effects of TUG-891 on neutrophil infiltration in the chronic ozone model prompted us to test the possibility that FFA4-induced anti-inflammatory effects might, at least in part, be mediated via modulation of leukotriene-mediated neutrophil infiltration. To assess this, we analyzed the transcript abundance of 5-lipoxygenase (5-LOX), the enzyme responsible for leukotriene B4 (LTB4) production, and the LTB4 receptors leukotriene B4 receptor 1 (BLT1) (high affinity) and BLT2 (low affinity). We found that mRNA encoding these proteins was unchanged by 3 weeks of ozone exposure and that TUG-891 treatment similarly had no effect on transcript abundance (fig. S6, A to C). Expression of LTB4 itself increased with chronic ozone treatment, and this increase was prevented by administration of TUG-891 (fig. S6D). We also considered that other factors may contribute to airway neutrophil infiltration, including increased production of the T helper 17 cell interleukin-17α (IL-17α) (2831); however, mRNA corresponding to this ligand was unchanged after chronic ozone exposure or treatment with TUG-891 (fig. S6E).

We detected FFA4 expression in BAL cells that also stained positive for CD11c, GR-1, and Siglec-F (fig. S7), markers of monocytes, macrophages, and dendritic cells, respectively. This result suggests that FFA4 activation might have a direct anti-inflammatory effect via these immune cells. FFA4 was not detected in CD3+ or B220+ cells, which suggests that FFA4 is not expressed by T lymphocytes (fig. S7).

Although the above data indicate that FFA4 has an anti-inflammatory role that could contribute indirectly to the reduced hyperresponsiveness as seen, for example, in Fig. 4I, our results in PCLS, where we observe relaxation of ASM on FFA4 agonism (see Fig. 3), point to a more direct role for FFA4 on ASM relaxation. We therefore tested whether acute FFA4 activation could reduce airway resistance in mice after hyperresponsiveness had already been induced by chronic exposure to ozone. FFA4 transcript abundance was not affected by chronic ozone exposure (fig. S8A) nor was FFA4 signaling (calcium mobilization) altered (fig. S8, B and C). In settings in which hyperresponsiveness had been induced in mice by chronic ozone exposure, administration of TUG-891 together with acetylcholine significantly reduced (P < 0.05) airway resistance (Fig. 4J), indicating that FFA4 agonism can produce ASM relaxation even in the context of an already- established inflammatory lung disease.

In an alternative disease model, we induced airway hyperresponsiveness by an 8-week exposure to cigarette smoke. Acute administration of TUG-891 by nebulization before treatment with a bronchoconstrictor resulted in a reversal of hyperresponsiveness (Fig. 5A) indicating that, similar to the chronic ozone model, FFA4 activation in the cigarette smoke model produced ASM relaxation. This procedure correlated with substantial inflammation as indicated by an increase in numbers of alveolar macrophages and neutrophils (Fig. 5, B and C). Chronic administration of TUG-891 before exposure to each cigarette prevented airway inflammation, as measured by a reduction in alveolar macrophages and neutrophils at termination of the study (Fig. 5, B and C), suggesting that FFA4 activation in the smoke model was anti-inflammatory.

Fig. 5 FFA4 agonism reduces increased airway resistance and lung inflammation induced by chronic cigarette exposure as well as airway resistance caused by exposure to HDMs.

