Synthesis of 9-Cinnamyl-9H-purine Derivatives as Novel TLR4/MyD88/NF-κB Pathway Inhibitors for Anti-inflammatory Effects

Abstract

The novel 9-cinnamyl-9H-purine skeleton, inspired by resveratrol and curcumin, was developed to avoid a pan-assay interference compound (PAINS) related to invalid metabolic pancreas activity (IMPS). It replaced the phenol group with purine analogues, the building blocks of DNA and RNA. Alterations to the hydroxyl group in the cinnamyl group, such as H, Me, or F substitutions, were made to impede its oxidation to a PAINS-associated quinone. Among the compounds tested, 5e significantly inhibited nitric oxide production in LPS-induced macrophages (IC₅₀: 6.4 vs 26.4 μM for resveratrol). 5e also reduced pro-inflammatory cytokine levels (IL-6, TNF-α, IL-1β) and lowered iNOS and COX-2 protein levels. Mechanistically, 5e disrupted the TLR4–MyD88 protein interaction, leading to the suppression of the NF-κB signaling pathway suppression. In an atopic dermatitis mouse model, 5e reduced ear edema and inflammation. These findings indicate that the novel 9-cinnamyl-9H-purine skeleton provides therapeutic insight into treating various human diseases by regulating inflammation.

Inflammation is a form of biological tissue defense against injury and infection, which causes the recruitment of leukocytes and plasma proteins to the affected tissue site. Inflammation-related molecules such as nitric oxide (NO), interleukin-1 (IL-1), interleukin-6 (IL-6), and tumor necrosis factor (TNFα) are produced by various immune cells, including macrophages, T cells, and B cells.^1,2 While acute inflammation is necessary for tissue repair and pathogen clearance, chronic inflammation can exacerbate several maladies, such as cardiovascular disease, cancer, Alzheimer’s disease, asthma, rheumatoid arthritis, and diabetes.³⁻⁷ In this regard, there is an urgent need for the development of anti-inflammatory agents with a novel skeleton that is devoid of side effects. One promising approach is the design and synthesis of novel drug candidates inspired by natural product chemical structures, which are significant drug discovery and development resources.

Resveratrol and curcumin are often extensively examined as natural products for their anti-inflammatory properties attributed to their NF-κB pathway inhibition ability.⁸ Resveratrol is a polyphenol phytochemical found in many plant species such as grapes, peanuts, and berries. It is produced in plants as a defense response to mechanical injury, fungal infection, and UV radiation.⁹ Resveratrol has various pharmacological effects, including anti-inflammatory, antioxidant, antimicrobial, antiaging effects, and cancer prevention and treatment. It alters numerous cell functions, but resveratrol’s biological activity mechanisms are poorly understood. In 2014, reports speculated that these phytochemicals perturb membranes and indiscriminately alter protein function.¹⁰ Many specific resveratrol biological targets were identified a year later, including NQO2 (alone and in interaction with AKT1), GSTP1, estrogen receptor beta, CBR1, and integrin αVβ. However, whether any were responsible for the observed effects in cells and model organisms was unclear.¹¹ Curcumin (2) is a diarylheptanoid, a prominent turmeric ingredient (Curcuma longa L.) that garnered interest over the past half-century because of its molecular target diversity, such as transcription factors, enzymes, protein kinases, growth factors, inflammatory cytokines, and receptors.¹²

However, resveratrol and curcumin are pan-assay interference compounds (PAINS) and an invalid metabolic pancreas (IMPS), which might produce masquerades as drug-like binding and yields false signals across a variety of assays.¹³⁻¹⁶ Due to curcumin’s false activity likelihood in vitro and in vivo, more than 120 clinical trials of curcuminoids have been attempted.¹⁷ However, most clinical trials have been unsuccessful. Curcumin derivative drawbacks in drug development are poor pharmacokinetic/pharmacodynamic (PK/PD) properties, low efficacy in several disease models, and toxic effects under certain testing conditions.¹⁸ Additionally, resveratrol’s limited bioavailability and stability hinder its clinical applications; thus, studies have designed and synthesized various resveratrol analogues with improved pharmacokinetic properties and efficacy to overcome these limitations.

