Synthesis and evaluation of novel, selective, functionalized γ-butyrolactones as sigma-2 ligands

The sigma-2 (σ2) receptor was discovered nearly 40 years ago and was recently identified as the Transmembrane Protein 97 (TMEM97, also known as MAC30 (Meningioma-associated protein)). Aberrant σ2 activity has been linked to diseases and conditions such as schizophrenia, Alzheimer’s disease, neuropathic pain, traumatic brain injury, and cancer. The utility of σ2 as a therapeutic target is currently under investigation in numerous laboratories. Herein, we report on the synthesis and evaluation of a series of novel, functionalized γ-butyrolactones that are potent σ2 receptor ligands.


Introduction
The discovery and characterization of the sigma receptors began in 1976 with W. R. Martin et. al.'s exploration of the impact of opioids on chronic spinal dogs. In these studies, they observed that the opioids morphine (1), ketocyclazocine, (2), and (rac)-SKF-100047 (3) (Fig. 1) produced different responses and hypothesized that each compound was interacting with a different receptor. They designated these receptors the µ-opioid receptor (morphine type, MOR), the κ-opioid receptor (ketocyclazocine type, KOR), and the σopioid receptor (SKF-100047 like) [1]. Follow-up studies conducted in the early 1980s using the individual enantiomers of SKF-100047 (3) demonstrated that the two enantiomers elicited physiological responses through different biochemical pathways. The opioid-mediated physiological response observed with (-)-SKF-100047 was determined to be the result of interactions with MOR and KOR. In addition, these studies revealed that (+)-SKF-100047 interacts with a previously unknown, non-opioid receptor that was designated the sigma receptor (σR) [2,3]. In 1993, W. D. Bowen et. al. determined that there were two sub-types of this receptor, which were designated sigma-1 (σ 1 ) and sigma-2 (σ 2 ) [4]. Three years later, Glossman H, et.al. cloned and expressed the mammalian σ 1 receptor in yeast cells [5], and in 2016 a crystal structure of the human σ 1 receptor was reported [6]. To date, there is no known natural ligand for this receptor.
The nature and function of the σ 2 receptor, on the other hand, remains the subject of intense research, but some progress has been made. The natural ligand of this receptor remains a mystery, but A.C. Krusea et. al. have demonstrated that the protein originally described as the σ 2 receptor is identical to Transmembrane Protein 97 (TMEM97, also known as MAC30 (Meningioma-associated protein)) [7]. The σ 2 receptor is present in lysosomes and the endoplasmic reticulum (ER) and there is evidence that cholesterol binds to this receptor [8]. Regulation of the Niemann-Pick protein NPC1 has also been suggested by H. Runz et al. [9]. Although the pharmacological role of σ 2 remains unclear, substantial effort has been devoted to the development of σ 2 binders based on the premise that aberrant σ 2 pharmacology contributes to the progression of diseases and conditions such as Alzheimer's disease [10][11][12], traumatic brain injury [13], neuropathic pain [14], schizophrenia [15], and cancer [16,17].
We recently reported a series of novel, selective γbutyrolactones sigma-2 ligands that included the identification of (4, Fig. 2). This compound was found to have moderate affinity for σ 2 (K i = 142 nM), excellent selectivity for this target over σ 1 (K i = 10,000 nM), and high stability in the presence of both mouse and human liver microsomes (MLM, HLM T 1/2 = 60 min) [18]. Herein we report follow up studies that describe the synthesis and characterization of a related series of novel γ-butyrolactones in which we explore (1) the impact of altering the length of the linker between the two ring systems, and (2) replacements of the aryl piperazine moiety with alternative ring systems.

Results and discussion
Synthesis of substituted γ-butyrolactones was conducted as shown in Scheme 1 utilizing novel methods developed in our laboratory. The synthesis of these compounds begins with the known alkenyl alcohols (5a, 5b), which were protected as the benzyl ethers using benzyl bromide, NaH, and tetrabutylammonium iodide in THF. The alkenes were then converted to the corresponding epoxide (5c, 5d) with mCPBA. Ring opening of epoxide (6) with the enolate of 2-ethyl-N,N-dimethylbutanamide via deprotonation with LDA provided the intermediate alcohol, which cyclized to form the γ-butyrolactone ring (7) in the presence of trifluoroacetic acid. Removal of the benzyl protecting group via hydrogenation in the presence of palladium on carbon provided the corresponding alcohol, which was then reacted with tosyl chloride in the presence triethyl amine to provide (8). Reaction of (8) with amines in refluxing THF provided the final target molecules (9). Alternatively, the previously reported γ-butyrolactone alcohol (10) was reacted with tosyl chloride in the presence triethyl amine, followed by reaction with amines in refluxing THF to provide the final target molecules (9).
Tables 1 and 2 describe the in vitro binding (K i at σ 1 and σ 2 ), physicochemical properties (MW, TPSA, LogP, solubility), and mouse liver microsomal (MLM) stability. All of the compounds are consistent with Lipinski's rule of 5 (MW, cLogP, TPSA) and have acceptable water solubility. In addition, TPSA and cLogP of the compounds are in a range that is indicative of BBB penetration. While the majority of compounds have low MLM stability, we were able to identify 3 compounds with MLM T 1/2 values > 10 min. Stability in MLM is an important factor, as future in vivo studies will be performed in rodents.
The structure activity relationship analysis of this series of compounds begins with an examination of the impact of length of the linker chain between the two rings systems (9a-9c). As indicated in Table 1, compounds with chain lengths of 2 (9a), 3 (9b), and 4 (9c) methylene units bind to σ 2 with moderate to high potency (σ 2 K i = 82, 7.7, and 12 nM), but selectivity versus σ 1 was low (σ 1 K I = 138.31, and 5.5 nM). Decreasing the size of the dialkyl side chains of the γ-butyrolactone (9d) lead to a nearly 10-fold decrease in σ 2 potency (K i = 753 nM) relative to (9a).
We next examined the impact of changes to the aryl piperazine region. Replacing the phenyl piperazine of (9a) with the corresponding 1-naphthyl piperazine (9e) led to a moderate increase in σ 2 potency (K i = 32 nM), as well as increase in selectivity over (σ 1 K I = 2167 nM) in comparison to (9a). Notably, this compound is the least soluble analog (sol = 47 µM), which is almost certainly the result of increased aromatic character of the aryl piperazine region. Employing heteroaromatic replacements for the aryl piperazine produced mixed results. While the 4-pyridine analog (4) is a moderate affinity σ 2 ligand (K i = 142 nM), with a high degree selectivity for this target over σ 1 (K i = 10,000 nM), the corresponding  (4) 4-pyrimidine analog (9f) had limited capacity to bind to σ 2 (KI = 10,000 nM) and low affinity for σ 1 (K i = 1017 nM).

