Synthesis and evaluation of resveratrol derivatives as fetal hemoglobin inducers

Abstract

Resveratrol (RVT) derivatives (10a-i) were designed, synthesized, and evaluated for their potential as gamma-globin inducers in treating Sickle Cell Disease (SCD) symptoms. All compounds were able to release NO at different levels ranging from 0–26.3%, while RVT did not demonstrate this effect. In vivo, the antinociceptive effect was characterized using an acetic acid-induced abdominal contortion model. All compounds exhibited different levels of protection, ranging from 5.9–37.3%; the compound 10a was the most potent among the series. At concentrations between 3.13–12.5 μM, the derivative 10a resulted in a reduction of 41.1–64.3% in the TNF-α levels in the supernatants of macrophages that were previously LPS-stimulated. This inhibitory effect was higher than that of RVT used as the control. In addition, the compound 10a and RVT induced double the production of the gamma-globin chains (γG+γA), compared to the vehicle, using CD34+ cells. Compound 10a also did not induce membrane perturbation and it was not mutagenic in the in vivo assay. Thus, compound 10a emerged as a new prototype of the gamma-globin-inducer group with additional analgesic and anti-inflammatory activities and proving to be a useful alternative to treat SCD symptoms.

INTRODUCTION

Sickle cell disease (SCD) is the most common inherited monogenic hemoglobinopathy worldwide. It is characterized by a single mutation (GAG to GTG), leading to the substitution of glutamic acid (Glu6) to valine (Val6) at the 6^th position in the β-globin chain. Some studies suggest that the disease afflicts around 300,000 newborns and the numbers are estimated to reach up to 400,000 by 2050; the majority of them being in sub-Saharan Africa [1,2].

The pathophysiology is complex and multifactorial. First, the hydrophobic interactions between the mutated β-chains induce a polymerization process. This damages the erythrocyte cytoskeleton that deforms itself into a sickle-shaped cell under low oxygen tension. After hemolysis caused by the fibrils, these cells release the heme group contained in the hemoglobin and sequester the nitric oxide (NO) in the vascular endothelium; this creates a ‘vasoconstriction profile’ in all the patients.

In addition, chronic inflammation promotes an increase in the pro-inflammatory cytokine levels and activates the pro adhesive states of the erythrocytes, leukocytes, and platelets and thus favors the vaso-occlusion process. This process is responsible for the main acute clinical complications of the disease such as stroke, sickle hepatopathy, gallstones, renal papillary necrosis, priapism, splenic sequestration/infarction, acute chest syndrome, ischemia, pain, septicemia among others, reducing the life expectancy of all the patients [3]. NO plays a key role in the pathophysiology of the SCD; once this pleiotropic mediator contributes as a vasodilator and induces fetal hemoglobin (HbF) production and reduces both the cellular adhesion to vascular endothelium and platelet aggregation [4]. We have previously demonstrated that NO donors such as furoxan (1,2,5-oxadiazole-2-N-oxide) derivatives, present a protective cardiovascular effect by inhibiting platelet aggregation, exhibiting analgesic/anti-inflammatory effects, and increasing the gamma-globin gene expression and HbF production [5,6].

Nowadays, the cruelest facet of the disease is the availability of a few therapeutic options for treatment. The only drugs that are currently available include glutamine, voxelotor, crizanlizumab, and hydroxyurea (HU). However, despite the beneficial effects of HU in the treatment of SCD, severe complications, such as myelosuppression, are associated with its long-term use. Moreover, it is estimated that around 30% of the SCD patients are non-responsive to HU treatment [7]. Among the validated strategies used to ameliorate the SCD symptoms, the increase of HbF production seems to be the most promising one. An inverse correlation between high HbF levels and low symptoms has been described [8,9].

Structural diversity is found among the HbF inducers, such as HU, metformin, DNA methyltransferase inhibitors (i.e., azacytidine, decitabine), histone deacetylase inhibitors (i.e., butyrate, panobinostat), thalidomide derivatives (i.e., pomalidomide, lenalidomide), and natural products among others [8]. Resveratrol (RVT) (trans-3,5,4’-trihydroxystilbene) is a natural phytoalexin known for its protective cardiovascular effect that possesses similar properties like HU in inducing erythroid differentiation. RVT induces differentiation of K562 cells and increases the HbF levels in the erythroid precursor cells isolated from sickle cell patients [10]. Moreover, RVT induces not only the γ-globin mRNA but also the α-globin and β-globin mRNAs in the human erythroid precursors [11]. Antiplatelet, anti-inflammatory, and analgesic activities are also useful properties described for RVT to treat SCD symptoms, which makes this stilbene a promising prototype for molecular optimization.

Therefore, in a continuing effort to develop new candidate drugs to treat SCD, we describe the synthesis and pharmacological evaluation of RVT hybrids with NO-donor properties. The combination of RVT properties associated with NO release by furoxan derivatives in all hybrid compounds could improve different aspects of the SCD, including the reduction of pro-inflammatory cytokines, pain, cellular adhesion, and promotion of HbF induction, thus representing a new alternative to the current therapy through these multiple approaches.

2. RESULTS

2.1. Chemistry

The synthetic routes for the synthesis of hybrid compounds (10a-i) are outlined in Scheme 1 and 2. All furoxan derivatives (4–6) were synthesized according to the methods previously described [12,13]. The treatment of furoxan derivatives (4–6) [5,6,12,13] with 4-hydroxystyrene in a dichloromethane medium, using 1,8-diazabicyclo [5.4.0]undec7-ene (DBU) as the base, provided the compounds (7–9) at yields ranging from 29–48% [12]. The coupling reaction between 4-methoxybenzenediazonium tetrafluoroborate and derivatives (7–9) in the presence of catalyst tris(dibenzylideneacetone)dipalladium(0) [Pd₂ (dba)₃] through Heck-Matsuda reaction [14] provided the final compounds (10a-c) at yields ranging from 51–70% (Scheme 1). It has been described that the Heck-Matsuda reaction exhibits several advantages because it is greener, faster, phosphine-free, and diastereoselective [15]. In addition, the reaction can be carried out at room temperature, an advantage for furoxan derivatives, which, in general, are unstable at high temperatures.

Scheme 1. Synthesis of compounds (10a-g).

a) NaNO₂, acetic acid, chloridric acid, dichloromethane, r.t., overnight; b) NaNO₂, sulfuric acid 60%, 1,2-dichloroethane, 50°C, 30 min; c) cl. water, NaOH, monochloroacetic acid, 110°C, 2h; c2. acetic acid, H₂O₂ 30%, r.t., 48h; c3. acetic acid, nitric acid, 110°C, 45 min; d) 4-hydroxystyrene (11), DBU, dichloromethane, r.t., overnight, e) 4-methoxybenzenediazonium tetrafluoroborate, Pd₂(dba)₃, sodium acetate, benzonitrile, r.t, 1h; f) CH₃P(C₆H₅)₃I, potassium tert-butoxide, tetrahydrofuran, r.t., 24h; g) l-bromo-3,5-dimethoxybenzene, palladium (II) acetate, triethanolamine, 110°C, 24hs; h) furoxan derivative (4–6), DBU, 1,2-dichloroethane, 55°C, 2–8h; i) 1,4-dioxane, tetrakis(triphenylphosphine)palladium(0), K₂CO₃, 90°C, 24h.

Scheme 2. Synthesis of compounds (10h-i).

a) trimethylphosphite, reflux, 48h; b) 4-iodophenol, DBU, dichloromethane, 45°C, 24h; c) 4- hidroxystyrene, palladium (II) acetate, triethanolamine, 65°C, 24h.

The conversion of 4-hydroxybenzaldehydes (11) to 4-hydroxystyrene (12a) through the Wittig reaction was performed using the ylide methylenetriphenylphosphorane (Ph₃PCH₂) in dry tetrahydrofuran medium at 47% of yields. The ylide was prepared using methyltriphenylphosphonium iodide and potassium tert-butoxide as the base, according procedures previously described [14]. The treatment of 4-hydroxystyrene (12a) with 1-bromo-3,5-dimethoxybenzene in the presence of triethanolamine as the solvent and palladium (II) acetate as the catalyst at 110 °C for 24 h provided the compound (13) [16]. The stereochemistry of this reaction produces isomer E with a yield of 41%. The ¹H NMR spectra for compound (13) revealed the signals as doublets at 6.94 (J = 16.5 Hz) and 7.16 (J = 16.5 Hz) ppm attributed to the double bond. The stilbene derivative (13) was treated with furoxan derivatives (4–6) in the presence of DBU and in a 1,2-dichloroethane medium to provide the derivatives (10d-f) at yields ranging from 30–60%.

