Structure–Activity Relationship Analysis of Fluoxetine for Suppression of Inflammatory Cytokine Production

Abstract

There is accumulating evidence that selective serotonin reuptake inhibitors (SSRIs), clinically used as antidepressants, have a beneficial effect on inflammatory diseases such as coronavirus disease 2019 (COVID-19). We previously compared the inhibitory effects of five U.S. Food and Drug Administration (FDA)-approved SSRIs on the production of an inflammatory cytokine, interleukin-6 (IL-6), and concluded that fluoxetine (FLX) showed the most potent anti-inflammatory activity. Here, we investigated the structure–activity relationship of FLX for anti-inflammatory activity towards J774.1 murine macrophages. FLX suppressed IL-6 production induced by the TLR3 agonist polyinosinic-polycytidylic acid (poly(I : C)) with an IC₅₀ of 4.76 µM. A derivative of FLX containing chlorine instead of the methylamino group lacked activity, suggesting that the methylamino group is important for the anti-inflammatory activity. FLX derivatives bearing an N-propyl or N-(pyridin-3-yl)methyl group in place of the N-methyl group exhibited almost the same activity as FLX. Other derivatives showed weaker activity, and the N-phenyl and N-(4-trifluoromethyl)benzyl derivatives were inactive. The chlorine-containing derivative also lacked inhibitory activity against TLR9- or TLR4-mediated IL-6 production. These derivatives showed similar structure–activity relationships for TLR3- and TLR9-mediated inflammatory responses. However, the activities of all amino group-containing derivatives against the TLR4-mediated inflammatory response were equal to or higher than the activity of FLX. These results indicate that the substituent at the nitrogen atom in FLX strongly influences the anti-inflammatory effect.

INTRODUCTION

Inflammatory reactions play key roles in maintaining the body’s homeostasis. For example, invading pathogens activate Toll-like receptors (TLRs), such as TLR3, TLR7, TLR8, and TLR9, which recognize nucleic acids derived from the pathogens.¹⁾ Activation of these pattern recognition receptors leads to the production of inflammatory cytokines that stimulate, recruit, and induce proliferation of immune cells, which function to eliminate the pathogens.

However, uncontrolled inflammatory responses can lead to the development or exacerbation of various inflammatory diseases with no established treatment, including sepsis, a systemic inflammatory syndrome triggered by infections.^2,3) Sepsis causes more than 10 million deaths annually, and has a mortality rate of more than 20%²⁾ as a result of multi-organ failure and septic shock due to “cytokine storm,” a pathology involving overproduction of inflammatory cytokines.³⁾ Steroidal anti-inflammatory drugs and antibody drugs have some therapeutic efficacy,^2,4,5) but more effective treatments are urgently required, especially for sepsis caused by viruses such as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).^1,4,5)

Selective serotonin reuptake inhibitors (SSRIs) are currently the most widely prescribed class of antidepressants,⁶⁾ and recent reports suggest that some SSRIs may be effective in treating coronavirus disease 2019 (COVID-19).^7–10) A possible reason for this is that SSRIs exhibit anti-inflammatory activity mediated by inhibition of TLRs, because we found that they suppress interleukin (IL)-6 production induced by stimulation of several TLRs. In particular, SSRIs strongly inhibit IL-6 production induced by activation of TLR3, which recognizes viral double-stranded RNA, in macrophages and dendritic cells.¹¹⁾ We conducted a comparative study of the anti-inflammatory effects of five clinically approved SSRIs, fluoxetine (FLX), paroxetine, fluvoxamine, sertraline, and escitalopram, and concluded that FLX might be the preferred SSRI for further development as a candidate to treat cytokine storms.¹¹⁾ Therefore, in this study, we investigated the structure–activity relationship of FLX for anti-inflammatory activity towards macrophages.

MATERIALS AND METHODS

Reagents

Fluoxetine (FLX) was purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Polyinosinic-polycytidylic acid (poly(I : C)) (a TLR3 agonist) was purchased from Tocris Bioscience (Minneapolis, MN, U.S.A.). Lipopolysaccharide (LPS) from Escherichia coli O55:B5 (a TLR4 agonist) was purchased from Sigma-Aldrich Co., LLC (St. Louis, MO, U.S.A.). CpG oligodeoxynucleotide (ODN; 1826) (a TLR9 agonist) was purchased from Novus Biologicals (Centennial, CO, U.S.A.).

