Exploring the functionalisation of the thieno[2,3-d]pyrimidinedione core: Late stage access to highly substituted 5-carboxamide-6-aryl scaffolds

The thieno[2,3-d]pyrimidinedione core is found as a component in a range of pharmaceutically active compounds, however, synthetic approaches to these scaffolds rely on access to functionalised, highly substituted thiophenes. Here we describe a new approach for the preparation of 5-carboxamide-6-aryl analogues that involves a two-step synthesis of the thieno[2,3-d]pyrimidinedione core from a readily available mercaptouracil derivative. Thio-alkylation with ethyl 3-bromopyruvate, followed by cyclisation and dehydration mediated by polyphosphoric acid allowed the scalable synthesis of the thieno[2,3-d] pyrimidinedione unit. The late-stage functionalisation of this core motif via bromination of the thiophene ring and a subsequent Suzuki-Miyaura reaction as the key steps permitted access to a novel library of 5-carboxamide-6-aryl analogues. The physicochemical properties of these compounds were determined, generating an insight into the potential bioavailability of these scaffolds. Based on these results, a selection of the novel 5-carboxamide-6-aryl analogues were tested as lactate uptake inhibitors of monocarboxylate transporters 1, 2 and 4 in Xenopus oocytes. © 2018 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Due to their similarity in structure to nucleic acid bases, pyrrolo [2,3-d]pyrimidine and thieno [2,3-d]pyrimidine scaffolds display a wide range of biological activities [1]. Within this structural class, the thieno [2,3-d]pyrimidinediones have been of particular interest, with pharmaceutical activity against a wide range of disease states. For example, thieno [2,3-d]pyrimidinediones bearing a N-piperazinylethyl moiety such as 1 (Fig. 1) are potent oral antihypertensive agents that can reduce systolic blood pressure [2]. In a programme to discover new inhibitors of the monocarboxylate transporter 1, AstraZeneca showed that thieno [2,3-d]pyrimidinediones with 5-carboxamide-6-alkyl substituents had potent immunomodulatory activity [3]. More recently, compounds from this series (e.g. AZD3965 2) have been shown to kill tumour cells reliant on glycolysis [4]. Other examples include thieno [2,3-d] pyrimidinediones with a 6-(p-methoxyureidophenyl) substituent, such as relugolix 3 [5]. These compounds are selective antagonists of the gonadotropin-releasing hormone receptor and are in phase 3 clinical trials for the treatment of endometrosis and prostrate cancer [6,7].
Despite the importance of thieno [2,3-d]pyrimidinediones, there are relatively few synthetic approaches that allow both the efficient preparation of the bicyclic core and further functionalisation of either ring. Common approaches to access thieno [2,3-d]pyrimidinediones include the reaction of 2-aminothiophenes with isocyanates [2,5] or the alkylation of mercaptouracil derivatives with a-halocarbonyl compounds, followed by cyclisation in the presence of Lewis acids such as titanium tetrachloride [3c, 8,9].
Building on our research programme that seeks to discover novel biologically active polycyclic scaffolds [10], we were interested in developing a scalable synthesis of a thieno [2,3-d]pyrimidinedione core that could then be further functionalised allowing late stage access to a diverse series of highly substituted analogues. We now report a three-stage synthetic approach for the rapid preparation of a small library of novel thieno [2,3-d]pyrimidinedione-5-carboxamide-6-aryl analogues. We have determined various physicochemical properties of these compounds, highlighting the potential of this scaffold for medicinal chemistry applications. The ability of these novel thieno [2,3-d]pyrimidinedione-5-carboxamide-6-aryl analogues to inhibit lactate uptake of monocarboxylate transporters 1, 2 and 4 in Xenopus oocytes is also presented.

Results and discussion
As outlined in Scheme 1, a three-stage approach was proposed for the preparation of the target thieno [2,3-d]pyrimidinedione-5carboxamide-6-aryl analogues. The first-stage involved the development of a reproducible, scalable synthesis of thieno [2,3-d]pyrimidinedione 5 from commercially available 6-chloro-3methyluracil (4). The aim was then to investigate the reactivity of this scaffold for the introduction of aryl groups at the 6-position. Finally, as thieno [2,3-d]pyrimidinediones bearing a 5-carboxamide moiety are often biologically active [1b,3], strategies for rapid conversion of the thiophene-carboxylate ester to various carboxamide groups were then explored.