The styrene derivative (12b) was treated with 3,5-dimethylisoxazole–4-boronic acid pinacol ester (14) in a medium containing 1,4-dioxane, potassium carbonate, and tetrakis (triphenylphosphine)palladium (0) as the catalyst to provide compound 15 at yields of 52%. Then, this styrene derivative (15) was treated with 4-methoxybenzenediazonium tetrafluoroborate, sodium acetate, benzonitrile, and tris(dibenzylideneacetone)dipalladium(0) as the catalyst to obtain the final compound 10 g at yields of 71% (Scheme 1).

The N-oxide subunit present in the compound 10a was reduced using trimethylphosphite, which led to the formation of the compound 10 h at yields of 75% (Scheme 2). Furoxan derivative (4) was treated with 4-iodophenol in a dry dichloromethane medium using DBU as the base to provide compound (16) at yields of 55%. In the second step, a Heck reaction allowed the coupling between 4-hydrostyrene (12a) and compound (16) to obtain the final compound 10i at yields of 53% (Scheme 2).

The chemical structures of all the final compounds (10a-i) were established by infrared (IR) spectroscopy, elemental analysis, and ¹H and ¹³C Nuclear Magnetic Resonance (NMR) spectroscopy. The NMR analysis of stereochemistry revealed that all the compounds (10a-i) were isomer E. The analysis of ¹H NMR spectra of all compounds (10a-i) showed signs attributed to the double bond as doublets with values ranging from 6.94 to 7.94 ppm and coupling constant of 16.5 Hz. All the final compounds were also analyzed by high-performance liquid chromatography (HPLC) and their chromatographic purity was confirmed to be higher than 98.5%.

2.2. Detection of nitrite through Griess reaction

Nitrite quantification is an indirect method to evaluate the nitric oxide (NO) released from the compounds (10a-i); it is based on the oxidation of NO in an aqueous medium. The amount of nitrite is measured through the Griess reaction after incubation of all compounds (10a-i) at 10⁻⁴ M for 1 h using an excess of L-cysteine (1:50) [5,6]. Table 1 summarizes all the results expressed as a percentage of nitrite (NO₂⁻; mol/mol).

Table 1.

NO-released data for compounds (10a-i).

Compounds	% NO2− (mol/mol) 5 mM of L-Cys
RVT	0a
DNS	10.7 ±1.1*b
10a	2.0 ±0.4*
10b	8.8 ±0.2*
10c	25.9 ±0.6*†
10d	1.8 ±0.3*
10e	9.3 ±0.5*
10f	26.3 ±1.3*†
10 g	0
10 h	0
10i	1.7 ±0.3*

^aRVT: resveratrol

^bDNS: isosorbide dinitrate (DNS possesses two ONO₂ groups that may release NO). The data are expressed as the means ±standard errors of the means. Significant differences between the experimental and control groups were evaluated by analysis of variance followed by Tukey’s Multiple Comparison Test.

^*p < 0.05 vs. RVT.

^†p < 0.05 vs. DNS.

After one hour of incubation, the compounds (10a-i) induced nitrite formations at levels ranging from 0–26.3%; the values were found to be 0 and 10.7% for the controls resveratrol (RVT) and isosorbide dinitrate (DNS), respectively. In the absence of L-cysteine, nitrite was not detected in the medium (not shown). Among the furoxan derivatives, those containing a phenylsulfonyl moiety attached to carbon at position 3 (10c and 10f) were more prone to release NO. Regarding this pattern of substitution, a direct relationship was observed between the electron-withdrawing substituents at the C3-position in furoxan and the nitrite levels, exhibiting the following order: phenylsulfonyl (10c and 10f) > phenyl (10b and 10e) > methyl (10a; 10d and 10i). Furazan derivative (10 h) and 3,5-dimethylisoxazole (10 g) were not able to release NO in this experiment.

2.3. Antinociceptive activity

A mice model displaying acetic acid-induced abdominal contortions was used to characterize the antinociceptive effect of compounds (10a-i) [17]. Table 2 exhibits the percentage of protection offered by compounds (10a-i) against chemical-induced abdominal constrictions after previous administration at 100 μmol/kg, per os. Dipyrone (Dip) and resveratrol (RVT) had a comparable analgesic effect of protecting the mice at levels of 34.7 and 39%, respectively (Table 2). It was found that among compounds (10a-i), the most potent was compound 10a, which protected the mice up to 37.3%. This value is comparable to those found for reference drugs Dip and RVT.

Table 2.

In vivo antinociceptive effects of compounds (10a-i), resveratrol (RVT) and dipyrone (Dip) using acetic acid-induced mice abdominal contortions. The results are expressed as the percentage of inhibition of the total writhing induced by acetic acid (n=6).

Compound	% protection (100 μmol/Kg, p.o.)
Dip	37.4 ±5.6*
RVT	39.0 ±5.2*
10a	37.3 ±4.5*
10b	7.7 ±1.7
10c	28.9 ±3.3*
10d	18.2 ±2.3*
10e	5.9 ±1.9
10f	20.5 ±3.1*
10 g	9.7 ±1.4*
10 h	15.8 ±3.2*
10i	16.8 ±2.9*

^*p < 0.01 compared to the negative control (saline) (ANOVA followed by Dunnett’s test).

Compounds (10b-i) have also shown analgesic activity, protecting the mice at levels ranging from 5.9–28.9%. Compounds 10c and 10f, which were able to release high levels of NO, also demonstrated the analgesic effect at values of 28.9 and 20.5%, respectively. Compound 10i exhibited weaker analgesic effects (16.8%) compared with its analog 10a (Table 2).

2.4. Quantification of proinflammatory cytokines in the murine macrophage supernatants

For the most antinociceptive compound (10a) and RVT, the levels of the pro-inflammatory cytokines TNF-α and IL-1β were quantified in the supernatants of murine macrophage cells culture that were previously stimulated with LPS (50 mg/mL). Initially, a dose-response curve was constructed to determinate the concentrations in which the viability of the cells in the presence of compound 10a and RVT was more than 75%. Compound 10a was added to the culture cells at viable concentrations of 12.5, 6.25, and 3.13 μM, while RVT was added at the concentrations of 50, 25, and 12.5 μM. Table 3 summarizes information on the inhibition of TNF-α and IL-1β production conferred by either compound. Compared to the negative control, the inhibition of TNF-α and IL-1β induced by resveratrol was not found to be significant at the concentrations tested. Thalidomide, used as the reference drug, inhibited 32.7% of TNF-α, but was unable to decrease the IL-1β at a concentration of 50 μM.

Table 3.

Percentages of inhibition of TNF-α and IL-1β determined by ELISA in the supernatant of macrophage mononuclear culture treated with LPS and co-incubated with RVT and compound 10a.

Concentrations (μM)	TNF-α inhibition (%)			IL-1β inhibition (%)
	RVT	10a	Thal	RVT	10a	Thal
3.13	-	41.1 ±8.3*	-	-	10.9 ±8.7	-
6.25	-	60.7 ±7.1*†	-	-	18.5 ±9.4	-
12.5	9.7 ±7.3	64.3 ±5.5*†	-	2.0 ±1.9	22.2 ±10.3	-
25	10.8 ±8.1	-	-	3.9 ±3.1	-	-
50	13.5 ±6.9	-	32.3 ±5.7*	12.0 ±5.6	-	0

Thalidomide (Thal) (50 μM) was used as the reference drug.

^*p < 0.05 (ANOVA followed by Tukey’s Multiple Comparison Test).

^*p < 0.01 vs. LPS (not shown).

^†p < 0.01 vs. Thal.

The percentage of LPS-induced TNF-α production that could be inhibited by compound 10a ranged from 41.1–64.3% and it was remarkable compared to the negative control. At concentrations of 6.25 and 12.5 μM, the inhibitory effect was superior to that of the reference drug thalidomide tested at 50 μM. However, compound 10a, was unable to inhibit the LPS-induced IL-1β production significantly when compared to the negative control (Table 3).

2.5. The induction of gamma-globin using CD34⁺ cells

The quantification of gamma-globin levels induced by compound 10a and RVT was performed using the CD34⁺ cells. Both compound 10a and RVT were evaluated at concentrations of 25 μM. This concentration was selected based on a previous work demonstrating the induction of gamma-globin at this particular concentration of RVT [11]. Figure 2 shows that both RVT and compound 10a double the levels of gamma-globin chains (γ^G+γ^A) compared to the vehicle. The differences between RVT and compound 10a were not significant. In addition, the HPLC quantification revealed that neither RVT nor compound 10a altered the levels of heme, delta, beta, and alfa chains (not shown).

Figure 2.

Effect of RVT and 10a on the production gamma globin chains, as evaluated by HPLC. The erythroid cells CD34⁺, differentiated in vitro, were treated with compounds resveratrol (RVT) and 10a at 25 μM for four days. Afterward, the cells were lysed, and the respective lysates were quantified through High Performance Liquid Chromatography method. The control sample received DMSO as a vehicle. The results are presented as mean ±standard error of the mean. * p <0.05 compared to control, n=3.