Mice were exposed to air or cigarette smoke for 8 weeks. (A) Transpulmonary (Rrs) airway resistance was measured in mice exposed to either control air or cigarette smoke treated for 3 min with nebulized TUG-1197 (0.364 mg/ml), followed by 3 min of nebulized methacholine (30 mg/ml). Mice were administered daily with TUG-891 (0.364 mg/ml) before each smoke exposure, and the effect of chronic TUG891 dosing on (B) macrophages and (C) neutrophil infiltration (in BAL fluid) was measured. Mice were sensitized with subcutaneous injection of either PBS or HDM extracts supplemented with Freunds complete adjuvant (CFA; 1:1 mixture 100 μg HDM) and then challenged with intranasal delivery of PBS or HDM (25 μg). Mice were euthanized and total cell counts (D) in BAL fluid were measured. CD4+ T cells (E) and eosinophils (F) were enumerated by flow cytometry. Mice were administered daily with either vehicle or TUG-891 (0.182 mg/ml) between the sensitization and final HDM challenge, and the prophylactic effect (G) of TUG-891 on lung resistance was measured. Data are means ± SD of three to nine animals. *P < 0.05, **P < 0.01, ***P < 0.001, ANOVA with Bonferroni post hoc test.

Last, we induced airway hyperresponsiveness in mice by sensitization to house dust mites (HDMs). After sensitization, immune cell counts in the BAL fluid (Fig. 5D), and specifically CD4+ T cells (Fig. 5E) and eosinophils (Fig. 5F), increased, consistent with the notion that this model reproduces important features of allergic asthma (32). Daily administration of TUG-891 after the initial sensitization and before the final HDM challenge significantly reduced (P < 0.05) airway hyperresponsiveness (Fig. 5G). Alongside the ozone and cigarette smoke models, these data indicate that FFA4 activation can ameliorate airway inflammation and hyperresponsiveness in the context of multiple respiratory disease models.

FFA4 is expressed in human airways and mediates bronchodilation

We detected FFA4 transcripts in human bronchial epithelial cells (HBECs) and human ASM derived from healthy donors by PCR and qPCR (Fig. 6, A and B). We confirmed the selectivity of a previously characterized in-house human (h) selective FFA4 antiserum (20) raised against a C-terminal peptide of hFFA4 (K342-R353). We detected hFFA4 as a broad band running at ~50 kDa only in lysates from hFFA4-transfected cells in Western blots of lysates prepared from Chinese hamster ovary (CHO) cells transfected with hFFA4 (Fig. 6C). This same antiserum detected hFFA4 in membrane and intracellular locations in immunofluorescent studies of hFFA4-transfected CHO cells (Fig. 6D). Using this antiserum, we investigated the expression profile of hFFA4 in isolated HBECs and ASM by flow cytometry (Fig. 6, E and F) and immunocytochemistry (Fig. 6, G and H) as well as in immunohistochemistry performed on human bronchial biopsies (Fig. 6, I to L, and fig. S9, A and B). These experiments established that hFFA4 was primarily expressed in the human airway epithelium with lower abundance in the ASM.

Fig. 6 FFA4 is expressed in HBECs and human smooth muscle cells and can be detected at both transcript and protein levels.

(A) Transcript abundances of FFA4 and 18S ribosomal RNA (rRNA) (as control) in bronchial epithelial cells (n = 5 donors) and ASM cells (n = 5 donors) using RT-PCR. A representative gel is shown. The expected sizes of PCR products are FFA4, 251 bp; 18S rRNA, 93 bp. (B) qRT-PCR of bronchial epithelial cells (n = 4 donors) and ASM cells (n = 5 donors) using ACTB housekeeping gene as a control. Data are means ± SEM. (C) Western blot of human FFA4 transiently transfected in human embryonic kidney–293 cells using an antiserum selective for FFA4 developed in-house (hFFA4-AB). (D) Immunofluorescence images of human FFA4 stably transfected in CHO-Flip-In cells (CHO-FFA4) and control nontransfected CHO cells (CHO-NT). (E and F) Example fluorescent histogram of FFA4 expression (black trace) in bronchial epithelial cells and ASM by flow cytometry versus isotype control antibody (gray shading); fold increase in geometric mean fluorescence intensity (GMFI) of anti-FFA4/isotype control antibody (95% confidence interval) 3.019 (2.042 to 3.996) (n = 11 donors, P < 0.01) and ASM cells 2.241 (1.662 to 2.819) (n = 14 donors, P < 0.001) by two-tailed Student’s t test against isotype control. PE, phycoerythrin. (G and H) Representative photomicrograph (×40 magnification) showing FFA4 expression (red; isotype control antibody inserts) in bronchial epithelial cells (n = 5 donors) and ASM cells (n = 4 donors) with nuclei stained blue by 4′,6-diamidino-2-phenylindole (DAPI) immunofluorescence. (I to L) Representative photomicrographs of normal human bronchial biopsy sections stained with (I) isotype control antiserum, (J) human FFA4-selective antiserum (4 μg/ml), and (K) zoom of selected region of the epithelium from (L). EC, epithelial cell layer; ASM, airway smooth muscle; LP, Lamina propria.

