Yaohong Shia,
Yutaro Saito*a,
Masashi Hotta
b,
Mayuko Akita
a,
Akiharu Uwamizu
c,
Junken Aokic,
Jun Kunisawa
bdefg and
Shinsuke Sando
*ah
aDepartment of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. E-mail: saito.y@chembio.t.u-tokyo.ac.jp; ssando@chembio.t.u-tokyo.ac.jp
bLaboratory of Vaccine Materials, Laboratory of Gut Environmental System, Microbial Research Center for Health and Medicine, National Institutes of Biomedical Innovation, Health and Nutrition (NIBN), 7-6-8 Asagi Saito, Osaka 567-0085, Ibaraki, Japan
cDepartment of Health Chemistry, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Bunkyo-ku, Tokyo, Japan
dDepartment of Microbiology and Immunology, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe, Hyogo 650-0017, Japan
eInternational Vaccine Design Center, The Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan
fGraduate School of Medicine, Graduate School of Dentistry, Graduate School of Pharmaceutical Sciences, Department of Science, The University of Osaka, 1-1 Yamadaoka, Suita, Osaka 565-0871, Japan
gResearch Organization for Nano and Life Innovation, Waseda University, Tokyo, Japan
hDepartment of Bioengineering, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
First published on 8th September 2025
Polyunsaturated fatty acids (PUFAs), fatty acids with multiple unsaturated carbon–carbon bonds, constitute a crucial class of lipids. While the vast diversity of PUFA species arises from their structural variations, most of them are poorly investigated due to their limited availability. Here, we utilize solid-phase synthesis of PUFAs, which we have recently developed, to construct a PUFA library. With the library, we investigate agonist activities for free fatty acid receptors 1 (FFAR1) and 4 (FFAR4), inhibition abilities against neutrophil pseudopod formation, and anti-inflammatory effects on the hypersensitivity mouse model. This study highlights the utility of PUFA solid-phase synthesis as a powerful approach for elucidating the relationship between the chemical structures of PUFAs and their biological functions.
The chemical structure of a typical PUFA can be divided into three moieties, as shown in Fig. 1A. Head moiety comprises a carboxyl group and a saturated fatty chain. Middle moiety consists of an unsaturated fatty chain with sequential methylene-interrupted Z-olefins. Tail moiety is a terminal alkyl chain. Here, we define these three structural components using the following parameters: head moiety: δ, representing the number of carbon atoms from the carboxyl carbon to the first olefin; middle moiety: π, indicating the number of olefins; and tail moiety: ω, denoting the number of carbon atoms from the terminal methyl group to the closest olefin. For instance, EPA (1) is expressed as [δ,π,ω] = [5,5,3].
![]() | ||
Fig. 1 Schematic illustration of the research background and key technologies in this work. (A) Chemical structures of PUFAs and their FFAR1 agonist activities. (B) Structures of FFAR1 (PDB ID: 4PHU) and FFAR4 (PDB ID: 8ID6), along with research progress of their agonists. (C) Schematic illustration of solid-phase synthesis of PUFAs.12 |
Free fatty acid receptors 1 (FFAR1, GPR40) and 4 (FFAR4, GPR120) are G protein-coupled receptors (GPCRs) with attractive attention as potential therapeutic targets for various diseases, including type-2 diabetes, autoimmune disorders, neurodegenerative diseases, and cancers.4,5 FFAR1/4 agonists have been extensively studied and remain challenging targets for drug development and disease prevention.6,7 These receptors are activated by a wide range of medium- and long-chain fatty acids.3,8–10 PUFAs are recognized as their natural agonists, contributing to health maintenance through dietary intake and supplementation.
