Cyclobutanes in Small‐Molecule Drug Candidates

Abstract

Cyclobutanes are increasingly used in medicinal chemistry in the search for relevant biological properties. Important characteristics of the cyclobutane ring include its unique puckered structure, longer C−C bond lengths, increased C−C π‐character and relative chemical inertness for a highly strained carbocycle. This review will focus on contributions of cyclobutane rings in drug candidates to arrive at favorable properties. Cyclobutanes have been employed for improving multiple factors such as preventing cis/trans‐isomerization by replacing alkenes, replacing larger cyclic systems, increasing metabolic stability, directing key pharmacophore groups, inducing conformational restriction, reducing planarity, as aryl isostere and filling hydrophobic pockets.

The 3D structure of cyclobutane opens opportunities for unique applications in medicinal chemistry. Improved synthetic methods and commercial availability now allow for more facile incorporation of this structural motif in small‐molecule drug candidates. Although its use has been increasing, it is still very much underused relative to other motifs in drug development. This review presents recent applications of the cyclobutane ring in small‐molecule drug development and the utility it can offer to medicinal chemists.

1. Introduction

1.1. Historical synopsis

Cyclobutane was first synthesized in 1907. ^[1] It is a colorless gas with no biological properties as such. Cyclobutane rings, although relatively rare, do however occur in natural products, most of which are found in plant and marine species. ^[2] The cyclobutane skeleton, present in sceptrins (e. g., 1, also isolated in acetylated form and as HCl salt) from Agelas sceptrum, contributes to its antimicrobial properties (Figure 1). ^[3]

Figure 1.

Cyclobutane 1, isolated from Agelas sceptrum sea‐sponge, thymine dimer 2 (R indicates the rest of the DNA helix), and carboplatin (3).

In addition, DNA bases can undergo [2+2] photodimerization upon UV irradiation and form cyclobutane pyrimidine dimers (2; Figure 1). Such DNA crosslinking can cause various adverse effects, often leading to skin cancer. ^[4] Introduction of cyclobutane rings in pharmacologically active compounds is relatively new, but inorganic‐based cyclobutane drugs have been around for a longer time, most profoundly in a drug inducing cell death in cancer cells by crosslinking DNA. This drug, carboplatin (3), is widely used in severe forms of cancer, including ovarian, testicular, cervical, head and neck cancers. ^[5] These compounds, ^[6] as well as oxetanes ^[7] and squaramide derivatives, ^[8] have already been extensively reviewed elsewhere and will not be discussed in the current review.

Synthetic efforts towards cyclobutane derivatives have since then progressed, improving their usefulness in drug development. As the synthesis of cyclobutane building blocks has been extensively reviewed elsewhere, ^[9] it will not be discussed in this review.

1.2. Properties of the cyclobutane ring

The cyclobutane ring is the second most strained saturated monocarbocycle after cyclopropane with a strain energy of 26.3 kcal mol⁻¹ (compared to 28.1 and 7.1 kcal mol⁻¹ for cyclopropane and cyclopentane, respectively).[ ¹⁰ , ¹¹ ] Strikingly, the strain energy drastically lowers when the cyclobutane ring is mono‐ or di‐substituted with methyl groups because of the Thorpe–Ingold effect (Figure 2). ^[10] With C−C bond lengths of 1.56 Å, the bond lengths are longer compared to ethane 1.54 Å. This lengthening is induced by 1,3 C−C non‐bonding repulsions, as the cross‐distance is only 2.22 Å. Comparatively, this interaction is not present in cyclopropane rings since every carbon is bound to two other carbon atoms, resulting in the shorter bond lengths of 1.53 Å. ^[12] Cyclobutane adopts a folded structure, reducing its bond angle slightly to 88° compared to the expected 90°, which increases the angle strain, but at the same time relieves the torsional strain. This balance of energies leads to the puckered conformation as the energetically most favorable structure (Figure 2). ^[13] As a result, the C−C bonds have slightly increased p‐character and the C−H bonds more s‐character. ^[14]

Figure 2.

A) Strain energies of various cycloalkanes. B) 3D structure of cyclobutane (i and ii). Newman projection along the C−C bond. Its structure slightly deviates from eclipsed (iii).

This increased s‐character is relatively subtle compared to cyclopropane. While cyclopropane readily reacts with sulfuric acid or undergoes bromination, cyclobutane does not. ^[14] This reactivity puts cyclobutane in between the very reactive cyclopropane and relatively inert cyclopentane and other alkanes.

2. Influence on pharmacological activity

This perspective will focus on illustrating how cyclobutanes as regular and spirocyclic fused rings, as substituents, and with different substitution patterns (1,2‐disubstituted up to octasubstituted) can favorably contribute to drug properties of small molecules. Rather limited knowledge is present in current literature regarding medicinal chemistry properties of cyclobutane scaffolds such as metabolic stability or their use in medicinal chemistry in general. Therefore, this review is meant to fill the gap in literature, showing the implications and opportunities cyclobutane rings can offer to improve druglike properties of compounds in hit‐to‐lead development, using literature from the time period 2010–2020.

An often‐described property that is commonly introduced by the inclusion of a cyclobutane ring is conformational restriction. Flexible ligands can suffer from an entropic penalty upon binding because of the freezing of rotatable bonds in the binding pocket. For example, a flexible ethyl linker can be replaced by a 1,3‐disubstituted cyclobutane to limit the number of possible conformations. In addition, conformational restriction by introduction of a fused ring system can block metabolically labile sites. ^[15] Furthermore, the use of saturated cyclobutane rings instead of planar aromatic rings correlates with stronger binding affinities because saturated molecules better complement spatial arrangements of target proteins. This saturation increase also correlates to higher water solubility and lower melting points, both of which are key to successfully develop a lead compound. ^[16] Moreover, the cyclobutyl ring can be used to direct key pharmacophore groups,[ ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² ] fill a hydrophobic pocket in the target enzyme,[ ¹⁷ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ ] prevent cis/trans‐isomerization,[ ³¹ , ³² , ³³ ] improve metabolic stability,[ ³⁴ , ³⁵ , ³⁶ ] replace aromatic groups as an aryl isostere, ^[37] conformationally restrict (part of) the molecule[ ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ ] or reduce planarity. ^[42]

The use of cyclobutanes in current drugs is limited compared to the use of other recurring structural moieties. As of January 2021, there were at least 39 (pre)clinical drug candidates containing a cyclobutane ring (DrugBank 5.1 database search), ^[43] some of which were discovered during the selected time and hence discussed in this review. The aim of this review is to provide an overview of different applications of cyclobutane rings in medicinal chemistry showing the positive effects this ring can have on the pharmacological properties of small molecules. The examples that follow in the next sections are ordered by disease area.

2.1. Cyclobutanes in anticancer compounds

MYC genes are a family of oncogenes encoding transcription factors involved in the regulation of apoptosis, cell metabolism, genome instability, cell growth and proliferation. As a consequence, MYC overexpression is often observed in cancer cells. ^[44] The interaction of MYC with WD repeat‐containing protein 5 (WDR5) is a key interaction in carcinogenesis and its interaction site has been recognized as a target for small‐molecule inhibitors. ^[45] By high throughput screening (HTS), Macdonald et al. ^[17] identified compound 4 as a hit showing inhibitory effects (Figure 3). Upon lead optimization, it demonstrated a high potency towards inhibition of the oncogenic function of MYC transcription factors. In this process, the cyclobutyl substituents (as well as other cycloalkanes) were originally installed on the phenolic ring as a tool for further growth along the edge of the WDR5 binding motif. It was, however, found that the cyclobutyl ring possessed optimal properties to complement the hydrophobic region of the binding pocket. It directs the nitrile group more towards the protein, with no observable directional interaction, but increased affinity nonetheless, which resulted in optimized structure 5 (Figure 3).

Figure 3.

WDR inhibitors, HTS hit 4 and optimized lead compound 5.

Cathepsin B is an enzyme of the lysosomal cysteine protease family. ^[46] Overexpression of this enzyme is prevalent in invasive and metastatic cancers. ^[47] In an effort to improve peptide linker stability of an antibody‐drug conjugate (ADC) that targets cancer cells where cathepsin B is active, Wei et al. ^[23] installed a cyclobutyl ring replacing a valine residue. It was hypothesized by a computational similarity search that this moiety would fit best in the hydrophobic binding pocket of cathepsin B. After successfully synthesizing numerous candidates, including compound 6 (Figure 4), it appeared that this cyclobutane‐containing linker showed greater selectivity towards cathepsin B over other enzymes, increasing its selectivity towards tumor cells compared to valine‐citrulline linker systems that are already in clinical trials. ^[48]

Figure 4.