Michael acceptors such as curcumin’s chemical structure display a full and broad bioactivity spectrum. However, their presumed indiscriminate reactivity often leads them to be overlooked in drug discovery. Although a proven approach targeted covalent modification using Michael acceptors, it requires a sophisticated design toward the target protein.¹⁹ Therefore, the Michael acceptor in this study was removed through chemical modifications to avoid its presumed indiscriminate reactivity.

In this work, 9-cinnamyl-9H-purines were designed and synthesized to avoid PAINS-related IMPS activities and to have the modified nucleoside’s properties (Figure 1). This foundation combines the cinnamyl moieties of resveratrol and curcumin with purine, building blocks for DNA and RNA. The nucleobase moiety in nucleosides and nucleotides is involved in several cellular processes, such as DNA and RNA synthesis, cell signaling, enzyme regulation, and metabolism.²⁰ Therefore, these useful purine rings were introduced to avoid PAINS rule complications. Applying this nucleobase moiety to resveratrol and curcumin structures was expected to overcome previously reported low efficacy in several disease models and toxic effects under certain testing conditions.^17,18 Furthermore, the resveratrol and curcumin’s hydroxyl group in the cinnamyl moiety was substituted with H, Me, or F to prevent its conversion to PAINS-related quinones through oxidation.

Figure 1.

Rationale for the design of the 9-cinnamyl-9H-purine derivatives.

In LPS-induced inflammation reaction, the interaction of toll-like receptor 4 (TLR4) and myeloid differentiation primary response 88 (MyD88) regulates nuclear factor-kappa B (NF-κB) signaling pathway.²¹ Activation of the TLR4/MyD88/NF-κB signaling pathway increases inflammatory molecules such as NO, IL-1, IL-6, and TNFα. In addition, cyclooxygenase-2(COX-2) and inducible NO synthase (iNOS) are increased by NF-κB activation.²² Aberrant NF-κB activation is associated with various ailments, including cancer, autoimmune disorders, and chronic inflammatory diseases.^23,24 Therefore, inhibition of the MyD88/NF-κB signaling pathway is considered as an effective way to develop therapeutic agents against inflammation diseases.

In this study, we report the design and synthesis of a novel 9-cinnamyl-9H-purine structure and structure–activity relationship by screening the inhibitory effects of LPS-induced NO production. Next, anti-inflammatory effects were evaluated by measuring the pro-inflammatory cytokines, COX-2, iNOS, and NF-κB activation. The mechanism of action was investigated by studying the TLR4/MyD88/NF-κB signaling pathway.

To synthesize the final 9-cinnamyl-9H-purine derivatives, the carboxylic acids (3b–3c) were converted to ethyl ester with catalytic sulfuric acid in ethanol under reflux condition for 5 h (Scheme 1). The resulting ester intermediates were reduced with diisobutylaluminum hydride in CH₂Cl₂ under −78 °C to afford allylic alcohol (4b–4c).²⁵ For purine introduction, the nucleobase moiety was achieved by treating allylic alcohol 4a–4c with 2,6-dichloropurine and 6-chloropurine at room temperature for 6 h under the Mitsunobu conditions^26,27 to establish the 9-cinnamyl-9H-purine skeletons 5a–5f.

Scheme 1. Synthesis of 9-Cinnamyl-9H-purine Derivatives 5a–5f.

Reagents and conditions: (a) (i) cat. H₂SO₄, EtOH, reflux, 5 h; (ii) DIBAL-H, CH₂Cl₂, −78 °C, 1 h; (b) 6-chloropurine or 2,6-dichloropurine, PPh₃, DIAD, THF, 0 °C to rt, 6 h.

Next, 6-amino analogues were introduced; the 6-chloropurine analogues (5a–5f) were treated with amine derivatives (NH₃, NH₂Me, and NH₂OH) to yield the final 6a–6f, 7a–7f, and 8d–8f, respectively (Scheme 2).