2.6. Membrane perturbation test

The compounds’ bilayer-modifying potency was determined using a gramicidin (gA)-based assay [18]. The assay relies on the movement of the heavy-ion quencher thallium (Tl⁺) into fluorophore-loaded large unilamellar vesicles (LUVs) doped with gramicidin, which forms bilayer-spanning channels by transbilayer dimerization. The conducting, dimeric channels have a hydrophobic length that is less than the bilayer’s hydrophobic thickness, meaning that channel formation leads to a local bilayer deformation, which has an energetic cost $\left(\Delta G_{\text{bilayer}}^{2\text{M}\to \text{D}}\right)$ that contributes the total free energy of dimerization $\left(\Delta G_{\text{total}}^{2M\to D}\right)$. The $\Delta G_{\text{bilayer}}^{2\text{M}\to \text{D}}$ contribution to $\Delta G_{\text{total}}^{2\text{M}\to \text{D}}$ varies with changes in bilayer physical properties (thickness, lipid intrinsic curvature, and elasticity), which change when amphiphiles partition into the bilayer/aqueous interface. That is, when biologically active molecules partition into lipid bilayers they will alter lipid bilayer properties, which may alter the contribution to the free energy of membrane protein conformational equilibria, as well as the gramicidin monomer⟷dimer equilibrium. The promiscuity of the bilayer-mediated changes in membrane proteins’ conformational equilibria are likely to cause indiscriminate changes in cell function, meaning that it becomes possible to test for undesired bilayer effects using gramicidin channels.

Figure 3 shows the effects on RVT and compounds 10a-j on gramicidin channel activity, quantified by the changes in fluorescence quench rate (Tl⁺ influx into fluorophore-loaded LUVs). Except for RVT, none of the compounds caused LUV instability at concentrations up to 100 μM. RVT destabilized the LUVs, resulting in ANTS leakage at 30 μM, which partly accounts for the error at that concentration (as well as the difficulty in getting results at 100 μM).

Figure 3.

Effects of compounds 10a-j and RVT on gramicidin channel activity, as quantified using a fluorescence quench assay; mean ± s.e.m., n = 3 (1, 3, 10, 30 and 100 μM); for RVT at 100 μM, n = 1.

RVT induced membrane perturbation at concentrations above 10 μM, where it shifted the fluorescence quench rate by a factor 2.3. In contrast, compounds 10ac and 10e-j did not cause changes in the quench rate at concentrations up to 100 μM (Figure 3), demonstrating that they produce minimal changes in bilayer properties. Compound 10d shifted the monomer⟷dimer equilibrium toward the left, indicating that it increased $\Delta G_{\text{bilayer}}^{2\text{M}\to \text{D}}$.

2.7. Evaluation of mutagenicity using a micronucleus assay in peripheral blood cells of mice

The mutagenicity induced by compound 10a and RVT was evaluated through a micronucleus assay in the peripheral blood cells of mice. The compounds were injected orally at different concentrations (25, 50, and 100 mg/Kg body weight of mice). Figure 4 shows the average frequencies of micronucleated reticulocytes (MNRETs) induced by RVT and compound 10a. In this assay, the reference drug cyclophosphamide (used as the control) increased the number of MNRETs up to 44.8 ±8.3. For RVT, the frequency of MNRETs was determined as being 2.2 ±1.3; 2.0 ±2.1, and 2.0 ±1.6 at concentrations of 25, 50, and 100 mg/kg, respectively. In addition, for compound 10a the frequencies of MNRETs were 1.6 ±0.5, 1.6 ±1.5, and 2.0 ±1.0 at concentrations of 25, 50, and 100 mg/kg, respectively. We did not observe differences between the negative control (vehicle – suspension containing 1% of carboxymethylcellulose (CMC) and 0.2% Tween), water, RVT, and compound 10a at all concentrations.

Figure 4.

Average frequency of micronucleated reticulocytes (MNRET) and standard deviation of 1000 cells obtained from mice treated with the positive control cyclophosphamide (C+; 50 mg/Kg), negative control (C–), water, resveratrol (RVT) and compound 10a (25, 50 and 100 mg/Kg). *p <0.05 (compared to the positive control).

3. DISCUSSION

It remains important to search for small molecules that can eliminate/minimize the symptoms, as out alternatives to treat SCD, since current advances such as hematopoietic stem cell transplantation and gene therapy exhibit several barriers [9]. Among the strategies, fetal hemoglobin (HbF) induction is a validated approach to mitigate SCD, with HU being the main representative drug used for infants and adults. However, HU exhibits several limitations, which justify the search for new HbF-inducing agents [2,7].

RVT has been reported as an interesting prototype drug for the treatment of SCD since it induces erythroid differentiation and increases the gamma-globin gene expression and HbF levels in the K562 cells and the erythroid precursor cells isolated from sickle cell patients [10–11]. Beyond gamma-globin and HbF production, RVT also exhibits an antioxidant effect, reduces the levels of reactive oxygen species (ROS), being useful not only for SCD, but also for beta-thalassemia [19–23]. In a double-blind randomized clinical trial, RVT showed a similar efficacy with HU in non-transfusion beta-talassemia intermedia patients to achieve the primary endpoint (change in Hb levels) after 6-month follow-up [23].

In the search for new HbF inducers, our research group has described several compounds with nitric oxide-donor properties active under in vitro and in vivo conditions [5,6,20,21]. Nitric oxide has a vital role in inducing the expression of gamma-globin gene through the activation of the soluble guanylyl cyclase (sGC); this leads to an increased expression of the gamma-globin gene in the primary erythroblasts and erythroleukemic cells [22]. Transcription factors such as c-Fox are activated by the NO/cGMP pathway and thus bind to the Hypersensitive Site 2 in the β-globin Locus Control region and to the Sp1 in the CACCC box [24, 25]. Because of the beneficial effects of both these compounds, it was hypothesized that the combination of RVT and its derivatives displaying NO-donor properties could be a valuable alternative to obtain new gamma-globin inducing agents.

In this work, furoxan derivatives were selected as NO-donor subunits to evaluate the contribution of different levels of NO donation on the production of gamma-globin. It is well established that NO release is dependent on the presence of cysteine, and different profiles can be achieved by furoxan due to the electronic influence of the substituents at the 3-position. Thus, electron-withdrawing substituents attached on the carbon at this position are more prone to release NO in the medium [5,6,12,26]. Studies have shown that cysteine acting as nucleophile add to carbon-3 in the furoxan derivatives, promoting the conversion to an open-ringed nitroso intermediate. This intermediate rearranges and releases NO from these heterocyclic systems [27,28]. This is the reason that absence of cysteine in the medium did not promote NO-release from furoxan derivatives in our experiment.

Indeed, using an indirect assay, we observed that the compounds (10a-i) generated different levels (ranging from 0–26.3%) of nitrite in the medium, while RVT did not demonstrate this effect. The nitrite quantified in this assay was resulted from oxidation of NO with oxygen and water. Compounds 10 g and 10 h are isosteres that lack the N-oxide subunit and do not release NO in the medium. These compounds were designed as negative controls aiming to evaluate the contribution of NO-release for the biological effects [5,6,12,26,29].

The multiple beneficial effects of NO on SCD go beyond HbF induction. This mediator acts directly on the cardiovascular system by decreasing the ‘vaso-constrictor’ profile in SCD patients, inhibits platelet aggregation, and reduces the adhesion molecules in the vascular endothelium. All these effects work synergistically on the reduction of the vaso-occlusion process [3,4,30]. Moreover, NO has an important contribution in analgesia, which is related to the activation of the NO-cGMP pathway [31–32].

Thus, we evaluated the antinociceptive effect of all the compounds after oral administration in the abdominal-contortions-model induced by acetic acid in mice [17]. RVT and its analogs have already been described to exhibit analgesic and anti-inflammatory effects [33–35]. Herein, the compounds (10a-i) have shown an antinociceptive effect by protecting the mice against chemical-induced writhes at values ranging from 5.9–37.3%. Except for the most active compound (10a), these values were inferior to that of dipyrone, and RVT (with analgesic effect at levels of 34.7 and 39%, respectively) used as the controls. We did not find a direct relationship between NO-release levels and the analgesic effect (Table 2). However, NO seems to contribute to the analgesic effect, since the percentage of protection for the furazan derivative (10 h) (analog of compound (10a)) was found to be only 15.8%. The dimethyloxazole analog (10 g) that does not release NO also exhibited weaker analgesic effects compared to the compound (10a). Acute pain, the hallmark of SCD, is the most common cause of hospitalization among patients [36]. The treatment of this chronic condition is frequently suboptimal, which decreases the quality of life of the patients [37]. Therefore, compounds capable of managing pain, such as compound (10a), are highly desirable in the treatment.