Cathepsin B ADC 6 .

Tankyrase (TNSK) enzymes control several cellular pathways that execute key functions such as mitosis, energy metabolism and cell fate. Inhibition can have therapeutic potential in selected cancers such as colorectal and non‐small cell lung cancers. ^[49] Anumala et al. previously reported ^[50] the potent tankyrase inhibitor 7, which displayed a poor pharmacokinetic (PK) profile. To enhance its PK profile as well as retain its potency, two moieties of known inhibitors were combined using various linkers. ^[18] Among 1,4‐phenylene, trans‐1,4‐cyclohexyl and trans‐1,3‐cyclobutyl linkers the latter provided the best balance between rigidity and flexibility. The shorter linker distance also directs the triazole moiety in a slightly different orientation compared to the six‐membered‐ring linkers, allowing the pyrimidine to engage in π‐π interactions more efficiently with a tyrosine residue in the binding pocket. This led also to an improved PK profile (compound 9; Figure 5).

Figure 5.

Tankyrase inhibitors 7 and 8, and optimized inhibitor 9.

Lapierre et al. ^[19] aimed to develop inhibitors for AKT proteins which are involved in cell proliferation, migration and anti‐apoptotic survival as observed in several human cancers. ^[51] A previous study indicated that optimizing the activity against AKT1, AKT2, and AKT3 was crucial for achieving biochemical potency and cellular inhibition in multiple tumor types. ^[52] Though potent, their imidazo[4,5‐b]pyridine series, including compound 10, exhibited low plasma and tumor exposures and therefore the authors aimed to improve the overall PK profile of the AKT inhibitor as well as improve in vivo activity. Installation of a cyclobutylamine on the benzylic position resulted in improved inhibition of AKT. The binding mode of this scaffold was investigated by co‐crystallization of 11 (ARQ092; Figure 6) with target AKT1 and revealed its binding mode. The benzylic amine engages in bidentate hydrogen bonds with Tyr272 and Asp274 residues in the binding pocket. This interaction is enabled by the positioning of the cyclobutane in a hydrophobic region of the binding pocket, which resulted in a highly potent analogue. Compound 11 demonstrated high enzymatic potencies against AKT1, AKT2, and AKT3 as well as tumor growth inhibition in human xenograft mouse models of adenocarcinoma and an improved PK profile. It is currently in phase I and II clinical trials for various cancers, Proteus syndrome, and PIK3CA‐related overgrowth spectrum.

Figure 6.

AKT inhibitor 10 and 11.

In cancer diagnostics, positron emission tomography (PET) is widely applied to locate tumors by visualizing the accumulation of biologically active molecules in tumor tissue. If the radiolabeled molecule is unstable in vivo, the visualization tools cannot distinguish between accumulation of the intact molecule or its metabolites. This instability can thus lead to increased background activity and misleading information. In existing therapeutics, often a ¹⁸F‐fluoroethyl conjugated to a heteroatom is used. These chemical moieties, both unlabeled and radiolabeled were prone to enzymatic degradation. ^[53] Franck and co‐workers ^[35] hypothesized that if the ¹⁸F‐alkyl chain was substituted for a ¹⁸F‐cycloalkyl group, it would increase the metabolic stability of the PET tracer molecule. The model chosen was a tyrosine‐based amino acid, as O‐alkylated tyrosines have been shown to be transported into cancer cells by the large amino acid transporter. ^[54] The authors replaced the ethyl linker from compound 12 by a trans‐cyclobutyl ring (compound 13; Figure 7). Compound 13 was transported into numerous different cancer cells and despite the bulkier cyclobutyl group the biological properties remained unchanged. In vitro stability of the radiolabel in human and rat plasma showed excellent stability displaying over 120 minutes of stability as well as the unlabeled moiety showing over 60 minutes of metabolic stability. This compound is a good starting point for further in vivo research and shows that the cyclobutyl linker exhibits improved in vivo stability compared to the ethyl‐linked radiolabel and might therefore be a potential candidate as a PET tracer.

Figure 7.

¹⁸F‐radiolabeled PET tracers 12 and 13.

Extracts from the South African willow tree Combretum caffrum have been used in traditional African medicine. ^[55] Extensive studies have identified compound 14 as a potent natural drug showing anti‐tumour activity. ^[56] Its activity is based on inducing apoptosis by selectively binding to the colchicine binding site of tubulin. This inhibits tubulin polymerization which then leads to cell cycle disruption. ^[57] Several analogues of compound 14 have already been tested in clinical studies. The core structure of these compounds is characterized by a highly oxygenated cis‐stilbenoid moiety. The substitution pattern on the aromatic rings was crucial for its activity, as well as the cis‐configuration of the alkene in order to efficiently direct the aromatic substituents to the tubulin target. To enhance the in vivo potency, Nowikow et al. ^[32] wanted to overcome alkene isomerization under physiological conditions by chemically locking the scaffold into the cis‐orientation. This was achieved by synthesizing saturated and unsaturated cis‐constrained carbocycles from cyclobutyl up to cyclohexyl analogues. Biological studies concluded that the larger carbocycle analogues showed lower potencies compared to cyclobutane and cyclobutene analogues. Cyclobutene 15 showed the highest activity towards numerous cancer cell lines, however, with relatively low selectivity. Cyclobutane derivative 16 exhibited comparable potency as the natural product for CCRF‐CEM and K562 cell lines with a high overall therapeutic index (TI) (Figure 8). These compounds therefore form a solid foundation for further in vivo evaluations and show how cyclobutanes and/or ‐butenes can be employed to conformationally lock compounds into their most active form.

Figure 8.

Natural product 14, its cis‐constrained analogues 15 and 16, and cis/trans‐1,3‐substituted analogue 17 (data refer to cis/trans isomers, respectively).

In a similar study, Malaschuk et al. ^[33] were following a comparable approach. They synthesized cis‐ and trans‐1,3‐disubstituted analogues 17 of antitumor natural product 14 to mitigate in vivo cis/trans‐isomerization (Figure 8). During in vitro cytotoxicity evaluations in HepG2 and SK−N‐DZ cell lines it was concluded, however, that their cytotoxicity was rather poor (micromolar range) compared to 15.

The p97 ATPase protein plays a key role in protein homeostasis. This protein facilitates degradation of polypeptides by the proteasome with energy supplied from the hydrolysis of ATP. Clinical success of other proteasome inhibitors in certain cancers suggests that p97 inhibitors may potentially become anticancer drugs. ^[58] In an effort to develop novel allosteric p97 inhibitors, Laporte and co‐workers ^[24] identified a 2‐phenylindole scaffold in a HTS screen. The indole moiety appeared crucial in structure activity relationship (SAR) studies, leading to optimized compound 18 (Figure 9). To further improve its activity, a SAR study was conducted on the side chain keeping the 2‐[3‐(piperidin‐1‐yl)phenyl]‐1H‐indole constant. The linker length did not show a significant effect, but replacement of the triazole by a piperazine as the terminal substituent yielded a 3‐fold increase in potency. The authors then chose to conformationally restrict the ethyl linker and opted to introduce a cyclobutyl ring after other restriction methods were unsuccessful. Interestingly, the cyclobutyl did not only conformationally restrict the scaffold, but also introduced a kink into the flexible linker. This structural change gave a remarkable 10‐time increase in potency compared to 18. This unexpectedly high potency was hypothesized to be driven by desolvation of the cyclobutyl group in a more buried binding mode. These compounds now provide a good starting point for further in‐depth studies and have the potential to be developed into a novel class of cancer therapeutics.

Figure 9.

P97 inhibitors, optimized HTS hit 18 and optimized lead 19.