Although some structure–activity relationship (SAR) studies have investigated the relationship between PUFAs and FFAR1/4, existing research has predominantly focused on PUFAs available from nature, leading to an insufficient understanding of SAR (Fig. 1B). Recent findings, however, have demonstrated that FFAR1/4 agonist activity is highly sensitive to subtle structural variations in PUFAs. For instance, Voest and colleagues reported that hexadeca-4,7,10,13-tetraenoic acid ([4,4,3], 3) exhibits potent FFAR1/4 agonist activity.11 We also found that FFAR1 agonist activity, specifically Gαq and Gα12/13 signalling, varies depending on carbon chain length and the number of olefins.12 Notably, [5,6,3] (4), which contains additional olefin and methylene group compared to EPA (1), exhibits higher FFAR1 agonist activity than 1, but lower than 3 (Fig. 1A). These findings underscore the necessity of a comprehensive examination of FFAR1/4 activation by PUFAs, extending beyond commonly available PUFAs to include structurally refined variants that may provide deeper insights into their SAR.
Recently, we developed a solid-phase synthesis approach for PUFAs.12 Solid-phase synthesis is widely used in the preparation of biopolymers such as peptides and nucleic acids, enabling efficient construction of highly diverse compound libraries.13–16 However, solid-phase synthesis has rarely been applied to lipids, including PUFAs. The developed method enables the efficient synthesis of PUFAs through streamlined operations and facilitates PUFA library construction containing diverse structures by controlling [δ,π,ω] (Fig. 1C). In contrast to conventional liquid-phase synthesis, which entails highly time-consuming and laborious processes, the solid-phase synthesis approach effectively overcomes these limitations.
In this study, we investigated FFAR1-dependent Gαq and Gα12/13 signalling, as well as FFAR4-dependent Gαq, Gα12/13, and Gαo signalling, utilizing a compound library comprising 30 PUFAs. Of these, 9 PUFAs were obtained from a commercial supplier, 3 PUFAs were previously reported,12 and 18 PUFAs were synthesized by the solid-phase synthesis. We selected several PUFAs exhibiting distinct activation profiles on FFAR1/4 and examined their effects on neutrophil pseudopod formation, a process known to be inhibited through FFAR1 activation.17
In Step 1, linear carboxylic acids (C5–12) bearing a terminal alkyne (6) were loaded onto polystyrene-based 2-chlorotrityl chloride resin (5) (Fig. 2, Step 1). The δ value is determined by the carbon number of the loaded carboxylic acid. Subsequently, the carbon chain was elongated by constructing a skipped diyne framework via the nucleophilic addition of copper acetylide to 3-(trimethylsilyl)propargyl bromide (8), followed by removing the trimethylsilyl group using Et3N·3HF or tetra-n-butylammonium fluoride (TBAF) to regenerate the terminal alkyne (Fig. 2, Step 2). Repeating Step 2, until the desired chain length is achieved, determines the π value. In this study, PUFAs with π = 2–5 were synthesized by repeating Step 2 up to 3 times. Since one alkyne is introduced in both Step 1 and Step 3, the π value is defined as the repetition number of Step 2 (m) plus 2.
Following Step 2, the ω value was defined through a capping reaction on 9 using either propargyl bromide 10 or tosylate with an aliphatic chain, in place of reagent 8 (Fig. 2, Step 3). Finally, the target PUFAs were obtained via hydrogen transfer reduction using IPrCuOtBu, tBuOH, and polymethylhydrosiloxane (PMHS), followed by cleavage from the solid support (Fig. 2, Step 4). While the previous study exclusively utilized tBuOH as a proton donor,12 this study also employed MeOH in some cases.
The PUFAs in this library are summarized in Table 1. This library encompasses PUFAs with [δ,π,ω] = [4–11,2–6,1–9]. While this library incorporates PUFAs with [δ,π,ω] = [4–6,3–6,2–6], which was previously investigated for their FFAR1 agonist activity,12 FFAR4 agonistic activities are newly provided, and the current library expands the explored chemical space (Fig. 3). This expanded library, including PUFAs with δ ≥ 7, π ≤ 3, and ω = 7, enables a broader analysis of PUFA structures and their associated biological functions.
