Attempts to Access a Series of Pyrazoles Lead to New Hydrazones with Antifungal Potential against Candida species including Azole-Resistant Strains

The treatment of benzylidenemalononitriles with phenylhydrazines in refluxing ethanol did not provide pyrazole derivatives, but instead furnished hydrazones. The structure of hydrazones was secured by X-ray analysis. The chemical proof was also obtained by direct reaction of 3,4,5-trimethoxybenzaldehyde with 2,4-dichlorophenylhydrazine. Newly synthesized hydrazones were tested against eight Candida spp. strains in a dose response assay to determine the minimum inhibitory concentration (MIC99). Five compounds were identified as promising antifungal agents against Candida spp. (C. albicans SC5314, C. glabrata, C. tropicalis, C. parapsilosis and C. glabrata (R azoles)), with MIC99 values ranging from 16 to 32 µg/mL and selective antifungal activity over cytotoxicity.


Introduction
Antifungal therapies evolved slowly during the early years of the twentieth century, with the development of antifungal agents lagging that of antibacterial agents. The current therapeutic arsenal for the systemic treatment of antifungal infections mainly includes polyenes, azoles, echinocandins and pyrimidine classes of compounds.
In 2020, the World Health Organization (WHO) set up the "WHO antifungal expert group on identifying priority fungal pathogens" posing a high risk to human health [1]. The azole-resistant fungal pathogens Candida spp. were identified as preoccupying pathogens of global public health importance. The main Candida species resistant to azoles drugs is Candida krusei (intrinsic resistance), while those with an acquired resistance are certain strains of Candida glabrata and rare strains of Candida tropicalis, Candida parapsilosis and Candida kefyr. The lack of therapeutic innovations or new chemical families under development for the discovery of therapeutic alternatives has reached a critical level.
Identifying new experimental drugs is challenging and expected more than ever. There are only a few different classes from a mechanistic point of view that are currently used to treat serious fungal infections.
The treatment of invasive candidiasis, which represent the most serious clinical forms in hospitals, is the subject of several international recommendations, in particular those published by the Infectious Diseases Society of America (IDSA) [2] and the European Society

Results and Discussion
In the frame of an ongoing medicinal chemistry program, an attempt to access new pyrazole derivatives of general structures 5 and 6, starting from benzylidenemalononitriles 3a-c, has been tried. The latter have been easily obtained by the Knoevenagel reaction of aldehydes 2a-c with malononitrile in the presence of piperidine in refluxing eth-

Results and Discussion
In the frame of an ongoing medicinal chemistry program, an attempt to access new pyrazole derivatives of general structures 5 and 6, starting from benzylidenemalononitriles 3a-c, has been tried. The latter have been easily obtained by the Knoevenagel reaction of aldehydes 2a-c with malononitrile in the presence of piperidine in refluxing ethanol (Schemes 1 and 2). They were next reacted with substituted phenylhydrazines 4a-n in ethanol at reflux. The expected pyrazolines 5a-o and/or 6a-o could not be obtained. The same operatory conditions were previously described to afford aminopyrazole derivatives [11,12]. However, in our case, no trace of the target heterocycle was detected in the crude by 1 H-NMR and TLC monitoring. Instead, in all cases, a different product was detected and isolated. The structure of the final product was resolved as a hydrazone derivative and secured by performing an X-ray on compounds 1e and 1i of the series ( Figure 2). In addition, a chemical proof was also obtained by reacting aldehyde 2a with hydrazine 4e, which provided the same hydrazone 1e. This confirmed the loss of the malononitrile unit during the reaction of benzylidenemalononitriles 3a-c with phenylhydrazines 4a-n, explaining undoubtedly the formation of hydrazones 1a-o (Scheme 3).