As an analgesic compound, RVT reduces the levels of the pro-inflammatory cytokines TNF-α and IL-1β, which act as pain mediators [38–40]. The direct action of TNF-α on the nociceptor is related to the maintenance of pain, and its inhibition decreases this effect [41]. On the other hand, IL-1β has been demonstrated to be a potent hyperalgesic agent in several peripheral tissues due to its action on upregulating the Nerve Growth Factor (NGF), an important mediator in both acute and chronic pain, at the post-transcriptional and transcriptional levels. Some pieces of evidence also suggest a direct effect of IL-1β on the nociceptor and activation of the signaling cascades, culminating in the release of nociceptive molecules including substance-P, interleukin–6, and prostaglandins [42].

Increased plasma levels of TNF-α and IL-1β are described for the SCD patients and associated with aggravating pain in these individuals [43,44]. The FDA-approved drug hydroxyurea does not decrease the levels of TNF-α, IL-1β, and KC in the supernatants of the monocyte cultures from sickle cell mice [20,21]. Thus, in order to investigate the ability of the most analgesic compound (10a) in reducing the levels of TNF-α and IL-1β, we performed their quantification in the supernatants of the macrophage culture cells previously stimulated with LPS (50 mg/mL). Our results showed that the percentage of TNF-α inhibition after LPS-stimulation demonstrated by compound (10a) ranged in the levels from 41.1–64.3% at concentrations between 3.13–12.5 μM. RVT did not decrease the TNF-α levels produced by the macrophages up to 50 μM, whereas thalidomide inhibited 32.7% at this concentration. None of the compounds reduced the levels of IL-1β at the test concentrations. These results suggest that not only the activation of the NO pathway but also TNF-α inhibition could be responsible for the analgesic effect of compound 10a.

The ability of compound 10a and RVT to induce the production of gamma-globin was measured by HPLC using the model of CD34⁺ cells. Both compounds, at a concentration of 25 μM, have shown similar activity by inducing twice the levels of the gamma-globin chains (γ^G+γ^A ) compared to the vehicle. Similar to literature reports, we observed a selective gamma-globin induction because the levels of heme, delta, beta, and alfa chains were not altered [10,11]. High levels of HbF are associated with a reduction in the disease severity, leading to improvement in the quality and expectancy of life [8,45].

Despite the beneficial effect of RVT on HbF induction, phenolic phytochemicals and their derivatives are known to have nonspecific effects, which may be mediated through the modification of membrane protein function. Being amphiphiles, some phytochemicals can adsorb at the bilayer/solution interface interface, thereby altering lipid bilayer properties. This effect results in modulation of membrane protein function, which in turn produces indiscriminate changes in signaling pathways [46].This may account for the multiple effects of certain phytochemicals. RVT, for example, perturbs the bilayer thickness and elasticity, as well as the lateral pressure profile in the interfacial region of lipid bilayers, which causes it to be promiscuous compound that also shifts the gramicidin monomer⟷dimer equilibrium toward the right [46]. We therefore decided to evaluate the bilayers-perturbing effects of all the compounds tested in this study using a fluorescence quench assay [18,46]. RVT induced membrane perturbation from 10 μM and the ratio rate/rate (control) were 2.3, 5.8 and 9.1 at concentrations of 10, 30 and 100 μM, respectively. In contrast, compounds 10a-j did not perturb the membrane, presumably because the polar furoxan group reduces their partition coefficient into the bilayer. The presence of the methoxy group reduces the polarity of the compounds 10a-j and hence their amphiphilic profiles and their subsequently reduced interference.

Moreover, the mutagenicity induced by compound 10a was evaluated through the micronucleus test in the peripheral blood cells of the mice [47]. Although some reports described that NO induces chromosomal mutations in the mammalian cells due to the formation of reactive species (i.e., peroxynitrite) or N-nitroso-compounds [48,49], we did not find any in vivo mutagenic effect after oral administration of compound 10a. For this compound, the frequencies of MNRETs were 1.6 ±0.5, 1.6 ±1.5, and 2.0 ±1.0 at concentrations of 25, 50, and 100 mg/kg, respectively. Furoxan derivatives have been described as non-mutagenic compounds, which was confirmed by our studies [5,6,29].

All of these results demonstrate that RVT derivatives with NO-donor properties could represent promising prototypes and an important alternative to treat SCD symptoms. Specifically, the compound 10a has emerged as a new prototype with NO-donor release properties, analgesic effect, ability to inhibit TNF-α inhibition and induce production of gamma-globin, and an in vivo non-mutagenic effect.

4. CONCLUSION

A new series of RVT derivatives with NO-donor properties (10a-i) was synthesized and characterized by analytical methods. All the compounds (10a-i) were able to release NO at different levels, ranging from 0–26.3%. RVT did not demonstrate NO-releasing properties. All these compounds (10a-i) exhibited an analgesic effect; compound 10a was the most active among them, protecting up to 37.3% of the contortions induced by acetic acid in mice. This value was similar to that of dipyrone and RVT when used as reference drugs. Compound 10a reduced the levels of TNF-α in macrophage-cultured cells after LPS-stimulation under in vitro conditions from 41.1–64.3% at concentrations. This reduction was achieved when 3.13–12.5 μM of compound 10a was used; this mechanism was related in part to its analgesic effect. Moreover, compound 10a at 25 μM concentration induced twice the amount of gamma-globin chains (γ^G+γ^A ) compared to the vehicle using the CD34⁺ cells. This effect was similar to that induced by RVT. Different from RVT, compound 10a did not induce any membrane perturbation. In addition, this promising compound was not mutagenic when evaluated through the micronucleus test in the peripheral blood cells of mice. Therefore, compound 10a emerged as a new prototype of an HbF-inducer with analgesic and anti-inflammatory actions and can be used as an alternative to hydroxyurea to treat SCD symptoms.

5. EXPERIMENTAL SECTION

Chemistry. General Information.

All reagents and solvents of purity grade were purchased from commercial suppliers. Distillation, followed by dehydration, was used to dry all solvents. The reactions were monitored through thin-layer chromatography (TLC) on a pre-coated silica gel 60 (HF–254; Merck) of a thickness of 0.25 mm and visualized under the UV light (265 nm). The purification of all compounds was performed on a chromatography column using silica gel (60 Å pore size, 35–75-μm particle size) as the stationary phase and a mixture of the following solvents in different proportions as the mobile phase: dichloromethane, ethyl acetate, hexane, and petroleum ether. Before starting the chemical analysis and biological assays, the purity was confirmed by HPLC to be more than 98.5%. The melting points (mp) were determined in open capillary tubes using an electrothermal melting point apparatus (SMP3; Bibby Stuart Scientific). Infrared (IR) spectroscopy (KBr disc) was performed on an FTIR–8300 Shimadzu spectrometer, and the frequencies were expressed in cm^–1. The nuclear magnetic resonance (NMR) spectra for ¹H and ¹³C of all the compounds were scanned on a Bruker Fourier with a Dual probe ¹³C/¹H (300-MHz) NMR spectrometer using deuterated solvents (dimethyl sulfoxide, acetone, and chloroform). Chemical shifts were expressed in parts per million (ppm) relative to tetramethylsilane. The signal multiplicities were reported as singlet (s), doublet (d), doublet of doublet (dd), and multiplet (m). Elemental analyses (C, H, and N) were performed on a Perkin-Elmer model 240C analyzer and the data were within ±0.4% of the theoretical values.

General procedures for the synthesis of compounds 7–9

Compounds 4–6 were synthesized according to previously described methodologies [12,13]. In a nitrogen atmosphere, an equivalent amount of 4-vinylphenol (4.2 mmol) and 1,8-diazabicycloundec–7-ene (DBU) (4.2 mmol) in anhydrous dichloromethane was kept under stirring conditions for 15 min. Subsequently, furoxan derivatives (4–6) in milligram concentrations (4.2 mmol) were solubilized in 7 mL of anhydrous dichloromethane and added drop-wise to the reaction medium. The reaction was stirred for 18–48 h at room temperature and monitored by TLC. The reaction medium was then diluted with 50 mL of dichloromethane and washed with saturated potassium carbonate solution (3 × 20 mL). The solvent was evaporated under reduced pressure, and the product was purified by column chromatography (flash silica; eluent: dichloromethane: petroleum ether, 7: 3 (v/v)) providing compounds 7–9 at yields ranging from 29–48% as described.

3-methyl-4-(4-vinylphenoxy)-1,2,5-oxadiazole-2-N-oxide (7).