Threonine tyrosine kinases (TTK) are often overexpressed in breast cancer cells. ^[59] Patients displaying high TTK levels generally have high tumor grades, usually resulting in poor clinical outcomes. ^[60] Inspired by a carbon‐to‐nitrogen switch from imidazo[1,2‐a]pyridazines to imidazo[1,2‐b]pyridazines in patent literature, Liu and co‐workers ^[31] reasoned a similar transformation for pyrazolo[1,5‐a][1,3,5]triazines would result in a more optimal PK profile as TTK inhibitors. During the SAR study, the authors aimed to modulate the physiochemical properties while maintaining in vitro potency and bioavailability. This goal was pursued by installing a hydroxy group on the solvent exposed region in the binding pocket and adding a weakly basic group to the aromatic hydrophobic core. By installing a cyclohexanol ring on this region it appeared that they were potent TTK inhibitors, but that only the cis‐isomer 20 (Figure 10) exhibited desirable oral exposure and cell activity. Unfortunately, in vivo this stereoisomer was rapidly converted into the trans‐isomer. The authors installed a cyclobutanol instead of a cyclohexanol that exhibited no in vivo isomerization, but at the cost of oral exposure. This analogue was modified to increase the steric bulk by the incorporation of a methyl group to obtain 21 (CFI‐402257) (Figure 10). This cis‐cyclobutanol analogue is a potent TTK inhibitor and exhibited the highest bioavailability in mice. This compound displays the use of the cyclobutyl ring for rigidity increase when stereochemical interactions are a major contributing factor towards oral exposures and activities. In addition, this analogue also exhibited TTK kinase selectivity over 262 other kinases and is currently in phase I and II clinical trials for breast and advanced solid cancers.

Figure 10.

TTk inhibitors 20 (values refer to cis/trans isomers, respectively) and 21.

Histone methyltransferases (HMTs) are epigenetic enzymes catalyzing methylation of histone lysines and arginines. They have emerged as therapeutic targets as this process modulates transcription ^[61] and its dysregulation is implicated in numerous forms of cancer. ^[62] Euchromatic histone methyltransferase 2, otherwise known as G9a, is an example of such an HMT which has been linked to various cancers. ^[63] Sweis et al. ^[64] searched for chemically distinct G9a inhibitors through screening of an in‐house compound library. Compound 22 stood out having submicromolar potency (IC₅₀=153 nm) and contained a spirocyclic cyclobutane ring that in SAR studies was found to be crucial for its potency towards G9a. Modification to a spirocyclic cyclopentane, cyclohexane or substitution to hydrogens resulted in potency drops of at least one order of magnitude. Spiro[cyclobutane‐1,3′‐indol]‐2′‐amine 23 (A‐366) (Figure 11) was found optimal during SAR studies and the authors conducted an X‐ray cocrystal study. Interestingly, the cyclobutyl moiety was shown to reside close to the polar Asp1078 residue, which could not be explained by the authors. This result nevertheless provided a potent inhibitor which was found to be selective over 21 other HMTs.

Figure 11.

G9a inhibitors 22 and 23.

Daigle and co‐workers ^[65] aimed to identify small‐molecule inhibitors for H3K79 HMT protein DOT1L. This enzyme is involved in processes such as DNA‐damage response, gene expression and cell cycle progression and has also been implicated in the development of mixed lineage leukaemia (MLL). ^[66] People with MLL often have a methylation of H3K79 that should not occur and therefore inhibiting the methylation might prevent the formation of MLL. ^[67] Compound 25 (EPZ‐5676, Figure 12) was identified as a potent small‐molecule inhibitor of DOT1L based on a structure‐guided design and optimization starting from known aminonucleoside inhibitor 24, whose pharmacokinetic profile was not optimal and was rendered unsuitable for clinical development. Even though the chemical rationale behind the transformation towards 25 was not published, it exhibited a strong potency towards DOT1L, inducing conformational changes as well as showing high selectivity over other protein methyltransferases, improved residence time and in vivo efficacy. This compound was submitted to clinical trials and is currently ongoing in phase I and II trials for treatment of leukaemia.

Figure 12.

DOT1L inhibitors 24 and 25.

2.2. Cyclobutanes in autoimmune disease research

Compound 27 (TAK‐828F, Figure 13 ) is a nuclear receptor retinoic acid receptor‐related orphan receptor gamma (RORγt) inhibitor. This protein plays an important role in regulating the immune response in human TH17 cells[ ⁶⁸ , ⁶⁹ ] and has been suggested to play a pathogenesis role in autoimmune diseases such as rheumatoid arthritis and psoriasis. ^[70] RORγt plays a crucial role in the function and differentiation of TH17 cells ^[69] and modulators thereof can be regarded as potential candidates for treatment of immunological diseases. Previously, Kono et al. had already reported on the favorable in vitro RORγt and PK properties of 26. ^[71] The aim was to retain or improve the attractive biological profile as well as lower its lipophilicity, since it was non‐optimal for drug delivery. The highly lipophilic trimethylsilyl group was swapped and cyclized to an indane ring. In addition, the tetrahydroquinolone was altered to a pyridine, with the addition of a methoxy substituent as well. These alterations reduced its lipophilicity substantially (log D=3.96 originally vs 3.39 after modification) and retained its in vitro potency. Lastly, the authors hypothesized that a more rigid linker would be beneficial as the entropy loss upon binding would be minimal. Multiple cyclobutane linked compounds were synthesized; the cis‐1,4‐cyclobutane linker yielding compound 27 was optimal upon in vitro testing, while the lipophilicity was only slightly increased compared to the flexible linker. Compound 27 also showed the highest plasma exposure and oral bioavailability in in vivo studies in mice at 1 mg/kg. ^[38] This compound has undergone phase I clinical studies with no follows up as of now.

Figure 13.

RORγt inhibitors 26 and 27.

Hirata and co‐workers ^[39] sought to identify novel RORγt inhibitors by means of a HTS. The initial screening hit showed a structurally unique molecule with moderate potency against RORγt (hLUC EC₅₀=1.7 μM, FRET EC₅₀=0.85 μM) with >20‐fold selectivity over five other nuclear receptors. This compound had several drawbacks such as modest time‐dependent human cytochrome (CYP) 3A4 inhibition and low liver microsomal stability. The authors hypothesized that by improving the ligand efficiency (LE) from 0.25 to at least 0.30, ^[72] and decreasing the lipophilicity would yield an improved drug profile. An additional strategy employed by the authors was to increase its fraction of saturated carbons (Fsp³), as a decrease in Fsp³ would result in an increased incidence of CYP inhibition. ^[73] After initial SAR studies and X‐ray co‐crystal analysis of intermediate 28 (Figure 14) the conclusion was made that the inhibitor was bound in a U‐shaped conformation and that van der Waals contact was important for its binding efficiency. Therefore, the authors modified the side chain of the 1,2,4‐triazole moiety by incorporating cyclobutane residues to stabilize their binding conformation and mask potential metabolic sites. These analogues exhibited the best LE values with slight increases in metabolic stability as well as their Fsp³ values with cis‐cyclobutane analogue 29 (Figure 14) being optimal. This derivative was subjected to in vivo experiments showing promising results that confirmed the authors original hypothesis.

Figure 14.

RORγt inhibitors optimized HTS hit 28 and optimized lead compound 29.

Compound 30 (PF‐04965842) is a selective intracellular Janus kinase 1 (JAK1) inhibitor. The JAK family of enzymes are central to inflammation and regulation of immunity. ^[74] This property makes them attractive drug targets for autoimmune diseases. ^[75] Based on the pre‐existing drug tofacitinib, Vazquez et al. ^[20] kept the pyrrolopyrimidine moiety, being central to the hinge binding, subsequently evaluating a range of diamine linkers. The result being that cis‐1,3‐cyclobutane diamine linkers tended to exhibit not only potencies in the low nanomolar range but also excellent selectivity in the JAK family of enzymes, as shown in Figure 15. The authors subsequently investigated the relation the cyclobutyl moiety had on the binding interaction within the JAK family. The puckered conformation of the cyclobutyl ring allowed for the sulphonamide NH to be in a position to form hydrogen bonds with Arg and Asn residues of JAK1. The trans‐isomer of 30 showed less favourable activities because this interaction cannot be fully achieved. Because the favourable activity profile, this compound is currently in phase III clinical trials for treatment of atopic dermatitis.

Figure 15.

Tofacitinib and JAK1 inhibitor 30.

