Pharmacological and SAR analysis of the LINS01 compounds at the human histamine H1, H2, and H3 receptors

Histamine is a transmitter that activates the four receptors H1R to H4R. The H3R is found in the nervous system as an autoreceptor and heteroreceptor, and controls the release of neurotransmitters, making it a potential drug target for neuropsychiatric conditions. We have previously reported that the 1‐(2,3‐dihydro‐1‐benzofuran‐2‐yl)methylpiperazines (LINS01 compounds) have the selectivity for the H3R over the H4R. Here, we describe their pharmacological properties at the human H1R and H2R in parallel with the H3R, thus providing a full analysis of these compounds as histamine receptor ligands through reporter gene assays. Eight of the nine LINS01 compounds inhibited H3R‐induced histamine responses, but no inhibition of H2R‐induced responses was seen. Three compounds were weakly able to inhibit H1R‐induced responses. No agonist responses were seen to any of the compounds at any receptor. SAR analysis shows that the N‐methyl group improves H3R affinity while the N‐phenyl group is detrimental. The methoxy derivative, LINS01009, had the highest affinity.


| INTRODUCTION
Histamine is a biogenic amine that has important clinical and pathophysiological roles in allergy, gastric acid secretion, and inflammatory diseases. Histamine effects occur by activation of histamine receptors of which there are four subtypes and all are G-protein-coupled receptors (GPCRs), namely H 1 R, H 2 R, H 3 R, and H 4 R. [1,2] The H 1 R is a G q -coupled receptor, and activation of this is important in allergy. There are many H 1 R antagonists in widespread clinical use for allergies such as hay fever (including oral agents, e.g., cetirizine, loratadine and chlorpheniramine, and topical preparations, e.g., mepyramine). The H 2 R is a G scoupled receptor. Activation is important in the production of gastric acid, and H 2 R antagonists are widely used to reduce gastric acid production (e.g., ranitidine). [1] The H 3 R is a G icoupled receptor and present in the CNS and on peripheral neurons. It is important for inhibiting the release of histamine (as an autoreceptor) but also negatively regulating the synthesis and release of many neurotransmitters. H 3 R antagonists have been investigated as potential treatments for several CNS disorders, including sleep disorders, schizophrenia, epilepsy, attention deficit, hyperactivity disorder, and even obesity. There is also potential for H 3 R antagonists to have effects in neurodegenerative disorders such as Alzheimer's disease, Parkinson's disease, and Huntington's chorea. [3][4][5][6] Recently, a H 3 R antagonist (pitolisant) has been approved for the maintenance of wakefulness in patients with narcolepsy. [7,8] The H 4 R is also a G i -coupled receptor and is most widely expressed in bone marrow and on white blood cells and shown to be an important target to treat inflammation, [2] although to date, no H 4 R antagonist is available for clinical use. There is a considerable overlap between the selectivity of ligands for the H 3 R and H 4 R, with many H 3 R ligands having high affinity for the H 4 R; however, some H 4 R selective ligands have been identified (e.g., JNJ-7777120). [1,2] We have recently developed a series of compounds, originally designed based on the H 4 R selective antagonist JNJ-7777120, the 1-(2,3-d ihydro-1-benzofuran-2-yl)methylpiperazine (LINS01) series ( Figure 1).
Initial studies examining [ 3 H]-histamine binding to H 3 R and H 4 R membrane preparations of the first four compounds suggested micromolar affinities for these compounds, with higher affinity for the H 3 R than H 4 R. [9] A further study of other LINS01 compounds in HEK293 cells transiently expressing the H 3 R or H 4 R suggested the LINS01 compounds are antagonists or weak inverse agonists but once again that they had a higher affinity for the H 3 R than the H 4 R. [10] Given the surprising finding that these LINS01 compounds had higher affinity for H 3 R than H 4 R, we decided to investigate the pharmacological profile of these ligands at the other histamine receptors, the H 1 R and H 2 R.