To investigate various thiophene-forming reactions, 6mercaptopyrimidinedione 7 was chosen as the key substrate. This was prepared in two steps from 6-chloro-3-methyluracil (4), by Nalkylation with isobutyl bromide under standard base-mediated conditions, followed by an efficient substitution reaction with sodium hydrosulfide (Scheme 2).
Previous reactions using mercaptopyrimidinedione 7 for the preparation of analogues of thieno [2,3-d]pyrimidinedione 5 have utilised a two-stage process involving S-alkylation with 2chlorocarbonyl compounds followed by cyclisation and dehydration with titanium tetrachloride [3a,c]. However, these approaches generally result in low overall yields (18e37%) [3a,c]. We attempted a similar transformation with 7 and ethyl 3-bromopyruvate. Although the S-alkylated intermediate 8 was formed cleanly (by 1 H NMR spectroscopy), cyclisation and dehydration with titanium tetrachloride gave only a 24% overall yield of thieno [2,3-d]pyrimidinedione 5 ( Table 1, entry 1). In addition, it was difficult to reproduce these results using this two-stage process. As a consequence of the poor yields and reproducibility of using titanium tetrachloride to form the thiophene ring, other methods were investigated. Thieno [2,3-d]pyrimidinediones have been prepared efficiently from mercaptopyrimidinedione 7, by deprotonation with sodium acetate, followed by reaction with a-halocarbonyl compounds [11]. However, application of this method with 7 and ethyl 3-bromopyruvate gave only a 10% yield of 5, as well as a number of side-products (entry 2) [12].
Ogura and co-workers demonstrated that mercaptopyrimidinediones could undergo S-alkylation with simple a-halocarbonyl compounds such as bromoacetone under neutral conditions [9]. Following isolation and purification of the S-alkylated mercaptopyrimidinediones, subsequent cyclisation with polyphosphoric acid (PPA) gave the corresponding thieno [2,3-d] pyrimidinediones in good yields. Using this approach as a starting point, conditions for the reaction of mercaptopyrimidinedione 7 with the highly electrophilic ethyl 3-bromopyruvate to form esterderived thieno [2,3-d]pyrimidinedione 5 were investigated. Reaction of 7 with ethyl 3-bromopyruvate under neutral conditions did form intermediate 8 cleanly (Table 1, entry 3). However, unlike the alkyl and aryl derived intermediates from the Ogura study [12], Salkylated mercaptopyrimidinedione 8 could not be isolated and purified due to significant decomposition. Therefore, intermediate 8 was directly converted to ester-derived thieno [2,3-d]pyrimidinedione 5 using the PPA-mediated cyclisation reaction. Initially, the cyclisation step was found to proceed at 100 C, generating thieno [2,3-d]pyrimidinedione 5 in modest yield (entry 3). Optimisation of this step included increasing the reaction temperature, resulting in shorter reaction times. At an optimal temperature of 145 C, this gave 5 in 54% overall yield from the one-pot, two step procedure (entry 5) [13]. More importantly, this approach was found to be readily reproducible and could be used for the multigram synthesis of 5. The next stage of this research programme involved aryl substitution of the 6-position of thieno [2,3-d]pyrimidinedione 5. A two-step strategy was proposed involving bromination of the thiophene ring, followed by a Suzuki-Miyaura reaction (Scheme 3) [14]. Bromination of 5 under standard conditions with N- Fig. 1. Pharmaceutically active thieno [2,3-d]pyrimidinediones. Scheme 1. Proposed approach to novel thieno [2,3-d]pyrimidinedione-5-carboxamide-6-aryl analogues from 6-chloro-3-methyluracil (4).
bromosuccinimide (NBS), in the presence of acetic acid gave the corresponding bromide 9 in 84% yield. Suzuki-Miyaura reaction of 9 was then performed with various boronic acids using Pd(PPh 3 ) 4 as the catalyst. To explore the electronic and steric limitations of this process, both electron-rich and electron-deficient phenylboronic acids bearing either ortho-or para-substituents were investigated. Despite the variations, consistently high yields were observed for all four analogues formed from this reaction.