Scheme 2. Synthesis Amine Derivatives 6–8.

Reagents and conditions: (a) NH₃, 2-propanol, 100 °C, 24 h; (b) NH₂-Me·HCl, Et₃N, EtOH, rt, 12 h; (c) NH₂OH·HCl, Et₃N, EtOH, rt, 18 h.

Next, we focused on the synthesis of cyclopropylated derivatives (10–12) using Simmons–Smith conditions.^28,29 Because, the inclusion of a cyclopropyl ring in a compound offers various advantages in the field of drug discovery, these benefits encompass enhancing potency, mitigating off-target effects, improving metabolic stability, promoting more favorable receptor binding through entropic considerations, and modifying the drug’s pK_a to decrease its P-glycoprotein efflux ratio.³⁰ Simmons–Smith cyclopropanation of 4c with diethyl zinc and iodomethane produced 9 in 88%. The introduction of 2,6-dichloropurine, followed by aminolysis with methylamine hydrochloride and ammonia, yielded the final products 11 (82%) and 12 (69%), respectively (Scheme 3).

Scheme 3. Synthesis Cyclopropylated Derivatives 10–12.

Reagents and conditions: (a) Et₂Zn, CH₂I₂, CH₂Cl₂, 0 °C to rt, 12.5 h; (b) 2,6-dichloropurine, PPh₃, DIAD, THF, 0 °C to rt, 6 h; (c) NH₂-Me·HCl, Et₃N, EtOH, rt, 12 h; (d) NH₃, 2-propanol, 100 °C, 24 h.

Scheme 4 depicts the synthesis of fluorinated purine analogues. Based on the most potent compound 5e, 2-fluorinated purine analogue 13 and 2,6-difluoropurine analogue 14 were designed and synthesized. Condensation of 4b with commercial 2-fluoro-6-chloropurine under Mitsunobu conditions yielded 13. Treatment of 13 with cesium fluoride, tetramethylammonium chloride, 18-crown-6 in MeCN yielded 2-fluoro-6-fluoropurine analogue 14.³¹

Scheme 4. Synthesis Fluorinated Derivatives 13–14a.

Reagents and conditions: (a) 2-fluoro-6-chloropurine, PPh₃, DIAD, THF, 0 °C to rt, 6 h; (b) CsF, (CH₃)₄NCl, 18-crown-6, MeCN, 0 to 60 °C, 12 h.

Macrophage cells are essential for initial inflammatory responses and LPS induces inflammatory mediator productions, such as NO, TNFα, IL-1β, and IL-6.³² For initial anti-inflammatory screening, 26 9-cinnamyl-9H-purine analogues were evaluated on NO production in LPS-stimulated RAW264.7 using the Griss reagent. LPS (500 ng/mL) significantly induced NO production, which was inhibited by the 9-cinnamyl-9H-purine analogues (Table 1). In comparison to resveratrol, 5e showed the most effective NO inhibition (Table 1). IC₅₀ of resveratrol and 5e were 26.4 and 6.4 μM, respectively (Figure 2A). However, 6a, 6b, 6c, 6d, 7b, and 12 did not show any inhibitory effect on NO production, whereas promoting its production in LPS-stimulated RAW264.7 cells. Using the MTS assay, we tested the cytotoxicity of resveratrol and 5e on RAW264.7 and 293T cells. Resveratrol and 5e did not show any cytotoxicity at concentrations of 30, 10, and 3 μM (data not shown).

Table 1. Inhibitory Effect of 9-Cinnamyl-9H-purine Analogues 5a–14 against NO Production.

