The synthesis and anti-inflammatory evaluation of 1,2,3-triazole linked isoflavone benzodiazepine hybrids

Copper catalyzed azide-alkyne cycloaddition was used for the first time to access a small series of eight novel 1,2,3-triazole linked isoflavone benzodiazepine hybrids. As part of this work, a previously unreported alkyne substituted pyrrolo[1,4]benzodiazepine was synthesized using a Sonogashira coupling reaction. Two previously unreported azide substituted isoflavones were also synthesized using TMS-azide as a key reagent. The eight new 1,2,3-triazole linked products


Introduction
Isoflavones such as daidzein 1 and formononetin 2 (see Figure 1) are naturally occurring phytoestrogens that have long been associated with a wide range of biological activities, 1,2 including anti-inflammatory 3 and neuroprotective properties. 4This has led to significant interest in the synthesis of this class of compounds and their investigation as potential leads for the development of treatments for neurodegenerative disorders such as Alzheimer's disease (AD). 4,5,6The neuroprotective properties of formononetin and their pharmacological origins have been evaluated. 5Daidzein has been shown in animal studies to induce improvements in cognitive function by correcting oxidative stress in neuronal cells in the brain, 6 and its hydroxylated metabolite, 6,7,4´trihydroxyisoflavone, has been shown to improve learning and memory via activation of the cholinergic system. 7It has been clearly established that neuroinflammation is a valid target in AD, and that the microglial cells produced in AD mediate neuroinflammation. 8,9The synthetic isoflavone derivative 3, shown in Figure 1, has been established to be an anti-neuroinflammatory compound with neuroprotective properties, and has shown the ability to improve cognitive ability in an animal model, making it a promising lead compound for the treatment of AD. 10 As part of an ongoing project that is looking at the synthesis and biological evaluation of novel antiinflammatory and neuroprotective small-ring heterocycles 11 and isoflavones, 12 and due to an interest in the chemistry of benzodiazepines, 13 we have been exploring the synthesis and anti-inflammatory properties of isoflavone benzodiazepine hybrid drugs, and report our results herein.Hybrid drugs seek to covalently link two pharmacophores in one molecule, and aim to produce improved pharmacodynamic and therapeutic profiles. 14ur choice of a benzodiazepine as the partner to an isoflavone is supported by reports that chronic benzodiazepine users have a lower brain amyloid load where a reduction in neuroinflammation has been identified as the potential pathway. 15Significantly, there are also reports that the imidazobenzodiazepines 4, 16,17 imidazobenzodiazepinones 5 18,19 and their pyrrolo-fused analogues 6 20 (Figure 1) offer the potential to aid learning and memory, and may limit the underlying causes of cognitive decline and age associated hyperactivity 19 associated with AD.
We favoured the use of a 1,2,3-triazole as the covalent link between the benzodiazepine and isoflavone moieties because of our familiarity with the use and synthesis of azides, 21 because of their common use as a linker using very well-known click-chemistry protocols, 14 and also because of the potential that they have to deliver interesting biological activity in their own right.The synthesis and use of hybrid drug candidates containing 1,2,3-triazoles has been recently and extensively reviewed, 22 and the synthesis of such compounds also features strongly in a recent review of potential drug candidates for Alzheimer's disease. 23We are aware that the antimicrobial activity of novel synthetic isoflavone 1,2,3-triazole hybrids has been evaluated, 24 that 1,2,3-triazole linked pyrrolobenzodiazepine dimers 25 and chalcone hybrids 26 have been synthesized and studied as DNA minor groove inhibitors 25,26 with significant anticancer potential, as have flavone pyrrolobenzodiazepine hybrids joined by an ether link, 27 and that hexahydrodibenzodiazepine 1,2,3-triazole hybrids have been reported to be butyrylcholinesterase inhibitors of interest in AD. 28 However, we are aware of no reports that have investigated benzodiazepines attached to isoflavones via a 1,2,3-triazole link.In this report, we focus on the synthetic methodology that was used to construct these target molecules.We also report some preliminary biological screening data, the full details of which will be reported elsewhere.

