Design and synthesis of fused tetrahydroisoquinoline-iminoimidazolines.

In the aim of identifying new privileged structures, we describe the 5-steps synthesis of cyclic guanidine compounds "tetrahydroisoquinoline-iminoimidazolines" derived from tetrahydroisoquinoline-hydantoin core. In order to evaluate this new minimal structure and the impact of replacing a carbonyle by a guanidine moiety, their affinity towards adenosine receptor A2A was evaluated and compared to those of tetrahydroisoquinoline-hydantoin compounds.


Introduction
In the field of medicinal chemistry, privileged structures are considered as a promising source of ligands able to interact with various targets. [1][2][3] In fact, these minimal structures constitute a powerful starting point to identify original compounds by playing with the nature of their substituents. They constitute an anchor point enabling to orientate substituents in various positions, giving access to potentially promising structures and to modulate their affinity and activity. Therefore, identification of novel privileged structures appears as an interesting challenge.
Our group described a series of tetrahydroisoquinoline-hydantoin (Tic-H 1, Figure 1) derived compounds with potent affinity for the sigma-1 receptor in the nanomolar range ensuing various therapeutical in vivo applications. [4][5][6][7][8][9] This heterocycle presents an interesting hydrogen bond acceptor group on the hydantoin cycle and is amenable to various decorations. In this study we proposed to identify a new potential privileged structure and decided to evaluate the consequence of replacing the hydantoin's (1) hydrogen bond acceptor by a hydrogen bond donor group resulting in the guanidine moiety of compound 2 ( Figure 1). This new heterocycle could be the central core for the design of new active compounds. Guanidines are present in a large variety of natural products with potent biological activities in many fields [10][11][12][13] and more specifically in the central nervous system area. 14,15 This approach was exemplified by preparing the guanidine derivative "Tic-guanidine" and its derivatives (2 to 13) in order to evaluate the affinity of this new series on the adenosine receptor A2AR. 16 In fact, in addition to its affinity towards sigma-1 receptor, Tic-H 1 showed a promising affinity constant towards adenosine receptor A2A (Ki = 44 µM). This receptor is widely expressed in the central nervous system. Expressed at different levels (neurons, astrocytes, microglial cells), [17][18][19][20][21] it acts at various levels of regulation. [22][23][24][25] Therefore, A2ARs are viewed as promising targets in various neurodegenerative diseases, mainly Parkinson's and Alzheimer's diseases. [26][27][28] Docking studies, based on the crystal structure of A2A bound to the high affinity antagonist (ZM241385), 29 showed that Tic-guanidine restored hydrogen bonds that were missing for Tic-H leading us to expect a better affinity of these compounds. 30 We therefore aim to develop a new series of Tic-guanidine compound. We set up an original and efficient chemical synthesis allowing pharmacomodulations. These latter concern the nature of the substituents on the original tricyclic structure Tic-guanidine (series A, figure 1) but also modifications by opening of the central cyclic core (series B, figure 1).  Supported by our docking studies, pharmacomodulations were envisaged: various Nsubstituents on the guanidine core (series A, figure 1) and opening of the central tricyclic core (series B, figure 1). Series A enables to evaluate the effect of the modification of the benzyl group in the hydrophobic upper pocket of the binding site. On the other hand, series B was prepared in order to evaluate the importance of the central core's nature and especially how its geometry impacts compounds' affinity. We wanted to establish whether a planar conformation was essential as described in many A2A's antagonists ref or if less restricted structures could improve affinity as compared to already published non planar A2A antagonists ref. Finally, B series would bring additional informations on structure activity relationships of our Tic-guanidine compounds.