	nitric oxide (NO) inhibition (%)
compd	20 μM	10 μM
Resveratrol	38.64 ± 0.82	26.56 ± 1.63
Bay11–7082 (Bay)	58.12 ± 0.23	44.81 ± 0.09
5a (R = H, X = H, Y = Cl)	14.78 ± 1.05	15.13 ± 0.62
5b (R = Me, X = H, Y = Cl)	12.07 ± 0.05	13.52 ± 1.78
5c (R = F, X = H, Y = Cl)	30.19 ± 0.90	11.59 ± 0.54
5d (R = H, X = Cl, Y = Cl)	34.73 ± 1.99	22.18 ± 0.47
5e (R = Me, X = Cl, Y = Cl)	60.39 ± 0.31	31.93 ± 0.88
5f (R = F, X = Cl, Y = Cl)	40.47 ± 0.69	29.30 ± 0.30
6a (R = H, X = H, Y = NH2)	–13.28 ± 0.29	–3.38 ± 0.62
6b (R = Me, X = H, Y = NH2)	–7.73 ± 0.25	0.48 ± 1.87
6c (R = F, X = H, Y = NH2)	–19.32 ± 0.41	–17.15 ± 0.28
6d (R = H, X = Cl, Y = NH2)	2.46 ± 0.47	–5.27 ± 1.26
6e (R = Me, X = Cl, Y = NH2)	11.23 ± 0.36	11.93 ± 0.83
6f (R = F, X = Cl, Y = NH2)	24.06 ± 0.68	13.88 ± 0.22
7a (R = H, X = H, Y = NHCH3)	33.41 ± 0.59	19.31 ± 0.63
7b (R = Me, X = H, Y = NHCH3)	–15.70 ± 0.71	–7.00 ± 0.75
7c (R = F, X = H, Y = NHCH3)	2.90 ± 1.35	0.97 ± 0.09
7d (R = H, X = Cl, Y = NHCH3)	34.73 ± 0.29	26.82 ± 0.81
7e (R = Me, X = Cl, Y = NHCH3)	34.62 ± 0.16	28.41 ± 0.14
7f (R = F, X = Cl, Y = NHCH3)	36.45 ± 0.29	28.16 ± 1.03
8d (R = H, X = Cl, Y = NHOH)	51.23 ± 3.04	57.36 ± 1.50
8e (R = Me, X = Cl, Y = NHOH)	40.91 ± 0.18	36.84 ± 0.62
8f (R = F, X = Cl, Y = NHOH)	17.63 ± 0.83	4.59 ± 0.66
10 (X = Cl, Y = Cl)	17.15 ± 0.24	4.11 ± 0.39
11 (X = Cl, Y = NHCH3)	11.93 ± 0.48	11.23 ± 1.31
12 (X = Cl, Y = NH2)	1.56 ± 0.38	–6.22 ± 0.78
13 (R = Me, X = F, Y = Cl)	60.48 ± 0.19	31.05 ± 0.75
14 (R = Me, X = F, Y = F)	12.13 ± 0.16	5.88 ± 0.74

^aLPS-stimulated NO production in RAW264.7 cells were evaluated using Griess Reagent [1% sulfanilamide in 5% H₃PO₄, 0.1% N-(1-naphthyl)-ethylenediamine dihydrochloride]. Inhibition% was expressed as means ± SEM from triplicates. Negative values in inhibition% indicate the activation effect on LPS-induced NO production.

Figure 2.

Effect of 5e on pro-inflammatory cytokine production. (A) NO production in LPS-stimulated macrophage cells. RAW264.7 cells were treated with resveratrol and 5e at different concentrations for 2 h and stimulated with LPS (500 ng/mL) for 24 h. The Griess reagent was used to measure the NO amount secreted in the cell culture supernatants. Experimental data were presented as means ± SEM (+++P < 0.001 compared with control group and *P < 0.05, **P < 0.01, ***P < 0.001 compared with LPS group). (B–D) Pro-inflammatory cytokine production in LPS-stimulated macrophage cells. RAW264.7 cells seeded into 48-well plates were treated with resveratrol, 5e, or positive drug Bay11-7082 (Bay) for 2 h. Next, the cells were incubated with LPS (500 ng/mL) for 24 h. (B) IL-6, (C) IL-1β, and (D) TNF-α in the cell supernatants were measured by ELISA. Experimental data were presented as means ± SEM (+++P < 0.001, ++P < 0.01 compared with control group, *P < 0.05, **P < 0.01, and ***P < 0.001 compared with LPS group).