2.3. Cyclobutanes in CNS disease treatment

Weiss and co‐workers ^[25] set out to identify potent β‐site amyloid precursor protein cleaving enzyme 1 (BACE1) inhibitors from a known set of inhibitors, aiming to optimize their suboptimal pharmacokinetics and central nervous system (CNS) partitioning. The BACE1 protein is considered as a therapeutic target for the treatment of Alzheimer's disease (AD), as it is involved in the formation of Aβ peptides. Previously reported hit 31 ^[76] (Figure 16) showed low nanomolar potency against BACE1 in addition to reducing Aβ₄₀ levels. This was further optimized by SAR studies to compound 32, having a better overall PK profile with lower rat clearance (CL) and metabolic stability in the form of human and rat liver CL_int, at the overall loss of some potency. This compound was co‐crystallized with BACE1 peptide to find outs its binding mode. The benzodioxolane ring system inhabits the S1 binding pocket. The spirocyclic cyclobutyl ring and the neopentyl group occupy the S1′ and the S2′ hydrophobic pockets respectively, resulting in a potent binding efficiency. Later efforts ^[26] further optimized this moiety based on its intrinsic clearance and CNS penetration resulting in 33.

Figure 16.

BACE1 inhibitors 31, 32, and 33.

Researchers from Pfizer ^[37] discovered two clinical candidates 35 (PF‐03654746) and 36 (PF‐03654764) as histamine 3 receptor (H₃R) antagonists. Though numerous H₁ and H₂ receptor antagonist pharmaceuticals have been developed, H₃ and H₄ receptor antagonists are overall less developed. ^[77] H₃ receptors have their highest densities in the prefrontal regions of the brain, where they play a central role in numerous functions including learning, arousal and wakefulness. ^[78] It is suggested that pharmaceutical antagonists of H₃ receptors can be therapeutic agents for attention deficit hyperactivity disorder (ADHD), AD, narcolepsy and other cognitive disorders. ^[79] Pharmaceuticals for these conditions are generally effective, but can also induce (toxic) adverse effects. The authors therefore sought to develop safer and more selective H₃ antagonists for CNS disorders. An early HTS identified compound 34 (Figure 17) and exhibited high affinity for human H₃ receptors (K_i=1.3 nM) but also a significant estimated unbound CL_int (CL_int, u=219 mL/min/kg). In addition, it was suspected that the biaryldiamine moiety might induce safety issues as similar structures are known to be genotoxic. ^[80] The authors searched for aryl isosteres to mimic the distance between the two basic groups and theorized that a cyclobutane linker might act as one. This hypothesis was confirmed by molecular modelling studies, where the two structures exhibited good overlap of the two basic groups. Furthermore, the authors theorized that the 3D volume of the cyclobutane ring would be less likely to induce DNA damage by binding in the minor groove of DNA. Therefore, one aryl group was substituted for a trans‐cyclobutyl ring. After numerous further optimization studies mostly regarding the PK and safety profiles, structures 35 and 36 identified as optimal compounds.

Figure 17.

H₃R inhibitors, initial HTS hit 34 and optimized compounds 35 and 36.

These leads displayed good absorption, distribution, metabolism and excretion (ADME) properties and displayed negative in vitro micronucleus assay results and no in vivo phospholipidosis (PL), thus displaying favorable safety properties. A handful of phase I and II trials were conducted for both leads, with no continuation to phase III as of yet.

Letavic et al. ^[81] also searched for novel preclinical H₃ receptor antagonists to develop into clinical candidates. An initial aryl‐oxynicotinamide showing moderate affinity for the H₃ receptor and no affinity for SERT inspired the authors to expand this series. SAR studies showed that a diazepane ring was preferred over a piperazine ring, with cyclic substituents on the diazepane moiety. Cyclobutyl derivatives 37 and 38 (Figure 18) exhibited an optimal balance between microsomal stability, rat plasma and brain concentrations and receptor occupancy. During further profiling studies, these compounds exhibited H₃ occupancy at low plasma concentrations, promoted wake and increased histamine release in rat studies in combination with a favorable PK profile. Compound 37 (JNJ‐39220675) has been submitted to clinical trials, but there was no continuation after phase II studies.

Figure 18.

H₃ antagonists 37 and 38.

Nirogi and co‐workers ^[82] aimed to identify novel H₃ receptor antagonists as well. Based on the challenges with H₃ receptor antagonists such as CYP inhibition, human ether‐a‐go‐go‐related gene (hERG) and PL, ^[83] the authors aimed to identify compounds that exhibited promising profiles regarding the earlier described properties while also maintaining in vivo activity. They identified a series of benzamide compounds with moderate affinities by means of an initial screening and set out to optimize these series. First, cyclization of the amide nitrogen yielded a more optimal balance of rodent half‐life and brain penetration. Numerous N‐alkyl substituents on both piperidine rings were investigated to test their activity. It was concluded that cyclic substituents on these positions were well tolerated though the most optimal combination was compound 40 (Figure 19) bearing a N‐cyclobutyl substituent on both piperidine rings. This compound was subjected to further in vitro testing and appeared both metabolically stable and exhibiting minimal hERG inhibition.

Figure 19.

H₃ antagonists 39 and 40.

Ontoria et al. ^[21] looked into novel small‐molecule inhibitors of the kelch‐like ECH associated protein 1 (KEAP1)/nuclear factor erythroid derived 2 (NRF2) interaction. This protein‐protein interaction is a key interaction in the NRF2‐antioxidant responsive element mechanism of defence against oxidative stress. Inhibiting the KEAP1/NRF2 interaction, and hence proteasomal degradation, promotes the intracellular concentration of NRF2. ^[84] This increase plays a key role in anti‐oxidant processes to prevent neurodegenerative diseases such as Huntington's and Parkinson's. ^[85] Previous research ^[86] identified an oligopeptide as the minimal binding sequence to the Kelch domain of KEAP1. A number of crystal structures of KEAP1 cocrystallized with this oligopeptide provided crucial structural information regarding the binding interactions with the protein. Based on this information, small‐molecule inhibitor tetrahydroisoquinoline 41 (Figure 20) was designed. This inhibitor however, which was shown to occupy only 3 of the 5 available binding sites, lacked binding affinity compared to small‐molecule inhibitors binding to all 5 available sites. In addition, 41 contains a carboxylic acid group and although favorable in this interaction, in general acidic moieties are avoided in CNS active pharmaceuticals as they commonly display poor blood‐brain barrier (BBB) penetration. The authors conducted a small SAR study on the glutamic acid position of the previously identified oligopeptide to gain more insight into the binding mode of the KEAP1 binding site. It was found that peptide derivatives showed a preference for non‐natural cyclobutylaniline (Cba) residues, where the ring expanded or ring‐contracted analogues both showed sharp decreases in affinity. In addition, a preference for carboxamides was found in this study. When docking the Cba peptide analogue and comparing it to 41, it was revealed that the cyclobutyl and the cyclohexyl occupied the same region of KEAP1. The cyclobutyl ring in the oligopeptide appeared to direct the primary amine in a favorable position in the binding pocket. Based on these findings, the authors constructed carboxamide‐substituted cyclobutyl derivatives of 41. The direct comparison of 41 with its ring‐contracted cyclobutyl carboxylic acid analogue resulted in a lower binding affinity due to reduced van der Waals contact. Its analogue though, both carboxamide ring‐contracted as well as hydroxyl substituted, resulted in similar binding affinities compared to 41. The authors speculated an alternative binding mode with KEAP1 was prevalent compared to 41. Therefore, X‐ray crystallography studies of 42 (Figure 20) with KEAP1 were performed which indeed resulted in an altered binding mode. The amide of the cyclobutyl derivative is directed in such a way it can form three different direct hydrogen bond interactions with Ser363, Asn414 and Arg415 in KEAP1’s P2 binding pocket. This compound now presents a more robust CNS drug candidate compared to previously identified acidic KEAP1 pharmaceuticals due to the projected BBB penetration.

Figure 20.

KEAP1 inhibitors 41 and 42.