The expression profile observed in human lung closely mirrored that seen in the mouse lung, further suggesting that the mouse airway responses reported here might mimic those of hFFA4 in human airways. We further tested this possibility by monitoring calcium mobilization in response to TUG-891 in human ASM and HBECs. TUG-891 (50 μM) caused a statistically significant increase in calcium mobilization in both ASM (P < 0.01) and HBECs (P < 0.05) (Fig. 7, A and B). We next assessed whether FFA4 in human ASM might mediate relaxation using a collagen gel preparation wherein human ASM contraction can be monitored by reduction in the diameter of the collagen gel (33). Treatment of ASM in this preparation with TUG-891 reversed both spontaneous contraction and contraction observed in the presence of the bronchoconstrictor acetylcholine (Fig. 7, C and D). These data demonstrate that FFA4 is expressed and functional in isolated human ASM cells and that activation of this receptor can promote human ASM relaxation ex vivo.

Fig. 7 Functional responses mediated by FFA4 in isolated human lung tissues and cells.

(A) Intracellular calcium (iCa2+) elevation in bronchial epithelial cells (n = 6 donors) and (B) ASM cells (n = 6 donors) in response to TUG-891 treatment or ionomycin (1.5 μg/ml) as a positive control. GMFI equates to total stimulated GMFI minus matched baseline minus vehicle control. Data are plotted as means ± SEM. (C) Representative gel photographs taken at 0 hours (basal) and 24 hours either uncontracted or precontracted with carbachol (CCh, 100 μM) followed by treatment of TUG-891 (50 μM). (D) Percentage contraction of collagen gels in the presence of ASM cells (n = 6 donors) either uncontracted or precontracted with carbachol (CCh, 100 μM) followed by treatment of TUG-891 (50 μM). Data are plotted as means ± SEM. (E) Concentration-response data for TUG-891–mediated relaxation of human airway strips precontracted with carbachol (100 μM). Data in (E) are means ± SEM of at least three independent experiments from healthy donors. *P < 0.05, **P < 0.01, ANOVA with Bonferroni post hoc test.

We extended these studies by testing human ASM contractile responses in intact lung tissue (undenuded bronchi) isolated from healthy participants using wire myography. In these experiments, TUG-891 produced a concentration-dependent (pEC50, 4.34 ± 0.13) relaxation of human airways that had been precontracted with acetylcholine (Fig. 7E).

Potential role of prostaglandins in the mechanism of FFA4-mediated bronchodilation