To investigate the anti-inflammatory effects of 5e, we measured pro-inflammatory cytokine quantities in the supernatants of LPS-stimulated cells using each ELISA kits, and the results were compared with those of a positive drug Bay11-7082 (Bay) and natural resveratrol. The results indicated that 5e exhibited the most significant IL-6 inhibition (Figure 2B). In addition, 5e showed more potent inhibition than resveratrol on IL-1β and TNF-α cytokine production (Figure 2C,D).

The effects of resveratrol and 5e were studied on the expression of iNOS and COX-2 enzymes using Western blot analysis. RAW264.7 cells were pretreated with resveratrol or 5e for 2 h and then treated with LPS for 20 h. The expression levels of iNOS and COX-2 proteins were significantly decreased with the treatment of 5e at concentrations of 10 and 30 μM (Figure 3).

Figure 3.

Effect of resveratrol and 5e on iNOS and COX-2 expression in LPS-stimulated macrophage cells. RAW264.7 cells were pretreated with resveratrol and 5e for 2 h and then incubated with LPS (200 ng/mL) for 20 h. The expression of iNOS and COX-2 proteins were confirmed by Western blot analysis and β-actin was used as a loaded protein control. The protein bands were quantified using ImageJ software.

NF-κB is an essential inflammation regulator and is translocated from the cytosol to the nucleus in the initial process of inflammation and then regulates the expression of various inflammation-related factors. First, the inhibitory effects of resveratrol and 5e effects were tested on NF-κB transcription activity using NF-κB-SEAP reporter plasmids. Figure 4A shows that 5e inhibited PMA-stimulated NF-κB transcription. Next, we identified NF-κB translocation in LPS-stimulated RAW264.7 cells using immunofluorescence staining. LPS caused NF-κB to translocate from the cytosol to the nucleus, which was inhibited by the treatment of resveratrol and 5e (Figure 4B). These results suggest that 5e inhibits transcription and translocation of the NF-κB protein.

Figure 4.

Effect of 5e on NF-κB transcription and translocation. (A) Effect of resveratrol and 5e on NF-κB transcriptional activity. The 293T cells were seeded into a 96-well plate coated with poly-d-lysine hydrobromide and transfected with pNF-κB-SEAP DNA using HilyMax overnight. Next, the cells were treated with resveratrol and 5e for 2 h and stimulated with PMA (50 ng/mL) for 24 h. Bay 11-7082 (Bay, 20 μM) was used as an NF-κB inhibitor. SEAP activity was determined using the QUANTI-Blue assay system. (B) Effect of resveratrol and 5e on NF-κB translocation to the nucleus. RAW264.7 cells were seeded into an 8-well glass chamber plate overnight. The cells were treated with resveratrol (30 μM), 5e (30 μM), or Bay (20 μM) for 2 h and stimulated with LPS for another 2 h. The cells were immunostained with a primary NF-κB p65 antibody and secondary Alexa Fluor 488-conjugated antibodies. After using an antifade agent containing DAPI, fluorescence images were acquired from the cells. Experimental data were presented as means ± SEM (+++P < 0.001 compared with control group, **P < 0.01 and ***P < 0.001 compared with PMA group).

Co-immunoprecipitation analysis was conducted to demonstrate the interaction of the TLR4-MyD88 proteins. LPS treatment caused the binding interaction of TLR4-MyD88, which was blocked by 5e treatment (Figure 5). The results showed that 5e significantly inhibited LPS-induced TLR4 and MyD88 complex formation (Figure 5).

Figure 5.

Effect of 5e on the LPS-induced Myd88/TLR4 interaction. RAW264.7 cells were treated with 5e at 30 μM for 2 h and then stimulated with LPS (500 ng/mL) for 24 h. Collected protein lysates were coimmunoprecipitated with MyD88 or TLR4 specific antibodies using protein A/G plus-agarose beads. The immunocomplexes were identified by Western blotting with anti- TLR4 or MyD88 or antibodies (IB, immno-blotting; IP, immune-precipitation).

