Identification of Transthyretin Tetramer Kinetic Stabilizers That Are Capable of Inhibiting the Retinol-Dependent Retinol Binding Protein 4-Transthyretin Interaction: Potential Novel Therapeutics for Macular Degeneration, Transthyretin Amyloidosis, and Their Common Age-Related Comorbidities

Abstract

Dissociation of transthyretin (TTR) tetramers may lead to misfolding and aggregation of proamyloidogenic monomers, which underlies TTR amyloidosis (ATTR) pathophysiology. ATTR is a progressive disease resulting from the deposition of toxic fibrils in tissues that predominantly presents clinically as amyloid cardiomyopathy and peripheral polyneuropathy. Ligands that bind to and kinetically stabilize TTR tetramers prohibit their dissociation and may prevent ATTR onset. Drawing from clinically investigated AG10, we designed a constrained congener (14) that exhibits excellent TTR tetramer binding potency, prevents TTR aggregation in a gel-based assay, and possesses desirable pharmacokinetics in mice. Additionally, 14 significantly lowers murine serum retinol binding protein 4 (RBP4) levels despite a lack of binding at that protein’s all-trans-retinol site. We hypothesize that kinetic stabilization of TTR tetramers via 14 is allosterically hindering all-trans-retinol-dependent RBP4–TTR tertiary complex formation and that the compound could present ancillary therapeutic utility for indications treated with RBP4 antagonists, such as macular degeneration.

Graphical Abstract

INTRODUCTION

In recent years, the circulating RBP4–TTR–all-trans-retinol transport complex has become a target for pharmacological intervention in ophthalmic diseases associated with enhanced accumulation of cytotoxic lipofuscin bisretinoids in the retina. Transport of all-trans-retinol (vitamin A, 1) (Figure 1) from the liver to tissues, including the retina, involves a transport complex composed of retinol binding protein 4 (RBP4) and transthyretin (TTR, thyroxine binding prealbumin).^1,2 Reports indicate that prevention of ternary complex formation between all-trans-retinol, RBP4, and TTR can be achieved via selective antagonists that compete with all-trans-retinol for binding at RBP4.^3–7 Prevention of holo-RBP4 (all-trans-retinol bound to RBP4) formation precludes complexation with TTR and subsequently leads to a lowering of serum RBP4 levels facilitated by rapid glomerular filtration of the protein due to its relatively low molecular weight (21 kDa).² Evidence suggests that pharmacological reduction of serum RBP4 levels induces a concomitant reduction in circulating all-trans-retinol that impedes ocular influx of the critical retinoid, resulting in a cessation of cytotoxic lipofuscin bisretinoid accumulation in the retina.^8,9 This approach to modulate retinal lipofuscin bisretinoid production is hypothesized to slow the progression of geographic atrophy in dry age-related macular degeneration (AMD) and Stargardt disease patients. We have shown that potent and selective RBP4 antagonists disrupt holo-RBP4–TTR ternary complex formation in vitro and significantly reduce serum RBP4 levels in vivo in rodents and nonhuman primates.^4–7,10,11 Furthermore, chronic oral administration of RBP4 antagonists in Abca4^−/− knockout mice, a model of excessive lipofuscinogenesis that recapitulates the Stargardt disease phenotype, led to a reduction in retinal cytotoxic bisretinoid accumulation with an ancillary stabilization of complement system protein expression in the retinal pigment epithelium (RPE).^9,10 Additional dosing studies in wild-type (WT) BALB/cJ mice revealed that RBP4 antagonist-induced reductions in circulating RBP4 levels correlated with partial reductions in bisretinoid precursor concentrations without disruption of visual cycle kinetics.¹⁰

Figure 1.

All-trans-retinol (vitamin A) (1) and bispecific RBP4 antagonist–TTR tetramer kinetic stabilizer (±)-2.

To date, only selective all-trans-retinol-competitive antagonists of RBP4 have been reported to block the formation of a ternary complex with TTR and lead to a reduction in circulating RBP4 levels in vivo. While selective RBP4 antagonists can be a safe and effective bisretinoid-lowering therapy for a majority of dry AMD and Stargardt disease patients, this class of compounds may potentially be counterindicated for a fraction of macular degeneration patients who are predisposed to diseases associated with TTR aggregation,^12–16 such as senile systemic amyloidosis (SSA).^17–19 Selective RBP4 antagonists would release unliganded TTR tetramer from the circulating holo-RBP4–TTR transport complex. This outcome may have serious implications as it has been previously suggested that in addition to transporting all-trans-retinol, the holo-RBP4–TTR complex may also serve to stabilize circulating TTR tetramers, as approximately 50% of TTR in circulation is bound to holo-RBP4.^20,21 Thus, the release of a significant pool of unliganded TTR tetramer induced by selective RBP4 antagonists may facilitate TTR amyloid fibril formation in susceptible individuals promoting TTR amyloidosis (ATTR)-related diseases. In an effort to address this potential liability, we recently identified and disclosed a series of potent RBP4 antagonists that also exhibit TTR tetramer kinetic stabilization activity. A standout analogue of this novel class of bispecific ligands, (±)-2, exhibited a favorable balance of in vitro RBP4 potency (RBP4 scintillation proximity assay (SPA) half-maximal inhibitory concentration (IC₅₀) = 80 nM and homogeneous time-resolved fluorescence (HTRF-fluorescence resonance energy transfer (FRET)) assay IC₅₀ = 0.250 μM) and TTR potency (TTR fluorescence polarization (FP) assay IC₅₀ = 2.85 μM), which was found to induce a robust and sustained lowering of serum RBP4 levels (>80%) upon oral dosing in mice and was also capable of decreasing the formation of high-molecular-weight TTR aggregates in an in vitro gel-based TTR aggregation assay.⁷ These data suggest that (±)-2 and its related analogues might hold promise as orally bioavailable therapeutics for the treatment of dry AMD and Stargardt disease that may also prevent potential ATTR comorbidities such as SSA in susceptible patients.

Concurrent with our recently reported bispecific RBP4 antagonist–TTR tetramer kinetic stabilizer work, we also conducted a medicinal chemistry campaign that focused on the identification of novel, potent, and selective TTR tetramer kinetic stabilizers for the treatment of ATTR-related diseases. During the course of this ancillary drug discovery campaign, we sought to determine whether our newly designed TTR tetramer stabilizers were also capable of inducing concomitant reductions in serum RBP4 levels by potentially antagonizing the formation of the holo-RBP4–TTR complex.

TTR Tetramer Kinetic Stabilizers.

TTR is a 55 kDa homotetramer composed of four β-sheet-rich, 127-residue polypeptide monomers that is largely synthesized in the liver for secretion into the blood.²² TTR tetramers possess two high-affinity binding sites for the thyroid hormone thyroxine (T4, 3) (Figure 2). However, less than 1% of circulating TTR carries T4, while another serum protein, thyroxine binding globulin (TBG), functions as its primary transporter in the blood.²² While TTR is not a primary carrier of T4 in the serum, it serves as the major transport protein for thyroxine in the central nervous system (CNS), where choroid plexus-derived TTR delivers T4 from the cerebrospinal fluid (CSF) to the choroid plexus and the brain.²³ Accumulating evidence suggests that TTR may play an auxiliary role in sequestering β-amyloid (Aβ) peptides within the CSF by promoting their clearance from the CNS to the periphery, potentially providing neuroprotective effects against Alzheimer’s disease.^24,25

Figure 2.

Thyroid hormone thyroxine (T4) (3).

The quaternary structure of TTR is of a homotetramer principally formed between two dimers.²² Formation of the structure begins with the association of two identical 127-residue β-sheet-rich polypeptide monomers that contain a total of eight antiparallel β-strands (denoted as β-strands A–H), with β-strand E presenting a short α-helix located at its terminus. The monomers associate via their edge β-strands (H-bond interactions involving the two edge β-strands H and F) yielding a dimer subunit, which further associates with a second dimer subunit in a back-to-back arrangement. The resulting dimer of dimer bears a large solvent channel, which passes between the two sheets and presents two identical C₂-symmetric binding sites for T4.²² The TTR dimer–dimer interface is relatively weak and its dissociation is the rate-limiting step in the overall TTR tetramer dissociation process.²⁶ The free dimer subunits may subsequently further dissociate into monomers that could potentially proceed to misfold and oligomerize. Oligomerization can eventually lead to aggregation and formation of toxic amyloid fibrils, which underlies the pathophysiology of ATTR.²⁶

Autosomal-dominant ATTR is a rare and progressive disease that involves severe organ damage due to the extracellular deposition of the aforementioned toxic TTR amyloid fibrils in tissues. The disease typically presents clinically as either TTR amyloid cardiomyopathy (ATTR-CM; can lead to arrhythmias, arterial fibrillation, and biventricular heart failure)^27,28 or peripheral polyneuropathy (ATTR-PN; can cause loss of sensation, tingling, numbness, or pain as well as damage to the autonomic nervous system)²⁹ and can arise from propathogenic monomers with inherited TTR mutations. Nonhereditary ATTR may emerge from wild-type TTR (WT-TTR) monomer misfolding in older individuals.³⁰ There are at least 77 TTR mutations associated with familial ATTR diseases, and these variants influence amyloidogenicity by either (1) reducing the thermodynamic stability of the TTR tetramer (i.e., the monomers are less likely to associate into a TTR tetramer and are more likely to misfold into an amyloidogenic intermediate), (2) reducing the kinetic barrier for tetramer dissociation (the TTR tetramer with the variant dissociates at a faster rate than WT-TTR with a concomitant increase in monomer aggregation rate), or (3) both thermodynamically and kinetically destabilizing the TTR tetramer.¹² The kinetically stable but thermodynamically destabilized variant V30M¹³ is predominantly associated with late-onset familial amyloid polyneuropathy (FAP) and is strongly pathogenic. The most common amyloidogenic TTR variant, V122I,¹⁴ presents at a relatively high frequency within the African-American population (approximately 3.4%) and is predominantly associated with familial amyloid cardiomyopathy (FAC). Its pathogenicity is attributed to its ability to kinetically destabilize the TTR tetramer and induce a dissociation rate that is approximately 2-fold faster than WT-TTR.³¹ The L55P mutation both thermodynamically and kinetically destabilizes tetramer formation and can aggressively promote early-onset ATTR-CM and ATTR-PN.¹⁶ Conversely, compound heterozygotes carrying a proamyloidogenic TTR mutation (e.g., V30M) and a disease-suppressing mutation that hyperstabilizes TTR tetramers, such as T119M or R104H,³² are reported to either develop a mild late-onset pathology or be completely protected against ATTR. The T119M variant kinetically stabilizes the TTR tetramer, whereas the R104H variant provides thermodynamic stability to the quaternary structure. This difference in mechanism of stabilization is crucial as the T119M variant is resistant to tetramer dissociation and aggregation and provides a greater level of protection against TTR aggregation in vitro relative to R104H. Finally, WT-TTR misfolding and aggregation that occur nongenetically with age are associated with SSA, a late-onset and prevalent form of ATTR that is estimated to affect 10–20% of individuals aged 80 years and older.¹⁹