Results and Discussion
In the frame of an ongoing medicinal chemistry program, an attempt to access new pyrazole derivatives of general structures 5 and 6, starting from benzylidenemalononitriles 3a-c, has been tried. The latter have been easily obtained by the Knoevenagel reaction of aldehydes 2a-c with malononitrile in the presence of piperidine in refluxing ethanol (Schemes 1 and 2). They were next reacted with substituted phenylhydrazines 4a-n in ethanol at reflux. The expected pyrazolines 5a-o and/or 6a-o could not be obtained. The same operatory conditions were previously described to afford aminopyrazole derivatives [11,12]. However, in our case, no trace of the target heterocycle was detected in the crude by 1 H-NMR and TLC monitoring. Instead, in all cases, a different product was detected and isolated. The structure of the final product was resolved as a hydrazone derivative and secured by performing an X-ray on compounds 1e and 1i of the series ( Figure  2). In addition, a chemical proof was also obtained by reacting aldehyde 2a with hydrazine 4e, which provided the same hydrazone 1e. This confirmed the loss of the malononitrile unit during the reaction of benzylidenemalononitriles 3a-c with phenylhydrazines 4a-n, explaining undoubtedly the formation of hydrazones 1a-o (Scheme 3 The structure of compounds 1e and 1i was demonstrated by single crystal X-ray diffraction method. According to X-ray crystallography, the two compounds are isostructural. They crystallize in the P21/c space group with close unit cell parameters (Table S1, see Supplementary Materials for full data). The asymmetric part of the unit cell ( Figure  2a,b) comprises one molecule of 1e and 1i as a crystallographic independent unit, respectively. There are no co-crystallized solvent molecules in both crystals. The structure of compounds 1e and 1i was demonstrated by single crystal X-ray diffraction method. According to X-ray crystallography, the two compounds are isostructural. They crystallize in the P21/c space group with close unit cell parameters (Table S1, see Supplementary Materials for full data). The asymmetric part of the unit cell ( Figure  2a,b) comprises one molecule of 1e and 1i as a crystallographic independent unit, respectively. There are no co-crystallized solvent molecules in both crystals.
(a) (b) Figure 2. A view of the asymmetric part in the crystal structure of 1e (a) and 1i (b) with atom labeling and thermal ellipsoids at 50% level.
As expected, compounds 1e and 1i feature similar crystal structure packing. Indeed, for both crystals, the main crystal structure motif is described as a one-dimensional supramolecular array running along the b axis, which is formed through intermolecular NH···O and CH···O hydrogen bonding. As an example, a view of the one-dimensional architecture in the crystal structure of 1i is shown in Figure 3.

Scheme 3.
Proposed mechanism for the formation of hydrazones 1a-o from benzylidenemalononitriles 3a-c upon reaction with hydrazines 4a-n.
Primary antifungal screening study by whole cell growth inhibition assays, using all the synthesized hydrazones 1a-o at a single concentration of 32 µg/mL, was realized in triplicate (n = 3). Hit confirmation of active compounds by whole cell growth inhibition assays was conducted as an eight-point dose response to determine the Minimum Inhibitory Concentration (MIC), in triplicate (n = 3). The inhibition of growth was measured against eight fungi strains: C. albicans SC5314, C. dubliniensis, C. glabrata, C. parapsilosis, C. albicans from mucoviscidosis patients (C. albicans (mucoviscidosis)), C. albicans resistant to echinocandins (C. albicans (R echinocandins)) and C. glabrata resistant to azoles (C. glabrata (R azoles)) ( Table 1). The fungal strains were obtained from Pôle de Biologie Pathologie Scheme 3. Proposed mechanism for the formation of hydrazones 1a-o from benzylidenemalononitriles 3a-c upon reaction with hydrazines 4a-n.
The structure of compounds 1e and 1i was demonstrated by single crystal X-ray diffraction method. According to X-ray crystallography, the two compounds are isostructural. They crystallize in the P2 1 /c space group with close unit cell parameters (Table S1, see Supplementary Materials for full data). The asymmetric part of the unit cell (Figure 2a,b) comprises one molecule of 1e and 1i as a crystallographic independent unit, respectively. There are no co-crystallized solvent molecules in both crystals.
As expected, compounds 1e and 1i feature similar crystal structure packing. Indeed, for both crystals, the main crystal structure motif is described as a one-dimensional supramolecular array running along the b axis, which is formed through intermolecular NH···O and CH···O hydrogen bonding. As an example, a view of the one-dimensional architecture in the crystal structure of 1i is shown in Figure 3.
A mechanism has been proposed for the formation of hydrazones 1a-o from benzylidenemalononitriles 3a-c (Scheme 3). The first step involved the classical nucleophilic attack of the marginal nitrogen of the hydrazine on the ethylenic carbon of the 2-cyano-3-aryl-acrylonitrile. The intermediate formed underwent a proton 1,3 migration. This allowed the formation of malononitrile as a leaving group and the formation of hydrazones (Schemes 2 and 3).