Phosphodiesterase (PDE) enzymes are involved in intracellular signalling of cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) by converting them to AMP and GMP. Their isoforms are highly localized with 10 out of 11 localized in the CNS. Therefore, controlling their activity in the CNS could potentially be beneficial towards neurodegenerative and neuropsychiatric conditions such as Alzheimer's and Parkinson's disease.[ ⁸⁷ , ⁸⁸ ] Hu and co‐workers ^[22] tried to identify PDE10A inhibitors, as its inhibition could present a novel target for the treatment of schizophrenia. ^[87] This enzyme is highly expressed in brain tissue, especially in the striatum, thought to be disregulated in schizophrenia and an antagonist may have therapeutic benefit. The authors previously identified a PDE10A inhibitor that exhibited low nanomolar potencies (IC₅₀=4.5 nM) with moderate rat clearances (Cl=0.53 L/h/kg) and a low bioavailability (F=10 %). ^[89] Focussing on reducing its metabolic liability, initial SAR studies led to compound 43 (Figure 21). After an X‐ray co‐crystal analysis with human PDEA10 it was concluded that the phenyl linker made no apparent interactions within the binding pocket, thereby posing an opportunity to introduce new interactions, while maintaining the linear orientation between the imidazo[4,5‐b]pyridine and the benzo[d]thiazol‐2‐amine units. After cyclohexyl linkers showed a reduction in potency, cyclobutyl derivatives were synthesized. Both cis‐ and trans‐linked cyclobutyl analogues 44 showed low nanomolar potency (Figure 21). Both isomers were used in X‐ray cocrystal structure studies in human PDE10A. While the cyclohexyl analogue was hypothesized to be too bulky to fit in the channel between Gln716 and Tyr683 residues in the binding pocket, the cyclobutyl derivatives fitted well and allowed imidazo[4,5‐b]pyridine and the benzo[d]thiazol‐2‐amine moieties to engage in critical hydrogen bonding interactions within the binding pocket. The cis‐isomer showed more potent binding, but its bioavailability was significantly lower, while the cyclobutyl series generally exhibited a satisfactory PK profile

Figure 21.

PDEA10 inhibitors 43 and 44 (values refer to cis/trans isomers, respectively).

Apoliprotein E (apoE) is a lipid carrier protein playing a major role in lipid homeostasis mostly manufactured in brain and liver tissues. ^[90] This protein has three isoforms namely apoE2, apoE3 and apoE4. The possession of apoE4 alleles have been shown to be a strong genetic risk factor for late‐onset Alzheimer's disease (LOAD). ^[91] The exact role of apoE4 is unknown, ^[92] but it was hypothesized that stabilizing the protein may impact its disease pathology. Petros et al. ^[27] aimed to identify such stabilizers by fragment‐based NMR screening. An initial screening identified cyclobutane scaffold 45 (Figure 22) with affinity towards the N‐terminal domain of apoE4. In addition, the scaffold stabilizes apoE4 and affects the kinetics of liposome breakdown. Next, the authors aimed to enhance its potency by structure based‐drug design. An X‐ray co‐crystal structure of 45 bound to apoE4 showed its binding mode and possible positions for potency enhancement. Besides other binding interactions, the cyclobutyl ring fits in a hydrophobic subpocket generated by Trp26, Leu30 and Ala153. On this basis the authors chose to extend off the central phenyl ring on the meta position. Addition of an unsubstituted phenyl ring already resulted in a 4‐fold potency increase, most likely due to increased van der Waals contact. After more SAR studies compound 46 (Figure 22) was identified, bearing hydrogen bonding capable groups, engaging in hydrogen bonds with Asp35. This lead showed similar favorable apoE4 protein stabilization but at a 5‐fold lower concentration.

Figure 22.

ApoE4 stabilizers 45 and 46.

2.4. Cyclobutanes in obesity research

Compound 48 (BMS‐814580) is a potent and selective melanin concentrating hormone 1 (MCHR1) inhibitor, an enzyme largely involved in energy homeostasis. Inhibition of MCHR1 has been shown to be pharmacologically relevant to decrease obesity in animal studies. ^[93] Ahmad and co‐workers ^[34] aimed to eliminate toxic effects of metabolites of the previously reported ^[94] lead 47 by improving its metabolic stability as it was found that the gem‐dimethyl substituent was prone to in vivo metabolic oxidation. Initially, a cyclobutyl substituent was installed with non‐satisfactory results, showing insufficient exposures. The difluorinated analogue 48 of the cyclobutyl drug was then synthesized to lower the cyclobutyl's susceptibility to metabolic oxidation and indeed showed major improvements in exposure and metabolic stability as depicted in Figure 23.

Figure 23.

MCHR1 inhibitors 47 and its transition to more robust and potent lead compound 48.

He et al. ^[95] set out to find small‐molecule inhibitors of the fat mass and obesity associated protein (FTO). FTO has been linked to obesity, ^[96] but much of its biology remains unknown. A virtual screening was performed to identify small‐molecule inhibitors and led to multiple catechol derivatives connected through a cyclobutyl linker to another aromatic substituent. These compounds did not yield the desired profile. Therefore, the scaffold was changed from catechol to resorcinol, in which the geometry of the alcohols was reasoned to mimic the substrate 3‐methyl‐adenine better and thus improve binding. These studies resulted in the optimized compound 49 (Figure 24). Binding studies revealed that a combination of hydrogen bonds and van der Waals forces are responsible for its specific recognition. The cyclobutylphenyl moiety of 49 establishes extensive and close hydrophobic contacts with the non‐conserved antiparallel β‐sheet of FTO, yielding a potent and novel FTO inhibitor.

Figure 24.

FTO inhibitor 49.

G‐protein coupled receptor subtype Y₄R selective agonists have been proposed as anti‐obesity agents. ^[97] Natural neurohormone ligands of the human family of Y_xR (x=1, 2, 4, 5) G‐protein coupled receptors are neuropeptide Y (NPY) and pancreatic polypeptide (PP). Potent and selective agonists and antagonists are available for all except for the Y₄R subtype. ^[98] Berlicki and co‐workers ^[99] aimed to identify potent and selective small‐molecule Y₄R inhibitors. The natural peptide hPP (50) is a Y₄R ligand with unsurpassed affinity and agonistic potency (K_i 0.53 nM, EC₅₀ 11 nM), albeit that it can also activate Y₁R and Y₅R. The authors set out to modify one of Y₄R's natural ligand agonists to enhance its selectivity towards being a more potent and selective agonist compared to the 2‐digit micromolar potent selective Y₄R agonist known to date of publication. The authors installed unnatural cyclic cycloalkane scaffolds into the peptide chains of 50 by replacing a glutamine residue, to induce enhanced selectivity and potency, but also to improve the stability towards protease degradation. A cyclopentyl candidate showed the most favorable profile, but cyclobutyl candidate 51 (Figure 25) also exhibited double digit nanomolar potency towards Y₄R, as well as selectivity over the other subtypes of the Y₄R family.

Figure 25.

Y₄R inhibitor 51 and natural peptide 50.

Clusters of differentiation 38 (CD38) is a type II membrane glycoprotein mostly located in immune cells but also expressed in bone and other major organ tissues such as liver, intestine, pancreas, muscle and brain. ^[100] Besides having receptor functions, it shows enzymatic activity towards redox factor nicotinamide adenine dinucleotide (NAD). As it is consumed, it is converted by CD38 to cyclic adenosine diphosphate ribose (cADPR) and ADPR. CD38 therefore plays a central role in NAD cofactor modulation. Artificially keeping NAD levels high by CD38 inhibition could have a positive effect on metabolic diseases such as obesity. Previously identified quinoline‐8‐carboxamide CD38 inhibitor 52 raised NAD levels in a diet‐induced obese mouse model with acceptable pharmacokinetics. ^[42] However, the authors intended to increase its potency further as well as reduce liabilities such as hERG inhibition. CD38 X‐ray co‐crystal structures of previously identified inhibitors revealed additional space in the binding pocket at the 2‐position of the quinoline that could be explored for the desired pharmaceutical properties. When exploring 1,4,5,6‐tetrahydropyrrolo[3,4‐c]pyrazole substituents at this position, the authors initially inserted geminal substituents to avoid inserting a chiral center. A gem‐dimethyl substituent at this position showed acceptable potencies. However, spirocyclic cyclopropyl and cyclobutyl derivatives appeared most potent for enzyme inhibition. Even though the cyclobutyl derivative 53 (Figure 26) did not exhibit the most ideal properties, the reduction of planarity by the cyclobutyl introduction might decrease the crystal lattice energy of the solids, resulting in an enhanced solubility according to the authors.

Figure 26.

CD38 inhibitors 52 and 53.