Currently available Food and Drug Administration (FDA)-approved approaches for treating ATTR-CM and ATTR-PN include two treatments that reduce circulating TTR levels (the antisense oligonucleotide inotersen³³ and the small interfering RNA (siRNA) patisiran³⁴) and the small molecule tafamidis (vyndaqel and vyndamax, 4)^35–40 (Figure 3) that binds to and stabilizes circulating TTR tetramers. Ligand binding at the T4 sites has been shown to kinetically stabilize TTR tetramers by increasing the dissociative energy barrier of the native tetrameric state. Due to the presence of two additional T4 transport proteins (TGB and albumin), the majority of TTR in circulation is not bound to TTR. Thus, drug discovery approaches to identify T4-competitive small molecules capable of kinetically stabilizing TTR tetramers have garnered significant interest as a therapeutic option for treating ATTR. Numerous structurally diverse scaffolds in addition to tafamidis have been reported to bind at the T4 site and stabilize TTR tetramers, and representatives of this class are highlighted in Figure 3. The two most advanced small molecule TTR tetramer stabilizers to date include the aforementioned FDA-approved tafamidis and clinically investigated AG10 (acoramidis, 5).^41–43 Tafamidis has been approved for the treatment of FAP and ATTR-CM. A phase III study with 441 ATTR-CM patients showed that administration of the drug reduced the risk of death by 30% and the rate of cardiovascular-related hospitalizations by 32% compared to placebo controls.⁴⁴ TTR stabilizer AG10 was reported to be well tolerated and demonstrated near-complete stabilization of TTR in a 28 day phase II proof-of-concept trial with ATTR-CM patients presenting symptomatic chronic heart failure.⁴⁵ Approximately 30% of patients with hereditary ATTR do not respond to tafamidis,⁴⁶ indicating a need for exploring additional classes of TTR tetramer stabilizers. Phase III clinical trials with AG10 for the treatment of ATTR-CM and ATTR-PN are currently ongoing. In addition, the repurposed FDA-approved nonsteroidal anti-inflammatory drugs (NSAIDs) diflunisal (dolobid, 6)⁴⁷ and catechol-O-methyl transferase (COMT) inhibitor tolcapone (tasmar, 8)⁴⁸ are also reported to exhibit TTR tetramer stabilization activity and have been investigated for clinical efficacy against ATTR-PN.

Figure 3.

Representative examples of various reported TTR tetramer stabilizer structural classes that bind at the T4 binding site. This sample set of TTR tetramer stabilizers include tafamidis (4),³⁵ AG10 (5),⁴³ diflunisal (6),⁴⁷ iododiflunisal (7),⁴⁹ tolcapone (8),⁴⁸ benzbromarone (9), diclofenac (10),⁵⁰ N-phenyl phenoxazine (11),⁵¹⁵¹ dibenzofuran (12),⁵² and bis-aryloxime ether (13).⁵³ The following compounds highlighted in the figure above are registered drugs for various indications: tafamidis (vyndaqel and vyndamax, 4) for treating ATTR-PN, diflunisal (dolobid, 6), an NSAID also used for ATTR, tolcapone (tasmar, 8), a COMT inhibitor for treating Parkinson’s disease, benzbromarone (9), a uricosuric agent and noncompetitive inhibitor of xanthine oxidase previously shown by us to act as a potent TTR ligand,^7,54,55 and diclofenac (voltaren, 10), an NSAID to treat pain, inflammatory disorders, and dysmenorrhea.

Nature of the Holo-RBP4–TTR Protein–Protein Interaction (PPI).

RBP4 is a single domain protein that contains an N-terminal coil, eight antiparallel β-strands (denoted as β-strands A–H), and a short α-helix within close proximity to the C-terminus.⁵⁶ The all-trans-retinol binding cavity resides within the core of the protein and consists of an eight-stranded up-and-down β-barrel. Formation of the aforementioned transport complex requires that all-trans-retinol be initially bound to its respective RBP4 binding cavity as apo-RBP4 poorly associates with TTR. The reported Protein Data Bank (PDB) high-resolution holo-RBP4–TTR complex X-ray crystal structures 1QAB⁵⁷ (RBP4–all-trans-retinol–TTR ternary complex) and 3BSZ⁵⁸ (RBP4–all-trans-retinol–TTR ternary complex bound to an anti-RBP Fab) show all-trans-retinol bound within its RBP4 binding cavity in a forward-facing pose (hand-in-glove-like fit) with its lipophilic β-ionone ring projecting into a vacuous hydrophobic and phenylalanine (Phe)-rich pocket located deep within the protein. The all-trans-retinol polyene chain traverses through a narrow β-barrel core region with the pendant terminal hydroxyl group projecting slightly out of the binding pocket and toward the solvent. These X-ray crystal structures show that the exterior RBP4 loops that surround the opening of the binding pocket (loops β3–β4 and β5–β6) present an interaction contact surface for docking with TTR.⁵⁸ Binding of all-trans-retinol to RBP4 induces conformational changes within these exterior loops that facilitate recognition of TTR, allowing holo-RBP4 to bind to TTR at a twofold axis of symmetry that is perpendicular to the T4 binding sites (Figure 4).⁵⁸ Furthermore, TTR features two equivalent sites for holo-RBP4 binding, providing a holo-RBP4 to TTR tetramer stoichiometry of 2:1. Finally, the structure of the holo-RBP4–TTR ternary complex fully encapsulates all-trans-retinol and conceals its binding site. However, the T4 binding sites are not occluded.⁵⁸

Figure 4.

Quaternary structures of TTR with T4-bound, holo-RBP4, and the holo-RBP4–TTR ternary complexes. (A) Ribbon diagram of the quaternary homotetrameric structure of TTR with T4 (3) bound within its respective binding site (PDB: 2ROX).⁵⁹ The two dimer subunits are each composed of two individual monomers (labeled A (blue-green), B (orange), C (light blue), and D (pink)). The dimers are associated back-to-back and present two C₂ symmetrical T4 binding sites running through the center of the protein. The binding of T4 is shown in a ball and stick format (black). (B) Ribbon diagram of holo-RBP4 (PDB: 1RBP). Holo-RBP4 is shown as yellow and all-trans-retinol is depicted in a ball and stick format (black). (C) The holo-RBP4–TTR complex (PDB: 1QAB) colored by chain and viewed from the front. The TTR tetramer is located at the center of the complex with two holo-RBP4 molecules (one holo-RBP4 molecule is shown as yellow and the other as light green) docked at a twofold axis of symmetry that is perpendicular to the T4 binding sites. All-trans-retinol is depicted in a ball and stick format (black), which is encapsulated by the complex and its binding site is blocked. The T4 binding site is not occluded.

Both holo-RBP4 and TTR each contribute 21 amino acids to their respective PPI interface domains, which are both largely hydrophobic at their cores and hydrophilic at their peripheries.⁵⁸ The TTR amino acids Val-20, Trp79, Leu82, Ile84, Pro113, and Tyr114 from subunits B and C form a hydrophobic patch, which strongly associates with the complementary RBP4 hydrophobic patch composed of residues Trp67, Leu63, Leu64, Val69, Phe96, and Leu97. The resulting PPI involves a holo-RBP4 molecule simultaneously interacting with three TTR subunits (A–C)⁵⁸ (Figure 5). Interacting hydrophilic residues from participating TTR subunit A are closer to the border of the contact surface of the protein. Four hydrogen bonds (H-bonds) between RBP4 and TTR subunit B also contribute to the PPI.⁵⁸ Notably, further stabilization of the complex is also achieved via a H-bond interaction between the hydroxyl group of all-trans-retinol and Gly83 of the TTR subunit B (EF loop).⁵⁸

Figure 5.

PPI interaction surfaces between holo-RBP4 and TTR. (A) Close-up view of the holo-RBP4–TTR PPI (PDB: 1QAB) colored by chain. Holo-RBP4 is shown as yellow and all-trans-retinol is depicted in a ball and stick format (black) and buried in the hydrophobic cavity of the RBP4 β-barrel. The alcohol moiety points toward the EF loop of TTR (subunit B) and engages in a H-bond interaction with Gly83. The TTR monomer subunits are labeled as A (blue-green), B (orange), C (light blue), and D (pink). The image shows how the loops β3–β4 and β5–β6 of holo-RBP4 provide a surface that fits into a crevice formed by the arrangement of three TTR subunits (A–C). (B) Molecular surface representation of holo-RBP4–TTR PPI that more clearly shows holo-RBP4 making contact with three TTR subunits (A–C) (PDB: 1QAB).