Results and Discussion
As discussed in the Introduction, we wished to attach isoflavones to benzodiazepines using an azide-alkyne cycloaddition.The previously unreported azides 10 and 11 were obtained as shown in Scheme 1.Thus, resorcinol and 4-nitrophenylacetic acid were reacted together in the presence of BF 3 .Et 2 O to give a deoxybenzoin intermediate 7 which was formylated and cyclized in the presence of DMF and methanesulfonyl chloride using a modified literature method 29 to give the known 29,30 isoflavone 8. Reduction to the corresponding previously reported 29,30 amine 9 was followed by diazotization and azidation with TMS-azide to give the previously unknown azide 10 which was reacted readily with acetic anhydride in pyridine to give the unknown acylated analogue 11.

Scheme 1. The Synthesis of Azido Isoflavones.
As reaction partners for the two isoflavone based azides 10 and 11, we required the alkyne substituted imidazobenzodiazepines 5 and 6 (R 2 = C≡CH), the syntheses of which are shown in Scheme 2. The imidazofused systems were selected due to the association of these compounds with cognitive enhancement and their potential usefulness 19 in treating the symptoms of AD, as discussed in the Introduction.Thus, isatoic anhydride 12 was reacted with the amino acids 13 and 14 to give the known benzodiazepinediones 15 31 and 16 32 which were brominated regioselectively to give the previously reported intermediates 17 33 and 18. 34 Condensation of 17 and 18 with ethyl isocyanoacetate 20 in the presence of chlorophosphate 19 proceeded without incident to furnish the known imidazobenzodiazepines 21 33,35 and 22. 35 The bromobenzodiazepines 17, 18, 21 and 22 were then subjected to Sonogashira couplings with the (trialkylsilyl)acetylenes 23 and 24, yielding the (trialkylsilyl)ethynyl-substituted benzodiazepines 25 -31.TBAF mediated deprotection produced the desired ethynyl-substituted benzodiazepines 32 -35.Compounds 26 -30 and 35 are previously unreported, whereas compounds 25, 36 31, 18 32, 37 33 19,36 and 34 18 have been reported.We chose to work with the known imidazo-fused systems 33 and 34 as reaction partners for the reasons described above.Compounds 32 and 35 were synthesized due to their ease of access from bromo-precursors 17 and 18, and to provide useful comparators for later biological testing.It is also of note that, as far as we are aware, the alkynyl pyrrolobenzodiazepine 35 together with its precursor trialkylsilyl alkynes 29 and 30 have not been reported before.Similarly, trialkylsilyl alkynes 27 and 28 are not known in the literature, although the alkyne 32 has been reported in the patent literature, where it was accessed via an alkyne substituted derivative of isatoic anhydride 12. 37 With the requisite alkynes 32 -35 and azides 10 and 11 in hand, eight new 1,2,3-triazoles 36 -43 were produced by reaction in the presence of a Cu(II)/ascorbate system, as shown in Scheme 3 and Table 1.The copper catalyzed azide-alkyne cycloadditions did not occur after reaction for two days at room temperature, showed no reaction after 24 hours at 50 °C, and showed only a small amount of product after 24 hours at 80 °C.It was found that 100 °C overnight (16 hours) was required for complete reactions to occur.No reaction was observed in the absence of copper, with toluene at reflux either returning the starting materials or, on prolonged heating, causing loss of the azide starting material by degradation.
The new triazole compounds were produced in sufficient purity (>95% by 1 H NMR) and amounts (10s of milligrams were produced, but only μLs of a 10 mM solution were used) for our biological testing requirements, which are described below.However, whilst all of the products were sufficiently soluble in DMSO to allow biological testing, the acetylated triazoles 40 -43 were not sufficiently soluble to allow 13 C NMR spectra to be recorded, and only just soluble enough to allow 1 H NMR data collection (see the Experimental Section and Supplementary Material).Nonetheless, for compounds 40 -43, the 1 H NMR data showed that the Cu(II)/ascorbate mediated cycloaddition reaction had clearly succeeded, a conclusion that was supported by the infra-red and HRMS data.All other compounds were fully characterized.The 1,2,3-triazoles 36 -43, together with some key precursors and the original natural products 1 and 2, were evaluated in vitro on LPS-stimulated BV2 mouse brain microglial cells as a preliminary screen for © AUTHOR(S) potential anti-inflammatory compounds, and the results are summarized in Table 2. Cell viability was performed using XTT assay 38 in order to determine if the compounds were toxic to the BV2 cells.BV2 mouse brain microglial cells are commonly used in place of primary microglia, 39 and have a key role in neuroinflammation. 40When microglia are activated, for example, with LPS (lipopolysaccharide), proinflammatory mediators such as nitric oxide are produced. 41Activated microglia have been found next to amyloid β plaques in Alzheimer's disease, 42 and are a recognized target for reducing inflammation and neurodegeneration.Nitric oxide production was measured using the Griess Assay. 43NO is produced in small amounts by normal cells (Table 2 -entries 1 and 2: 20% NO production), but when the cells are LPS-activated, NO production increases significantly (Table 2, entry 3: 100% NO production).Table 2 shows that two of the novel hybrids (39 and 43), both derived from the imidazobenzodiazepine 33, showed reasonable cell viability (> 60%) and significant inhibition of NO production (< 20%), when compared to the known anti-inflammatory natural products daidzein 1 and formononetin 2, and so were selected for further pharmacological evaluation.The alkynyl benzodiazepines 32 -35 were not cytotoxic but did not inhibit LPS-induced NO production when compared to daidzein 1 and formononetin 2. The previously unreported azide 10 also showed good inhibition of NO production when compared to daidzein 1 and formononetin 2, and so was selected for further study.We will report the results and full details of these ongoing pharmacological studies elsewhere, as it is beyond the remit of this journal.