2.5. Cyclobutanes in antiviral compounds

Hepatitis C virus (HCV) is a pathogen affecting millions of people worldwide. ^[101] Numerous efforts to produce antivirals have been made, but room for improvement is still vast. Beaulieu and co‐workers ^[102] have actively investigated allosteric inhibitors of the main viral RNA polymerase non‐structural protein 5B (NS5B) in one of its pockets “Thumb 1” vital to its RNA synthesis abilities. In an effort to late stage optimize known inhibitors the authors found that substituted indole–cinnamic acid derivatives were optimal for this goal (Figure 27). When optimizing the diamide linker between the two moieties it was found that cyclobutyl or cyclopentyl had significantly improved absorption irrespective of the substituents on the indole scaffold. While a gem‐dimethyl linker had improved cell‐based potency, the cyclobutyl linker showed improved oral bioavailability and an overall balanced profile of potency, in vitro ADME properties and in vivo rat exposure. This observation was found consistent across numerous indole–cinnamic acid derivatives and allowed for development of clinical lead compounds 54 (BILB1941) and 55 (BI207524) (Figure 27 ). Compound 54 was submitted to clinical trials, but no further studies were performed after phase II. Despite its attractive potency compared to 54, compound 55 was not submitted to clinical trials since in preclinical studies was found that a genotoxic aniline metabolite formed in human and rat liver microsome assays. In later optimization efforts, LaPlante et al ^[103] further elaborated on these chemotypes. Continued conformational restrictions were applied in the molecule guided by NMR, producing new scaffolds that showed optimized properties. The main change applied was cyclization of one of the diamides to a 1‐methylimidazole moiety as in compound 56 (BI207127). This compound was submitted to clinical trials and is currently under investigation in phase III.

Figure 27.

NS5B thumb pocket 1 inhibitors 54, 55, and 56.

The influenza virus remains a global challenge as seasonal outbreaks keep afflicting large human populations. Additionally, pandemic outbreaks such as the 2009 H1N1 swine flu or the H5N1 bird flu are examples of the potential severity of the influenza strain and therefore drug discovery efforts remain of high interest. Targeting viral replication pathways remains a suitable strategy for small‐molecule inhibitors. The current standard of care for influenza infections utilizes this in the form of neuraminidase inhibitors. ^[104] Farmer et al. ^[28] previously identified ^[105] the novel 7‐azaindole 57 (Figure 28) inhibitor from a phenotypic cell protection assay screen showing survival benefits in a mouse model when administered 48 h post infection. ^[105] These inhibitors inhibit the influenza polymerase‐B2 (PB2) heterotrimeric complex essential to viral RNA replication. In addition to replacing the chloride for a fluoride atom in previous research, optimization efforts focused on the alanine diethylamide side chain. After docking and X‐ray studies on the PB2 binding pocket the authors conjectured that synthesizing derivatives containing tert‐butyl side groups would fill the hydrophobic pocket defined by three phenylaniline residues. The best analogue showed two‐digit nanomolar affinities, but the authors suggested based on the X‐ray structures of the tert‐butyl analogues that the pocket could be filled by even larger substituents. After numerous iterations, it was observed that cyclohexyl derivatives showed promising activities, leading to the synthesis of more cycloalkyl substituents. The methyl‐spirocyclobutyl compound appeared the most potent one, while removal of the methyl led to a 100‐fold decrease in activity, displaying the stringent balance in shape and size for filling this hydrophobic pocket. This result led to a potent cyclobutyl containing candidate 58 (Figure 28) with a potential as a therapeutic.

Figure 28.

PB2 inhibitors 57 and 58.

2.6. Cyclobutanes in antidiabetes compounds

Glycogen synthase kinase 3 (GSK‐3) is a serine/threonine protein kinase that phosphorylates and inactivates glycogen synthase, limiting the formation of glycogen. GSK‐3 activity and expression are found to be higher in type 2 diabetics. ^[106] Therefore the inhibition of GSK‐3 in type 2 diabetes patients might be a reasonable target to find pharmaceuticals effective against this condition. ^[107] Seto and co‐workers ^[108] previously identified the potent and selective GSK‐3β quinolone inhibitor 59 (Figure 29). These inhibitors, however, were not as potent in cell‐based tests. The authors hypothesized that this lack in potency was because of insufficient cell permeability. They then merged the quinolone moiety of the previous inhibitor 59 with another inhibitor known in literature to arrive at a 6–6–7 quinolone ring system. This compound was subjected to a SAR study also involving a spirocyclic substituent. It was concluded that cyclobutyl derivative 60 (Figure 29) exhibited GSK‐3β inhibitory activity in both cell‐free and cell‐based assays. In addition, compound 60 decreased the plasma glucose concentration in a dose‐dependent manner in an oral glucose tolerance test in mice. Based on these promising results, this compound will be subjected to further biological testing.

Figure 29.

GSK‐3β inhibitors 59 and 60.

Intracellular enzyme glucokinase (GK) is present in both liver hepatic and β‐cells and is responsible for the first step in glucose utilization being the conversion of glucose into glucose‐6‐phosphate. ^[109] In both tissues GK plays a role in regulating glucose utilization and production. ^[110] Glucokinase activators (GKAs) influence glucose binding (K _m or S_0.5) and the kinetic profile (V _max) of the phosphorylation reaction. ^[111] Therefore GKAs play a central role in regulating blood glucose concentrations and therefore can be employed as possible therapeutics for type 2 diabetes. ^[112] Du et al. ^[40] previously identified ^[113] a series of GKAs inhibitors and pursued the identification of a structurally distinct series. While doing so, the authors aimed to keep S_0.5>0.6 mM and 0.8<V _max<1.3 based on data analysis to mitigate adverse effects. The authors pursued a methyl urea‐substituted pyridine series, including compound 61 (Figure 30). After some initial SAR studies, the authors also included cyclic alcohols at C‐5 as tertiary alcohols had been identified to prevent secondary metabolism. In addition, the introduction of cyclic moieties may decrease the degrees of freedom in the molecule, resulting in improved pharmacokinetic properties. ^[114] Cyclohexanols showed some potency but outside the V _max range set by the authors. The values were more encouraging for cyclobutanols. A methyl was added to create a tertiary alcohol, of which the trans‐variant 62 (AM‐9514) (Figure 30) was most potent and in the desirable parameter range. It showed favorable in vitro properties as well as desirable pharmacokinetics in mouse and dog.

Figure 30.

GKAs inhibitors 61 and 62.

2.7. Cyclobutanes in miscellaneous disease areas

Compound 63 (AM2389) is a selective cannabinoid 1 (CB1) agonist. Modulation of the CB1 and CB2 receptors is a possible treatment for conditions such as pain and inflammation. ^[115] It was previously identified that the phenolic hydroxyl and the lipophilic side chain are pharmacophores within the (−)‐Δ⁸‐tetrahydrocannabinol (THC) (Figure 32) core structure that determines its potency. Nikas and co‐workers ^[29] investigated the effect of modifying the C3 side lipophilic tail on hexahydrocannabinol (HHC) derivatives towards their potency and selectivity. The authors previously identified C1′ cyclic substituents to be useful moieties for achieving potency in (−)‐Δ⁸‐THC derivatives. Using the same strategy, numerous derivatives bearing cyclopentyl or cyclobutyl groups on the C1′ position were synthesized to assess the effect of conformational restriction towards potency and selectivities of HHC derivatives. According to their hypothesis, the cyclopentyl and cyclobutyl rings increased the potency towards CB1 and CB2 receptors significantly. The cyclopentyl moiety was most potent, though the cyclobutyl still had a high potency, as well as high selectivity (Figure 31). The authors conducted a modelling study of 63 in implicit water for CB1 binding. It was found that the cyclobutyl ring could engage in optimal interactions at the putative receptor‐binding site. In vivo experiments showed that 63 has a slow onset (60–90 min at 0.1 mg/kg in rats), a long duration of action, and a high potency in assays of antinociception and hypothermia. When compared to morphine, the analgesic action of 63 was 100‐fold higher.

Figure 32.

CB1 and CB2 agonist 64, its de‐esterified analogue and (−)‐Δ⁸‐THC.

Figure 31.

CB1 and CB2 inhibitors HHC and 63.