The mechanism of action for selective all-trans-retinol-competitive RBP4 antagonists has been attributed to the significant and unfavorable conformational changes they induce in loops β3–β4 and β5–β6 that preclude facile association with the TTR PPI domain.⁶⁰ We questioned whether ligands that bind at the T4 site of TTR could confer unfavorable changes to the protein’s PPI domain and allosterically antagonize its ability to associate with holo-RBP4 and disrupt complex formation. Such ligands could potentially induce serum RBP4 and all-trans-retinol reductions, thereby presenting RBP4 antagonist pharmacological activity with significant TTR tetramer stabilization capability. Prior to the work reported herein, TTR has not been considered as a drug target for indications that may be treated via RBP4 antagonists, such as Stargardt disease and dry AMD. Thus, such drugs may potentially prevent the formation of amyloid fibrils in patients who, in addition to dry AMD and Stargardt disease, may also carry proamyloidogenic mutations in the TTR gene.

In our work reported herein, we initiated a structure-based drug design effort to identify a novel class of TTR tetramer kinetic stabilizers that would be assessed for their ability to both stabilize TTR tetramers in vitro and reduce circulating levels of RBP4 in vivo. We selected AG10 as a benchmark scaffold from which to develop a novel series of ligands as (1) the compound has been reported to effectively bind to and stabilize WT-TTR and the proamyloidogenic V122I-TTR variant and (2) the compound is also reported to be more potent and selective for stabilizing TTR tetramers in buffer and human serum than tafamidis, despite both compounds exhibiting similar TTR binding affinities (K_d for AG10 = 4.8 ± 1.9 nM; K_d for tafamidis = 4.4 ± 1.3 nM).^42,43 Our primary goals were to identify novel chemical matter capable of (1) retaining or improving the observed in vitro TTR tetramer binding potency observed for AG10, (2) exhibiting suitable pharmacokinetics (PK) properties to allow for adequate TTR tetramer stabilization and prevention of aggregation in vivo upon oral administration, and (3) presenting a favorable absorption distribution metabolism and excretion (ADME) profile and no limiting off-target pharmacology that would preclude drug development. In addition, recently published phase I human PK data for AG10 revealed that the major metabolic pathway for the compound involves acyl glucuronidation of its carboxylic acid.⁶¹ Thus, we also sought to include analogues within our sample set that may potentially circumvent this metabolic pathway, which is generally reported to be associated with increased incidences of idiosyncratic toxicity.

We utilized the X-ray crystallographic data PDB 4HIQ⁴³ in our structure-based drug design efforts, which shows AG10 occupying both T4 binding sites of V122I-TTR tetramer (Figure 6A). AG10 was reported to bind to V122I-TTR with negative cooperativity $\left(K_{d_1}=4.8\text{nM},K_{d_2}=314\text{nM}\right)$.⁴³ The C₂ symmetrical T4 binding sites are subdivided into inner and outer binding cavities containing three pairs of halogen binding pockets (HBP1, HBP1′, HBP2, HBP2′, HBP3, and HBP3′) that the iodine atoms of thyroxine occupy. Within the innermost region of the T4 binding cavity, residues Ser117 and Ser117′ engage in bridging H-bond interactions with a resident structural water molecule. These residues are within close proximity to the T4 inner binding cavity HBP3 and HBP3′ pockets. The lipophilic HBP2 and HBP2′ pockets are located between the inner and outer binding cavities, and the HBP1 and HBP1′ pockets are housed near the outer region of the binding cavity. The ionizable residues Lys15/Lys15′ and Glu54/Glu54′ reside near the periphery of the T4 binding site opening.

Figure 6.

X-ray crystallographic structure showing AG10 (5) bound to V122I-TTR (PDB: 4HIQ). (A) Ribbon diagram of the quaternary homotetrameric structure of TTR with AG10 bound in the T4 binding sites with both symmetry-related binding conformations shown. The TTR dimers are shown as dimer subunits A and B, and each monomer subunit of the tetramer (labeled 1–4) is shown with its secondary structural elements and colored differently. The monomer components of dimer A are shown in cyan (monomer 1) and orange (monomer 2). The monomer components of dimer B are shown in pink (monomer 3) and yellow (monomer 4). The dimers are associated back-to-back creating a large channel through the center of the tetramer that presents two C₂ symmetrical T4 binding sites. AG10 is shown in stick format (gray). (B) TTR tetramer kinetic stabilizer AG10 is shown in a ball and stick format (dark gray), with contacting residues labeled and illustrated in stick format (white). The binding pose of AG10 (5) positions the benzoic acid motif bent out of the plane with the rest of the molecule in a nearly orthogonal orientation. (C) Ligand interaction diagram of AG10 (shown as blue) derived from PDB 4HIQ, which features the HBPs with their associated localized amino acids. The diagram further illustrates key binding interactions observed in the X-ray crystallographic data, including pivotal H-bonds and salt-bridge interactions. The pyrazole head group of AG10 projects deep within the inner binding cavity of the T4 binding pocket where it engages in two H-bond interactions with Ser117 and Ser117′. The pyrazole 3,5-dimethyl groups of AG10 occupy HBP3 and HBP3′. The aliphatic chain of AG10 comes in contact with the residues lining HBP2/HBP2′. The fluorine atom of AG10 projects into a hydrophobic region of HBP1 in the outer T4 binding cavity. The carboxylic acid of AG10 is sandwiched between Lys15 and Lys15′. The H-bond interactions are depicted as blue dashed lines, with the atomic distances labeled in angstrom.

The occupancy of AG10 within the T4 site involves a forward binding mode, with the 3,5-dimethyl-1H-pyrazole group projecting deep within the inner cavity where it engages in two H-bond interactions with Ser117 and Ser117′ (Figure 6B,C).^42,43 The two pyrazole methyl groups occupy the neighboring HBP3 and HBP3′ pockets, and the n-propyloxy linker resides within the center of the binding site channel where it makes van der Waals contacts with the residues lining HBP2 and HBP2′. The fluorine atom of the benzoic acid fragment projects toward the lipophilic HBP1 pocket (Figure 6B,C),^42,43 while the carboxylic acid is positioned in close proximity to the opening of the T4 binding site where it is sandwiched between the two symmetry-related ε-amino groups of Lys15 and Lys15′. The interactions between the carboxylic acid of AG10 and Lys15/Lys15′ induce partial closure of the T4 binding site opening and effectively shield the compound from the solvent.^42,43

The reported isothermal titration calorimetry (ITC) thermograms for tafamidis and AG10 show that although both compounds present similar changes in Gibbs free energy of binding (ΔG, where ΔG = ΔH − TΔS) (ΔG for AG10 = −11.34 kcal/mol; ΔG for tafamidis = −11.34 kcal/mol), the compounds differed significantly with regard to their relative enthalpic and entropic contributions.^42,43 The binding of AG10 was found to be largely enthalpically driven (ΔH = −13.60 kcal/mol and TΔS = −2.26 kcal/mol), while tafamidis demonstrated equal enthalpic and entropic contributions to binding (ΔH = −5.00 kcal/mol and TΔS = 6.39 kcal/mol). The larger enthalpic contribution for AG10 is attributed to the H-bond interactions between its pyrazole head group and Ser117 and Ser117′, which are not present for tafamidis. Importantly, these additional H-bond interactions of AG10 with Ser117 and Ser117′ appear to mimic the direct H-bond interaction observed between these two residues in the T119M-TTR disease-suppressing variant.^42,43 The Ser117 and Ser117′ residues of T119M-TTR are within close proximity (~2.8 Å) to form an intramolecular H-bond interaction between them. This pivotal H-bond interaction helps to draw the dimer subunits closer and fortifies the aggregate molecular interactions between them. This is hypothesized to significantly increase in the T119M-TTR dimer–dimer dissociation energy barrier, thus providing superior kinetic stabilization relative to WT-TTR quaternary structure, which does not contain this key Ser117/Ser117′ H-bond interaction. Indeed, the T119M variant displays a 40-fold slower dissociation rate relative to WT-TTR. Thus, the ability of AG10 to mimic this critical Ser117/Ser117′ binding interaction of T119-TTR may explain why it demonstrates superior potency to tafamidis with regard to stabilizing TTR tetramers in buffer and human serum despite exhibiting a similar binding affinity.^42,43

Our medicinal chemistry strategy utilized the X-ray crystallographic data of PDB 4HIQ to construct a computational docking model that would assist with the design of novel scaffolds with the goal of (1) maintaining the pivotal H-bond interactions with Ser117/Ser117′ and key van der Waals interactions observed for AG10 and (2) potentially improving entropic contributions to binding by reducing the number of rotatable bonds and conformational flexibility presented by AG10. The model was generated with Molecular Operating Environment (MOE) 2020.09 (Chemical Computing Group, Inc., Montreal, Quebec, Canada, http://www.chemcomp.com) and utilized the protein structure preparation tool followed by Protonate3D, as implemented in MOE. Our initial docking analysis involved piperazine analogue 14 (Figure 7A), a novel and conformationally constrained congener of AG10, and the data generated was primarily used to determine whether the proposed compound was predicted to have good docking orientations by positioning it within the T4 binding cavity relative to AG10. The docking model images shown in Figure 7 present AG10 and 14 bound with one-half of the C₂-symmetric T4 binding site removed to visualize the binding poses and interactions for the compounds. The model predicted that 14 would exhibit good convergence with AG10 in the T4 binding site, as evidenced in the observed calculated root-mean-square deviation (RMSD) value of 1.7 Å. Figure 7B,C shows compound 14 presenting a similar binding pose as AG10 in the T4 site, with both compounds exhibiting a forward binding mode projecting their respective 3,5-dimethyl-1H-pyrazole head groups into the inner binding cavity and oriented orthogonally. The 3,5-dimethyl-1H-pyrazole of 14 is proposed to reside within the inner T4 binding cavity, where it is predicted to engage in a H-bond interaction with Ser117 and van der Waals interactions between its methyl group and HBP3 in a similar manner to AG10. Our docking model also predicts that the piperazine ring serves as a suitable replacement for the n-propyloxy linker, as it is shown to effectively place the 3,5-dimethyl-1H-pyrazole head group within the T4 inner binding cavity while also presenting the benzoic acid in an optimal position to engage in electrostatic interactions with Lys15 and Lys15′ residing near the opening of the binding site. Finally, our model also shows the fluorine atom of 14 positioned to occupy HBP1 in a similar manner to the AG10 aryl fluoride (Figure 7C).