Scheme 3.
Proposed mechanism for the formation of hydrazones 1a-o from benzylidenemalononitriles 3a-c upon reaction with hydrazines 4a-n.
Primary antifungal screening study by whole cell growth inhibition assays, using all the synthesized hydrazones 1a-o at a single concentration of 32 µg/mL, was realized in triplicate (n = 3). Hit confirmation of active compounds by whole cell growth inhibition assays was conducted as an eight-point dose response to determine the Minimum Inhibitory Concentration (MIC), in triplicate (n = 3). The inhibition of growth was measured against eight fungi strains: C. albicans SC5314, C. dubliniensis, C. glabrata, C. parapsilosis, C. albicans from mucoviscidosis patients (C. albicans (mucoviscidosis)), C. albicans resistant to echinocandins (C. albicans (R echinocandins)) and C. glabrata resistant to azoles (C. glabrata (R azoles)) ( Table 1). The fungal strains were obtained from Pôle de Biologie Pathologie Primary antifungal screening study by whole cell growth inhibition assays, using all the synthesized hydrazones 1a-o at a single concentration of 32 µg/mL, was realized in triplicate (n = 3). Hit confirmation of active compounds by whole cell growth inhibition assays was conducted as an eight-point dose response to determine the Minimum Inhibitory Concentration (MIC), in triplicate (n = 3). The inhibition of growth was measured against eight fungi strains: C. albicans SC5314, C. dubliniensis, C. glabrata, C. parapsilosis, C. albicans from mucoviscidosis patients (C. albicans (mucoviscidosis)), C. albicans resistant to echinocandins (C. albicans (R echinocandins)) and C. glabrata resistant to azoles (C. glabrata (R azoles)) ( Table 1). The fungal strains were obtained from Pôle de Biologie Pathologie Génétique, Centre Hospitalier Universitaire (CHU) de Lille, France. Fluconazole was used as the positive reference in the assay. Samples were prepared in DMSO and water to a final testing concentration of 32 µg/mL and in triplicate (n = 3), keeping the final DMSO concentration to a maximum of 1% DMSO. All the sample preparation was performed using liquid handling robots. Only five hydrazones, 1c, 1d, 1i, 1k and 1l, displayed notable antifungal activity against tested Candida spp., with MIC values between 16 and 32 µg/mL. The results are presented in Table 1. All other synthesized hydrazones, 1a, 1b, 1e-h, 1j and 1m-o, were less active, with MIC values > 32 µg/mL (data not shown). Active hydrazones 1c, 1d, 1i, 1k and 1l share the same 3,4,5-trimethoxyphenyl unit and have the particularity of a monosubstitution on the other phenyl ring. Generally, the ortho-substitution by an electro-withdrawing group (F, Cl and Br) in 1i, 1d and 1c, respectively, was the most favorable chemical modulation for the antifungal activity in the Molecules 2021, 26, 5861 6 of 14 current work. The comparison of halogens reveals that the chlorine atom in hydrazone 1d was the most active, especially on C. glabrata (MIC = 16 µg/mL) compared to fluoro and bromo congeners 1i and 1c (MIC = 32 µg/mL) ( Table 1). The trifluoromethyl substituent in hydrazone 1a resulted in dramatical loss of the antifungal potential (MIC > 32 µg/mL). The para-bromo substitution in hydrazone 1k slightly decreased the antifungal potential on C. glabrata (R azoles) (MIC > 32 µg/mL) but conserved the notable antifungal effect on C. glabrata (MIC = 16 µg/mL) and on C. tropicalis (MIC = 32 µg/mL) ( Table 1). To be noted, only para-substituted compounds 1k and 1l displayed inhibition activity against C. tropicalis (MIC = 32 µg/mL), while the ortho-substituted analogs were less active (MIC > 32 µg/mL) ( Table 1). The electron-donating substituents by inductive or mesomeric effect (Me or OMe) in the ortho position of the phenyl unit were not tolerated on