To evaluate the in vivo pharmacologic activity of 5e, DNCB-induced atopic dermatitis mice were used. DNCB caused ear swelling and atopic dermatitis (AD)-like skin lesions in mice that peaked on day 7 after sensitization. Treatment with 0.5% 5e suppressed DNCB-induced ear swelling and AD-like skin lesions (Figure 6A,B). Mouse ear specimens were assessed through H&E staining and microscopic analysis. 5e significantly reduced DNCB-induced epidermal and dermal thickness and infiltration of inflammatory cells compared with DNCB-induced AD-like skin lesions (Figure 6C,D).

Figure 6.

Effect of 5e in an atopic dermatitis mouse model. (A) Representative photographs illustrating the effect of 5e on DNCB-induced atopic dermatitis (AD) in mouse ears. (B) Ear thickness was measured daily with a digimatic micrometer. Data are presented as mean ± SEM (+++P < 0.001, +P < 0.05 compared with control group, *P < 0.05, **P < 0.01, and ***P < 0.001 compared with DNCB group). (C) Histological analysis of ear section. Representative photomicrographs showed of ear sections stained with H&E. (D) Histopathologic findings were assessed semiquantitatively on a 0–2 scale (0 = no change compared with negative control group; 2 = maximum change).

Figure 7 illustrates the summarized structure–activity relationships (SAR) of synthesized 9-cinnamyl-9H-purine analogues regarding nitric oxide inhibition. We concluded that the hydrophobic and electron-withdrawing chlorine atom at position 6 increased biological activities. And fluorine and chlorine atoms at position 2 improved potency. In addition, when comparing 5d–5f, the compound with a methyl group at the para position exhibited the most potent biological activity. However, this feature was not manifested upon comparison of 5a–5c, 6d–6f, and 8a–8c. Further structure–activity relationship studies of the para position are expected to be necessary.

Figure 7.

Summarized structure–activity relationships of 9-cinnamyl-9H-purine analogues concerning the inhibition of nitric oxide production.

This study established anti-inflammatory potential of a novel 9-cinnamyl-9H-purine analogue, an inhibitor of the TLR4/MyD88/NF-κB signaling pathway. The core structure integrates elements of resveratrol and curcumin with purine, a foundational component of DNA and RNA, thereby embodying the beneficial properties of modified nucleosides while circumventing potential issues with PAINS-related IMPS. Alterations to the hydroxyl group in the cinnamyl group, such as H, Me, or F substitutions, were made to impede its oxidation to a PAINS-associated quinone. In vitro and in vivo studies on atopic dermatitis models confirmed that these newly designed molecules were free of IMPS activities. Notably, among the 9-cinnamyl-9H-purine analogues, 5e displayed the most pronounced anti-inflammatory activity, effectively inhibiting the TLR4-MyD88 protein interaction. Additionally, 5e attenuated DNCB-induced ear swelling and symptoms resembling atopic dermatitis in vivo. These results provide therapeutic insight into treating various human diseases by mediating inflammatory processes. Further insights into the mechanism of action along with comprehensive pharmacokinetic and pharmacodynamics studies will be reported elsewhere.

Glossary

Abbreviations Used

NO

nitric oxide

LPS

lipopolysaccharide

COX-2

cyclooxygenase 2

iNOS

inducible nitric oxide synthase

Bay

Bay11-7082

Cel

celecoxib

IC₅₀

half-maximal inhibitory concentrations

IL-6

interleukin 6

TNF-α

tumor necrosis factor

TLR4

interleukin 1β IL-1β) toll-like receptor 4

MyD88

myeloid differentiation primary response 88

AD

atopic dermatitis

PVDF

polyvinylidene fluoride membrane

SDS-PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

NF-κB

nuclear factor-κB

PMA

phorbol 12-myristate 13-acetate

SEM

standard error of the mean

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00437.

• Experimental procedures and product characterization, biology, and copies of ¹H and ¹³C NMR spectra (PDF)

Author Contributions

^# Linh Pham and Rui Jiang contributed equally to this work.

The authors declare no competing financial interest.