Figure 7.

Putative binding mode of piperazine compound 14 within the T4 binding site. (A) Proposed TTR tetramer kinetic stabilizer 14, which contains a piperazine core. (B) Overlay of docked poses of piperazine 14 (magenta) and AG10 (orange) within the TTR T4 binding cavity using PDB 4HIQ (obtained from the Protein Data Bank); contacting residues are labeled and illustrated in stick format; H-bond interactions are depicted as blue dashed lines. (C) Molecular surface with the overlay of docked poses of piperazine 14 (magenta) and AG10 (orange) within the TTR T4 binding cavity using PDB 4HIQ. The protein surface is colored by hydrophobicity (green) and hydrophilicity (purple) using contact preferences; contacting residues are labeled and illustrated in stick format; H-bond interactions are depicted as blue dashed lines. (D) Legend for the ligand interaction diagrams in the binding pose using MOE. (MOE) 2020.09 (Chemical Computing Group, Inc., Montreal, Quebec, Canada, http://www.chemcomp.com).

As outlined in Figure 8, our medicinal chemistry plan focused on exploring the aforementioned computational docking results as we conducted structure–activity relationship (SAR) campaigns that focused on three key regions of lead AG10 and piperazine 14. This included exploring the SAR of (1) alternative conformationally constrained cyclic core linkers, (2) isosteres and replacements for the 3,5-dimethyl-1H-pyrazole “head group region”, and (3) variation of substituents and substitution patterns on benzoic acid “bottom-group region” appendage. Based on our docking model, we designed and synthesized illustrative examples of constrained analogues of AG10 that initially incorporated a piperazine linker and later included fused bicyclic and spirocyclic piperazine ring isosteres. We subsequently investigated the SAR of a series of analogues that featured a 3,5-dimethyl-1H-pyrazole head group with a piperazine core linker while presenting differing benzoic acid substituents. Previously reported SAR campaigns for AG10 suggest that positioning the carboxylic acid meta to its n-propyloxy linker was optimal for electrostatic binding interactions with the Lys15 and Lys15′ residues residing near the periphery of the T4 binding site opening. Thus, we fixed the carboxylic acid at the same position for our series of analogues while exploring alternative substituents and substitution patterns to the fluorine of AG10. We also explored the SAR of a small set of isosteres and replacements for the 3,5-dimethyl-1H-pyrazole head group of AG10, which projects deep into the T4 binding site and engages in H-bond interactions with Ser117 and Ser117′. Finally, we investigated the SAR effects of a carboxylic acid replacement and bioisostere that could mitigate or circumvent potential undesirable acyl glucuronidation metabolism.

Figure 8.

Medicinal chemistry work plan for the identification of novel and conformationally constrained TTR tetramer kinetic stabilizers.

CHEMISTRY

The synthetic route for a series of substituted 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoic acid analogues (14, 20a–i), highlighted in Scheme 1, begins with a palladium(0)-catalyzed amination between tert-butyl piperazine-1-carboxylate (15) and a corresponding substituted methyl bromobenzoate. The resulting aryl amines (16a–j) were N-Boc-deprotected with trifluoroacetic acid (TFA) in dichloromethane (CH₂Cl₂) at 0 °C, yielding 17a–j. Amines 17a–j were subsequently alkylated with 3-chloropentane-2,4-dione to give dione intermediates 18b–j in situ, which then underwent condensation with hydrazine hydrate (N₂H₄·H₂O) to provide the respective dimethyl pyrazoles 19a–j in one pot (the exception to this sequence was dione 18a, which was isolated). The methyl esters of 19a–j were hydrolyzed with anhydrous lithium hydroxide (LiOH), and the lithium carboxylate salts were carefully neutralized with 2 N aqueous hydrochloric acid (HCl) to yield the desired carboxylic acids 14 and 20a–i. Compound 14 was subsequently converted to the bis-hydrochloride salt 14a via treatment with a solution of 2 N HCl in diethyl ether (Et₂O).

Scheme 1.

^aReagents and conditions: (a) substituted methyl bromobenzoate, 2-dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl (X-Phos), Tris-(dibenzylideneacetone)dipalladium(0)-chloroform adduct (Pd₂(dba)₃·CHCl₃), cesium carbonate (Cs₂CO₃), 1,4-dioxane, stirred at reflux for 16 h under an atmosphere of N₂ or heated in a sealed vessel at 110 °C, 16 h; (b) TFA, CH₂Cl₂, 0 °C to room temperature (rt), 3 h; (c) 3-chloropentane-2,4-dione, N,N-diisopropylethylamine (i-Pr₂NEt), tetrahydrofuran (THF), 0 °C to rt, 16 h; (d) N₂H₄·H₂O (64–65% in H₂O), THF, rt, 1 h; (e) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 3 h; (ii) 2 N aqueous HCl; (f) 2 N HCl in Et₂O, 0 °C to rt, 3 h.

The iodo analogue of 14 (26) was prepared via the route depicted in Scheme 2. The synthesis begins with a nucleophilic substitution reaction (S_NAr) displacement between methyl 3-fluoro-4-nitrobenzoate and 15, proving intermediate 21. Aryl nitro compound 21 was readily hydrogenated to the corresponding aniline 22, which was subsequently converted to the aryl iodide via a Sandmeyer reaction involving 0.25 N aqueous sulfuric acid (H₂SO₄), H₂O, and potassium iodide (KI). The Sandmeyer reaction conditions served to produce the desired conversion to the aryl iodide as well as induce removal of the Boc group in one pot, affording 23 directly from 22. Construction of the 3,5-dimethyl-1H-pyrazole head group was conducted in the same manner outlined in Scheme 1. The desired iodo-substituted benzoic acid 26 was achieved upon hydrolysis of ethyl ester 25 and neutralization of the resulting lithium carboxylate salt with 2 N aqueous HCl.

Scheme 2.

^aReagents and conditions: (a) ethyl 3-fluoro-4-nitrobenzoate, potassium carbonate (K₂CO₃), CH₃CN, 60 °C, 16 h; (b) H₂ (1 atm pressure), 10% Pd/C, CH₃OH, rt, 24 h; (c) (i) 0.25 N aqueous H₂SO₄, sodium nitrite (NaNO₂), H₂O, 0 °C, 30 min; (ii) KI, urea, rt, 1 h; (d) 3-chloropentane-2,4-dione, i-Pr₂NEt, THF, 0 °C to rt, 16 h; (e) N₂H₄·H₂O (64–65% in H₂O), THF, rt, 1 h; (f) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 3 h; (ii) 2 N aqueous HCl.

Fused bicyclic and spirocyclic piperazine ring isostere analogues of 14 (32, (±)-38, and 44) were prepared and their corresponding syntheses are shown in Scheme 3. The preparation of each analogue of this sample set began with a palladium(0)-catalyzed amination between the corresponding N-Boc-protected piperazine ring isostere (tert-butyl (3aR,6aS)-hexahydropyrrolo[3,4-c]pyrrole-2(1H)-carboxylate (27), (±)-tert-butyl 2,7-diazaspiro[4.4]nonane-2-carboxylate ((±)-33), and tert-butyl 2,6-diazaspiro[3.3]heptane-2-carboxylate (39)) and methyl 3-bromo-4-fluorobenzoate. TFA-promoted N-Boc deprotection followed by 3,5-dimethyl-1H-pyrazole construction was conducted using the same three-step process as outlined in Scheme 1. LiOH-mediated hydrolysis of the corresponding methyl esters followed by neutralization of the lithium carboxylate salts with 2 N aqueous HCl provided the desired piperazine ring isostere analogues 32, (±)-38, and 44.

Scheme 3.

^aReagents and conditions: (a) tert-butyl (3aR,6aS)-hexahydropyrrolo[3,4-c]pyrrole-2(1H)-carboxylate (27), (±)-tert-butyl 2,7-diazaspiro[4.4]-nonane-2-carboxylate ((±)-33), or tert-butyl 2,6-diazaspiro[3.3]heptane-2-carboxylate (39), methyl 3-bromo-4-fluorobenzoate, X-Phos, Pd₂(dba)₃·CHCl₃, Cs₂CO₃, 1,4-dioxane, sealed vessel at 110 °C, 16 h; (b) TFA, CH₂Cl₂, 0 °C to rt, 3 h; (c) 3-chloropentane-2,4-dione, i-Pr₂NEt, THF, 0 °C to rt, 16 h; (d) N₂H₄·H₂O (64–65% in H₂O), THF, rt, 1 h; (e) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 3 h; (ii) 2 N aqueous HCl.

The previously described dione intermediate 3-(4-(2,4-dioxopentan-3-yl)piperazin-1-yl)-4-fluorobenzoate (18a) also afforded a handle for the construction of isoxazole 46 (Scheme 4), an analogue bearing a 3,5-dimethyl-1H-pyrazole head group isostere. Condensation of 18a with hydroxyl amine hydrochloride (NH₂OH·HCl) provided methyl 3-(4-(3,5-dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoate (45), which was hydrolyzed with anhydrous LiOH followed by neutralization with 2 N aqueous HCl to give the final product 3-(4-(3,5-dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid (46).

Scheme 4.

^aReagents and conditions: (a) NH₂OH·HCl, CH₃OH, rt, 16 h; (b) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 1 h; (ii) 2 N aqueous HCl.