any of the tested fungi (MIC > 32 µg/ mL for hydrazones 1b, 1g and 1m). The di-or polysubstitution of the same phenyl ring by both electron-withdrawing or electron-donating groups in hydrazones 1e, 1f, 1j and 1o were not favorable for antifungal activity (MIC > 32 µg/mL). Finally, the replacement of the 3,4,5-trimethoxyphenyl unit by a 4-nitrophenyl or 4-bromophenyl moiety in hydrazones 1m-o also abolished the antifungal activity against all tested Candida spp. (MIC > 32 µg/mL). The 3,4,5-trimethoxyphenyl unit seemed essential to maintain the biological activity on C. albicans SC5314, C. glabrata, C. parapsilosis and C. glabrata resistant to azoles (MIC values of 16 and 32 µg/mL, respectively). Diminished potential was registered on the clinical isolates of C. albicans (mucoviscidosis) and C. albicans resistant to echinocandins. Newly synthesized hydrazones generally showed more pronounced antifungal activity on strains of C. glabrata, including C. glabrata that is resistant to azoles (Table 1).
To verify the mammalian cytotoxicity of the newly identified antifungals presented herein, compounds 1c, 1d, 1i, 1k and 1l were tested against human embryonic kidney cells (HEK293) at ten different concentrations (0.06 to 32 µg/mL) (Figure 4). The highest concentration of 32 µg/mL tested corresponded to the concentration at which the compounds exhibit potent antifungal activity. Since compounds were dissolved in 0.1% DMSO in the stock solution for this assay, DMSO was used as a negative reference in the same test. As depicted in Figure 4, the concentration of 0.1% of DMSO is devoid of cytotoxic effect and is safe for compound solubilization, while the concentration of 20% of DMSO displayed high toxicity. This denotes the importance of the concentration of DMSO used to dissolve the experimental drugs so as not to have distorted effects due to the solvent. All tested compounds showed no toxicity in viable kidney HEK293 cells.
In addition, compounds 1c, 1d, 1i, 1k and 1l have also been selected and evaluated for cell growth inhibition activity on the NCI-60 cancer cell lines panel. Molecules were tested at 3.65 µg/mL (10 µM) concentration (compound 1c), 3.21 µg/mL (10 µM) concentration (compound 1d), 3.04 µg/mL (10 µM) concentration (compound 1i), 3.65 µg/mL (10 µM) concentration (compound 1i) and 3.00 µg/mL (10 µM) concentration (compound 1k) and did not show any notable cytotoxic effect. The full one-dose mean graphs for antifungal agents 1c, 1d, 1i, 1k and 1l are available in the supplementary data section associated with this article. exhibit potent antifungal activity. Since compounds were dissolved in 0.1% DMSO in the stock solution for this assay, DMSO was used as a negative reference in the same test. As depicted in Figure 4, the concentration of 0.1% of DMSO is devoid of cytotoxic effect and is safe for compound solubilization, while the concentration of 20% of DMSO displayed high toxicity. This denotes the importance of the concentration of DMSO used to dissolve the experimental drugs so as not to have distorted effects due to the solvent. All tested compounds showed no toxicity in viable kidney HEK293 cells. In addition, compounds 1c, 1d, 1i, 1k and 1l have also been selected and evaluated for cell growth inhibition activity on the NCI-60 cancer cell lines panel. Molecules were tested at 3.65 µg/mL (10 µM) concentration (compound 1c), 3.21 µg/mL (10 µM) concentration (compound 1d), 3.04 µg/mL (10 µM) concentration (compound 1i), 3.65 µg/mL (10 µM) concentration (compound 1i) and 3.00 µg/mL (10 µM) concentration (compound 1k) and did not show any notable cytotoxic effect. The full one-dose mean graphs for antifungal agents 1c, 1d, 1i, 1k and 1l are available in the supplementary data section associated with this article.