Preparation of Fluoxetine Derivatives (FLX-b, e, g, h, i, and j)

FLX derivatives were synthesized from 3-chloro-1-phenylpropan-1-one via FLX-a according to the protocol reported by Silvestri and coworkers.¹²⁾ The spectroscopic data for FLX-a were identical with those reported.¹²⁾ The spectroscopic data of FLX-c, d, and f were reported in our previous paper.¹¹⁾ Newly synthesized products were fully characterized by IR spectroscopy (IR), ¹H- and ¹³C-NMR spectroscopy, and high-resolution mass spectrometry (HR-MS) using an electrospray-ionisation quadrupole time-of-flight (ESI/QTOF) instrument.

FLX-b: Yellow oil; IR (neat) ν_max = 3408, 3020, 2401, 1643, 1516, 1475, 1421, 1329, 1215, 1167, 1124 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 7.43 (d, J = 9.2 Hz, 2H), 7.30 (m, 5H), 6.90 (d, J = 8.8 Hz, 2H), 5.41 (dd, J = 8.4, 4.4 Hz, 1H), 3.02 (t, J = 7.2 Hz, 2H), 2.79 (t, J = 8.0 Hz, 2H), 2.38 (m, 2H), 1.72 (sext, J = 7.6 Hz, 2H), 0.90 (t, J = 7.6 Hz, 3H); ¹³C-NMR (100 MHz, CDCl₃) δ: 159.8, 139.4, 129.0 (2C), 128.3 (2C), 126.9 (q, J = 4 Hz, 2C), 125.8, 124.2 (q, J = 269 Hz), 123.3 (q, J = 33 Hz), 115.8 (2C), 77.7, 50.7, 44.9, 35.2, 20.0, 11.3; HR-MS (ESI/QTOF) m/z: Calcd for C₁₉H₂₃F₃NO ([M + H]⁺) 338.1726. Found 338.1726.

FLX-e: White solid; Mp = 126 °C; IR (KBr) ν_max = 3440, 2958, 2802, 1614, 1589, 1518, 1471, 1456, 1425, 1408, 1333, 1309, 1248, 1165, 1119, 1110 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 7.44 (d, J = 8.8 Hz, 2H), 7.37–7.28 (m, 5H), 6.90 (d, J = 8.4 Hz, 2H), 5.42 (dd, J = 8.0, 4.4 Hz, 1H), 3.11 (t, J = 7.2 Hz, 2H), 2.70 (d, J = 7.2 Hz, 2H), 2.50 (hep, J = 7.2 Hz, 2H), 2.09 (quin, J = 6.8 Hz, 1H), 0.97 (d, J = 6.8 Hz, 6H); ¹³C-NMR (100 MHz, CDCl₃) δ: 159.8, 139.5, 129.0 (2C), 128.3 (2C), 126.9 (q, J = 4 Hz, 2C), 125.8 (2C), 124.2 (q, J = 270 Hz), 123.4 (q, J = 33 Hz), 115.9 (2C), 78.1, 56.0, 45.6, 35.3, 26.2, 26.0; HR-MS (ESI/QTOF) m/z: Calcd for C₂₀H₂₅F₃NO ([M + H]⁺) 352.1883. Found 352.1880.

FLX-g: White solid; Mp = 61 °C; IR (KBr) ν_max = 3401, 3020, 1616, 1518, 1495, 1454, 1423, 1329, 1250, 1215, 1178, 1165, 1122, 1113 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 7.42 (d, J = 8.4 Hz, 2H), 7.28 (m, 10H), 6.88 (d, J = 8.8 Hz, 2H), 5.34 (q, J = 4.4 Hz, 1H), 3.78 (t, J = 13.6 Hz, 2H), 2.80 (m, 2H), 2.22 (m, 1H), 2.04 (m, 1H); ¹³C-NMR (100 MHz, CDCl₃) δ: 160.5, 141.0, 139.9, 128.8 (2C), 128.4 (2C), 128.2 (2C), 127.8 (2C), 127.0, 126.7 (q, J = 4 Hz, 2C), 125.7, 124.4 (q, J = 270 Hz), 122.7 (q, J = 32 Hz), 115.7 (2C), 78.6, 53.8, 45.4, 38.7; HR-MS (ESI/QTOF) m/z: C₂₃H₂₃F₃NO ([M + H]⁺) 386.1726. Found 386.1727.