Yellow powder; yield: 48%; mp 65.2 – 67 °C. IR Vmax (cm⁻¹; KBr pellets): 3113 (C-H _aromatic), 2927 (C-H _alkyl), 1637 (C=C _alkene), 1479 (C=C _aromatic), 1382 (N-O _oxide). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.48 (2H; d; J_orto = 8.7 Hz), 7.26 (2H; d; J_orto = 8.7 Hz), 6.71 (1H; dd; J_cis = 10.8 and J_trans = 17.7 Hz), 5.74 (1H; d; J_gem = 0.6 Hz and J_trans = 17.7 Hz), 5.28 (1H; d; J_cis = 10.8 Hz), 2.21 (3H; s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 162.86, 152.23, 135.93, 135.59, 127.80, 119.71, 114.82, 106.61, 7.18. (MW: 218.21; C₁₁H₁₀N₂O₃; Rf : 0.35, eluent petroleum ether: dichloromethane 7: 3 (v / v)).

3-phenyl-4-(4-vinylphenoxy)-1,2,5-oxadiazole 2-N-oxide (8).

White powder; yield: 35%; mp 86.7 – 88.8 °C. IR Vmax (cm⁻¹; KBr pellets): 3064 (C-H _aromatic), 2933 (C-H _alkyl), 1633 (C=C _alkene), 1440 (C=C _aromatic), 1334 (N-O _oxide). ). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 8.20 (2H; dd; J_orto = 8.1 Hz and J_meta = 1.8 Hz), 7.56 – 7.53 (3H; m), 7.5 (2H; d; J_orto = 8.8 Hz), 7.35 (2H; d; J_orto = 8.8 Hz), 6.73 (1H; dd; J_cis = 10.8 Hz and J_trans = 17.7 Hz), 5.75 (1H; dd; J_trans = 17.7 Hz), 5.30 1H; dd; J_cis = 10.8 Hz). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 161.96, 152.09, 136.13, 135.49, 130.76, 129.02, 127.73, 126.42, 122.07, 120.16, 114.83, 107.94. (MW: 280.28; C₁₆H₁₂N₂O₃; Rf : 0.17, eluent petroleum ether: dichloromethane 9: 1 (v / v)).

3-(phenylsulfonyl)-4-(4-vinylphenoxy)-1,2,5-oxadiazole 2-N-oxide (9).

White powder; yield: 29%; mp 112.2 – 115 °C. IR Vmax (cm⁻¹; KBr pellets): 3095 (C-H _aromatic), 2922 (C-H _alkyl), 1622 (C=C _alkene), 1446 (C=C _aromatic), 1355 (N-O _oxide). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 8.12 (2H; d; J_orto = 9 Hz), 7.78 – 7.86 (1H; m), 7.64 – 7.69 (2H; m), 7.47 – 7.57 (2H; m), 7.26 – 7.29 (2H; m), 6.73 (1H; dd; J_cis = 10.9 Hz and J_trans = 17.7Hz), 5.76 (1H; d; J_trans = 17.7 Hz), 5.32 (1H; d; J_cis = 10.9 Hz). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 158.56, 152.08, 138.05, 136.49, 135.96, 135.52, 129.92, 128.79, 127.82, 120.09, 115.16, 110.89. (MW: 344.34; C₁₆H₁₂N₂O₅S; Rf : 0.90, eluent petroleum ether: dichloromethane 2: 8 (v / v)).

General procedures for the synthesis of compounds 10a-c.

A reaction mixture containing tris(dibenzylideneacetone)dipalladium(0) [Pd₂ (dba)₃] (4% mol), sodium acetate (1.74 mmol), N-oxide derivative (7 or 8 or 9) (0.58 mmol), and dry benzonitrile (2 mL) was stirred under a nitrogen atmosphere at room temperature for 10 min. Subsequently, the reactive 4-methoxybenzenediazonium tetrafluoroborate (0.87 mmol) was slowly added to the reaction mixture. The reaction was then maintained under stirring conditions at room temperature for 1 h and monitored by TLC. The product was isolated by diluting the reaction medium with approximately 80 mL of dichloromethane. The organic phase was washed with distilled water (4 × 15 mL) and dried with sodium sulfate or magnesium sulfate. After filtration, the organic phase was concentrated under reduced pressure to produce the crude, which was purified by column chromatography (flash silica; eluent: dichloromethane: petroleum ether gradient, 1:1 up to 8: 2 (v/v)) providing yields of 51–70% for the compounds 7–9 ranging as described.

(E)-4-(4-(4-methoxystyryl)phenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (10a).

White powder; yield, 65%; mp 159 – 162 °C. Rf : 0.30, eluent petroleum ether: dichloromethane 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3067 (C-H _aromatic), 2929 (C-H _alkyl), 1635 (C=C alkene), 1539 (C=N _furoxan), 1602 and 1481 (C=C _aromatic), 1381 (N-O _oxide), 1253 (C-O-C _ether). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.54 (2H; d; J_ortho = 8.8 Hz), 7.45 (2H; d; J_ortho = 8.7 Hz), 7.28 (2H; d; J_orto = 8.8 Hz), 7.02 (2H; t; J_trans = 16.5 Hz), 6.91 (2H, d, J_ortho = 8.7 Hz), 3.84 (3H; s), 2.21 (3H;s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 162.67, 159.41, 151.52, 135.87, 129.64, 128.97, 127.71, 127.46, 124.94, 119.61, 114.09, 106.42, 55.25, 6.98. Calculated for C₁₈H₁₆N₂O₄. C, 66.66; H, 4.97; N, 8.64. Found: C, 66.43; H, 4.95; N, 8.62.

(E)-4-(4-(4-methoxystyryl)phenoxy)-3-phenyl-1,2,5-oxadiazole 2-N-oxide (10b).

White powder; yield, 70%; mp 176.4 – 179.5 °C. Rf : 0.41, eluent petroleum ether: dichloromethane 6: 4 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3066 (C-H _aromatic), 2974 and 2839 (C-H _alkyl), 1606 (C=C _alkene), 1514 (C=C _aromatic), 1336 (N-O_oxide), 1255 (C-O-C _ether). ¹H NMR (300 MHz, DMSOd₆, δ ppm) δ: 8.09 (2H; dd; J_ortho = 8.1 Hz), 7.69 (2H; d; J_ortho = 8.7 Hz), 7.64–7.59 (3H; m), 7.55 (2H; d; J_ortho = 8.7 Hz), 7.50 (2H; d; J_ortho = 8.7 Hz), 7.24 (1H; d; J_trans = 16.2 Hz), 7.13 (1H; d; J_trans = 16.2 Hz), 6.96 (2H; d; J_ortho = 8,7 Hz), 3.78 (3H;s). ¹³C NMR (75MHz, DMSOd₆, δ ppm) δ: 162.19, 159.12, 151.64, 135.84, 130.94, 129.53, 129.14, 128.79, 127.91, 127.60, 126.68, 124.89, 120.44, 114.23, 55.19. Calculated for C₂₃H₁₈N₂O₄. C, 71.49; H, 4.70; N, 7.25. Found: C, 71.42; H, 4.74; N, 7.21.

(E)-4-(4-(4-methoxystyryl)phenoxy)-3-(phenylsulfonyl)-1,2,5-oxadiazole 2-N-oxide (10c).

Yellow powder; yield, 51%; mp 175.4 – 176.8 °C. Rf : 0.22, eluent petroleum ether: dichloromethane 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3051 (C-H _aromatic), 2956 and 2841 (C-H _alkyl), 1631 (C=C _alkene), 1537 (C=C _aromatic), 1357 (N-O_oxide), 1257 (C-O-C _ether), 1163 (S=O). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 8.12 (2H; d; J_ortho = 8.4 Hz), 7.79 (1H; t; J_ortho = 8.4 Hz), 7.66 (2H; t; J_ortho = 8.4 Hz), 7.55 (2H; d; J_ortho = 9 Hz), 7.47 (2H; d; J_ortho = 8.8 Hz), 7.30 (2H; d; J_ortho = 9 Hz), 7.00 – 7.13 (2H; m), 6.92 (2H; d; J_ortho = 8.8 Hz), 3.85 (3H, s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 159.34, 158.16, 151.26, 137.76, 136.26, 135.50, 129.49, 129.13, 128.39, 127.58, 127.28, 124.68, 119.77, 113.96, 110.51, 55.08. Calculated for C₂₃H₁₈N₂O₆S. C, 61.33; H, 4.03; N, 6.22. Found: C, 61.29; H, 4.06; N, 6.25.

General procedures for the synthesis of compounds 10d-f.

Compounds 12a-b and 13 were synthesized according to previously described methodologies [16]. A mixture containing 4-(3,5-dimethoxystyryl)phenol (13) (0.97 mmol), DBU (1.33 mmol), and 7 mL of dry 1,2-dichloroethane was stirred under a nitrogen atmosphere at the room temperature for 15 min. Afterward, the N-oxide derivative (4 or 5 or 6) (1 mmol) was added in the medium. The reaction was kept under stirring conditions at 55°C for 2–8 h and monitored by TLC. The product was isolated by diluting the reaction medium with approximately 60 mL of 1,2-dichloroethane. The organic phase was washed with distilled water (4 × 15 mL) and dried with sodium sulfate or magnesium sulfate. After filtration, the organic phase was concentrated under a reduced pressure to produce the crude, which was purified by column chromatography (flash silica; eluent: dichloromethane: petroleum ether, 7:3 (v/v)) providing compounds 10d-f at yields ranging from 30–60% as described.