It has been known for some time that through an interplay between EP1 (contraction) and EP2 (relaxation) prostanoid receptors, prostaglandins can regulate lung function via both contraction and relaxation of ASM (34). Among the prostanoids reported to have an impact on lung function is prostaglandin PGE2 (34). PGE2 is released from various lung cell types including ASM (3537), airway epithelium (38), and immune/inflammatory cells of the lung (39) and can mediate bronchodilation via EP2 receptors (4042). The bronchodilation properties of a number of GPCR ligands including bradykinin, substance P, adenosine 5′-triphosphate, and proteases that activate proteinase-activated receptors such as PAR2 are mediated by the release of PGE2 (36, 43, 44). Therefore, in addition to directly acting on ASM, we tested the possibility that FFA4 might also mediate bronchodilation via the release of PGE2. Incubation of PCLS derived from wild-type mice with TUG-891 resulted in a significant release of PGE2 (P < 0.05), whereas the same concentration of TUG-891 applied to PCLS from FFA4-KO(βgal) mice generated no significant (P > 0.05) PGE2 release (Fig. 8A). We confirmed that PGE2 administration to precontracted PCLS derived from wild-type mice resulted in relaxation of ASM (Fig. 8B). Moreover, the EP2-selective antagonist, PF-04418948 (45), significantly (P < 0.01) blocked the relaxation of ASM in response to TUG-891 in PCLS (Fig. 8C). TUG-891 was used in this set of experiments because our in vitro studies indicated that it was the more potent of the two compounds (Fig. 2, A, C, and E). Last, we measured PGE2 concentrations in bronchoalveolar lavage fluid (BALF) from mice simultaneously exposed to ozone and treated with TUG891. We observed a significant increase (P < 0.05) in PGE2 after chronic administration of TUG891 (Fig. 8D).

Fig. 8 FFA4-mediated airway relaxation in mouse is partly dependent on PGE2 and the prostanoid receptor EP2.

(A) PGE2 release from PCLS obtained from the lungs of wild-type mice and FFA4 KO(βgal) mice detected by enzyme-linked immunosorbent assay. (B) Effects of PGE2 (500 nM) treatment of PCLS precontracted with carbachol (CCh). (C) TUG-891 response on PCLS pretreated with the selective EP2 receptor antagonist PF-04418948 (10 μM) for 1 hour. (D) PGE2 release into the bronchoalveolar lavage fluid (BALF) of mice exposed to normal air or chronically exposed to ozone and treated with either vehicle (1% dimethyl sulfoxide) or TUG891 (100 μM). Data are means ± SD of at least n = 4 animals. *P < 0.05, **P < 0.01, ANOVA with Bonferroni post hoc test.

Previous work has highlighted species differences in the roles of prostanoid receptor subtypes in mediating PGE2-dependent airway relaxation, with EP2 being the predominant subtype in the mouse and EP4 the key subtype in human (46, 47). We confirmed expression of all four known EP prostanoid receptors (EP1–4) in isolated HBECs (fig. S10A). However, only EP2–4 were expressed in human ASM (fig. S10B).

To further investigate the mechanism by which FFA4 mediates increased PGE2 production in mouse, we performed gene expression analysis in lung tissues obtained from the 3-week ozone exposure model. mRNA abundances of the majority of genes involved in the biosynthesis of PGE2, including PGE2 synthases and cyclooxygenase-1 (COX1), were not substantially altered, whereas there was a small but significant (P < 0.05) up-regulation of cyclooxygenase-2 (COX2) mRNA (fig. S11, A to E). This suggested that FFA4 activation may induce expression of COX2 to increase PGE2 concentrations. In the 3-week ozone exposure, model expression of the EP2 receptor was not significantly altered (fig. S11F), highlighting that this pathway would still be expected to be operational in the disease state.

Because the FFA4 agonists appeared to produce a response via EP2 receptors, we assessed whether either TUG-891 or TUG-1197, which represent very different chemotypes, could directly activate EP2 receptors. TUG-891 (10 μM) produced no apparent activation of the EP2 receptor (table S1). TUG-891 also lacked agonist function at other EP prostanoid receptors, including the EP1 receptor (table S1). Equally TUG-891 did not block effects of endogenously generated prostanoids at any of the EP prostanoid receptors (table S2). There was a similar lack of effect of 10 μM TUG-1197 as an agonist (table S1) or antagonist (table S2) on each of the EP1–4 prostanoid receptors. These data provide further support that the effect of ligands with potency at FFA4, in both ex vivo and in vivo studies, is mediated by FFA4 and not via off-target activity at, for example, prostanoid receptors.