Compound 64 is an agonist of CB1 and CB2 receptors. Seeking to expand the medicinal toolbox of cannabinergic drugs, Sharma et al. ^[36] developed novel (−)‐Δ⁸‐THC derivatives. The authors aimed to solve the fact that cannabinoid receptor pharmaceuticals metabolites often show adverse or unpredictable side effects. ^[116] It was speculated that if a biologically labile substituent was introduced on the 3‐side tail of (−)‐Δ⁸‐THC it would predictably and controlled metabolized in a way that can be chemically tuned. Using this so‐called soft‐drug approach, an ester was introduced at the 2′‐position of the alkyl tail. Using modifications at the 1′‐position the lability could be tuned. This position was functionalized with methyl (both R and S diastereoisomers), gem‐dimethyl and a cyclobutane substituent. Though the potencies towards CB1 and CB2 were comparable to other potent CB1 and CB2 agonists, the cyclobutyl scaffold showed the most favorable kinetics. Its mouse and rat plasma half‐lives were significantly longer compared to the other derivatives, ensuring a larger therapeutic window. Molecular modelling studies revealed that the cyclobutyl ring optimally occupied a putative subpocket of the CB1 receptor. The larger van der Waals radius of the cyclobutyl group is proposed to also hinder its enzymatic cleavage by esterases (Figure 32). In addition, the esterase cleavage product of 64 has no significant affinity for CB receptors. In this way the cyclobutyl ring was used both as a tool to fill the enzymatic pocket as well as slow down the enzymatic metabolism.

Saavedra and co‐workers ^[117] sought to identify novel antipsychotics to battle schizophrenia. Current antipsychotics are limited in their ability to combat some of the negative and neurocognitive effects of schizophrenia. ^[118] The authors aimed to design potent blockers that showed selectivity on dopamine receptor 3 (D3R) over D2R receptors as well as having 5‐HT6R blocking potency. Based on previous findings in literature, the authors designed cyclobutaindole derivatives hypothesizing that these compounds would exhibit the desired pharmacological profile. Having executed numerous SAR studies, compounds 65 and 66 (Figure 33) showed optimal results. These lead compounds behaved as 5‐HT6R ligands as well as exhibiting D3R selectivity over D2R.

Figure 33.

5‐HT₆ receptor ligands 65 and 66.

Visceral Leishmaniasis (VL) is caused by L. donovani and L. infantum. The current treatment is effective, but far from ideal, since it expensive, accompanied by pain and often toxic side effects. ^[119] Sijm et al. ^[120] aimed to develop novel potent compounds showing anti‐leishmanial activity, but with a smaller risk of toxic side effects. In an initial library screening, hexahydrophthalazinone 67 was identified as a hit that formed the starting point for a SAR study. Although other candidates showed higher potency, candidate 68 (Figure 34) exhibited lower cytotoxicity towards human MR5‐C cells. This compound is a viable candidate for further studies and optimization on anti‐leishmanials.

Figure 34.

Anti‐leishmanial inhibitors 67 and 68.

Mycobacterium tuberculosis (Mtb) is a bacterium responsible for tuberculosis (TB) and has a significant presence among human populations. Therapeutic agents for its treatment have been developed, but significant side effects are often observed during the long period that the patients have to take the medication. ^[121] This results in unwanted consequences such as non‐compliance, relapses and the emergence of multidrug‐resistant strains. ^[122] Therefore, novel therapeutics are constantly being developed and studied. The sturdiness and drug‐resistant properties of the mycobacteria is mostly because of the thick lipid cell wall that consists for a large part of mycolic acids. Existing Mtb treatments focus on inhibiting the enzymes responsible for the biosynthesis of mycolic acid. It is known that mycobacteria incorporate C₁₆ and C₁₈ fatty acids into their mycolic acid synthesis, as well as modified ones. ^[123] Therefore, Sittiwong and co‐workers ^[124] hypothesized that specifically functionalized fatty acids might hijack this pathway and limit mycobacterial growth. The authors synthesized carbocyclic derivatives of decanoic and oleic acids with different functionalities. The cyclobutane derivatives showed significant inhibitory anti‐mycobacterial activity against Mtb, 69 and 70 (Figure 35) appeared more active than clinically used TB drug D‐cycloserine (CDC1551 & H37Rv week MIC: 4/(39) and 8/(78) μM, respectively) and 69 proved equally, if not more potent than isoniazid (CDC1551 & H37Rv week MIC: 4/(29) and 8/(58) μM, respectively). In a similar study performed by the same group, Zinniel et al. ^[125] identified that the same group of cyclobutane derivatives of decanoic and oleic acids exhibit similar behaviour towards mycobacterium avium paratuberculosis (Map). While both C₁₀ and C₁₈ derivatives showed efficacy in the Mtb study, only C₁₈ analogues showed significant activity efficacy against Map. Compounds 69 and 70 will therefore be a good starting point for biological evaluations and might be useful starting tools for development of therapeutics targeting Mycobacterium species. ^[126]

Figure 35.

Mtb inhibitors 69 and 70.

Compound 72 is a sulfonamide indazole derivative β₃‐adrenergic receptor (AR) inhibitor. Wada et al. ^[30] began searching for inhibitors of the β‐AR subfamily of enzymes as potential therapeutics against overactive bladder. ^[127] Aiming to improve the PK properties of the previously reported lead 71 ^[128] as well as maintaining its selectivity, it was found that large alkane substituents on the sulfonamide sulfur resulted in higher potencies and selectivities. The cyclobutyl substituent gave the best balance between potency, selectivity and ADME properties. A molecular dynamics simulation was run to predict its binding mode. It was found that the cyclobutyl ring of 72 (Figure 36) fits the hydrophobic pocket of the target enzyme particularly well. This pocket was also partially responsible for the selectivity over α‐AR enzymes lacking a significant hydrophobic pocket compared to β₃‐AR.

Figure 36.

β₃‐AR inhibitors 71 and 72.

Zhang and co‐workers ^[41] attempted to reduce the negative side effects of propofol, the most widely used intravenous general anaesthetic in clinical use. ^[129] Numerous propofol derivatives have already been synthesized to improve the overall profile. The authors noted that meta‐modification yielded suitable positions for modifications from literature. ^[130] Installation of a methyl group on the meta‐position yielded a good therapeutic index as well as acceptable potency. Moving forward the authors decided to fuse the ortho‐ and meta‐substituents into a cyclobutane ring. This fusion, giving rise to compound 73, both reduced the molecular weight and increased the scaffolds’ rigidity. Doing so already improved the therapeutic index 2‐fold with a retention in ED₅₀ compared to propofol for the γ‐aminobutyric acid A receptor (Figure 37). From further SAR studies it was concluded that this moiety generally yielded drug candidates with improved therapeutic indices and comparable or enhanced ED₅₀ values. Finally, introduction of a hydrogen bond acceptor on the tertiary carbon improved the pharmacological profile as well, yielding optimized structure 74 (Figure 37).

Figure 37.

Propofol, cyclized analogue 73, and optimized lead 74.

3. Conclusions

This review highlights the role of cyclobutyl rings in (pre)clinical drug candidates and is summarized in Table 1 below. Based on this review, researchers could use rational design for applications of a cyclobutane ring in small‐molecule drug development such as conformational restriction, increase in rigidity, filling of a hydrophobic pocket and directing key pharmacophores. Other uses of a cyclobutane ring such as improvement of metabolic stability, as an aryl isostere, improved solubility or affinity alteration may be less straightforward. The examples in this review do however illustrate its use in these areas. This potentially gives the medicinal chemist one more tool for tackling these challenges in the development of small‐molecule drug candidates in the future and in addition also potentially allows for greater insight into using a cyclobutane ring for these applications.

Table 1.

Summary of contributions of a cyclobutyl ring towards drug properties discussed in this article.