FLX-h: Colorless oil; IR (neat) ν_max = 3643, 3020, 1616, 1589, 1518, 1454, 1421, 1327, 1215, 1165, 1126 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 7.52 (d, J = 8.0 Hz, 2H), 7.43 (d, J = 8.8 Hz, 2H), 7.39 (d, J = 8.0 Hz, 2H), 7.30 (m, 5H), 6.87 (d, J = 8.4 Hz, 2H), 5.35 (dd, J = 8.4, 4.8 Hz, 1H), 3.84 (dd, J = 16.0, 14.0, Hz, 2H), 2.79 (m, 2H), 2.21 (m, 1H), 2.01 (m, 1H); ¹³C-NMR (100 MHz, CDCl₃) δ: 160.5, 144.4, 140.9, 129.2 (q, ²J (C,F) = 32 Hz), 128.8 (2C), 128.2 (2C), 127.8 (2C), 126.4 (q, J = 4 Hz, 2C), 125.7, 125.3 (q, J = 4 Hz, 2C), 124.3 (q, J = 270 Hz), 124.2 (q, J = 270 Hz), 122.9 (q, J = 32 Hz), 115.6 (2C), 78.4, 53.4, 45.4, 38.9; HR-MS (ESI/QTOF) m/z: Calcd for C₂₄H₂₂F₆NO ([M + H]⁺) 454.1600. Found 454.1598.

FLX-i: Colorless oil; IR (neat) ν_max = 3332, 3016, 2954, 2935, 2837, 1614, 1587, 1514, 1454, 1423, 1329, 1250, 1217, 1178, 1163, 1119, 1113 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 7.42 (d, J = 8.8 Hz, 2H), 7.28 (m, 5H), 7.20 (d, J = 8.4 Hz, 2H), 6.88 (d, J = 8.4 Hz, 2H), 6.82 (d, J = 8.4 Hz, 2H), 5.33 (dd, J = 8.4, 4.8 Hz, 1H), 3.78 (s, 3H), 3.71 (t, J = 13.8 Hz, 2H), 2.79 (m, 2H), 2.21 (m, 1H), 2.01 (m, 1H); ¹³C-NMR (100 MHz, CDCl₃) δ: 160.5, 158.6, 141.0, 132.2, 129.3 (2C), 128.7 (2C), 127.7(2C), 126.7 (q, J = 4 Hz, 2C), 125.7, 124.3 (q, J = 269 Hz), 122.6 (q, J = 32 Hz), 115.7 (2C), 113.7 (2C), 78.5, 55.2, 53.3, 45.3, 38.8; HR-MS (ESI/QTOF) m/z: Calcd for C₂₄H₂₅F₃NO₂ ([M + H]⁺) 416.1832. Found 416.1831.

FLX-j: Yellow oil; IR (neat) ν_max = 3604, 3020, 2401, 1614, 1518, 1477, 1425, 1329, 1215, 1178, 1165, 1113 cm⁻¹; ¹H-NMR (400 MHz, CDCl₃) δ: 8.54 (d, J = 2.4 Hz, 1H), 8.49 (dd, J = 4.8, 1.6 Hz, 1H), 7.62 (dt, J = 8.0, 2.0 Hz, 1H), 7.43 (d, J = 8.8 Hz, 2H), 7.35–7.24 (m, 5H), 7.20 (dd, J = 7.6, 4.8 Hz, 1H), 6.88 (d, J = 8.8 Hz, 2H), 5.34 (dd, J = 8.0, 4.8 Hz, 1H), 3.79 (s, 2H), 2.80 (m, 2H), 2.27–2.17 (m, 1H), 2.06–1.99 (m, 1H); ¹³C-NMR (100 MHz, CDCl₃) δ: 160.5, 149.7, 148.5, 140.9, 135.8, 135.6, 128.8 (2C), 127.9 (2C), 126.8 (q, J = 4 Hz, 2C), 125.8 (2C), 124.4 (q, J = 269 Hz), 122.8 (q, J = 32 Hz), 115.7 (2C), 78.5, 51.3, 45.5, 38.9; HR-MS (ESI/QTOF) m/z: C₂₂H₂₂F₃N₂O ([M + H]⁺) 387.1679. Found 387.1679.