![]() | ||
Fig. 3 3D-plot of [δ,π,ω] of each PUFA in the PUFA libraries. The orange circles represent PUFAs only in the current work. The blue circles represent PUFAs only in our previous work.12 The red circles represent PUFAs in both libraries. |
![]() | ||
Fig. 4 Activation of FFAR1 and FFAR4 by PUFAs, measured using the TGFα shedding assay. (A) Bar graphs of the activation of FFAR1/4. TUG: TUG-891, a synthetic agonist for FFAR1 and 4.19 Concentration of TUG: 1.0 μM. Concentration of PUFAs: 3.0 μM. Each point represents an average value obtained from individual experiments (six replicates). Bar graphs show mean values, with error bars representing the standard deviation (SD), n = 3. (B) Heat-maps of FFAR1/4 agonist activity shown in (A). The numbers in each box represent mean values of normalized AP-TGFα release, indicating FFAR1/4 agonist activities. |
The results revealed that agonist activity varied depending on subtle differences in carbon chain length (δ), degree of unsaturation (π), and terminal structure (ω) (Fig. 4A).
Among the δ = 4 series, PUFA [4,2,9] (C16), which shares the same δ and π values but has a longer ω than the smallest PUFA in this library, [4,2,2] (C9), exhibited higher agonist activity for FFAR1 (P < 0.05 (FFAR1), [4,2,2] vs. [4,2,9]). PUFAs [4,3,6] and [4,4,3], which have the same carbon number C16 with [4,2,9], also exhibited higher agonist activities for both FFAR1 and 4 than [4,2,2] (P < 0.0001 (FFAR1), P < 0.0001 (FFAR4), [4,2,2] vs. [4,3,6]; P < 0.0001 (FFAR1), P < 0.0001 (FFAR4), [4,2,2] vs. [4,4,3]). This observation aligns with the known preference of FFAR1 and FFAR4 for long-chain fatty acids.4,5
Conversely, among PUFAs with the same total carbon number (C16), [4,3,6], which possesses one additional double bond and a shorter ω value (by three carbons), exhibited higher FFAR1 agonist activity than [4,2,9], suggesting a preference for increased unsaturation (P < 0.0001 (FFAR1), [4,2,9] vs. [4,3,6]). However, [4,3,9] (C19), which has an extended carbon chain, demonstrated reduced activity (P < 0.0001 (FFAR1), [4,3,6] vs. [4,3,9]), whereas [4,4,3] (C16), which maintains the same chain length as [4,3,6] but with an additional double bond (π = 4), displayed comparable activity levels.
Among the δ = 5 series, PUFA [5,4,2] (C16) exhibited higher FFAR1 agonist activity than [5,2,9] (C17), [5,3,3] (C14), [5,4,6] (C20), and [5,5,2] (C19) (P < 0.05, [5,4,2] vs. [5,2,9]; P < 0.05, [5,4,2] vs. [5,3,3]; P < 0.0001, [5,4,2] vs. [5,4,6]; P < 0.0005, [5,4,2] vs. [5,5,2]). This observation is consistent with the trend observed in the δ = 4 series, where C16 PUFAs demonstrated the highest activity. However, the results also suggest that carbon chain length alone does not solely determine activation potency. As mentioned above, [4,3,6] (C16) and [4,4,3] (C16) have higher FFAR1 agonist activities than [4,2,9] (C16). Additionally, in the comparison of PUFAs [4,2,9] (C16) and [5,4,2] (C16), the latter, which possesses a longer δ, higher degree of unsaturation, and a shorter ω, exhibited greater FFAR1 activation (P < 0.01 (FFAR1), [4,2,9] vs. [5,4,2]). Similarly, while [5,3,3] (C14), which shares the same π value as [5,3,6], displayed comparable activity with [5,3,6], [5,4,2] (C16) demonstrated a more pronounced FFAR1 agonist effect (P < 0.05 (FFAR1), [5,3,3] vs. [5,4,2]).