Having demonstrated efficient functionalisation of the C-6 position of the thieno [2,3-d]pyrimidinedione core through incorporation of various aryl groups, the final stage required preparation of the C-5 carboxamide. Initial studies began with the synthesis of morpholine carboxamide 14 (Scheme 4). Rapid hydrolysis of ethyl ester 10 was achieved using sodium hydroxide in ethanol. Coupling of the resulting carboxylic acid with morpholine was then attempted using various standard coupling reagents (EDCI, HBTU and EDCI/HOBt), however, all of these reactions gave low yields (9e32%) of 14. It was proposed that the low yields were due to the steric hindrance associated with the highly substituted thiophene ring and the subsequent slow reaction with the bulky coupling agents. It was believed that this could be overcome by using a smaller and more reactive acid chloride intermediate. Therefore, the carboxylic acid was converted to the acid chloride under mild conditions using oxalyl chloride and DMF. Without purification, this was treated with morpholine, resulting in the isolation of carboxamide 14 in 48% yield over the three steps (Scheme 4). Following the development of a straightforward approach to access the thieno [2,3-d]pyrimidinedione-5-carboxamide-6-aryl scaffold, the scope of this three step transformation was explored with structurally different amines, for the formation of carboxamides bearing cyclic and acyclic groups. As Weinreb amides are commonly biologically active due to an ability to hydrogen bond to biological targets [3], a series of these were also formed. Overall, all Table 1 Optimisation of the synthesis of thieno [2,3-d]pyrimidinedione 5.
Scheme 4. Preparation of the thieno [2,3-d]pyrimidinedione 5-carboxamide-6-aryl library. Isolated yields of 14e25 over the three steps are shown. a Hünig's base was also used for the amidation step.
four Suzuki-Miyaura products 10e13 were easily converted via the three step sequence to the corresponding carboxamides. Yields were generally good for preparation of the morpholine carboxamides 14e17 and the Weinreb amides 22e25. Lower yields were observed over the three steps for the diethyl carboxamide analogues 18e21. In comparison to the other amines, this is likely due to the less rigid and less reactive nature of diethylamine.
With the successful synthesis of a library of novel thieno [2,3-d] pyrimidinedione-5-carboxamides, we were interested in evaluating how the incorporation of aryl groups at the C-6 position might affect the physicochemical properties. For compounds that might find application in binding to neurological receptors by penetrating the blood brain barrier, permeability across the plasma membrane is important. Similarly, for compounds that transport across cell membranes through passive diffusion, evaluation of the membrane partition coefficient is crucial. Therefore, the partition coefficient (log P), permeability (P m ), the membrane partition coefficient (K m ) and the percentage of plasma protein binding (%PPB) of all twelve thieno [2,3-d]pyrimidinedione-5-carboxamides were evaluated using established HPLC methods ( Table 2). Previous work by Tavares et al., of ten biologically active compounds, established the limits of each of these parameters (log P < 4, P m < 0.5, K m < 250, %PPB < 95%) [15]. Based on these criteria, the physicochemical properties of the majority of these compounds were found to be excellent and well within the acceptable limits. Despite the incorporation of an aryl moiety and the resulting increase in lipophilicity, these compounds possess properties that should allow effective transport through cell membranes. Due to the relatively lipophilic diethyl carboxamide group, compounds 18 and 19 were found to have the highest plasma protein binding (%PPB). While these values are above the acceptable limit, there are medicinally important compounds with similar plasma protein binding that still demonstrate good bioavailability [13]. Therefore, all of these structural classes are of interest for further development.