Cell Culture

Cell culture was performed as described previously.¹¹⁾ J774.1 cells (RIKEN BioResource Center, Ibaraki, Japan) were grown in RPMI-1640 medium (FUJIFILM Wako Pure Chemical Corporation), supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco, MA, U.S.A.), 100 U/mL penicillin and 100 µg/mL streptomycin in a humidified atmosphere of 5% CO₂ in air at 37 °C.

Enzyme-Linked Immunosorbent Assay (ELISA)

The concentration of IL-6 was measured by ELISA, as described previously.¹¹⁾ J774.1 cells were pre-incubated with FLX or its derivatives (0.1–30 µM) for 30 min, then incubated with poly(I : C) (5 µg/mL), LPS (1 µg/mL) or CpG ODN (200 ng/mL) for 6 h in an atmosphere of 5% CO₂ in air at 37 °C.

The culture supernatant was harvested, and IL-6 was measured by ELISA. Ninety-six-well plates were coated with purified anti-mouse IL-6 monoclonal antibody (mAb) (1 : 2000) (eBioscience, San Diego, CA, U.S.A.), incubated at 4 °C overnight, washed with PBS containing 0.05% Tween-20, and blocked with PBS containing 1% bovine serum albumin (BSA). The plates were incubated for 1 h at room temperature, washed again, and incubated overnight at 4 °C with culture supernatant and recombinant mouse IL-6 (BioLegend, San Diego, CA, U.S.A.), in order to obtain a standard curve. The plates were washed again, incubated with anti-mouse biotin-conjugated IL-6 mAb (1 : 2000) (eBioscience) for 1 h at room temperature, washed and incubated with avidin-horseradish peroxidase (FUJIFILM Wako Pure Chemical Corporation) for 10 min at room temperature. Next, the plates were washed again, and 3,3′,5,5′-tetramethylbenzidine (TMB) (FUJIFILM Wako Pure Chemical Corporation) was added. When an appropriate color reaction was observed, 2.5 M H₂SO₄ was added to stop the reaction. The absorbance at 450 nm was measured with a Wallac 1420 ARVO Fluoroscan (Wallac, Turku, Finland).

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Bromide (MTT) Assay

After removal of the culture supernatant to measure cytokine production, MTT reagent suspended in RPMI-1640 supplemented with 10% FBS was added to J774.1 cells in the 96-well plates. When the color reaction due to the reduction of MTT was observed to be sufficient in the control group, the stopping solution was added. The absorbance of each well was read on a WALLAC 1420 ARVO Fluoroscan (570 nm).

Calculation of IC₅₀ and CC₅₀

The IL-6 levels were calculated from the standard curve obtained as described above, and the inhibitory ratio in each sample with respect to the control group was calculated. Cell viability was calculated from the ratio of the absorbance measured in MTT assay of each sample to that of the control group. The values of the IC₅₀ or the 50% cytotoxic concentration (CC₅₀) were determined by linear interpolation from the adjacent two points in dose-response plots of the mean values of data obtained in three or more independent experiments.

Calculation of the Lowest Energy Conformations of FLX, FLX-f, and FLX-g

Conformational analyses of FLX, FLX-f, and FLX-g were performed using the MFF94s conformer search algorithm.^13–15) Conformers within an energy range of 1.36 kcal/mol were optimized using density functional theory (DFT) calculations at the B3LYP/6-31G(d,p) level with the default solvation model (water), implemented in the Gaussian 09 program package.¹⁶⁾ The lowest energy conformer of each compound was determined by comparing the sum of the electronic and zero-point energies of the conformers of each compound. The populations of the conformers of each compound at 27 °C (300 K) were simulated according to the relative energies for a Boltzmann population distribution.

Statistics

Values are given as the mean ± standard error (S.E.). Multiple groups were compared using one-way ANOVA followed by Dunnett’s test (control group vs. FLX- or FLX-a-treated group) as implemented in the GraphPad Prism version 9.0 statistical package (GraphPad Software, San Diego, CA, U.S.A.). The criterion of significance was set at p < 0.05.