(E)-4-(4-(3,5-dimethoxystyryl)phenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (10d).

White powder; yield, 40%; mp 115.7 – 116.9 °C. Rf : 0.50, eluent petroleum ether: dichloromethane 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 2924 and 2852 (C-H _alkyl), 1637 (C=C _alkene), 1593 (C=C _aromatic), 1379 (N-O _oxide), 1228 (C-O-C _ether). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.56 (2H; d; J_ortho = 8.7 Hz), 7.31 (2H; d; J_ortho = 8.7 Hz), 7.09 (1H; d; J_trans = 16.2 Hz), 7.01 (1H; d; J_trans = 16.2 Hz), 6.68 (2H; d; J_meta = 2.4 Hz), 6.43 (1H; t; J_meta = 2.4 Hz), 3.85 (6H; s), 2.23 (3H; s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 162.54, 160.87, 151.87, 138.79, 135.25, 129.36, 127.78, 127.49, 119.59, 106.34, 104.69, 100.02, 55.23, 7.33. Calculated for C₁₉H₁₈N₂O₅. C, 64.4; H, 5.12; N, 7.91. Found: C, 64.7; H, 5.15; N, 7.87.

(E)-4-(4-(3,5-dimethoxystyryl)phenoxy)-3-phenyl-1,2,5-oxadiazole 2-N-oxide (10e).

Yellow powder; yield, 60%; mp 101.7 – 102.2 °C. Rf : 0.75, eluent petroleum ether: dichloromethane 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3084 (C-H _aromatic), 2966 and 2833 (C-H _alkyl), 1604 (C=C _alkene), 1504 (C=C _aromatic), 1336 (N-O _oxide), 1215 (C-O-C _ether). ¹H NMR (300 MHz, DMSOd₆, δ ppm) δ: 8.09 (2H; dd; J_ortho = 8.6 Hz), 7.72 (2H; d; J_ortho = 8.7 Hz), 7.67 – 7.59 (3H; m), 7.54 (2H; d; J_ortho = 8.7 Hz), 7.34 (1H; d; J_trans = 16.5 Hz), 7.21 (1H; d; J_trans = 16.5 Hz), 6.80 (2H; d; J_meta = 2.1 Hz), 6.43 (1H; t; J_meta = 2.1 Hz), 3.78 (6H; s). ¹³C NMR (75MHz, DMSOd₆, δ ppm) δ: 162.13, 160.70, 152.06, 138.94, 135.36, 130.95, 129.14, 128.02, 127.72, 126.68, 121.75, 120.47, 108.35, 104.57, 100.06, 55.25. Calculated for C₂₄H₂₀N₂O₅. C, 69.22; H, 4.84; N, 6.73. Found: C, 69.28; H, 4.88; N, 6.77.

(E)-4-(4-(3,5-dimethoxystyryl)phenoxy)-3-(phenylsulfonyl)-1,2,5-oxadiazole 2-N-oxide (10f).

White powder; yield, 30%; mp 125.5 – 132.1 °C. Rf : 0.67, eluent petroleum ether: dichloromethane 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3066 (C-H aromatic), 2935 and 2839 (C-H alkyl), 1593 (C=C alkene), 1500 (C=C aromatic), 1373 (N-oxide), 1296 (C-O-C ether), 1151 (S=O). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.85 (2H; d; J_ortho = 8.2 Hz), 7.68 (1H; t), 7.54 (2H; d; J_ortho = 8.2 Hz), 7.41 (2H; d; J_ortho = 9 Hz), 6.94 (4H; m), 6.63 (2H; d; J_meta = 2.2 Hz), 6.40 (1H; t; J_meta = 2.2 Hz), 3.82 (6H; s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 161.14, 153.22, 148.94, 138.99, 136.41, 135.47, 134.39, 129.95, 129.30, 128.68, 127.73, 122.77, 119.61, 104.81, 100.33, 55.33. Calculated for C₂₄H₂₀N₂O₇S. C, 59.99; H, 4.20; N, 5.83. Found: C, 60.02; H, 4.17; N, 5.80.

Synthesis of compound (E)-4-(4-(4-methoxystyryl)phenyl)-3,5-dimethylisoxazole (10g).

The compound 10 g was prepared in two steps from 4-bromostyrene (12b). First, compound 3,5-dimethyl-4-(4-vinylphenyl)isoxazole (15) was synthesized through the treatment of the 4-bromostyrene (12b) (1.37 mmol) (commercial available reagent), 3,5-dimethylisoxazole-4-boronic acid pinacol ester (14) (1.5 mmol), tetrakis (triphenylphosphine)palladium (0) (0.07 mmol), and potassium carbonate solution (2.75 mmol in 0.5 mL distilled water) in a dioxane (8 mL) medium. The reaction was kept under stirring at 90°C for 24 h and monitored by TLC. The solvent was concentrated under reduced pressure, and the medium was diluted with approximately 60 mL of dichloromethane. The organic phase was washed with distilled water (4 × 15 mL) and dried with sodium sulfate. After filtration, the organic phase was concentrated under reduced pressure to produce the crude, which was purified by column chromatography (flash silica; eluent: hexane: ethyl acetate, 7:3 (v/v)) providing the compound 15 at yields of 55%.

In the second step, a reaction mixture containing tris(dibenzylideneacetone)dipalladium(0) [Pd₂ (dba)₃] (4% mol), sodium acetate (1.74 mmol), 3,5-dimethyl-4-(4-vinylphenyl)isoxazole (15) (0.58 mmol), and dry benzonitrile (2 ml) was stirred under a nitrogen atmosphere at the room temperature for 10 min. Subsequently, the reactive 4-methoxybenzenediazonium tetrafluoroborate (0.87 mmol) was slowly added to the reaction mixture. The reaction was stirred at room temperature for 3 h and monitored by TLC. When necessary, additional 4-methoxybenzenediazonium tetrafluoroborate was added to enhance the styrene (15) consumption. The product was isolated by diluting the reaction medium with approximately 100 mL of dichloromethane. The organic phase was washed with distilled water (4 × 15 mL) and dried with sodium sulfate. After filtration, the organic phase was concentrated under reduced pressure to produce the crude, which was purified by column chromatography (flash silica; eluent: hexane: ethyl acetate, 7:3 (v/v)) providing the compound 10 g at a yield of 41%.

3,5-dimethyl-4-(4-vinylphenyl)isoxazole (15).

Yellowish powder; yield, 55%. ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.48 (2H; d), 7.22 (2H; d), 6.74 (1H; dd; J_cis = 10.7 Hz and J_trans = 17.5 Hz), 5.81 (1H; d; J_trans= 17.5 Hz), 5.30 (1H; d; J_cis = 10.7Hz), 2.39 (3H; s), 2.27 (3H; s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 165, 159, 137, 136, 129, 128, 125, 116, 114, 11, 9.

(E)-4-(4-(4-methoxystyryl)phenyl)-3,5-dimethylisoxazole (10g).

Yellowish powder; yield, 41%; mp 158.4 – 164.4 °C. Rf : 0.65, eluent hexane: ethyl acetate 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3022 (C-H _aromatic), 2931 (C-H _alkyl), 1598 (C=C _alkene), 1508 (C=C aromatic), 1249 (C-O-C _ether). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.56 (2H; d; J = 8.3 Hz), 7.48 (2H; d; J = 8.8 Hz), 7.24 (2H; d; J = 8.3 Hz), 7.12 (1H; d; J_trans = 16.1 Hz), 7.00 (1H; d; J_trans = 16.1 Hz), 6.92 (2H; d; J = 8.8 Hz), 3.84 (3H; s), 2.43 (3H; s), 2.30 (3H; s). ¹³C NMR (75MHz, DMSOd₆, δ ppm) δ: 165.33, 159.62, 137.08, 129.44, 129.31, 128.96, 127.96, 126.71, 125.94, 114.34, 55.50, 11.81, 11.04. Calculated for C₂₀H₁₉NO₂. C, 78.66; H, 6.27; N, 4.59. Found: C, 78.65; H, 6.30; N, 4.55.

Synthesis of compound (E)-3-(4-(4-methoxystyryl)phenoxy)-4-methyl-1,2,5-oxadiazole (10h).