Chemical synthesis
Synthesis of Tic-guanidine derivatives of series A and B was achieved thanks to a unique chemical pathway as depicted in Scheme 1. This enabled us to access desired compounds starting from commercially available amino acids and therefore to easily access various pharmacomodulations. We optimized the synthetic pathway of compound 2 starting from commercially available Boc-protected L-tetrahydroisoquinoline carboxylic acid I whose chemistry is well mastered in our group. 7,9,31 Key step of this synthesis was the final formation of the cyclic guanidine. Various methods are described in the literature to access guanidines 12 and cyclic guanidines 32-34 but could not be applied to our strategy. Indeed, we previously described the synthesis of Tic-thiohydantoins, which could be an interesting intermediate for the synthesis of these Tic-guanidines. But their chemical and enantiomeric instability precluded their use in this study. 35 For final formation of guanidine cycle, direct cyclisation of the free diamine in the presence of cyanogen bromide [36][37][38][39][40][41][42] was unsuccessful.
Some adjustments were then required and we finally chose to maintain Boc-protection in order to selectively functionalize the free amine of compounds 2b-13b with cyanogene bromide. 43 Expected guanidines were obtained after Boc-deprotection and subsequent cyclization of intermediate.

Synthesis of series
A started from Boc-protected L-tetrahydroisoquinoline carboxylic acid I.
Corresponding aldehyde was obtained via reduction of Weinreb amide II 44,45 using LiAlH4 in THF at 0°C. 44 Reductive amination with appropriate amine in the presence of sodium triacetoxyborohydride in CH2Cl2 gave expected Boc-monoprotected derivatives 2b-6b. 46 Free amine of this last compound was then functionalized using cyanogen bromide in ethanol to give corresponding nitrile derivatives. 43    Some of our compounds were able to crystallize. X-ray spectroscopy thus enabled us to confirm the structure and enantiopurity of compounds 2 and 3 (figure 5) but also compounds 7, 9 and 11 whose absolute configuration was maintained. 47 On the other hand, crystallographic data showed compound 8 was present as a racemic mixture (for details, see supplementary informations). This compound differs from the other ones, as it is N-methylated. N-methylation was achieved following a protocol described in the literature to be a non-epimerizing route. 48 However, in our case, the basic conditions of methylation led to a complete racemization.   Concerning the tricyclic structures (2 to 6), we therefore assume the replacement of carbonyle moiety of 1 by a hydrogen bond donor does not give us expected improved affinity for the A2AR.

Cytotoxicity of Tic-guanidine compounds
Cytotoxicity assays have been conducted on SY5Y cells and showed no toxicity of our compounds at 100 µM (Table 1).

Conclusion
As Tic-H core was of interest for the design of various biologically active compounds, we proposed the Tic-guanidine core as a new privileged structure. This work therefore presented a new and efficient synthesis of guanidine cycles derived from amino acids. We applied this concept to the design of A2AR ligands. Unfortunately, binding results established that our initial hypothesis was not confirmed: replacement of the hydrogen bond acceptor moiety carbonyl of compound 1 by the hydrogen bond donor guanidine did not improve the affinity for A2AR. Other decorations are needed to improve the affinity for A2AR. Of particular importance is the lack of cytotoxicity of this new scaffold. Thus valorization of this new scaffold for other receptors is currently under evaluation.

General information
Chemicals and solvents were obtained from commercial sources, and used without further purification unless otherwise noted. Reactions were monitored by TLC performed on Macherey-Nagel Alugram® Sil 60/UV254 sheets (thickness 0.2mm). Purification of products was carried out by either column chromatography or thick layer chromatography. Column chromatography was carried out on using Macherey-Nagel silica gel (230-400 mesh). Thick layer chromatography was performed on glass plates coated with Macherey-Nagel Sil/UV254 (thickness 2 mm), from which the pure compounds were extracted with the following solvent
Cells were seeded at 2,000 cells per well onto 96-well plates in DMEM medium. Cells were starved for 24 hours to obtain synchronous cultures, and were then incubated in culture medium that contained various concentrations of test compounds, each dissolved in less than 0.1% DMSO. After 72 hours of incubation, cell growth was estimated by the colorimetric MTT (thiazolyl blue tetrazolium bromide) assay.