RESULTS AND DISCUSSION

First, the cytotoxicity and anti-inflammatory activity of FLX and FLX-a were evaluated in J774.1 cells (Fig. 1). FLX-a has a chlorine substituent in place of the methylamino group (Fig. 1A). Neither FLX nor FLX-a showed cytotoxicity towards J774.1 cells at concentrations below 30 µM (Fig. 1B). FLX inhibited IL-6 production induced by the TLR3 agonist poly(I : C) with an IC₅₀ value of 4.76 µM (Fig. 1C, Table 1), whereas FLX-a lacked inhibitory activity (Fig. 1C, Table 1). These results indicate that the methylamino group of FLX is required for the inhibition of IL-6 production.

Fig. 1. Role of the Amine Structure in the Inhibitory Effect of FLX on IL-6 Production Induced by a TLR3 Agonist

(A) Structural formulae of FLX and its derivative, FLX-a, lacking the amine. (B, C) J774.1 cells were pre-incubated with FLX or FLX-a (1–30 µM) for 0.5 h, and then incubated with poly(I : C) (5 µg/mL) for 6 h. (B) Cell viability was evaluated by MTT assay. (C) IL-6 was measured by means of ELISA. Error bars indicate ± S.E. (B, C: n = 10–11, three independent experiments). Statistical analysis was performed by one-way ANOVA followed by Dunnett’s test. Significant differences between control and test groups are indicated by * (p < 0.05), or *** (p < 0.001). FLX: fluoxetine.

Table 1. Inhibitory Activity of FLX and Its Derivatives Towards IL-6 Production Mediated by TLR3 and Effect on Cell Viability

Compounds	IC50 for IL-6 production (µM)	CC50 (µM)
FLX	4.76	>30
FLX-a	>30	>30
FLX-b	5.15	>30
FLX-c	8.73	9.78
FLX-d	28.9	>30
FLX-e	12.9	>30
FLX-f	>30	>30
FLX-g	27.2	>30
FLX-h	>30	>30
FLX-i	8.87	20.5
FLX-j	5.22	28.8

The IC₅₀ value for inhibition of IL-6 production and the CC₅₀ value were determined from the dose-response curves.

Next, to investigate the influence of the substituent on the amino nitrogen atom of FLX on the anti-inflammatory activity, nine FLX derivatives were synthesized (Fig. 2), and their effects on IL-6 production and cell viability were examined (Figs. 3, 4, Table 1).

Fig. 2. Structural Formulae of FLX and Its Derivatives Examined in This Study

Fig. 3. Inhibition by FLX and Its Derivatives of IL-6 Production Induced by a TLR3 Agonist

J774.1 cells were pre-incubated with FLX or its derivatives (0.1–30 µM) for 0.5 h, and then incubated with poly(I : C) (5 µg/mL) for 6 h. IL-6 was measured by ELISA. The inhibitory ratio of each derivative relative to the control is shown (A: FLX-b, B: FLX-c, C: FLX-d, D: FLX-e, E: FLX-f, F: FLX-g, G: FLX-h, H: FLX-i, I: FLX-j). Error bars indicate ± S.E. (n = 8–14, five independent experiments). FLX: fluoxetine.

Fig. 4. Cytotoxicity of FLX and Its Derivatives

J774.1 cells were pre-incubated with FLX or its derivatives (0.1–30 µM) for 0.5 h, and then incubated with poly(I : C) (5 µg/mL) for 6 h. Cell viability was evaluated by MTT assay. The effect of each derivative on cell viability relative to the control is shown (A: FLX-b, B: FLX-c, C: FLX-d, D: FLX-e, E: FLX-f, F: FLX-g, G: FLX-h, H: FLX-i, I: FLX-j). Error bars indicate ± S.E. (n = 8–14, five independent experiments). FLX: fluoxetine.

FLX-b bearing an n-propyl substituent showed similar anti-inflammatory activity to FLX without marked cytotoxicity (IC₅₀ = 5.15 µM, CC₅₀ >30 µM). FLX-c bearing an n-hexyl substituent showed a slightly lower activity than FLX, and significantly reduced cell viability (IC₅₀ = 8.73 µM, CC₅₀ = 9.78 µM). These results suggest that replacement of the methyl group at the nitrogen atom in FLX by a hydrophobic alkyl group may lead to more potent cytotoxicity.