The LINS01 compounds (1a-1i) were synthesized as previously reported by our group (Figure 2, see Supporting Information). [9][10][11] In case of the nonsubstituted phenols (to obtain compounds 1a-d), the commercial 2-allylphenol was used as starting material. In case of R′-substituted molecules, the corresponding phenols 2e-h were used as starting material, which was allylated with allyl bromide in conventional or ultrasonic conditions, with good to excellent yields. The yields in conventional or ultrasonic methods are comparable; however, the ultrasonic methodology saved reaction time and thus was considered advantageous. The allyl-phenyl ethers 3e-h were then thermally isomerized (>200°C) to access the 2-allylphenols (4a-h) with good yield using a microwave-assisted Claisen rearrangement approach, using DMF as a solvent. The dihydrobenzofurans (5a-h) were obtained through iodine-promoted cyclization from the corresponding ortho-allylphenols in an eco-friendly medium, using water or a mixture of ethanol-water as solvent. Finally, the iodinated heterocycle was used to alkylate the N-substituted piperazines in aprotic solvent (THF), with potassium carbonate as base, with moderate yields. A novel microwave-assisted methodology was employed to avoid the considerable excess of 1-phenylpiperazine to obtain compounds 1d and 1i, leading to moderate to good yields (50%-70%). Although this method gave comparable yields to the conventional methodology, it also saved reaction time and required less 1-phenylpiperazine (1.1 eq.) indeed, then comprising the green chemistry principles. [12] The spectroscopic data for the final compounds (see Supporting Information) and intermediates are in accordance with the literature reports. [9][10][11] None of the LINS01 molecules were previously evaluated in H 1 R and H 2 R, and the compound 1i was not evaluated in H 3 R before. This is also the first report on the affinity data for these compounds through reporter gene assays.
Histamine stimulated a response that was 2.3 ± 0.1-fold over basal in the H 1 -SRE-luciferase cells (pEC 50 6.78 ± 0.04; n = 8, Figure 3a,b). This response was readily inhibited by mepyramine, a histamine H 1 R antagonist, causing a parallel rightward shift of the histamine concentration response and yielding a pK D value of 8.53 ± 0.05 (n = 9). ICI162846 (an H 2 R antagonist) and JNJ-7777120 (an H 4 R antagonist) were not able to antagonize the histamine response at concentrations of up to 10 μM. Clobenpropit (an H 3 R/H 4 R antagonist) caused a small rightward shift of the histamine response (pK D 5.01 ± 0.11; n = 9). Thus, histamine responses are readily antagonized by mepyramine, but not ICI182846, clobenpropit, or JNJ-7777120, confirming the presence of the H 1 R in this cell line. [13] F I G U R E 1 Structure of LINS01 compounds F I G U R E 2 Reagents and conditions.