Conclusions
In summary, a series of substituted lactones with drug-like physicochemical properties (MW, TPSA, cLogP) have been investigated as potential selective σ 2 ligands. We have determined that increasing the length of the linker chain from two (9a) to four carbons (9c) leads to increase s 2 potency, but selectivity over s 1 decreases. In addition, we have demonstrated that s 2 potency and selectivity for s 2 over s 1 Is maintained when the phenyl ring of the aryl piperazine is replaced with 1-napthylene (9e) or 4-pyridine (4), but replacement with a 4-pyrimidine (9f) leads to a significant lose of s 2 potency. Replacement of the piperazine ring with bioisosteres such as homopiperazine (9g) 2,6-diazaspiro [3.3]heptane (9h), and octahydropyrrolo [3,4-c]pyrrole (9i) was tolerated with respect to s 2 potency, but s 1 selectivity was substantially decreased. Incorporation of tetrahydroisoquinolines (9k-9n) in place of the aryl piperazine also produced high potency s 2 binders, but naphthyridine analogs examined to date had limited s 2 binding capacity. We anticipate these studies will help us further evaluate the potential value of this series for the identification of novel therapeutic agents for the treatment of diseases associated with abnormal σ 2 activity. Future studies will be focused on the identification of highly potent, selective, novel s 2 binders that have improved MLM stability.

Experimental methods and materials
Reagents were purchased from Fisher Scientific, VWR International, Sigma Aldrich, and Combi-Blocks, Inc. Chromatographic purification of compounds (normal phase and reverse phase) were carried out on a Teledyne Isco
Sigma-2 receptor binding assay K i values for test compounds for the sigma-2 receptor were determined using a filtration assay in a 96 well polypropylene plate using membranes prepared from HEK293T cells stably transfected with the sigma-1 receptor or PC12 cells. The membranes were prepared from cultured cells rinsed with PBS, lysed in cold 50 mM Tris-HCL (pH 7.4), centrifuged at 20000 × g, pellets resuspended in buffer and then stored at −80 C until used. In a final volume of 250 uL of assay buffer (50 mM Tris-HCl, 10 mM MgCl 2 , 1 mM EDTA, pH 7.4) the membranes were incubated with 5-7 nM [ 3 H]-1,3-di-(2-tolyl)guanidine ([ 3 H]-DTG, K d = 9.9 nM) and test compound (11 concentrations) at room temperature for 90 min. Nonspecific binding was defined with 10 uM haloperidol. Membranes were then collected by rapid filtration on to filter mats pretreated with 0.3% polyethyleneimine, washed 4x with cold assay buffer, dried, microscintillant added and then counted in a Microbeta scintillation counter. IC 50 values were determined using a three-parameter non-linear curve fitting program in Prism 4.0 (GraphPad Software). K i values were calculated from the IC 50 values using the Cheng-Prusoff equation [17]. The reference standard haloperidol had a K i = 13.9 nM.
Sigma-1 receptor binding assay K i values for test compounds for the sigma-1 receptor were determined using the sigma-2 method except that membrane from HEK-293 cells stably transfected with the sigma-1 receptor or PC12 cells were used and 2-10 nM [ 3 H]-Pentazocine (K d = 6.5 nM) was the radioligand. Nonspecific binding was defined with 10 uM haloperidol. The reference standard haloperidol had a K i = 3.54 nM.

Aqueous solubility (pH 7.4) assay
Compounds were assessed for their solubility at pH 7.4 using the commercially available Millipore MultiScreenTM Solubility filter system (Millipore, Billerica, MA). Analysis was performed by liquid chromatography tandem mass spectrometry (LC/MS/MS).

Microsomal stability assays
Test compounds were assessed for microsomal stability by incubating them at 37°C in the presence of mouse or human liver microsomes and an NADPH regenerating system as described by Yang et al. [20]. Microsomal protein content was adjusted to give accurate rates of substrate consumption. Analysis was performed by Liquid Chromatographytandem mass spectrometry (LC/MS/MS) analysis.