FLX-d is a secondary amine, which possesses two propyl groups at the nitrogen atom. This compound had lower inhibitory activity than FLX or FLX-b (IC₅₀ = 28.9 µM). FLX-e, which has an isobutyl group showed slightly lower activity than FLX and FLX-b (IC₅₀ = 12.9 µM). These results suggest that bulky N-substituents are unfavorable for the activity.

FLX-f, which has a phenyl group instead of the methyl group in FLX, showed a substantial loss of activity compared to FLX (IC₅₀ > 30 µM). The lone pair of electrons at the nitrogen atom in this compound can be delocalized over the benzene ring, so the basicity and nucleophilicity of FLX-f should be lower than those of FLX. Thus, the basicity and/or nucleophilicity of the amino group in FLX are important for the anti-inflammatory activity.

FLX-g, which has a benzyl group, showed inhibitory activity (IC₅₀ = 27.2 µM), but FLX-h, which has a 4-trifluorobenzyl group, lacked activity (IC₅₀ > 30 µM). The electron-withdrawing 4-trifluorobenzyl group would decrease the basicity and nucleophilicity through its inductive effect, and this may account for the decrease in inhibitory activity. On the other hand, FLX-i, which has a 4-methoxybenzyl group, was more potent than FLX-g (IC₅₀ = 8.87 µM), presumably because the electron-donating 4-methoxybenzyl group increases the basicity and nucleophilicity. Taken together, these results suggest that the nucleophilicity of the nitrogen atom is important for the anti-inflammatory action of FLX. However, FLX-i was quite cytotoxic (CC₅₀ = 20.5 µM). Interestingly, FLX-j, which has a 3-picolyl group, showed similar activity to FLX (IC₅₀ = 5.22 µM), with weak cytotoxicity (CC₅₀ = 28.8 µM). The electron-withdrawing picolyl group may act as a base or hydrogen bond acceptor.

Next, the most stable conformers and their highest occupied molecular orbitals (HOMOs) of unprotonated FLX, FLX-f, and FLX-g were calculated using density functional theory (DFT) calculations at the B3LYP/6-31G(d,p) level (Fig. 5). The HOMOs of FLX, which exhibits potent anti-inflammatory activity, are preferentially localized on the amine moiety. In contrast, the HOMOs of FLX-f, which lacks inhibitory activity, are spread over the aniline moiety. In addition, the HOMOs of FLX-g, which has weak activity, appear at both the amine moiety and the benzene ring. These results support the idea that the basicity and/or nucleophilicity of the amino group in FLX are important for the anti-inflammatory activity.

Fig. 5. Stable Conformers and HOMOs of Unprotonated FLX (A), FLX-f (B), and FLX-g (C) Calculated at the DFT/B3LYP/6-31G(d,p) Level with the Default Solvation Model (Water)

The Boltzmann population of each conformer was simulated at 300 K based on its relative energy.

We also investigated the effects of FLX and its derivatives on IL-6 production induced by CpG ODN, a TLR9 agonist and LPS, a TLR4 agonist. FLX-a had lower inhibitory activity than FLX towards TLR9- and TLR4-mediated IL-6 production (Figs. 6A, B). As regards TLR9-mediated IL-6 production, derivatives in which the methyl group was replaced with a highly hydrophobic alkyl group (FLX-b, FLX-c, FLX-d, and FLX-e) showed similar or inferior inhibitory activity to FLX (Fig. 6C), though FLX-f showed quite potent activity at both concentrations examined. Compared to FLX, FLX-g and FLX-h had lower inhibitory activity, while FLX-i showed similar inhibitory activity, and FLX-j showed higher activity than FLX. In the case of TLR4-mediated IL-6 production, on the other hand, most derivatives other than FLX-a showed inhibitory activity similar to or higher than that of FLX (Fig. 6D). FLX-c was found to have particularly potent inhibitory activity, suggesting that the highly hydrophobic n-hexyl moiety at the amino group may enhance the inhibitory effect on the TLR4 activation-induced inflammatory response. These results suggest that the amino group is important for the inhibitory effect of FLX on TLR4- and TLR9-mediated inflammatory responses, as well as the TLR3-mediated inflammatory response.