Among the δ = 6 series, [6,3,3] exhibited the highest FFAR1 agonist activity. Interestingly, no significant difference was observed in FFAR4 agonist activity in our assay. The C18 PUFA series, [6,2,9], [6,3,6], and [6,4,3] displayed comparable activity. This exemplified a case in which differences in the degree of unsaturation (π) and a terminal segment (ω) do not impact agonist activities. By contrast, a distinct difference was observed in PUFAs with δ = 9, as shown by the comparison between [9,2,6] and [9,3,3] (P < 0.0005 (FFAR1), P < 0.005 (FFAR4), [9,2,6] vs. [9,3,3]). In PUFAs with δ = 10, increasing ω enhanced FFAR1 agonist activity in comparison of [10,2,1] and [10,2,2] (P < 0.05 (FFAR1), [10,2,1] vs. [10,2,2]), whereas [10,2,3], with longer terminal chain (ω = 3), exhibited decreased activity (P < 0.05 (FFAR1), [10,2,2] vs. [10,2,3]). Further elongation of the tail moiety to ω = 6, [10,2,6], caused a larger decrease of both FFAR1/4 agonist activities (P < 0.0001 (FFAR1), P < 0.0001 (FFAR4), [10,2,2] vs. [10,2,6]).
In this study, no universal structural features were identified that directly correlate FFAR1/4 agonist activity with the chemical structures of PUFAs. However, among PUFAs with δ ≤ 9, there was a tendency for compounds with higher unsaturation (larger π, smaller ω) and chain lengths around C16 to exhibit relatively higher agonist activity. In contrast, for δ ≥ 10, compounds with shorter chain lengths (around C15), lower unsaturation (smaller π), and more compact terminal structures (smaller ω) often showed higher activity. Notably, exceptions to these trends were also observed.
Under these experimental conditions, a correlative trend was observed between FFAR1 and FFAR4 agonist activities, where PUFAs with potent FFAR1 activation also exhibited enhanced FFAR4 activation (Fig. 4B). Previous studies investigating naturally abundant PUFAs suggested that FFAR4 preferentially recognizes longer-chain PUFAs compared to FFAR1.3–5 When comparing [6,3,3] (C15) and [9,3,3] (C18), which differ in chain length (δ) but share identical π and ω values, [9,3,3] (C18) showed significantly higher activation of FFAR4 than [6,3,3] (C15), whereas FFAR1 was similarly activated by both PUFAs (P > 0.99 (FFAR1), P < 0.01 (FFAR4), [6,3,3] vs. [9,3,3]). However, when [5,3,3] (C14), [6,3,3] (C15), [9,3,3] (C18), [10,3,3] (C19), and [11,3,3] (C20) were compared, both FFAR1 and FFAR4 were activated to a comparable extent. Overall, in this study, utilizing a detection system that measures Gαq, Gα12/13, and Gαo signal in cells transiently overexpressing receptors, no clear difference in chain-length preferences between FFAR1 and FFAR4 was observed. To uncover a more detailed relationship between the structural features and FFAR1/4 preference, further investigations using other evaluation systems and an expanded library are necessary. Nevertheless, this analysis revealed that distinct activation profiles exist among individual PUFAs and that specific chemical structures can confer strong agonist activity for both FFAR1 and FFAR4.
![]() | ||
Fig. 5 Investigation of the anti-inflammatory effect of PUFAs. (A) Chemical structures of EPA (1), 17,18-EpETE (12), and [4,4,3] (3). (B) Fluorescent images showing the inhibition of neutrophil pseudopod formation. PUFA concentration: 100 nM. Green: phalloidin indicating the pseudopod formation, Blue: DAPI. Scale bars: 20 μm. N.I.: non LTB4-induction control. (C) Plots of pseudopod formation of neutrophils. Bar graphs represent averages of values obtained in six independent experiments. Error bars indicate the SD. ***P < 0.001, n.s.: not significant. N.I.: non LTB4-induction control. N.C.: negative control group. P.C.: positive control group. The control data (N.I., Vehicle, 1, and 12) are adopted from those described in our previous report.12 [4,4,3] (3) was examined in parallel with these controls under the same experimental settings. (D) Schematic representation of the experimental outline. (E) Plots of ear swelling. Data are combined from three independent experiments, and each point represents data from an individual mouse. Horizontal bars indicate medians. Statistical significance was evaluated by one-way analysis of variance with post-hoc Tukey test for multiple comparisons. N.I.: non DNFB-induction control. P.C.: positive control group. ****P < 0.0001, n.s.: not significant. (F) Plots of pseudopod formation of neutrophils. Bar graphs indicate averages of values obtained in six independent experiments. Error bars indicate the SD. ****P < 0.0001, n.s.: not significant. N.I.: non LTB4-induction control. |
To test this hypothesis, we first evaluated the effect of [4,4,3] using neutrophil pseudopod formation assay (Fig. 5B and C). In this assay, pseudopod formation was induced by leukotriene B4 (LTB4) stimulation in neutrophils isolated from mouse bone marrow, following pre-treatment with PUFAs or EtOH (vehicle control). After cell fixation, pseudopod formation was assessed via F-actin staining. Contrary to our expectations, [4,4,3] did not exhibit pseudopod inhibition even at 100 nM, a concentration at which 12 exerts a significant inhibitory effect.