Property	Contribution	Compound example, drug target and pharmacological area
		Compound example	Drug target[Ref.]	Pharmacological area
Binding pocket exploitation	Fits well in hydrophobic pocket and directs nitrile group		WDR5 inhibitor [17]	Cancer
	Fits well in hydrophobic pocket		Cathepsin B substrate [23]	Cancer
	Directs triazole motif towards more efficient interactions of the pyrimidine scaffold		TNKS1/2 inhibitor [18]	Cancer
	Directs amine towards favourable interactions in binding pocket		Allosteric AKT1/2/3 inhibitor [19]	Cancer
	Fits well in hydrophobic pocket driven by desolvation.		P97 ATPase inhibitor [24]	Cancer
	Puckered conformation allows for favourable interactions of sulphonamide in binding pocket		JAK1 inhibitor [20]	Autoimmune disease
	Cyclobutyl motif fits well in hydrophobic pocket		BACE1 inhibitor[ 25 , 26 ]	AD
	Cyclobutyl motif directs amide towards favourable interactions		KEAP1/NRF2 interaction inhibitor[21	Parkinson's, Huntington's
	Cyclobutyl fits well in narrow binding hinge and directs substituents		PDE10A inhibitor [22]	Schizophrenia
	Cyclobutyl fits well in hydrophobic binding pocket		APOE4 stabilizer [27]	LOAD
	Cyclobutane ring establishes extensive van der Waals contact in binding pocket		FTO inhibitor [95]	Obesity
	Methylcyclobutyl substituent fills hydrophobic pocket well		Influenza PB2 [28]	Influenza virus
	Cyclobutyl engages in interactions in putative binding site		CB1 agonist [29]	Inflammation
	Fit hydrophobic pocket		ß3‐AR agonist [30]	Overactive bladder
Chemical stability	Locks scaffold in biologically active cis‐conformation Inducing increased potency		Tubulin polymerization inhibitor [32]	Cancer
	Locked into cis‐conformation but lower potency		Tubulin polymerization inhibitor [33]	Cancer
	Prevents cis/trans isomerization of the alcohol		TTK inhibitor [31]	Cancer
Metabolic stability	Increased		Cancer cells [35]	Cancer
	Increased		MCHR1 inhibitor [34]	Obesity
	Increased, cyclobutyl slows esterase cleavage		CB1 & CB2 agonist [36]	Inflammation
Aryl isostere	Simulates shape and aryl distances well		H3 receptor antagonist [37]	ADHD, AD, narcolepsy
Conformational restriction	Rigid linker increases binding efficiency		RORγt inverse agonist [38]	Rheumatoid arthritis, psoriasis
	Rigid linker increases binding efficiency		RORγt inhibitor [39]	Rheumatoid arthritis, psoriasis
	Cyclization and reduction of degrees of freedom improves PK profile		GK activator [40]	Type 2 Diabetes
	Increase in rigidity improves TI		GABAA receptor substrate [41]	Anaesthesia
Reduction of planarity	Decreases crystal lattice energy, resulting in faster dissolution		CD38 inhibitor [42]	Obesity
Potency alteration	Increased		G9a inhibitor [64]	Cancer
			DOT1L inhibitor [65]	Cancer
	Increased		GSK‐3β inhibitor [108]	Type 2 Diabetes
Binding affinity alteration	Increased		H3 receptor antagonist [81]	ADHD, AD, narcolepsy
Binding affinity	Increased		H3 receptor antagonist [82]	ADHD, AD, narcolepsy
Selectivity	Increased		Y4R agonist [99]	Obesity
Absorption and plasma levels	Increased		NS5B Thumb pocket 1 inhibitor[102, 103]	Hepatitis C
			5‐HT6R antagonists [117]	Schizophrenia
Cytotoxicity	Lowered		Anti‐leishmanial [120]	Visceral Leishmaniasis
Mycobacterium toxicity	Increased		Anti Mtb and Map mycobacterials[ 124 , 125 ]	TB

Even though commercial availability and improved synthesis methods have increased the accessibility of cyclobutane rings, it is still not the easiest moiety to include in a small‐molecule drug candidate, especially if the cyclobutyl needs to be incorporated into the core structure of the drug candidate. Therefore, it might not be applied as much compared to other motifs in drug development. However, its unique structure and properties as demonstrated in this review and continuous efforts in synthesis methods might render it a prominent member in the medicinal chemist's toolbox in the not‐too‐distant future.

List of abbreviations

AD

Alzheimer's disease

ADC

antibody‐drug conjugate

ADHD

attention deficit hyperactivity disorder

ADME

absorption distribution metabolism excretion

ADP

adenosine di‐phosphate

Ala

alanine

apoE

apolipoprotein E4

AR

adrenergic receptor

ARE

antioxidant responsive element

Arg

arginine

Asn

asparagine

Asp

aspartic acid

BACE1

β‐site amyloid precursor protein cleaving enzyme

BBB

blood‐brain barrier

cADPR

cyclic ADP‐ribose

CB1

cannabinoid 1

Cba

cyclobutylanaline

CD38

cluster of differentiation 38

Cit

citrulline

CL

clearance

CNS

central nervous system

CPE

phenotypic cell protection

CYP

cytochrome P450

D3R

dopamine receptor 3

DNA

deoxyribonucleic acid

EHMT2

euchromatic histone methyltransferase 2

EM

electron microscopy

FRET

fluorescence resonance energy transfer

Fsp³

fraction saturated carbons

FTO

fat mass and obesity associated protein

GABA_A

gamma aminobutyric acid A

GK

glucokinase

GLN

glutamine

GS

glycogen synthase

GSK‐3β

glycogen synthase kinase 3β

H₃R

histamine receptor 3

HCV

hepatitis C virus

hERG

human ether‐a‐go‐go‐related gene

HHC

hexahydrocannabinol

HLM

human liver microsomal clearance

HMT

histone methyltransferase

HTS

high throughput screening

HUVEC

human umbilical vein endothelial cells

IVMN

in vitro micronucleus

JAK1

janus kinase 1

KEAP

kelch‐like ECH associated protein

LAT

large amino acid transporter

LE

ligand efficiency

Leu

leucine

LOAD

late‐onset Alzheimer's disease

Map

mycobacterium avium paratuberculosis

MCH

melanin concentrating hormone

MCHR1

highly efficacious melanin concentrating hormone receptor 1

min

minutes

MLL

mixed lineage leukaemia

Mtb

mycobacterium tuberculosis

NAD

nicotinamide adenine dinucleotide

NPY

neuropeptide

NRF2

nuclear factor erythroid‐derived 2

NSB5

nonstructural protein 5B

OAB

overactive bladder

PDE

phosphodiesterase

PET

positron emission tomography

PK

pharmacokinetic

PL

phospholipidosis

PP

pancreatic polypeptide

PROS

PIK3CA‐related overgrowth spectrum

RLM

rat liver microsomal clearance

RNA

ribonucleic acid

RORγt

nuclear receptor retinoic acid receptor‐related orphan receptor gamma

SAR

structure activity relationship

Ser

serine

TB

tuberculosis

THC

tetrahydrocannabinol

TI

therapeutic index

Trp

tryptophan

TTK

threonine tyrosine kinase

Tyr

tyrosine

UV

ultraviolet

Val

valine

Conflict of interest

The authors declare no conflict of interest.

Biographical Information

Marnix R. van der Kolk obtained his MSc in Chemistry cum laude from Radboud University in 2021. During his master's work he performed research on synthetic organic pigments at Radboud University, and was also involved in the development of base metal catalyzed reactions in the Cardenas research group at the Universidad Autónoma de Madrid. He is currently working on a medicinal chemistry project at Radboud University.

Biographical Information

Mathilde A. C. H. Janssen received her MSc in Chemistry from Radboud University in 2020. During her master's work she focused on carbohydrate inhibitors at Radboud University as well as covalent inhibitors at Acerta Pharma B.V. (member of the AstraZeneca group). Currently, she is a PhD candidate working on an NWO TTW project; her goal is to generate novel 3D fragments by using unique synthetic chemistry methodologies such as high‐pressure chemistry and flow chemistry.

Biographical Information

Floris P. J. T. Rutjes received his PhD from the University of Amsterdam in 1993 with Profs. W. N. Speckamp and H. Hiemstra and conducted postdoctoral research in the group of Prof. K. C. Nicolaou at The Scripps Research Institute, La Jolla (USA). In 1995 he was appointed assistant professor in Amsterdam, and in 1999 he became full professor in organic synthesis at Radboud University. Awards include the Gold Medal of the Royal Netherlands Chemical Society (KNCV, 2002), the AstraZeneca award for research in organic chemistry (2003), and Most Entrepreneurial Scientist of the Netherlands (2008). He is currently director of the Institute for Molecules and Materials at Radboud University and President of the European Chemical Society (EuChemS).

Biographical Information

Daniel Blanco Ania began his independent career as a teacher at Academia Blanco (his own private school of Organic Chemistry) in 1994. He then received his PhD from Radboud University (Nijmegen) in 2009 with Prof. F. P. J. T. Rutjes and Dr. Hans W. Scheeren. After his PhD, he conducted postdoctoral research at Radboud University and at the Spanish National Research Council. In 2016, he became staff scientist at Radboud University. In 2019, he was awarded the “Golden Teacher of the Year” award of the Royal Netherlands Chemical Society (KNCV). His research interests include the use of metal‐ and organocatalysis in organic synthesis, photochemistry, flow chemistry, and crop protection.

M. R. van der Kolk, M. A. C. H. Janssen, F. P. J. T. Rutjes, D. Blanco-Ania, ChemMedChem 2022, 17, e202200020.