Fig. 6. Inhibition by FLX and Its Derivatives of IL-6 Production Induced by a TLR9 Agonist or a TLR4 Agonist

(A, C) J774.1 cells were pre-incubated with FLX or its derivatives (3 or 10 µM) for 0.5 h, and then incubated with CpG ODN (200 ng/mL) for 6 h. (B, D) J774.1 cells were pre-incubated with FLX or its derivatives (10 or 30 µM) for 0.5 h, and then incubated with LPS (1 µg/mL) for 6 h. IL-6 was measured by ELISA. (C, D) The inhibitory ratio of each derivative relative to the control is shown. Error bars indicate ± S.E. (n = 12–15, four or more independent experiments). Statistical analysis was performed by one-way ANOVA followed by Dunnett’s test. Significant differences between control and test groups are indicated by *** (p < 0.001) or **** (p < 0.0001). FLX: fluoxetine.

In our previous study, we compared the IL-6 production-inhibitory activity of five FDA-approved SSRIs and found that the serotonin reuptake-inhibitory ability of SSRIs did not correlate with the activity to inhibit IL-6 production.^11,17) Also, pretreatment with serotonin had no effect on IL-6 production induced by poly(I : C).¹¹⁾ These data suggest that the conventional role of SSRIs is unrelated to the anti-inflammatory effect, although we did not examine the serotonin reuptake-inhibitory activity of the derivatives used in this study.

FLX is metabolized by multiple CYP450 isozymes,¹⁸⁾ and norfluoxetine, the N-demethylated FLX metabolite generated by CYP450s, is the most important contributor to FLX’s pharmacological effects.^17,18) This metabolite and the parent compound have comparable K_i values for inhibition of serotonin uptake.¹⁷⁾ It has also been reported that both compounds inhibit the Kv3.1 channel,¹⁹⁾ T-type calcium channels,²⁰⁾ and CYP2D6,¹⁸⁾ as well as inducing apoptosis in microglia,²¹⁾ indicating that they have similar pharmacological activities. It was reported that the trifluoromethylphenyl ring moiety is not necessary for the inhibition of lipopolysaccharide-induced nitric oxide production by FLX in BV-2 murine microglial cells.²²⁾ On the other hand, the inhibitory activity towards enterovirus 2C protein was abolished in an amine-free fragment of FLX.²³⁾ Thus, the amine structure may be important for the pharmacological activities of FLX.

In conclusion, our results indicate that the amine structure of FLX is critical for the anti-inflammatory activity. In particular, the basicity and/or nucleophilicity of the amino group in FLX appear to be of importance for the anti-inflammatory activity. Although additional optimization is still needed, our findings provide new insight into the structural requirements for the anti-inflammatory activity of FLX.

Acknowledgments

This work was supported in part by a Tokyo University of Science Grant for the president’s research promotion (to MT, SA, KKu). This work was also supported in part by JST SPRING, Grant Number: JPMJSP2151, and JSPS KAKENHI, Grant Number: 23KJ1966 (JSPS Research Fellowship) (to YT).