Conclusions
A series of 8 novel 1,2,3-triazole linked benzodiazepine substituted isoflavones was produced in 56 -91% yield using a copper catalyzed azide-alkyne cycloaddition reaction between two previously unreported azidoisoflavones and four alkynyl-1,4-benzodiazepines.The alkynes were produced in good to excellent overall yields using Sonogashira couplings.The cell viability and inhibition of NO production in LPS-stimulated BV2 cells were determined for the novel triazole products and their precursors, and the results were compared to the parent natural products daidzein and formononetin.Two of the new 1,2,3-triazole linked benzodiazepine substituted isoflavones, both derived from the same imidazobenzodiazepine, and one of the azido-isoflavones, showed useful levels of activity (inhibition of NO production in LPS-stimulated BV2 cells greater than daidzein and much greater than formononetin) and are candidates for further pharmacological evaluation, the results of which will appear elsewhere.

Figure 1 .
Figure 1.Isoflavones and Benzodiazepines of Interest in AD Research.

Table 2 .
In vitro cell viability and NO production for Triazoles 36 -43 and Selected Precursors

Table 2 .
Continued a For cell viability, the % values are reported relative to negative control cells DMSO (-DMSO, 100%) that contained the amount of DMSO used to add the compound solutions in the cell medium.b For NO production, the % values are reported relative to LPS-stimulated BV2 cells (+, 100%).c Negative control, -: BV2 cells were incubated for 24 h only with serum free medium RPMI.d LPS-stimulated cells, + : BV2 cells were incubated for 24 h with 100 ng/mL LPS in the medium.e Compounds: BV2 cells were incubated with the compounds at a concentration of 20 µM (final concentration in well) and stimulated with 100 ng/mL LPS.Values are expressed as mean ± SEM (%) of minimum three experiments; SEM is standard error of mean.MW: molecular weight.

9, 29,30 16, 44 18, 34 22, 35 31, 18 34, 18 15, 31 17, 23 32, 37 21, 35 25 36
36d 3336are known and were synthesized as described in the Supplementary Information using the relevant literature method.The compounds reported below are all previously unreported. Rections were magnetically stirred, heated on a paraffin oil bath, and monitored on Merck TLC silica gel 60 F 254 aluminium sheets.Visualization of spots was accomplished using a UV lamp (254 or 365 nm) and/or staining with potassium permanganate solution.Column chromatography was performed on silica gel (Aldrich, technical grade, pore size 60 Å, 40-63 μm particle size) using the solvent mixtures indicated (volume to volume ratios).Unless otherwise indicated, reactions were conducted under an atmosphere of dry nitrogen.Agilent 6210 TOF MS (Dual ESI source), Agilent 6530 Q-TOF MS (Jet Stream ESI source), Agilent 1290 HPLC + 6530 Q-TOF (Dual AJSESI source +ve) or Agilent 7890A-5975C (EI-GCMS) and spectra were recorded in positive mode.FT-IR spectra were recorded on a Thermo Nicolet 380 FT-IR Spectrometer with Diamond ATR (neat sample).Melting points were recorded using a Stuart SMP10 melting point apparatus.