As highlighted above, thieno [2,3-d]pyrimidinediones have significant biological activity, particularly against monocarboxylate transporters (MCT), with certain compounds showing potent immunomodulatory activity or the ability to kill tumour cells that rely on glycolysis [3,4]. For these reasons, a selection of nine of the thieno [2,3-d]pyrimidinediones prepared in this current study were tested as inhibitors of lactate uptake of MCT1, MCT2 or MCT4 in Xenopus oocytes (Fig. 2) [16]. As well as control (blank) experiments, AR-C155858 (AR-C in Fig. 2), a commercially available inhibitor of MCTs was also tested as a standard [3c]. While none of the compounds showed any reduction in lactate uptake against MCT1, morpholine carboxamide 15 showed significant activity against MCT2 (p 0.01). For several of the compounds, a tendency to inhibit MCT2 was also observed, however the reduction in transport activity did not significantly differ from the control cells. The morpholine carboxamides also showed a slight inhibitory effect on MCT4, with compounds 14 and 16 showing~10% reduction in uptake. Although these compounds are not significantly active in inhibiting lactate uptake (each compound was tested at 1 mM), this study has demonstrated the structure activity relationship of this series and identified thieno [2,3-d]pyrimidinedione morpholine carboxamide derivatives as potential scaffolds for further development as lactate uptake inhibitors of MCTs.  The lactate uptake for each compound shown is the mean ± SEM of eight independent experiments (n ¼ 8). The asterisks refer to the values of control cells (white bars). *p 0.05, ***p 0.001.

Conclusions
In summary, a flexible and concise approach for the preparation of novel thieno [2,3-d]pyrimidinedione scaffolds has been developed. In particular, a reliable and scalable route for the synthesis of a thieno [2,3-d]pyrimidinedione core bearing a C-5 ester moiety has been achieved from a mercaptouracil derivative by an alkylation reaction with ethyl 3-bromopyruvate followed by an acid-mediated cyclisation. The product of this process has served as a key intermediate to explore the introduction of aryl groups at the C-6 position via bromination and a Suzuki-Miyaura reaction. Despite the highly substituted nature of the thiophene ring, transformation of the C-6 substituted esters to the corresponding amides was readily achieved via the formation of acid chlorides under mild conditions. Determination of several important physicochemical properties of this small library of thieno [2,3-d]pyrimidinedione-5-carboxamide-6-aryl analogues showed that despite the introduction of aryl units, the majority of these compounds still retain the potential ability to transport through cell membranes. Finally, a selection of compounds prepared in this study was tested for the ability to inhibit lactate uptake against various monocarboxylate transporters, with the morpholine carboxamide analogues showing most activity. Work is currently underway to further develop this class of thieno [2,3-d]pyrimidinediones for a range of biological applications, as well as explore further synthetic applications of this functionally rich bicyclic core.

General methods
All reactions were performed under an atmosphere of air unless otherwise stated. All reagents and starting materials were obtained from commercial sources unless otherwise stated. Brine refers to a saturated solution of sodium chloride. All dry solvents were purified using a PureSolv 500 MD solvent purification system. Flash column chromatography was carried out using Merck Feduran Si 60 (40e63 mm). Macherey-Nagel aluminium-backed plates pre-coated with silica gel 60 were used for thin layer chromatography and were visualised under a UV lamp. 1 H NMR and 13 C NMR spectra were recorded on a Bruker DPX 400 spectrometer or Bruker 500 spectrometer with chemical shift values in ppm relative to trimethylsilane or residual chloroform as standard. J values are reported in Hz. The assignment of 1 H NMR spectra is based on COSY and HSQC experiments and 13 C NMR spectra is based on DEPT experiments. Infrared spectra were recorded using a Shimadzu FTIR-84005 spectrometer directly as either a solid or liquid and mass spectra were obtained using a JEOL JMS-700 spectrometer or a Bruker MicroTOFq high-resolution mass spectrometer. Melting points were determined on a Gallenkamp melting point apparatus. All physicochemical analyses were performed using a Dionex Ultimate 3000 series, and data acquisition and processing performed using Chromeleon 6.8 Chromatography software [14]. Standard and test compounds were dissolved in 1:1 organic/aqueous phases, and prepared to a concentration of 0.5 mg/mL. The HPLC system was set to 25 C, and UV detection achieved using a diode array detector (190e800 nm). Analysis was performed using 5 mL sample injections.