This reaction was adapted according to a previous methodology [50]. A solution containing (E)-4-(4-(4-methoxystyryl) phenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (10a) (20 mmol) in trimethyl phosphite (30 mL; 0.25 mol) was stirred and kept under reflux conditions for 48 h. The reaction was monitored by TLC. Later, the medium was cooled and added to the water/ice mixture (60 mL), and the pH was adjusted to 7 with 4M NaOH. Later on, the aqueous phase was washed with ethyl acetate (4 × 20 mL) and dried with sodium sulfate. After filtration, the organic phase was concentrated under reduced pressure to produce the crude, which was purified by column chromatography (flash silica; eluent: hexane: ethyl acetate, 7:3 (v/v)) providing compound 10 h with a yield of 75%.

(E)-3-(4-(4-methoxystyryl)phenoxy)-4-methyl-1,2,5-oxadiazole (10h).

White powder; yield, 75%; IR Vmax (cm⁻¹; KBr pellets): 3066 (C-H aromatic), 2974 – 2839 (C-H alkyl), 1606 (C=C alkene), 1514 (C=C aromatic), 1336 (N-O), 1255 (C-O-C ether). ¹H NMR (300 MHz, CDCl₃d, δ ppm) δ: 7.51 (2H; d; J = 8.2 Hz), 7.44 (2H; d; J = 8.5 Hz), 7.30 (2H; d; J = 8.2 Hz), 7.02 (1H; t; J = 15.9 Hz), 6.95 (1H; t; J = 15.9 Hz), 6.90 (2H; d; J = 8.5 Hz), 3.83 (3H; s), 2.38 (3H, s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 163.4, 159.4, 153.2, 144.62, 135.4, 129.8, 128.7, 127.7, 127.5, 125.1, 119.0, 114.1, 55.3, 7.4. Calculated for C₁₈H₁₆N₂O₃. C, 70.12; H, 5.23; N, 9.09. Found: C, 70.08; H, 5.19; N, 10.03.

(E)-4-(4-(4-hydroxystyryl)phenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (10i)

Compound 10i was prepared in two steps from 3-methyl-4-nitro-1,2,5-oxadiazole 2-N-oxide (4). First, in a round bottom flask, 4-iodophenol (2.27 mmol), DBU (2.27 mmol; 340 μL), and 5 mL of anhydrous dichloromethane were added. This mixture was kept under stirring conditions and a nitrogen atmosphere (protected from light) at 45°C for 15 min. Then, 3-methyl-4-nitro-1,2,5-oxadiazole 2-N-oxide (4) (2.27 mmol), previously solubilized in anhydrous dichloromethane, was slowly added dropwise in the medium. The reaction was kept under stirring conditions at 45 °C, protected from light for 24 h, and monitored by TLC. Afterward, the product was isolated by diluting the reaction medium with approximately 100 mL of dichloromethane, which was washed with distilled water (4 portions of 15 mL). The organic phase was dried with sodium sulfate, filtered, and concentrated under reduced pressure to produce the crude. The purification was performed by column chromatography (stationary phase: silica gel; mobile phase: hexane: dichloromethane, 7: 3 (v/v)) providing the compound (16) with 55% of yield.

In the second step, a mixture containing 4-vinylphenol (12a) (1.26 mmol), triethanolamine (1 mL), palladium acetate (10.5 mg; 4.5% mol), and 4-(4-iodophenoxy)-3-phenyl-1,2,5-oxadiazole 2-N-oxide (16) was kept under stirring conditions and a nitrogen atmosphere, protected from light at 65^oC for 24 h. Afterward, the reaction was cooled, and 10 mL of distilled water was added, leading to the precipitation of compound (10i). The solid was filtered and left to dry at room temperature. The purification was performed by column chromatography (stationary phase: silica gel; mobile phase: hexane: dichloromethane, 7: 3 (v/v)) providing compound (10i) as a white solid with 53% of yield.

4-(4-iodophenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (16).

White powder; yield, 55%; mp 103 – 104.5 °C. IR Vmax (cm⁻¹; KBr pellets): 3091 (C-H _aromatic), 2900 (C-H _alkyl), 1499 (C=C _aromatic), 1384 (N-O _oxide). ¹H NMR (300 MHz, DMSOd₆, δ ppm) δ: 7.82 (2H; d; J_orto = 9 Hz), 7.24 (2H; d; J_orto = 9Hz), 2.09 (3H; s). ¹³C NMR (75MHz, CDCl₃d, δ ppm) δ: 162.54, 152.64, 139.16, 121.8, 106.43, 90.22, 7.16.

(E)-4-(4-(4-hydroxystyryl)phenoxy)-3-methyl-1,2,5-oxadiazole 2-N-oxide (10i).

Grey powder; yield, 53%; mp 216 – 218 °C. Rf : 0.45, eluent hexane: ethyl acetate 7: 3 (v / v). IR Vmax (cm⁻¹; KBr pellets): 3527 (O-H), 3010 (C-H _aromatic), 2821 (C-H _alkyl), 1539 (C=C alkene), 1514 (C=C _aromatic), 1381 (N-O _oxide). ¹H NMR (300 MHz, Acetoned₆, δ ppm) δ: 7.67 (2H; d; J_orto = 9 Hz), 7.48 (2H; d; J_orto = 8.7 Hz), 7.40 (2H; d; J_orto = 9 Hz), 7.20 (1H; d; J_trans = 16.3 Hz), 7.08 (1H; d; J_trans = 16.3 Hz), 6.87 (2H; d; J_orto = 8.7 Hz), 2.18 (3H; s). ¹³C NMR (75MHz, Acetone_d6, δ ppm) δ: 164.04, 158.33, 152.87, 136.92, 129.98, 129.62, 128.76, 128.24, 124.96, 120.59, 116.36, 107.62, 7.07. Calculated for C₁₇H₁₄N₂O₄. C, 65.80; H, 4.55; N, 9.03; O, 20.62. Found: C, 65.77; H, 4.52; N, 9.07.

Pharmacology

Animals.

Adult male Swiss albino mice (25 – 35 g body weight) were used in the experiments. They were housed in single-sex cages under a 12 h light/12 h dark cycle (lights on at 06:00 hrs) in a controlled-temperature room (22 ±2 °C). The animals had free access to food and water. The experiments were performed after approval by the local Institutional Ethics Committee (CEUA Process No. 08/2017; 17/2015; 76/2015; 77/2015; FCM UNICAMP CAAE: 45878215.8.0000.5404). All experiments were performed in accordance with the current guidelines for the care of laboratory animals and the ethical guidelines for the investigation of experimental pain in conscious animals.

Humans.

For the experiments using CD34+ cells, the biological material was donated by volunteers and the experiment was performed in accordance with the current and ethical guidelines for experimentation using biologic material from humans. The procedure was previously approved by the local Institutional Ethics Committee (FCM UNICAMP CAAE: 45878215.8.0000.5404).

Quantification of Nitrite by the Griess Reaction

The levels of nitrite resulting from the oxidation of NO in the aqueous medium were quantified through the Griess reaction after incubation with an excess of L-cysteine (1:50) according to previously published methods [5,12,26]. The experiments were performed in triplicate and repeated three times on different days. No production of nitrite was observed in the absence of L-cysteine. The results were expressed as a percentage of nitrite (% NO₂⁻; mol/mol). Statistical analysis was carried out using ANOVA followed by Tukey’s Multiple Comparison Test at a significance level of p < 0.05.

Antinociceptive Activity

The in vivo antinociceptive effect was evaluated on the basis of acetic acid-induced writhes according to described methods [17]. The experiment was performed after approval by the local Institutional Ethics Committee (CEUA Process No. 17/2015 and 76/2015). Swiss adult male mice weighing between 18–23 g were used in the experiment. All compounds (10a-i), including the references drugs RVT and dipyrone, were administered orally at 100 μmol/Kg using a suspension in 5% gum arabic prepared in saline (vehicle). One hour after the oral pre-treatment with the test compounds, the acetic acid solution (0.6%, 0.1 mL/10 g) was administered through the i.p. route After 10 min of injection, the number of writhes in each animal was registered for 20 min. The negative control group received the same volume of saline. The results were expressed as mean ±SEM of the percentage inhibition compared to the negative control, using eight animals per group. Statistical analysis was carried out using ANOVA, followed by Dunnett’s test at a significance level of p < 0.01.