We next examined whether [4,4,3] could exert anti-inflammatory effects in vivo. To evaluate its potential anti-inflammatory properties, we employed a hypersensitivity model in mice (Fig. 5D and E). In this experiment, mice received topical application of PUFAs or EtOH (vehicle control) on the abdominal skin, followed by sensitization with 1-fluoro-2,4-dinitrobenzene (DNFB) on Day 0. On Day 5, hypersensitivity was induced by applying DNFB to the ear 30 minutes after topical administration of PUFAs or EtOH, and ear swelling was measured on Day 7.
When 17,18-EpETE (12) was administered at 100 ng per treatment, ear swelling was significantly reduced compared to the vehicle group, consistent with previous reports.12,17 However, no significant difference was observed between the [4,4,3] and the vehicle group, indicating that despite the FFAR1 agonist activity, [4,4,3] exhibits no anti-inflammatory properties.
Previous studies have suggested that FFAR1 agonist activity plays a crucial role in mediating its anti-inflammatory effects.17 To further explore this, we examined the inhibitory effects of [5,4,2], [5,5,2], [6,3,3], [6,4,3], [10,2,3], and [10,2,6] on neutrophil pseudopod formation, as they have structural variations and exhibit different patterns of FFAR1/4 agonist activity (Fig. 5F).
At a concentration of 100 nM, none of the tested PUFAs produced a significant reduction in neutrophil pseudopod formation. Interestingly, contrary to our expectations, the PUFA with FFAR1/4 activation ability used in this study did not exhibit inhibitory effects. These findings may suggest that factors other than Gαq, Gα12/13, and Gαo signalling through FFAR1/4 contribute to neutrophil pseudopod formation and the anti-inflammatory response in this hypersensitivity model.
Despite exhibiting varying degrees of FFAR1/4 agonist activity, none of the tested PUFAs showed anti-inflammatory effects. Although it has been suggested that the anti-inflammatory effect is mediated by activation of FFAR1,17 our results suggest that the suppression of neutrophil pseudopod formation and its associated anti-inflammatory response is not solely governed by FFAR1 activation. Instead, alternative signalling pathways independent of FFAR1 agonism and/or FFAR1-mediated signals beyond Gαq and Gα12/13, measured in our TGFα shedding assay, may play a critical role.
Future research will require a comprehensive analysis of the structural determinants of PUFA-mediated anti-inflammatory effects through an expanded PUFA library, where chemical structures are precisely defined. In addition to the structural features described here, the effects of other structural factors, such as cis/trans-configurations, should also be investigated. In particular, it would be of interest to investigate whether the observed biological activities of individual PUFAs arise from their intrinsic agonist activity toward FFARs or are influenced by physicochemical properties such as solubility and aggregation behavior. Computational approaches, including docking-based binding energy estimation, could offer further insights.
This study highlights the utility of PUFA solid-phase synthesis as a powerful approach for elucidating the relationship between PUFA chemical structures and their biological functions.
This journal is © The Royal Society of Chemistry 2025 |