Because we established a link here between FFA4 bronchodilation and PGE2, we wanted to test whether PGE2 might also be involved in the FFA4-mediated anti-inflammatory response observed in the mouse disease models. PGE2 concentrations were increased in BALF after chronic administration of a FFA4 agonist in the ozone model (Fig. 8D), a response that correlated with an increase in COX2 transcription (fig. S11, A to E), the enzyme response for prostaglandin biosynthesis including PGE2. We tested the effects of the COX2 inhibitor celecoxib on the FFA4-mediated anti-inflammatory response by monitoring changes in number of cells, TNFα concentrations, and IL-1β transcripts in BALF from mice chronically exposed to ozone. In these experiments, celecoxib alone was anti-inflammatory and phenocopied the TUG-891 anti-inflammatory response (fig. S12, A to C). The coadministration of celecoxib and TUG-891 provided an anti-inflammatory effect equivalent to celecoxib or TUG-891 alone (fig. S12, A to C). Hence, unlike the bronchodilation response where it was possible to define a role for PGE2 in the FFA4 response, we were not able to conclude from these experiments a specific role for PGE2 in FFA4-mediated anti-inflammation.


We present evidence that the GPCR FFA4 is expressed and active in the murine lung where it mediates ASM relaxation under normal physiological conditions, and in the context of respiratory disease models, FFA4 mediates both ASM relaxation and generates an anti-inflammatory effect. That these mouse studies have applicability to human disease is supported by our data establishing FFA4 expression and function in human ASM and HBECs and demonstrating the relaxation of human ASM in response to the FFA4 agonist TUG-891. Together, our data support the conclusion that the long-chain free fatty acid receptor FFA4 might be a tractable pharmacological target for the treatment of human airway diseases.

It is important to note the FFA4-selective agonist TUG-1197 and the FFA1/FFA4 coagonist TUG-891 were equally effective in mediating ASM relaxation in mouse lung ex vivo and in vivo experiments, whereas in human cell– and tissue-based studies, we found that TUG-891 was more effective. Although this might be explained by the fact that TUG-891 is slightly more potent than TUG-1197 at FFA4, it might also suggest that, in addition to FFA4, FFA1 may also contribute to the airway responses reported here, particularly in human lung where the expression of functional FFA1 has been reported (48). In contrast to our data, however, an ex vivo study has suggested that FFA1 might augment acetylcholine-mediated ASM contraction rather than mediate ASM relaxation (48). As our data support the opposite possibility, we would argue that there is a potential for dual FFA1/FFA4 agonists to be more effective agents at mediating ASM relaxation than FFA4-selective agonists. Hence, testing dual agonists alongside FFA4-selective agents in experimental medicine studies will determine which approach is desirable ahead of human clinical trials. It will also be important to reduce the possibility of off-target adverse responses before human trials by increasing the potency of our current tool compounds. That off-target activity of our current ligands may exist is indicated by the small calcium transient and a possibility of a relaxation response to TUG-1197 in PCLS derived from FFA4-KO(βgal) mice.

We provide evidence that one mechanism of action of FFA4 in promoting ASM relaxation in mouse is via receptor-mediated release of the prostaglandin PGE2 that subsequently acts at the EP2 prostanoid receptor to mediate relaxation. A similar mechanism of ASM relaxation has been reported for other GPCRs (36, 43, 44) raising the question of whether it might be beneficial to directly target EP2 to deliver ASM relaxation in a therapeutic context. However, prostanoid biology in the lung is complex. For example, EP4 is considered the most predominant subtype in humans (46, 49), and therefore, targeting specifically EP2 would be less likely to be effective in humans. Moreover, PGE2 has multiple modes of action in that it can activate four different receptor subtypes (EP1 to EP4), which transduce a variety of processes, including bronchodilation, cough, microvasculature leaks, and regulation of pulmonary blood vessel tone (47, 50, 51). Our data also show that mRNA encoding each of the four subtypes can be detected in human ASM and HBECs. Hence, the complexity of prostanoid biology makes directly targeting this system in airway disease extremely challenging.