Scheme 5 presents the routes starting with aryl piperazine intermediate 17a that enabled the preparation of both tetrazole head group analogue 48 and 2-trifluoromethylphenyl head group analogue 50. Treatment of 17a with cyanogen bromide (CNBr) at 0 °C in THF in the presence of N,N-diisopropylethylamine (i-Pr₂NEt) afforded an intermediary cyanopiperazine, which underwent cycloaddition with sodium azide (NaN₃) in N,N-dimethylformamide (DMF) at 120 °C to give tetrazole 47. Subsequent LiOH saponification of the methyl ester of 47 and neutralization of the lithium carboxylate salt with2 N aqueous HCl gave the desired tetrazole 48. The preparation of 2-trifluoromethylphenyl analogue 50 began with a palladium(0)-catalyzed amination between 17a and 1-bromo-2-(trifluoromethyl)benzene to give methyl ester intermediate 49. The aforementioned saponification and neutralization process was conducted with 49 affording the desired 2-trifluoromethylphenyl head group analogue 50.

Scheme 5.

^aReagents and conditions: (a) (i) CNBr, i-Pr₂NEt, THF, 0 °C to rt, 1 h, (ii) ammonium chloride (NH₄Cl), NaN₃, DMF, 120 °C, 12 h; (b) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 1 h; (ii) 2 N aqueous HCl; (c) 1-bromo-2-(trifluoromethyl)benzene, X-Phos, Pd₂(dba)₃·CHCl₃, Cs₂CO₃, 1,4-dioxane, sealed vessel, 110 °C, 16 h; (d) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 1 h; (ii) 2 N aqueous HCl.

The preparation of the homologated variant of 14, analogue 52, is shown in Scheme 6. A sodium triacetoxyborohydride (NaBH(OAc)₃)-mediated reductive amination between amine 17a and 3-methyl-1H-pyrazole-4-carbaldehyde provided methyl ester 51, which upon saponification and subsequent 2 N aqueous HCl neutralization provided the desired homologated analogue 52.

Scheme 6.

^aReagents and conditions: (a) 3-methyl-1H-pyrazole-4-carbaldehyde, NaBH(OAc)₃, acetic acid (HOAc), 1,2-dichloroethane, 60 °C, 12 h; (b) (i) anhydrous LiOH, CH₃OH, THF, H₂O, rt, 3 h; (ii) 2 N aqueous HCl.

Replacement of the carboxylic acid of 14 with a primary carboxamide (53) and the bioisostere tetrazole (54) was also explored (Scheme 7). Production of primary carboxamide 53 involved amide bond formation between ammonium chloride and 14 using the peptide coupling agent O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU). The bioisostere tetrazole 54 was furnished from 53 using tetrachlorosilane-sodium azide (NaN₃). Tetrachlorosilane-mediated dehydration of primary carboxamide 53 afforded the corresponding nitrile in situ, which underwent 1,3-dipolar cycloaddition with NaN₃ to afford tetrazole 54 in one pot.

Scheme 7.

^aReagents and conditions: (a) NH₄Cl, HBTU, i-Pr₂NEt, DMF, rt, 18 h; (b) NaN₃, tetrachlorosilane, CH₃CN, 80 °C, 18 h.

RESULTS AND DISCUSSION

Structure–Activity Relationships.

Compound binding potency (IC₅₀ values) of novel compounds for unliganded TTR tetramers isolated from human plasma was measured using an established fluorescence polarization assay designed for TTR kinetic stabilizers.⁴¹ The assay measured competitive displacement of a fluorescent probe (fluorescein isothiocyanate (FITC) coupled to a diclofenac analogue via a PEGylated linker) from TTR.

The less favorable entropic binding energy observed with the reported ITC thermogram of AG10 relative to tafamidis (AG10 TΔS = −2.26 kcal/mol; tafamidis TΔS = 6.39 kcal/mol)^42,43 may potentially be attributed to the conformational flexibility of the n-propyloxy core linker, which contains five rotatable bonds. Thus, our initial goal was to prepare a conformationally constrained congener of AG10 by installing a piperazine ring as a replacement for the n-propyloxy linker. We hoped to potentially further improve potency relative to AG10 by maintaining its key enthalpic contributions to binding while ameliorating a putative entropic penalty by reducing the number of rotatable bonds within the molecule’s linker region. Indeed, piperazine 14 did exhibit excellent potency at TTR (TTR FP IC₅₀ = 220 nM) that was approximately 2-fold better than tafamidis (TTR FP IC₅₀ = 410 nM) and comparable to that of AG10 (TTR FP IC₅₀ = 160 nM) (Table 1).

Table 1.

TTR FP and RBP4 SPA In Vitro Assay Data for Piperazine-Containing TTR Tetramer Kinetic Stabilizers Bearing Various Benzoic Acid Substituents (Bottom-Group Region SAR)^a

Compound	Benzoic Acid Bottom Group Motif (R)	TTR FPa IC50 (μM)d	RBP4 SPAb IC50 (μM)d	RBP4-TTR HTRFc IC50 (μM)d
Tafamadis (4)	---	0.41	>30	0.981
AG10 (5)	---	0.16	>30	0.512
14		0.22	>3	0.845
20a		0.78	>30	NDe
20b		0.30	>3	0.583
20c		0.21	>3	NDe
20d		0.26	>3	3.56
20e		0.20	>3	1.74
20f		0.55	>3	NDe
20g		1.44	>3	NDe
20h		0.62	>3	NDe
20i		0.60	>3	24.01
26		0.19	ND	2.65

^aIC₅₀ values for the TTR FP assay obtained in the presence of a fixed, 25 μM concentration of fluorescein isothiocyanate (FITC)-coupled TTR FP probe.

^bIC₅₀ values for the RBP4 SPA assay obtained in the presence of a fixed, 10 nM concentration of ³H-retinol.

^cIC₅₀ values for the HTRF assay obtained in the presence of 1 μM concentration of retinol.

^dFor compounds tested multiple times (more than twice), the IC₅₀ data is represented as the mean ± standard deviation (SD). For those compounds that were only tested twice, the IC₅₀ data is shown as the mean of two independent experiments and not as the mean ± standard deviation.

^eND, not determined.

With 14 in hand, we next sought to explore the SAR effects of varying the substituents and substitution patterns of the benzoic acid bottom-group motif. We prepared a focused sample set of follow-up analogues to 14 that replaced its para-fluorine atom with alternative groups that might better occupy the neighboring HBP1 pocket, stabilize a putative forward binding pose, and potentially further improve potency (20a–e). The HBP1 pocket, and its twofold symmetry-related HBP1′ pocket, resides near the outer binding cavity and is occupied by one of the iodine atoms of T4. They are hydrophobic and are composed of residues Leu17, Thr106, and Val121 as well as the methylene groups of Lys15 and Glu54. Our docking model predicted that the fluorine of 14 would project toward HBP1 in a similar manner as the fluorine of AG10. Therefore, we sought to initially explore this region of the scaffold to see if substituents other than fluorine could lead to improved potency via better HBP1 occupancy and more robust van der Waals binding interactions. To that end, our initial sample set included substituents that were larger and more lipophilic than fluorine (Table 1). With the exception of des-fluoro analogue 20a (TTR FP IC₅₀ = 780 nM), all of the analogues of this sample set were well tolerated and exhibited potencies within a twofold range of 14. Indeed, para-substituted analogues 20b–e are equipotent to 14. We also replaced the fluorine of 14 with the larger, more lipophilic, and halogen bond-donating iodine that is featured on T4 (26), and this analogue was also found to be an active TTR ligand within our portfolio, exhibiting comparable potency relative to the parent fluoro compound 14 and AG10 (26 TTR FP IC₅₀ = 190 nM). We plan to follow up on this emerging SAR and test future analogues presenting larger and more sterically bulky substituents (e.g., isopropyl, tert-butyl, cyclopropyl, cyclobutyl, etc.) and polar substituents (i.e., oxetane, nitrile, etc.) to further probe the capacity of HBP1 and to determine if such substituents can lead to more potent scaffolds within our series.

The improved potency observed for several analogues within our series relative to tafamidis may suggest the compounds’ benefit from increased enthalpic contributions to binding via putative H-bonds between their respective pyrazole head groups and Ser117/Ser117′ in a similar manner as AG10, which would be consistent with the binding pose of 14 predicted by our computational docking model. The lack of further improvement in potency for the series relative to AG10 may be attributed to several factors. For example, potential entropic benefits for binding gained may be slightly offset by subtle adverse changes to enthalpic contributions. It is possible that the more electron-rich amino pyrazole head group of our series presents a slightly weaker H-bond-donating group relative to the pyrazole of AG10, which could affect the strength of the Ser117/Ser117′ H-bond interaction. Furthermore, replacing the aryl ether of AG10 with the aryl amine for our compounds may alter the pK_a of the carboxylic acid and modulate the strength of its interactions with Lys15/Lys15′. Additionally, our piperazine-containing scaffolds may be unable to attain a fully suitable bioactive conformation to allow for the optimal pyrazole head group and/or carboxylic acid bottom-group binding interactions due to the reduced flexibility of the core. The increased rigidity imposed by the piperazine core may actually not confer a significant entropic benefit and may prevent optimal occupancy required for fortified H-bond, van der Waals, and hydrophobic effect contributions. Alternatively, it may also be possible that the potencies exhibited by AG10 and 14 and 26 may simply represent the upper limit for ligand binding affinity at the T4 site.