(a) Phenol, allyl bromide (2 eq.), K 2 CO 3 (2 eq.), acetone, 60°C, 4 hr or conventional 12 hr; (b) 200°C, MW (300 W max., 300 psi max.), DMF, 1 hr; (c) I 2 (1.1 eq.), water, 4-8 hr; (d) piperazines (NH, N-Me, N-allyl, N-Ph, 2-4 eq.), K 2 CO 3 (1 eq.), THF, 80°C, 24 hr or 1-phenylpiperazine (1.1 eq.), MW (300 psi max, 300 W max), 120°C, 1.5 hr Histamine also stimulated a response (pEC 50 7.50 ± 0.03; n = 8) in the H 2 -CRE-SPAP cells that was 3.8 ± 0.1-fold over basal (Figure 3c,d). This response was inhibited by the H 2 R antagonist ICI162846 (pK D 8.71 ± 0.04; n = 9), whereas neither mepyramine nor clobenpropit nor JNJ-7777120 was able to cause a rightward shift of the histamine concentration-response curve. This confirms the presence of the H 2 R in this cell line. [13] Histamine did not stimulate a response, nor inhibit the basal response, on its own in the CHO-H 3 -CRE-SPAP cells, indicating the absence of a G s -coupled stimulatory histamine receptor. [14] Histamine did however inhibit forskolin-induced CRE-SPAP production (pEC 50 6.66 ± 0.04; n = 9) in a manner consistent with the stimulation of a G i -coupled histamine receptor (Figure 3e,f). This histamine G i -induced agonist response was inhibited by clobenpropit (pK D 9.18 ± 0.06; n = 9). As expected, neither mepyramine nor ICI162846 was able to inhibit the histamine responses in this cell line. JNJ-7777120 cause a small rightward shift of the histamine concentration-response curve (pK D 5.45 ± 0.07; n = 4), very similar to that previously reported for the human H 3 R (pK D 5.29) and very different from that for the histamine H 4 R (pK D 8.40). [15] Thus, the presence of the histamine H 3 R was confirmed in this cell line. [14] None of the LINS01 compound series stimulated any agonist responses in any of the cell lines ( Figure 4). With the exception of 1a, the LINS01 compounds were able to inhibit the H 3 R histamine-induced response to yield measurable pK D values. The compound 1g had the highest affinity (pK D 7.18) followed by 1h (pK D 6.99; Figure 4, Table 1). Increasing concentrations of 1 caused progressive rightward shifts of the histamine concentration-response curve in keeping with competitive antagonism at the H 3 R (Figure 4e). Three compounds did cause a small rightward shift of the histamine response in CHO-H 1 -SRE-luciferase cells giving measurable pK D values. Interestingly, 1a, that did not have any measurable affinity for the H 3 R, had a very weak affinity for the H 1 R. None of the LINS01 ligands had any measurable affinity for the H 2 R.
As the LINS01 molecules were originally designed to be similar to the H 4 R selective antagonist JNJ-7777120, [9] it was expected that these molecules would also be H 4 -selective. Indeed, two molecules (LINS01005-1d and LINS01007-1e) appear to have a dose-dependent anti-inflammatory effect in a murine mouse model of asthma that correlated with the H 4 R affinity. [9,10] However, these studies have also shown that most of LINS01 molecules actually have a higher affinity for the H 3 R than the H 4 R. [9,10] It was therefore important to know the pharmacological profile of these compounds for the human H 1 R and H 2 R.
The LINS01 compounds, together with mepyramine (selective H 1 R antagonist), ICI 162864 (selective H 2 R antagonist), clobenpropit (H 3 R/H 4 R antagonist), and JNJ-7777120 (H 4 R antagonist), were studied using the functional readout of reporter gene assays in CHO cell lines stably expressing each of the human H 1 R, H 2 R, and H 3 R. For the G q -coupled H 1 R, an SRE reporter was used, whereas for the G s -coupled H 2 R and G i -coupled H 3 R (both of which affect cAMP), a CRE reporter was used. [13,14] Reporter gene assays have several advantages. They can be used to examine responses from receptors without the need for large extracellular or intracellular tags, which have the potential to alter the ligand-receptor binding or receptor-effector coupling, respectively. The longer agonist incubation times (here, 5 hr at 37°C) means that the interaction between the ligand and the receptor is highly likely to reach equilibrium allowing the use of pharmacological analyses such as the rightward shift of agonist concentration responses (Gaddum equation) [13,14] to assess ligand affinity (often not possible with very short-term second messenger responses such as calcium). Thus, the efficacy of ligands (ability to stimulate a response) and the affinity of antagonist ligands can be assessed within one assay. Also unlike second messenger assays (e.g., calcium and cAMP), being downstream responses, reporter gene assays also read out from several different signaling cascades (e.g., cAMP and MAP kinase). [16] None of the LINS01 compounds stimulated any agonist responses at any of the histamine receptors studied. Although each of the selective antagonists inhibited histamine responses with high affinity at their respective receptors, no inhibition of histamine responses was seen with the LINS01 compounds in the H 2 R-expressing cells. There is therefore no evidence for any ligand-receptor interaction for the LINS01 compounds with the H 2 R. Three of the LINS compounds had a measurable affinity, albeit low, for the human H 1 R. Of these, 1a is most interesting in that this compound was the one compound from the LINS01 series that had no affinity for the H 3 R. The compound 1a therefore appears to have H 1 R selectivity (although the H 1 R affinity is very poor). All compounds with the exception of 1a antagonized histamine responses yielding measurable pK D values.