Our study points to FFA4 as having a dual effect in the context of inflammatory airway disease. The first effect centers on an action on the ASM resulting in bronchodilation with the potential of symptomatic relief in conditions featuring airway restriction, such as asthma. Although our collagen gel contraction studies on human airways point to a direct effect of FFA4 on ASM, the relatively high expression of this receptor on epithelial cells means that an indirect effect cannot be ruled out. It is possible, for example, that the airway epithelium is the origin of FFA4-mediated PGE2 release that subsequently leads to ASM relaxation. Second, our data point to a role for FFA4 in the regulation of inflammation consistent with reports that FFA4, expressed on various immune cells including macrophages (13, 52), can act as a negative modulator of inflammation in the context of metabolic disease (13, 14, 53). Exposure of mice to repeated doses of ozone resulted in a profound inflammatory response, evidenced by an increase in immune cells (neutrophils and macrophages) in airway spaces and a corresponding elevation of gene expression for the macrophage proinflammatory cytokine TNFα, and the neutrophil chemotactic factor, KC/CXCL1. Intranasal pretreatment of mice with TUG-891 before each exposure to ozone reduced the expression of these proinflammatory cytokines and prevented immune cell infiltration. Thus, our data indicate that in addition to an effect on ASM contraction with potential for symptomatic treatment of inflammatory airway disease, FFA4 activation could also have an anti-inflammatory property that would provide for disease modification.

It is not clear from our study whether the dual action of FFA4, namely, relaxation of ASM and anti-inflammation, is via a shared mechanism. For example, we provide evidence that ASM relaxation is in part mediated by the release of PGE2 acting on EP2 receptors. However, we currently have no evidence for the involvement PGE2 release in the anti-inflammatory response. The detection of FFA4 expression in CD11c+, GR-1+, and Siglec-F+ immune cells in the lung might indicate that FFA4 has a direct role in regulating immune cell response and therefore an anti-inflammatory mechanism distinct from that which mediates ASM relaxation. Despite these unresolved questions regarding the mechanism and cell types involved in mediating FFA4 dual effects on ASM and inflammation, our study reveals the therapeutic potential of targeting FFA4 in inflammatory lung disease.

A further important question that remains to be addressed is the origin of endogenous LCFAs that regulate lung FFA4. It is noteworthy that we observed no difference in the “basal” airway resistance between wild-type and FFA4-KO mice, indicating that more sensitive approaches and possibly additional mouse models with modulated endogenous lung fatty acid metabolism will be required to fully appreciate the role of endogenous free fatty acids. However, the omega-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) have been reported to be enriched in airway mucosa (54) where they are proposed to act as precursors for the biogenesis of lipid mediators such as the resolvins that resolve inflammatory responses (55) including lung inflammation (30) via interacting with members of the GPCR family including the chemerin-1 and chemerin-2 receptors. Whether this enriched pool of dietary-derived omega-3 fatty acids, which can potentially be augmented by the recruitment of circulating DHA and EPA at the sites of inflammation (56) or that can be released from membranes by the activity of phospholipase A2 (57), might be the source of endogenous LCFA activators for FFA4 is a possibility that is currently under investigation.

Our data demonstrate that expression of FFA4 in human airways is similar to that observed in mouse and that FFA4 agonism promotes relaxation isolated human ASM and of human airway strips. In this way, we provide proof of concept that pharmacological activation of lung FFA4 can generate in vivo efficacy. The precise pharmacological properties of a clinically effective hFFA4 agonist have yet to be fully established. In particular, it is unclear whether a full or partial agonist, a dual FFA4/FFA1 agonist, or indeed a G protein–biased ligand that might avoid potential challenges such as receptor desensitization and tachyphalaxis (20), would be most efficacious. Our study does, however, provide the first evidence that FFA4 can promote airway relaxation and resolve airway inflammation and thus may be a target for the treatment of airway diseases associated with bronchoconstriction and inflammation.