We next investigated whether moving the fluorine of 14 from the para position relative to the carboxylic acid to the corresponding ortho and meta positions would have an effect on TTR potency (20f–h). Interestingly, the SAR for this group demonstrated a preference for fluorine substitution para to the carboxylic acid. The ortho-fluorinated analogue 20f (TTR FP IC₅₀ = 550 nM) exhibited a >2-fold loss in potency relative to 14, while the alternate ortho-substituted fluorinated analogue 20g exhibited a greater loss in potency (~6.5-fold) (TTR FP IC₅₀ = 1.44 μM). The meta-fluoro analogue 20h (TTR FP IC₅₀ = 620 nM) was nearly equipotent with 20f. The relatively modest decrease in potency observed for 20f and 20h may be attributed to their respective fluorine atoms being unable to optimally project into and fully occupy the HBP1′ or HBP1 pocket, potentially leading to weaker van der Waals interactions and destabilized binding poses. Contrarily, the substitution pattern of 20g may potentially project the fluorine atom into an area that does not allow it to occupy either HBP1 or HBP1′ and/or engage in any favorable binding interactions, thus leading to the more significant loss in potency exhibited. Finally, we prepared and tested the disubstituted analogue 20i that incorporated a methyl group ortho (2-position) to the carboxylic acid, which could serve as a potential blocking group to hinder a possible acyl glucuronidation metabolic pathway. Gratifyingly, the compound was active and exhibited good potency (TTR FP IC₅₀ = 600 nM). As previously stated, uridine 5′-diphospho-glucuronosyltransferase (UGT) acyl glucuronidation has been identified as the predominant metabolic pathway of AG10.⁶¹ As acyl glucuronidation is reported to be associated with an increased incidence of idiosyncratic toxicity,⁶² the diminishment or elimination of this metabolic pathway should be considered during the early stages of drug discovery to potentially remove or hinder the probability of encountering this liability. Although AG10 has been found to be safe and well tolerated clinically, we wanted to investigate early-stage analogues that may potentially hinder an acyl glucuronidation metabolic pathway for our series. Thus, the introduction of a methyl group within close proximity to the nucleophilic carboxylate moiety was anticipated to sterically encumber and reduce its accessibility to UGT and the UDP-glucuronic acid cofactor, potentially impairing or prohibiting acyl glucuronidation.⁶³ Future metabolite profiling and structural elucidation studies will be conducted to investigate whether 14 and/or 20i are subject to acyl glucuronidation in vivo and if there are significant differences observed with regard to the extent of this type of metabolism between the two analogues.

In addition to TTR binding, we also tested our compounds, together with tafamidis and AG10, in the in vitro RBP4 scintillation proximity (SPA) assay, which is used to measure binding potency of all-trans-retinol-competitive ligands at RBP4. Neither tafamidis, AG10, nor any of our novel TTR ligands exhibited activity at RBP4 in this assay (RBP4 SPA IC₅₀ for all compounds tested >3 μM). We also tested our most potent TTR ligands in the RBP4–TTR HTRF assay,^3–7 which is typically used to determine the all-trans-retinol-competitive RBP4 antagonist potency for inhibiting the holo-RBP4–TTR complex formation. Interestingly, despite a lack of activity in the RBP4 SPA assay, we did observe activity in the RBP4–TTR HTRF assay for some of the compounds tested from this portfolio. Tafamidis, AG10, 14, 20b, 20d, 20e, and 26 were all found to exhibit single-digit to submicromolar IC₅₀ potency values in this assay, providing evidence that the compounds are also capable of hindering holo-RBP4–TTR complexation. Interestingly, compound 20i exhibited very weak activity in this assay, despite presenting TTR potency that is comparable to the aforementioned HTRF-active compounds.

We next explored the SAR effects of a focused sample set of three piperazine ring isostere core linker analogues (Table 2). The (3aS,6aS)-octahydropyrrolo[3,4-c]pyrrole fused bicyclic analogue 32 exhibited an order of magnitude loss in potency relative to 14 (TTR FP IC₅₀ = 2.82 μM), while racemic 2,7-diazaspiro[4.4]nonane (±)-38 was found to be 5-fold less potent (TTR FP IC₅₀ = 1.08 μM). 2,6-Diazaspiro[3.3]heptane 44 was inactive at TTR. None of the piperazine isostere analogues demonstrated activity in the in vitro RBP4 SPA assay. The diminishment in potency or complete loss of activity for these compounds could be attributed to either the longer lengths or unfavorable geometries (or both) of the linkers, which prohibited the formation of compatible binding poses within the T4 site. Due to their poor binding at TTR, these compounds were not profiled for activity in the RBP4–TTR HTRF assay.

Table 2.

TTR FP and RBP4 SPA In Vitro Assay Data for Conformationally Constrained Piperazine Isostere TTR Tetramer Kinetic Stabilizer Analogues (Core Scaffold Region SAR)

compound	TTR FPa IC50 (μM)c	RBP4 SPAb IC50 (μM)c
14	0.22	>3
32	2.82	>3
(±)−38	1.08	>3
44	>30	>3

^aIC₅₀ values for the TTR FP assay obtained in the presence of a fixed, 25 μM concentration of fluorescein isothiocyanate (FITC)-coupled TTR FP probe.

^bIC₅₀ values for the RBP4 SPA assay obtained in the presence of a fixed, 10 nM concentration of ³H-retinol.

^cFor compounds tested multiple times (more than twice), the IC₅₀ data is represented as the mean ± standard deviation. For those compounds that were only tested twice, the IC₅₀ data is shown as the mean of two independent experiments and not as the mean ± standard deviation.

A concise exploration of 3,5-dimethyl-1H-pyrazole head group SAR for analogue 14 was also conducted (46, 48, 50, and 52) (Table 3). Previous reports indicate very tight SAR for this region as exchanging the methyl groups of the 3,5-dimethyl-1H-pyrazole of AG10 (5) with ethyl groups or methylation of one of the pyrazole nitrogen atoms was not well tolerated.²⁹ Thus, we prepared a focused sample set that was designed to further probe the nature of the T4 inner binding cavity. Replacement of the 3,5-dimethyl-1H-pyrazole of 14 with a 3,5-dimethylisoxazole ring (46) was not well tolerated and leads to a >20-fold loss in potency (TTR FP IC₅₀ = 5.3 μM). This observation may be attributed to the installation of a poorer H-bond-accepting isoxazole ring system, which may have led to significantly weaker putative H-bonds with Ser117 and Ser117′ relative to 14. Incorporation of a tetrazole group (48) led to a complete loss in TTR activity, which may be attributed to (1) the contrasting polarity of the tetrazole with the hydrophobic macromolecular interior of the inner T4 binding site and/or (2) the lack of pendant methyl groups to occupy the adjacent HBP3 and HBP3′ pockets. Interestingly, replacing the 3,5-dimethyl-1H-pyrazole of 14 with a 2-trifluoromethylphenyl head group (50) also led to a diminishment in potency (TTR FP IC₅₀ = 1.7 μM), though not as severe as was observed for isoxazole 46 and tetrazole 48. The 2-trifluoromethylphenyl head group of 50 is significantly more hydrophobic than the 3,5-dimethyl-1H-pyrazole, and it is incapable of H-bond interactions with Ser117 and Ser117′. Furthermore, the X-ray crystal structure PDB 4HIQ indicates that there is insufficient space between the symmetry-related side chains of Ser117 to accommodate the size of the 2-trifluoromethylphenyl ring. We speculate that the relatively moderate loss in potency observed for 50 may be due to an alternative binding pose similar to that reported for TTR stabilizer 12,⁵¹⁵¹ which could involve the benzoic acid of 50 projecting inward and toward HPB3 and Ser117, while the 2-trifluoromethylphenyl ring is positioned closer to the opening of the binding site where it may engage HPB2 and/or HBP3. It is unclear at this time as to why this compound exhibits moderate potency and we plan to obtain an X-ray crystal structure of it cocrystallized with TTR to determine its mode of binding. The 3-methyl-1H-pyrazol-4-yl analogue 52 was found to be inactive.

Table 3.

TTR FP and RBP4 SPA In Vitro Assay Data for TTR Tetramer Kinetic Stabilizer Analogues Bearing 3,5-Dimethyl-1H-pyrazole Replacements (Head Group Region SAR)

Compound	Head Group (R)	TTR FPa IC50 (μM)c	RBP4 SPAb IC50 (μM)c
14		0.22	>3
46		5.3	>3
48		>30	>3
50		1.7	>3
52		>30	>30

^aIC₅₀ values for the TTR FP assay obtained in the presence of a fixed, 25 μM concentration of fluorescein isothiocyanate (FITC)-coupled TTR FP probe.

^bIC₅₀ values for the RBP4 SPA assay obtained in the presence of a fixed, 10 nM concentration of 3H-retinol.

^cFor compounds tested multiple times (more than twice), the IC₅₀ data is represented as the mean ± standard deviation. For those compounds that were only tested twice, the IC₅₀ data is shown as the mean of two independent experiments and not as the mean ± standard deviation.

We then examined two replacements for the carboxylic acid of 14, which included primary carboxamide 53 and the bioisostere tetrazole 54. Carboxamide 53 exhibited good potency and was approximately 3-fold less potent than 14. This is presumably due to the fact that the primary carboxamide, although capable of serving as a H-bond donor and acceptor that can form strong charge-reinforced H-bonds with Lys15/Lys15′ and Glu54/Glu54′, is nonionizable and unable to engage in stronger electrostatic interactions that the carboxylic acid of 14 is capable of. The tetrazole of 54 is ionizable at physiological pH and serves as a suitable bioisostere of the carboxylic acid, thus likely allowing it to participate in similar electrostatic interactions with Lys15/Lys15′, which may explain why it is equipotent with 14. As an additional beneficial attribute, the tetrazole of 54 may also serve to potentially hinder acyl glucuronidation. Interestingly, despite conferring TTR potency and selectivity that is comparable to 14 (TTR FP IC₅₀ = 0.26 nM; RBP4 SPA IC₅₀ > 3 μM), tetrazole 54 exhibited an approximate 10-fold loss in potency in the RBP4–TTR HTRF assay (RBP4–TTR HTRF IC₅₀ = 9.58 μM) (Table 4).

Table 4.

TTR FP and RBP4 SPA In Vitro Assay Data for Primary Carboxamide and Tetrazole TTR Tetramer Kinetic Stabilizer Analogues (Bottom-Group Region SAR)

compound	TTR FPa IC50 (μM)d	RBP4 SPAb IC50 (μM)d	RBP4–TTR HTRFc IC50 (μM)d
14	0.22	>3	0.845
53	0.67	>3	NDe
54	0.26	>3	9.58

^aIC₅₀ values for the TTR FP assay obtained in the presence of a fixed, 25 μM concentration of fluorescein isothiocyanate (FITC)-coupled TTR FP probe.