Looking into the SAR on the H 3 R affinity, the importance of the N-alkyl group (R) is clear. The unsubstituted compound (1a) had no affinity for the H 3 R although did for the H 1 R. Substitution with either N-methyl (1b) or N-allyl (1c) groups led to improved H 3 R affinity, suggesting there is an additional hydrophobic interaction of the R group in the binding site. The N-phenyl group (1d) also decreased H 3 R affinity, possibly due to steric hindrance or by decreasing the basicity of the nitrogen. Thus, the N-alkyl group is important for H 3 R affinity, and possible lack of it is important for H 1 R affinity.
Several H 3 R ligands reported in the literature present the N-arylpiperazine moiety such as DL-80 [17] which presents medium to low affinity to H 3 R, but with certain selectivity over H 4 R. However, these compounds present a longer linker than the LINS01 compounds, and thus, the position of the most basic nitrogen changes accordingly. Docking studies [18,19] suggest that a salt bridge between a basic nitrogen and Glu206 is a key interaction to the antagonist activity. Figure 5 shows the proposed interactions of several antagonists with H 3 R and the hypothetical interaction of 1. The alignment of UCL2190, DL-80, and 1g ( Figure 5) suggests that the nitrogen supposed to interact with Glu206 and the aromatic rings of these molecules can occupy the same positions in space, supporting this hypothesis.
We have previously shown that only compounds 1e and 1f showed any measurable affinity for the H 4 R, suggesting that the R′ group may play the role in the H 4 R affinity. Compound 1e also presented similar pK i for both receptors (~6.0). [10] The high homology between these two receptors explains the nonselectivity of several ligands from the literature. It is known that H 3 R and H 4 R share two highly conserved amino acids in the binding site, an aspartate (Asp114 and Asp94, respectively) and an glutamate (Glu206 and Glu182, respectively), which interact with the primary amine and imidazole, respectively, of histamine. This explains the higher affinity of histamine for these receptors than for H 1 R and H 2 R. [20] The interaction with Glu206 generally determines the affinity of H 3 R ligands and possibly the low affinity of these ligands to H 1 R and H 2 R, since Glu206 is not present in these receptors. By observing the present results, we believe that the more basic nitrogen of LINS01 compounds interacts with Glu206 by analogy. The results also suggest that some volume in the R′ region is tolerated by H 3 R. As can be noted in Table 1, compounds 1e-1h present a substituent in R′ and the binding affinities kept similar. The presence of the bulky t-butyl group in 1h maintained the binding affinity value ~7.0, corroborating this hypothesis. However, the more polar methoxy group led to better affinity (pK D 7.18), suggesting that polar atoms can improve the potency without increasing lipophilicity indeed. Molecules with excessive lipophilicity can lead to pharmacokinetic issues in the development stage and thus should be avoided at the initial stages. [21] Ligand efficiency metric analysis (see Supporting Information) is a suitable approach to evaluate the contribution of each atom of a given molecule to the binding affinity. [21,22] The concept of ligand efficiency (LE) arouses from the need to measure whether a potency of a ligand derives from an adequate fit with the target or simply by doing many nonspecific interactions with it. [23] The literature considers that a good LE value to a lead-like molecule is at least >0.3, while market drugs present an average LE value >0.45. [21,22] All LINS01 compounds presented LE >0.3 regarding H 3 R affinity and thus can be considered promising compounds for further development. Compounds 1b, 1f, and 1g must be highlighted due to their high LE values (>0.52), and with exception of the N-phenyl derivatives 1d and 1i, the LINS01 molecules present LE values comparable to drugs that are doing well in the market. The fit quality (FQ) is a size-independent measurement of LE, that is, it allows direct comparison of the values independently from the molecular size. [24] Again, the LINS01 compounds present very good values for FQ and are considered interesting lead compounds.