Study design

Studies were primarily designed to characterize the pharmacological properties of FFA4 ligands, investigate FFA4 expression in the lung and immune cells, and define the effects of FFA4 activation in lung function in healthy lungs and in the context of inflammatory lung disease. A range of in vitro assays using cell lines, ex vivo assays using mice and human lung tissues/cells, and whole-body in vivo assays were performed. For in vitro assays, at least three independent experiments were carried out, and in the case of primary cells and isolated tissues, at least three donors were tested if not indicated differently. All participants gave written informed consent, and the study was approved by the Leicester Research Ethics Committee (reference number 4977). For in vivo studies, sex- and age-matched animals were randomly assigned to drug or vehicle treatment groups. Animal numbers for each study type were determined by the investigators on the basis of previous experience with the disease models or from pilot studies. Mice were humanely euthanized at defined study endpoints, and all experimental procedures were performed under a Project License from the British Home Office, UK, under the Animals (Scientific Procedures) Act 1986.

Ozone exposure

For the acute ozone exposure protocol, wild-type adult male C57BL6/N mice (8 to 12 weeks old) and FFA4-KO(βgal) mice (a kind gift from AstraZeneca as described previously) (58) were administered with either medical air or 3 parts per million (ppm) ozone for 3 hours in a sealed Perspex container (EMB104, EMMS). An ozone concentration of 3 ppm was continuously monitored with an ozone probe (ATi Technologies). Twenty-four hours after exposure, mice were used for experiments. For subchronic treatment, mice were administered intranasally with either vehicle or FFA4 agonist (TUG-891, 100 μM, 25 μl volume) 1 hour before ozone or control air treatments. Mice received 3 ppm ozone produced from an ozonizer (Model 500 Sander Ozoniser) for 3 hours, twice a week, for a total duration of 3 weeks. Control animals received filtered air only for the equivalent exposure period. In some experiments, mice were exposed to ozone for 3 weeks and then administered TUG-891 at the end, during the lung function measurements.

HDM mouse model

Adult male Balb/c mice (8 to 12 weeks old from Envigo) were sensitized on day 0 with subcutaneous injection of 100 μg in 100-μl volume of HDM extract (HDM, Citeq Biologics) supplemented with Freund’s Complete Adjuvant (1:1 mixture, Sigma-Aldrich). Control mice were given an injection with equivalent volume of phosphate-buffered saline (PBS) (Life Technologies). Mice were then administered TUG-891 daily (500 μM, 25-μl volume) for 15 days via intranasal delivery under general anesthetic. On day 15, mice were challenged with 25 μg of HDM in 25-μl volume delivered intranasally under anesthetic condition. Mice were used for experiments 48 hours after challenge.

Cigarette mouse model

Six- to eight-week-old female wild-type mice (Australian BioResources) were exposed to 12 3R4F cigarettes (University of Kentucky) using a custom-designed and purpose-build nose-only smoke system (CH Technologies) twice per day, five times per week for 8 weeks, or room air as previously described (5961). In some experiments, 100 μM TUG-891 was administered daily before cigarette smoke exposure.

Statistical analysis

Data analysis and curve fitting were carried out using the GraphPad Prism software Version 7.0 (GraphPad Software Inc.). Concentration-response data were fitted to three-parameter sigmoidal concentration-response curves. Normality tests were performed using Shapiro-Wilk, and nonnormally distributed data (P < 0.05) were analyzed using a nonparametric two-tailed Mann-Whitney U test when comparing two groups or a Kruskal-Wallis test with Dunn’s multiple comparison posttest when assessing more than two groups. Statistical analysis of normally distributed data was carried out using standard t tests, one-way analysis of variance (ANOVA) with Tukey’s post hoc analysis, or two-way ANOVA combined with Bonferroni post hoc analysis as appropriate. Unless otherwise stated, data presented represent means ± SEM of three independent experiments.


Materials and Methods

Fig. S1. FFA4 activation does not cause airway contraction in murine PCLS.