^bIC₅₀ values for the RBP4 SPA assay obtained in the presence of a fixed, 10 nM concentration of ³H-retinol.

^cIC₅₀ values for the HTRF assay obtained in the presence of a 1 μM concentration of retinol.

^dFor compounds tested multiple times (more than twice), the IC₅₀ data is represented as the mean ± standard deviation. For those compounds that were only tested twice, the IC₅₀ data is shown as the mean of two independent experiments and not as the mean ± standard deviation.

^eND, not determined.

We next evaluated our most potent compounds from our SAR campaign for their corresponding ADME properties. The in vitro pharmacological profiles presented in Table 5 show that all of the compounds possess excellent kinetic aqueous solubility in phosphate-buffered saline (PBS) (pH = 7.4), excellent microsomal stability with low intrinsic clearance (CL_int) values across multiple species, and suitable % plasma protein binding (PPB) values that largely indicate relatively low fraction unbound in human, rat, and mouse. Compound 14 emerged from this collection as the lead because it was among the most potent in the series that presented the best cytochrome P450 (CYP) inhibition profile. Further analysis of 14 revealed that the compound did not demonstrate any limiting off-target pharmacology at the human ether-a-go-go (hERG) channel (hERG IC₅₀ > 30 μM; five-point dose–response (n = 2), electrophysiological patch-clamp (QPatch) assay) or at the nuclear peroxisome proliferator-activated receptor-γ receptor (PPARγ) (PPARγ EC₅₀ > 30 μM). PPARγ agonists may potentially induce adverse effects, which include risk for fluid retention, weight gain, bone loss, and congestive heart failure.⁶⁴ Our previously reported RBP4 antagonist BPN-14136 (structure not shown)^5,10,11 was found to exhibit modest agonist activity for the receptor (PPARγ IC₅₀ = 3.6 μM in the agonist-induced corepressor NCoR release assay). Due to the fact that 14 shares some structural complementarity with BPN-14136 and with reported PPARγ agonists, we surveilled our novel TTR tetramer kinetic stabilizer series for this undesirable off-target pharmacology.

Table 5.

In Vitro ADME Profile for Selected TTR Tetramer Kinetic Stabilizer Compounds

compound	TTR FP IC50 (μM)	solubility (μM)a	microsomal CLint (mL/(min mg))b			liver microsomal stability (% remaining at 30 min)c				predicted half-lifed (H, R, M, monkey)	CYP inhibition (% inhibition at 10 μM) 2C9, 2C19, 2D6, 3A4	%PPBe
			H	R	M monkey	HLM	RLM	MLM	monkey LM			H	R	M
14	0.22	184.5	<0.0231			93	98	102	94	>60 min for all species	02C9—10.4% 2C19—3.1% 2D6—8.1% 3A4—(−)2.2%	83	97	93
20b	0.30	200	<0.0231			108	101	102	102	>60 min for all species	2C9—1.4% 2C19—4.2% 2D6—8.0% 3A4—19.6%	75	78	ND
20d	0.21	200	<0.0231			101	101	97.5	96.7	>60 min for all species	2C9—17.1% 2C19—2.3% 2D6—5.9% 3A4—14.8%	86	91	ND
20c	0.26	200	<0.0231			97.8	96.6	96	100	>60 min for all species	2C9—56.2% 2C19—0.06% 2D6—7.34% 3A4—(−)5.2%	97	97	ND
20e	0.20	200	<0.0231			103	104	98	100	>60 min for all species	2C9—25.2% 2C19—1.85% 2D6—10.0% 3A4—19.5%	91	89	ND
20i	0.60	200	<0.0231			90	105	96	95.2	>60 min for all species	2C9—1.1% 2C19—(−)3.0% 2D6—9.2% 3A4—31.7%	87	93	ND
54	0.26	200	H = 0.0184; R, M, monkey <0.0231			71.4	98.2	91.5	95.7	H = 75.4 min; R, M, monkey <60 min	2C9—26.3% 2C19—(−)4.8% 2D6—0.8% 3A4—18.9%	92	94	ND

^aKinetic solubility measured in PBS (pH = 7.4). When the observed mean solubility is greater than 200 μM, the mean value is adjusted to the maximum assay concentration, which is 200 μM.

^bMicrosomal intrinsic clearance (CL_int); H, human; R, rat; M, mouse; cyno, cynomolgus monkey.

^cLiver microsomal metabolic stability, % of parent drug remaining after a 30 min incubation in the presence of the microsomes; HLM, human liver microsomes; RLM, rat liver microsomes; MLM, mouse liver microsomes; monkey LM, monkey liver microsomes. %Remaining values shown as higher than 100% may be potentially attributed to error.

^dWhen the calculated half-life is longer than the duration of the experiment, the half-life is expressed as greater than the longest incubation time.

^e%PPB, plasma protein binding; H, human; R, rat; M, mouse.

Analogue 14 Inhibits Formation of High-Molecular-Weight Forms of TTR.

The ability of TTR ligands to act as kinetic stabilizers of TTR tetramers can be assessed in vitro using a low pH-induced TTR aggregation sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) assay.⁶⁵ A 72 h incubation of TTR tetramers at pH = 4.0 initiates its dissociation into monomeric intermediates that misassemble and oligomerize into amyloid fibrils and other high-molecular-weight forms.^65,66 To confirm the ability of analogue 14 to act as a kinetic stabilizer of TTR tetramers, we tested its ability to suppress low pH-mediated TTR aggregate formation using the previously published protocol. Tafamidis was used as a positive control in the aggregation experiments. As shown in Figure 8A, following 72 h of incubation of TTR tetramers with dimethyl sulfoxide (DMSO) at pH = 4 at 37 °C, the amount of high-molecular-weight forms of TTR is greatly increased in comparison to the DMSO control incubated for 72 h at pH = 7.5. Consistent with its ability to act as a kinetic TTR stabilizer, treatment with analogue 14 (50 μM) as well as with tafamidis (50 μM), significantly inhibited the formation of high-molecular-weight forms of TTR (Figure 9A). Higher intensities of the TTR monomer and dimer bands in samples treated with 14 and tafamidis when compared to that of DMSO correlated with a corresponding reduction in TTR aggregates induced by 14 and tafamidis. Quantification of Western blot band intensities established a 3.1-fold reduction in the amount of aggregates in the presence of 14, while a 3.0-fold reduction was induced by tafamidis (Figure 9B). Significant increase in the dimer band intensities and appreciable increase in the intensity of TTR monomer bands conferred by 14 and tafamidis were associated with inhibition of the low pH-induced TTR aggregate formation (Figure 9C,D). The results of the conducted aggregation experiments confirmed that 14 can act as a TTR tetramer kinetic stabilizer.

Figure 9.

Analogue 14 reduces the formation of high-molecular-weight TTR forms in the acid-induced aggregation assay. TTR protein (5 μg) was aggregated using acetate buffer (pH = 4.0) and incubated for 72 h at 37 °C. TTR tetramer concentration during the incubation was 9 μM. After incubation in the presence of DMSO, 50 μM tafamidis, and 50 μM 14 and cross-linking with glutaraldehyde, samples were subjected to SDS-PAGE followed by Western blot analysis of TTR. The representative blot of at least three independent experiments is presented (A). Bar graphs represent the pixel volume means ± SD of the scanned bands on the immunoblots in arbitrary units for TTR high-molecular-weight aggregates (B), dimers (C), and monomers (D). Statistical significance was determined by one-way analysis of variance (ANOVA) with the Holm–Š_idák post hoc test; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001; ****, p ≤ 0.0001 compared to TTR aggregation + DMSO group (pH 4.0); ^#, p ≤ 0.05; ^##, p ≤ 0.01; ^###, p ≤ 0.001; ^####, p ≤ 0.0001 compared to TTR without aggregation group (pH = 7.5).

In Vivo Activity: PK Characteristics of 14 in Mice.

The PK properties of 14 were determined via studies conducted with CD-1 male mice (2 mg/kg intravenous (IV) and 5 mg/kg oral (PO)) (Table 6). The compound exhibited a favorable plasma clearance (0.354 L/(h kg)) and a good half-life of 5.08 h. Rapid oral absorption was observed (T_max = 0.42 h) with a maximal concentration (C_max) of 1563 ng/mL. Compound 14 presented a low volume of distribution (V_ss = 2.19 L/kg), good overall exposure (area under the curve (AUC)_INF was 16 073 h ng/mL), and excellent oral bioavailability (%F = 103%). The observed slightly >100% oral bioavailability for 14 may be the result of its low plasma clearance and potential nonlinear pharmacokinetics over the dose range tested or it may be due in part to enterohepatic recycling.

Table 6.

In Vivo PK Data for Analogue 14 Following IV and PO Administration in CD-1 Mice^a

route	dose (mg/kg)	C0b (ng/mL)	CLc (L/(h kg))	t1/2d (h)	Vsse (L/kg)	AUClastf (h ng/mL)	AUCINFg (h ng/mL)	%Fj
IV	2	1622 (267)	0.354 (0.139)	5.08 (0.814)	2.18 (0.585)	6214 (2321)	6242 (2303)	NA
route	dose (mg/kg)	Cmaxh (ng/mL)	Tmaxi (h)	t1/2d (h)	Vsse (mL/kg)	AUClastf (h ng/mL)	AUCINFg (h ng/mL)	%F j
PO	5	1563 (115)	0.42 (0.14)	5.38 (0.624)	NA	16 040 (2778)	16 073 (2783)	103 (17.8)

^aData are represented as the mean with standard deviation in parentheses (mean (SD)). Dosing groups consisted of three drug naïve adult male CD-1 mice. IV administration: test article was administered at the 2 mg/kg dose; test article vehicle = 3% dimethylacetamide (DMA)/45% poly(ethylene glycol) 300 (PEG300)/12% ethanol/40% sterile water; PO administration: test article was administered at the 5 mg/kg dose, vehicle = 2% Tween 80 in 0.9% saline.