Derivations of LE concept considering the lipophilicity also appeared, such as the lipophilicity-corrected ligand efficiency (LELP) and the lipophilic ligand efficiency (LLE), in order to ponder the LE with the lipophilicity (mainly measured by clogP, see Supporting Information). [25,26] Excessive hydrophobicity should be avoided during the drug discovery process because it limits the water solubility (which directly impacts on pharmacokinetics) and favors nonspecific interactions with biological proteins. The optimum LELP values are considered in the range from −10 to +10, while LLE values should be between 5 and 7. [21] Once again, the N-phenyl derivatives 1d and 1i showed LELP values outside the desired values, indicating that the phenyl group led to excessive lipophilicity. This lipophilicity can be verified by the high clogP values for these molecules. Considering that the lipophilicity usually increases during the development, [25] these molecules are not promising for further modifications, or these modifications should consider more hydrophilic groups. The LLE value of 1g is within the optimal value, indicating that the insertion of methoxy group in R′ led to improved affinity and contributed to reduce the lipophilicity of the molecule at the same time, leading to a clogP value similar to that from the nonsubstituted molecule 1b.
Group efficiency (GE) is also a metric value to evaluate the contribution of an inserted group to the potency. [27] LINS01 molecules present a common dihydrobenzofuranylmethylpiperazine core with structural modifications in the basic nitrogen (R) and in the aromatic ring (R′). As can be noted in Tables 1 and 2, modifications in the R group lead to important variations in the binding affinities. The presence of N-methyl group led to important improvement in the affinity to H 3 R (GE = 2.29), indicating that this group should be maintained to good H 3 R affinity. The N-allyl group also increased the affinity, but as it presents three heavy atoms, its efficiency was lower than N-methyl (GE = 0.85). In counterpart, the Nphenyl group of 1d presented low GE value. The higher the GE value, the higher the contribution of the group to increase the potency. For example, the addition of a group with one heavy atom (HA = 1) and a GE value of 0.52 in a molecule with molecular weight <300 means an increase of 2.3-fold in   (8) <5.00 (9) ICI162846 <5.00 (9) 8.71 ± 0.04 (9) <5.00 (9) Clobenpropit 5.01 ± 0.11 (9) <5.00 (8) 9.18 ± 0.06 (9) JNJ-7777120 <5.00 (4) <5.00 (4) 5.45 ± 0.07 (4) Note. pK D values for ligands as determined in reporter assays from the rightward parallel shift of a histamine concentration-response curve. Values are mean ± SEM of n separate determinations.
potency. [21,27] This means that the N-methyl group contributes very importantly to the binding affinity, while the N-phenyl group neither, and thus, its insertion in the molecule do not justify. The contribution of the R′ group in the potency, as indicated by GE values, is less prominent. In the case of chlorine atom in 1e, negative GE value was observed, stating the negative contribution of this atom to the affinity. The best contribution was given by the methoxy group of 1g (GE = 0.35).
In summary, the dihydrobenzofuranyl-methylpiperazine (LINS01) molecules are selective H 3 R antagonists with no intrinsic activity at either the human histamine H 1 R, H 2 R, or H 3 R and no, or extremely poor, affinity for the H 1 R or H 2 R. Although the best ligands (1g and 1h) have greater than 100fold H 3 R selectivity (over H 1 R or H 2 R), their affinities for the human H 3 R (66 and 100 nM) clearly show potential scope for further improvement with different chemical modifications. As future directions, this chemical scaffold should be used to design novel derivatives in order to increase the affinity to H 3 R taking advantage on the selectivity profile of such compounds.