Fig. S2. TUG-891-mediated relaxation in precontracted PCLS was reduced by the FFA4 antagonist AH7614.

Fig. S3. FFA4 agonist compound A mediates ASM relaxation in precontracted PCLS.

Fig. S4. FFA4 agonism mediates relaxation in airways precontracted with serotonin.

Fig. S5. FFA4 agonist TUG-891 reduces airway resistance in wild-type C57BL6/N mice.

Fig. S6. FFA4 agonism did not alter transcript abundance of 5-Lox, BLT1, BLT2, or IL17a but increased expression of LTB4 in a chronic ozone exposure model.

Fig. S7. FFA4 is expressed in Gr-1+, CD11c+, and Siglec-F+ immune cells.

Fig. S8. FFA4 is normally expressed and fully functional in ozone-exposed mice.

Fig. S9. Expression of FFA4 in human lung.

Fig. S10. PGE2 receptors are expressed in HBECs and human ASM cells.

Fig. S11. FFA4 agonism up-regulates COX-2 but not other PGE2-related transcripts in a chronic ozone model.

Fig. S12. FFA4 activation mimics the anti-inflammatory effects of a COX-2 inhibitor but is not altered by the presence of COX-2 inhibition.

Table S1. FFA4 agonists do not act as off-target agonists at prostanoid receptors.

Table S2. FFA4 agonists do not act as off-target antagonists at prostanoid receptors.

Data file S1. Raw data from figures.

References (6265)


Acknowledgments: We thank E. Kostenis (University of Bonn) for FR900359 and both C. Azevedo and B. Shimpukade (University of Copenhagen) for assistance with compound synthesis. We acknowledge the BSU facilities at the Cancer Research UK Beatson Institute and the Biological Services and Flow Core at the University of Glasgow. We thank M. Gaellman and A. Ryan for administrative support of the Tobin and Milligan laboratories. Funding: This study is funded by the following Biotechnology and Biosciences Research Council (BBSRC) grants BB/K019864/1 (to G.M.) and BB/K019856/1 (to A.B.T.) and by the Medical Research Council (MRC), grant MR/R00305X/1 (to G.M. and A.T.B.). C.H.W. and K.F.C. were funded by MRC grant G1001367/1 and TU by The Danish Council for Strategic Research (grant 11-116196). Funding and fellowships for C.D. and P.M.H. are from NHMRC 1079187, 1059238, and the Rainbow Foundation. P.M.H. was also funded by a fellowship and grants from the NHMRC (1079187 and 1175134). R.P. received funding from The Royal Society and Welcome Trust Institutional Strategic Support Fund. This paper also contains independent research funded by the National Institute for Health Research (NIHR) Leicester Respiratory Biomedical Centre. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health, UK. Author contributions: A.B.T. and G.M. conceived and led the project. A.B.T., G.M., and R.P. wrote the paper. R.P. designed and coordinated the study. C.E.B. designed and led the human airway studies. C.D., P.M.H., R.P., and R.Y.K. designed and performed the cigarette study. R.P., K.F.C., C.H.W., and J.G.L. contributed to the design and performed the ozone experiments. E.A.-C., M.R.T., Z.D., A.G.M.A., and E.E. performed the pharmacology study. L.C. and D.K. performed the human airway study. T.U. generated FFA1/FFA4 ligands. Competing interests: During the period of the study, G.M. received payment to institution grants and personal fees from Sosei-Heptares, Bayer, Galapagos NV, and Confo-therapeutics, unrelated to the current study. G.M. and T.U. are directors of Caldan Therapeutics. C.E.B. received payment to institution grants and personal fees from GSK, AZ, Novartis, Sanofi, Regeneron, BI, Chiesi, Roche/Genentech, Mologic, 4DPharma, and Gossamer, also unrelated to the current study. The other authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials. All reagents and materials used in this study may either be purchased or obtained from the corresponding authors.

Stay Connected to Science Translational Medicine

Navigate This Article