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Khanmohammadi S, Rezaei N. Role of Toll-like receptors in the pathogenesis of COVID-19. J. Med. Virol., 93, 2735–2739 (2021).
• 2) Rudd KE, Johnson SC, Agesa KM, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet, 395, 200–211 (2020).
• 3) Chousterman BG, Swirski FK, Weber GF. Cytokine storm and sepsis disease pathogenesis. Semin. Immunopathol., 39, 517–528 (2017).
• 4) Rommasi F, Nasiri MJ, Mirsaeidi M. Immunomodulatory agents for COVID-19 treatment: possible mechanism of action and immunopathology features. Mol. Cell. Biochem., 477, 711–726 (2022).
• 5) Gordon AC, Mouncey PR, Al-Beidh F, et al. Interleukin-6 receptor antagonists in critically Ill patients with covid-19. N. Engl. J. Med., 384, 1491–1502 (2021).
• 6) Luo Y, Kataoka Y, Ostinelli EG, Cipriani A, Furukawa TA. National prescription patterns of antidepressants in the treatment of adults with major depression in the U.S. between 1996 and 2015: a population representative survey based analysis. Front. Psychiatry, 11, 35 (2020).
• 7) Lenze EJ, Mattar C, Zorumski CF, Stevens A, Schweiger J, Nicol GE, Miller JP, Yang L, Yingling M, Avidan MS, Reiersen AM. Fluvoxamine vs. placebo and clinical deterioration in outpatients with symptomatic COVID-19: a randomized clinical trial. JAMA, 324, 2292–2300 (2020).
• 8) Németh ZK, Szűcs A, Vitrai J, Juhász D, Németh JP, Holló A. Fluoxetine use is associated with improved survival of patients with COVID-19 pneumonia: a retrospective case-control study. Ideggyogy. Sz., 74, 389–396 (2021).
• 9) Reis G, Dos Santos Moreira-Silva EA, Silva DCM, et al. Effect of early treatment with fluvoxamine on risk of emergency care and hospitalisation among patients with COVID-19: the TOGETHER randomised, platform clinical trial. Lancet Glob. Health, 10, e42–e51 (2022).
• 10) Oskotsky T, Maric I, Tang A, Oskotsky B, Wong RJ, Aghaeepour N, Sirota M, Stevenson DK. Mortality risk among patients with COVID-19 prescribed selective serotonin reuptake inhibitor antidepressants. JAMA Netw. Open, 4, e2133090 (2021).
• 11) Takenaka Y, Tanaka R, Kitabatake K, Kuramochi K, Aoki S, Tsukimoto M. Profiling differential effects of 5 selective serotonin reuptake inhibitors on TLRs-dependent and -independent IL-6 production in immune cells identifies fluoxetine as preferred anti-inflammatory drug candidate. Front. Pharmacol., 13, 874375 (2022).
• 12) Silvestri R, Artico M, La Regina G, Di Pasquali A, De Martino G, D’Auria FD, Nencioni L, Palamara AT. Imidazole analogues of fluoxetine, a novel class of anti-Candida agents. J. Med. Chem., 47, 3924–3926 (2004).
• 13) Gotō H, Ōsawa E. Corner flapping: a simple and fast algorithm for exhaustive generation of ring conformations. J. Am. Chem. Soc., 111, 8950–8951 (1989).
• 14) Gotō H, Ōsawa E. Further developments in the algorithm for generating cyclic conformers. Test with cycloheptadecane. Tetrahedron Lett., 33, 1343–1346 (1992).
• 15) Gotō H, Ōsawa E. An efficient algorithm for searching low-energy conformers of cyclic and acyclic molecules. J. Chem. Soc., Perkin Trans. 2, 2, 187–198 (1993).
• 16) Frisch MJ, Trucks GW, Schlegel HB, et al. Gaussian 09, Revision D.01. Gaussian, Inc., Wallingford, CT. (2009).
• 17) Hiemke C, Härtter S. Pharmacokinetics of selective serotonin reuptake inhibitors. Pharmacol. Ther., 85, 11–28 (2000).
• 18) Deodhar M, Rihani SBA, Darakjian L, Turgeon J, Michaud V. Assessing the mechanism of fluoxetine-mediated CYP2D6 inhibition. Pharmaceutics, 13, 148 (2021).
• 19) Choi BH, Choi JS, Yoon SH, Rhie DJ, Min DS, Jo YH, Kim MS, Hahn SJ. Effects of norfluoxetine, the major metabolite of fluoxetine, on the cloned neuronal potassium channel Kv3.1. Neuropharmacology, 41, 443–453 (2001).
• 20) Traboulsie A, Chemin J, Kupfer E, Nargeot J, Lory P. T-type calcium channels are inhibited by fluoxetine and its metabolite norfluoxetine. Mol. Pharmacol., 69, 1963–1968 (2006).
• 21) Dhami KS, Churchward MA, Baker GB, Todd KG. Fluoxetine and its metabolite norfluoxetine induce microglial apoptosis. J. Neurochem., 148, 761–778 (2019).
• 22) Park J-Y, Kim S-W, Lee J-K, Im WB, Jin BK, Yoon SH. Simplified heterocyclic analogues of fluoxetine inhibit inducible nitric oxide production in lipopolysaccharide-induced BV2 cells. Biol. Pharm. Bull., 34, 538–544 (2011).
• 23) Bauer L, Manganaro R, Zonsics B, Strating JRPM, Kazzi PE, Lopez ML, Ulferts R, Hoey C, Maté MJ, Langer T, Coutard B, Brancale A, Kuppeveld FJM. Fluoxetine inhibits enterovirus replication by targeting the viral 2C protein in a stereospecific manner. ACS Infect. Dis., 5, 1609–1623 (2019).