Cellular viability and quantification of TNF-α levels in the macrophage supernatant

Experiments to determine the macrophage viability and measure the TNF-α levels in the supernatant were conducted according to the previously described method [51]. The experiment was performed after approval by the local Institutional Ethics Committee (CEUA Process No. 77/2015). The stimulated macrophages were collected after three days of administration of a solution of 3% thioglycollate in the peritoneal cavity of the male Swiss mice. For the viability assay, an amount of 100 μL containing 5 × 10⁶ cells/mL suspended in the RPMI–1640C medium was added to the 96-well microplates. After incubation for 1 h at 37 °C in an atmosphere containing 5% CO₂ (Forma Scientific, Marietta, OH), the non-adherent cells were removed, and the supernatants were discarded. For the adherent cells, stimulation with lipopolysaccharide (Escherichia coli 0111 B) (LPS) (10 mg/mL) was performed, and the plates were incubated at 37° C under 5% CO₂ atmosphere for 24 h. Afterward, an amount of 100 μL of a solution containing 3-(4,5-dimethylthiazol-2-yl)–2,5-diphenyltetrazolium bromide (MTT) (1 mg/mL) was added to each well, and they were incubated at the same conditions for additional three hours. The content was discarded, and 100 μL of isopropyl alcohol was added to each well. The absorbance was measured at 540 nm using a microplate reader spectrophotometer (BioTek^®). The results conducted in triplicate after three independent experiments were expressed as the percentage of viable macrophage cells. The TNF-α and IL-1β levels were characterized by ELISA in the culture supernatants of the 96-well microplates at those concentrations in which the cellular viability was more than 75%. All quantifications were performed using the BD OptEIA ELISA kit (BD Biosciences, San Diego, CA) according to the manufacturer’s instructions. Thalidomide (50 μM) and RVT (12.5–50 μM) were used as reference drugs. The results conducted in triplicate after three independent experiments were expressed as the means ±SEM of the percentage of inhibition compared to the negative control (vehicle). One-way ANOVA with Tukey’s Multiple Comparison Test as a post hoc test was performed using the software GraphPad InStat version 3.00 for Windows 95 (GraphPad Software, San Diego, California, U.S.A.). The statistical significance was defined as a p-value less than 0.01.

2.6. Induction of gamma-globin using CD34⁺ cells

The procedure was previously approved by the local Institutional Ethics Committee (FCM UNICAMP CAAE: 45878215.8.0000.5404). The quantification of gamma-globin (γ^A+γ^G) chains induced by compound 10a and RVT was performed using the CD34⁺ cells [52, 53]. The separation of the mononuclear cells from the peripheral blood samples was performed by centrifugation using a Ficoll-Hypaque gradient. After immunomagnetic separation, the CD34⁺ cells (conditions: 5% CO₂ at 37°C) were cultured for 13 days in the Iscoves’s Modified Dulbecco’s Medium [IMDM; GIBCO™, Invitrogen Corporation, USA]. This medium was prepared with the following components: sodium bicarbonate [Merck, Germany], L-glutamine [GIBCO™, Invitrogen Corporation, USA], alpha-thioglycerol [Sigma Aldrich, St. Louis, MO] penicillin/streptomycin [Sigma Aldrich], fungizone [Sigma Aldrich], and supplemented with bovine serum albumin [BSA] [US Biological, Swampscott, MA], 7.5% sodium bicarbonate [Sigma Aldrich], a solution of liposomes [cholesterol, oleic acid, and phosphatidylcholine dipalmitoyl— Sigma Aldrich], apo-transferrin [Sigma Aldrich], and fetal bovine serum [FBS; 10% – days 0 through 7 and 30% – from 7 to 13 days; GIBCO™, Invitrogen Corporation, USA]). Moreover, some cytokines responsible for contributing toward erythroid differentiation were used as described: 50 ng/mL SCF (R&D Systems, USA), 5 ng/mL interleukin-3 (IL-3; R&D Systems), and 2 units/mL of erythropoietin (Epo; Eprex3000, Vetter Pharma, Ravensburg, Germany). The viability and the cell numbers were determined by trypan blue staining. The compound 10a and RVT were added in the medium on day 9. All the cells were collected four days later (day 13). Approximately, 1 × 10⁷ cells were pelleted and suspended in the saline buffer. Sterile water was then added to the pellet, and it was homogenized and frozen. After one hour, the tubes were centrifuged, and the supernatant was stored at–80 °C for analysis of the globin chains through HPLC (HPLC Alliance 2695–Waters^®).

The flow cytometry analysis was performed to monitor the erythroid differentiation. The following cell surface-specific antibodies were used in the experiment: anti-transferrin receptor (FITC-conjugated) (CD 71), antiglycophorin A (PE-conjugated), and anti-fetal hemoglobin (FITC-conjugated).⁵² These antibodies were commercially acquired from Caltag Laboratories (Burlingame, CA). In the flow cytometry experiments, the cellular content was 1 × 10⁵ cells in a final volume of 300 μL in phosphate-buffered saline. The data were quantified using a FACS Calibur instrument (Becton-Dickinson, USA), and 10,000 events were acquired for analysis using the BD CellQuest Pro Software (Becton Dickinson^®).

Gramicidin-Based Fluorescence Assay.

In order to determine the ability of compounds (10a-i) and RVT to perturb lipid bilayer properties, we used a gramicidin-based fluorescence assay using methods previously described previously [18,46,53]. In brief, large unilamellar vesicles (LUVs), loaded with intravesicular ANTS were prepared from DC_22:1PC and gramicidin (weight ratio 1000:1, corresponding to a ~2000:1 molar ratio) using freeze-drying, extrusion and size-exclusion chromatography; the final lipid concentration was 4–5 mM; the suspension was stored in the dark at 12.5 °C for a maximum of 7 days. The size distribution was determined using dynamic light scattering using an Anton Paar Litesizer TM 500 instrument; the average diameter was 133 nm, with an average Polydispersity index of 7.6 % indicating that the samples are monodisperse. Before use, the LUV-ANTS stock was diluted to 200–250 μM lipid with NaNO₃ buffer (140 mM NaNO₃, 10 mM HEPES, pH 7).

The time course of fluorescence quenching was measured using a SX.20 Stopped-Flow Spectrometer (Applied Photophysics, Leatherhead, UK). The excitation wavelength was at 352 nm and the fluorescence emission above 455 nm was recorded a high-pass filter. The fluorescence was normalized to the initial buffer value. The instrument has a dead time of < 2 ms, and the next 2–100 ms segment of each fluorescence quench trace was fitted to a stretched exponential, which is a computationally efficient way to represent a sum of exponentials with a distribution of time constants, reflecting the LUV size distribution as well the fluctuations in the number of gramicidin channels in the LUV membranes:

$$F(t)=F(\infty )+(F(0)-F(\infty ))\cdot \exp \left\{-{\left(t/{\tau }_0\right)}^{\beta }\right\}$$

where F(t) denotes the fluorescence intensity at time, t, τ₀ is a parameter with units of time, and β (0 < β ≤ 1, where b = 1 denotes a homogenous sample) provides a measure of the LUV dispersity. The quench rate (Rate), the rate of Tl⁺ influx, was determined at 2 ms:

$$Rate={\frac{\text{β}}{{\tau }_0}\cdot {\left(\frac{t}{{\tau }_0}\right)}^{\text{β}-1}|}_{2\text{ms}}$$

The Rate for each experiment represents the average quench rates of the trials with Tl⁺. The average Rate was normalized to the rate in control experiments (Rate_cntrl) without modulator. The reported values are averages from three or more experiments.

Statistical analysis was carried out using ANOVA followed by Dunnett’s test at a significance level of p < 0.05.

Evaluation of Mutagenicity Using a Micronucleus Test in the Peripheral Blood Cells of Mice [47].

The experiment was performed after approval by the local Institutional Ethics Committee (CEUA Process No. 08/2017). In this assay, the control and test-compound groups, each contained ten Swiss male mice weighing between 25−30 g. RVT and compound 10a were given orally to the mice at the following concentrations: 25, 50, and 100 mg/Kg body weight. Cyclophosphamide, used as the positive control, was given through the i.p. route at 50 mg/Kg body weight. A suspension containing 1% carboxymethylcellulose (CMC), 0.2% Tween, and plain filtered water was given orally and considered as the negative controls. After 30 h of administration of the compounds and the controls, the mice were sacrificed, and their blood was collected and added to pre-stained laminas containing acridine orange. An amount of 1000 reticulocytes/animal/lamina was counted, and the frequency of the micronucleated cells was recorded. The average frequencies of the cell micronuclei were determined for each group and expressed as means ±SEM. Statistical analysis was performed using ANOVA followed by Tukey’s method at a significance level of p < 0.05.

Figure 1.

Design of the hybrid RVT derivatives.

ABBREVIATION

DBU

1,8-diazabicyclo[5.4.0]undec-7-ene

DIP

Dipyrone

DMAP

4-dimethylaminopyridine

DNS

isosorbide dinitrate

FDA

U.S. Food and Drug Administration

HbF

Fetal hemoglobin

HbS

Sickle hemoglobin

HPFH

hereditary persistence of HbF

HU

hydroxyurea

IL-1β

interleukin-1 beta

LPS

Lipopolysaccharide

NGF

Nerve Growth Factor

NO

Nitric oxide

RBC

Red blood cells

RVT

resveratrol

SCD

Sickle Cell Disease

sGC

Soluble guanylyl cyclase

TNF-α

Tumor Necrosis Factor alpha