Materials and Methods
Starting materials are commercially available and were used without further purification (suppliers: Carlo Erba Reagents S.A.S., Val-de-Reuil, France; Thermo Fisher Scientific Inc., Illkirch, France; Tokyo Chemical Industry Co. Ltd., Zwijndrecht, Belgium; and Sigma-Aldrich Co., Saint-Quentin-Fallavier, France). Melting points were measured on an MPA 100 OptiMelt ® apparatus (Stanford Research Systems, Sunnyvale, CA, USA) and are uncorrected. Nuclear magnetic resonance (NMR) spectra were acquired at 400 MHz for 1 H-NMR, and at 100 MHz for 13 C-NMR on a Varian 400-MR spectrometer (Varian, Les

Materials and Methods
Starting materials are commercially available and were used without further purification (suppliers: Carlo Erba Reagents S.A.S., Val-de-Reuil, France; Thermo Fisher Scientific Inc., Illkirch, France; Tokyo Chemical Industry Co., Ltd., Zwijndrecht, Belgium; and Sigma-Aldrich Co., Saint-Quentin-Fallavier, France). Melting points were measured on an MPA 100 OptiMelt ® apparatus (Stanford Research Systems, Sunnyvale, CA, USA) and are uncorrected. Nuclear magnetic resonance (NMR) spectra were acquired at 400 MHz for 1 H-NMR, and at 100 MHz for 13 C-NMR on a Varian 400-MR spectrometer (Varian, Les Ulis, France) with tetramethylsilane (TMS) as internal standard, at room temperature (RT) or at 500 MHz for 1 H-NMR, and at 125 MHz for 13 C-NMR on a Bruker Avance III 500 MHz spectrometer (Bruker, Mannheim, Germany) with tetramethylsilane (TMS) as internal standard, at room temperature (RT). Chemical shifts (δ) are expressed in ppm relative to TMS. Splitting patterns are designed: s, singlet; d, doublet; dd, doublet of doublets; t, triplet; q, quadruplet; quint, quintuplet; m, multiplet; sym m, symmetric multiplet; br s, broaden singlet; br t, broaden triplet. Coupling constants (J) are reported in Hertz (Hz). The designation Ph1 in the 13 C-NMR spectra description corresponds to the phenyl ring coming from the starting hydrazine, and Ph2 is the phenyl ring linked to -CH=Ngroup. Thin layer chromatography (TLC) was realized on Macherey Nagel silica gel plates (Macherey Nagel, Hoerdt, France) with fluorescent indicator and were visualized under a UV-lamp at 254 nm and 365 nm. Elemental analyses (C, H, N) of new compounds were determined on a Thermo Electron apparatus (Thermo Fisher Scientific Inc., Illkirch, France) by "Pôle Chimie Moléculaire-Welience", Faculté des Sciences Mirande, Dijon, France.

General Procedure for the Synthesis of Benzylidenemalononitriles (3a-c)
A mixture of aldehyde (2a-c, 5.4-25.5 mmol, 1 equiv), malononitrile (5.4-25.5 mmol, 1 equiv) and 0.5 mL of piperidine in ethanol, was stirred at reflux for 6-8 h. After cooling the reaction medium to room temperature, the obtained precipitate was filtered, washed with ethanol, and then recrystallized from ethanol to afford the pure expected product (3a-c).

X-ray Crystallography
X-ray diffraction measurements for 1e and 1i were carried out with a Rigaku Oxford Diffraction XCALIBUR E CCD diffractometer (Rigaku Europe SE, Frankfurt, Germany) equipped with graphite-monochromated MoKα radiation. The unit cell determination and data integration were carried out using the CrysAlis package of Oxford Diffraction (Rigaku Europe SE, Frankfurt, Germany) [18]. The structures were solved by intrinsic phasing using Olex2 [19] software with the SHELXT [20] structure solution program and refined by full-matrix least-squares on F 2 with SHELXL-2015 [20], using an anisotropic model for non-hydrogen atoms. All H atoms attached to carbon were introduced in idealized positions (d CH = 0.96 Å) using the riding model, with their isotropic displacement parameters fixed at 120% of their riding atom. The positions of H atoms for NH groups were determined from Fourier synthesis maps and verified through the hydrogen bonds parameters. Table S1 provides a summary of the crystallographic data together with refinement details for compounds. The geometric parameters are summarized in studies were conducted using a 48-h exposure protocol using the sulforhodamine B (SRB) assay [22].

Conclusions
An attempt to isolate pyrazoles derivatives, by reacting benzylidenemalononitriles with hydrazines in refluxing ethanol, did not provide the target heterocyclic systems, as expected and as previously reported, but instead hydrazones whose structure was secured by both chemical evidence and X-ray studies. Indeed, the direct reaction of 3,4,5trimethoxybenzaldehyde with 2,4-dichlorophenylhydrazine furnished the same hydrazone as that obtained from benzylidenemalononitrile with hydrazine. This study identified five hydrazones as promising antifungal agents (MIC 99 values ranging from 16 to 32 µg/mL) against Candida spp. These compounds showed a CC 50 (concentration at 50% cytotoxicity) value against HEK293 cells at >32 µg/mL and against NCI-60 cancer cell lines panel at >10 µM, demonstrating selective antifungal activity over cytotoxicity. Compared to known acylhydrazones and hydrazones previously reported in the literature, a selection of which is available in Figure 1, the newly synthesized hydrazones are positioned as very promising experimental molecules with antifungal activity on Candida spp.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. Supplementary Materials include X-Ray crystallography and NCI-60 data.