^bObserved initial concentration of compound in blood at time zero.

^cTotal body clearance.

^dApparent half-life of the terminal phase of elimination of compound from blood.

^eVolume of distribution at steady state.

^fArea under the blood concentration versus time curve from 0 to the last time point that the compound was quantifiable in blood.

^gArea under the blood concentration versus time curve from 0 to infinity.

^hMaximum observed concentration of compound in blood.

ⁱTime of maximum observed concentration of compound in blood.

^jBioavailability; F = (AUC_INFpo × dose_iv)/AUC_INFiv × dose_po.

In Vivo Activity: RBP4-Lowering and PK–Pharmacodynamics (PD) Correlations of Analogue 14 in Mouse.

Taking into account that the levels of serum RBP4 depend on formation of the RBP4–TTR–all-trans-retinol complex in circulation, we wanted to establish that a selective TTR tetramer binding ligand can allosterically antagonize the retinol-dependent RBP4–TTR interaction and potentially be capable of reducing circulating RBP4 levels in vivo. We evaluated the effect of a single oral dose of 14 on dynamics of serum RBP4 levels in mice (Figure 10). After a single 25 mg/kg oral dose of 14, a maximal 66% reduction in serum RBP4 levels was observed 2 h following the dose administration (Figure 10A). Given that T4 and synthetic TTR ligands may stimulate the secretion of certain TTR forms,⁶⁷ we assessed the effect of oral administration of 14 on circulating levels of transthyretin. No changes in serum TTR levels in response to 14 dosing were noted (data not shown). The dynamics of serum RBP4 changes induced by 14 demonstrated a very good correlation between the presence of the compound in circulation after oral dosing (Figure 10B) and a reduction in serum RBP4. The maximal RBP4 reduction seen at the 1 and 2 h time points correlated very well with rapid oral absorption (Table 6) that led to the high concentration of 14 in the blood at time points of the maximal serum RBP4 reduction. Compound clearance by 24 h matches well with the lack of the PD response at this time point.

Figure 10.

Pharmacokinetic and pharmacodynamic properties of 14 in mice. (A) Serum RBP4 levels following a single 25 mg/kg oral administration of 14. (B) Blood compound levels following administration of a single oral 5 mg/kg dose of 14. Data are presented as means ± SD. Three mice per treatment group were used in the PK–PD study.

To test whether members of the general class of TTR ligands may be able to reduce serum RBP4, we conducted a series of single oral dose experiments in Balb/c mice with established TTR tetramer stabilizers (tafamidis and AG10) as well as with our new analogues 14, 20c, and 20e. Analogues 20c and 20e were chosen as they exhibited equivalent potency to 14 in the TTR FP assay and exhibited similar ADME profiles relative to 14. A bis-hydrochloric salt of 14 (14a) was used in these experiments due to an observed improved solubility in the dosing vehicle (0.9% NaCl, 2% Tween 80). Despite variability in baseline levels of serum RBP4 between study groups, all tested compounds induced a significant reduction in circulating levels of RBP4 in comparison to the vehicle (Figure 11). Analogue 14a showed superior RBP4-lowering efficacy in comparison to AG10 and tafamidis (81% serum RBP4 reduction at the 6 h time point), while tafamidis seems to have a longer duration of action (Figure 11). We plan to obtain PK data for analogues 20e, 20c, and 14a to generate PK–PD correlations in a similar manner to 14. Given that we and others previously established a very good correlation between pharmacological reduction of serum RBP4 and efficacy in the Abca4^−/− mouse model of enhanced retinal lipofuscinogenesis,^10,68,69 our RBP4-lowering data for TTR ligands may indicate that members of a general class of TTR tetramer stabilizers could have utility in treating retinal diseases associated with accumulation of RPE lipofuscin.

Figure 11.

Effect of TTR ligands on circulating levels of serum RBP4 in Balb/c mice. Serum RBP4 levels were measured following a single oral administration of (A) AG10, (B) tafamidis, (C) 14a, (D) 20e, and (E) 20c. Percent changes of serum RBP4 in comparison to the baseline levels (100%) are shown in (F) for test compounds as well as for the dosing vehicle (0.9% NaCl, 2% Tween 80). The oral dose was 25 mg/kg for AG10, 14a, 20e, and 20c; the tafamidis dose was 50 mg/kg. Data represented as the mean ± SD. Three to eight mice per compound treatment group were used in the study.

It is currently unclear how 14 and other TTR ligands inhibit RBP4–TTR–all-trans-retinol complex formation, which seems to be the likely cause of serum RBP4-lowering. Indeed, tafamidis, AG10, 14, and 20e were all found to exhibit good potency in the RBP4–TTR HTRF assay, providing evidence that the compounds are in fact hindering complex formation. One possible mechanism of action may be that the binding of kinetic stabilizers such as 14 at the T4 site causes conformational changes to occur within the distal PPI interface domain that are unfavorable for holo-RBP4 association. For example, the aforementioned TTR residues Ile84, Val-20, and Ala-81 play pivotal roles in the formation of the PPI interface domain and mutations involving TTR Ile84 have deleterious consequences with regard to complex formation with holo-RBP4.⁵⁴ The K_d for the holo-RBP–TTR complex is approximately 0.4 μM, whereas the I84S mutation results in negligible TTR affinity for holo-RBP4 and abrogates complex formation.⁵⁴ Indeed, individuals expressing variants of TTR whereby Ile84 is replaced with either Ser or Asn have substantially lowered plasma concentrations of RBP4.⁵⁴ It may be possible that ligand binding at the T4 site causes a conformational shift that impedes Ile84 and/or other residues within the TTR hydrophobic patch from facilitating docking to holo-RBP4, in much the same manner as RBP4 antagonist prevents loops β3–β4 and β5–β6 to adopt favorable conformations that promote association with the PPI domain of TTR. The emerging RBP4–TTR HTRF SAR is also intriguing. Of the compounds tested in this assay, the two that were found to be the least potent (methylated analogue 20i and tetrazole 54) present either a steric blocking group ortho to or a bioisosteric replacement of the carboxylic acid of 14. We plan to determine if 20i, carboxamide 53, and tetrazole 54 also induce a significant lowering of serum RBP4 in mice and we will generate additional amide analogues and isosteres of 53, carboxylic acid isosteres of 14, and analogues of 20i featuring various blocking substituents beyond a methyl group to further probe TTR FP and RBP4–TTR HTRF SAR effects.

It may also be possible that in addition to binding at the T4 site, compound 14 may also be binding at another-to-be-determined site on TTR and/or RBP4 that results in antagonism of holo-RBP4–TTR complex formation. We plan to address questions concerning the putative mechanism of action for holo-RBP4–TTR complex disruption by 14 via conducting a series of studies designed to further elucidate the nature of the compound’s mode(s) of binding. This includes the generation of an X-ray crystal structure of TTR cocrystallized with 14 to confirm that the compound is binding within the T4 site in the pose predicted by our docking model. Such data, coupled with saturation transfer difference (STD) nuclear magnetic resonance (NMR) experiments,⁷⁰ may also allow us to determine if 14 binding at the T4 site imparts conformational changes within the TTR hydrophobic patch and PPI domain that disfavor association with holo-RBP4. An X-ray crystal structure may also determine if 14 is binding at any other potential sites at TTR in addition to the T4 binding site. We also plan to conduct ITC studies with 14. Generation of ITC data with 14 and TTR will help elucidate the relative magnitude of enthalpic and entropic contributions to binding for 14 relative to tafamidis and AG10, which could guide subsequent drug design campaigns for the series. Furthermore, we will also conduct ITC studies with 14 and RBP4 to probe whether the compound is potentially disrupting the holo-RBP4–TTR complex formation via binding to an allosteric site on the protein that overlaps with or is located separately from the all-trans-retinol orthosteric binding site. Such potential mechanisms of allosteric binding at RBP4 may not be detected with our RBP4 SPA binding assay if they do not disrupt all-trans-retinol binding.

In general, our current data proves that there is a very good PK–PD relationship between 14 exposure and serum RBP4-lowering activity in mice. In our previous work, we proved a direct correlation between serum RBP4-lowering induced by different classes of selective RBP4 antagonists and bisretinoid-lowering efficacy in the Abca4^−/− mouse model of Stargardt disease.^9,10 Based on the very good in vivo RBP4-lowering activity exhibited by 14 and its bis-hydrochloride salt 14a, it is expected that this compound will be efficacious in suppressing the formation of cytotoxic lipofuscin bisretinoids in the retina, and studies with the Abca4^−/− mouse model are ongoing. The data presented herein justifies further evaluation of selective TTR ligands as a class of potential therapeutics for the treatment of Stargardt disease, dry AMD, and other conditions characterized by enhanced accumulation of lipofuscin in the retina.

CONCLUSIONS

Inherited Stargardt disease and dry AMD are the forms of macular degeneration that may potentially be treated by pharmacological inhibition of lipofuscin bisretinoid synthesis in the retina. Only one class of compounds, competitive antagonists of all-trans-retinol binding to RBP4, was currently known to block the all-trans-retinol-dependent RBP4–TTR interaction and reduce cytotoxic bisretinoid production in the animal models of excessive lipofuscin accumulation. However, the use of selective RBP4 antagonists may not be optimal in patients who, in addition to macular degeneration, may also be predisposed to genetic or sporadic forms of ATTR. The holo-RBP4–TTR interaction stabilizes a portion of TTR tetramers circulating in the bloodstream, thus preventing formation of TTR amyloid fibrils and ATTR. Selective RBP4 antagonists release the TTR tetramer from the serum holo-RBP4–TTR complex, and the release of a pool of unliganded TTR may be associated with destabilization of TTR tetramers. We hypothesized that ligands that selectively bind to TTR and not to RBP4 may also be capable of allosterically antagonizing all-trans-retinol-dependent RBP4–TTR ternary complex formation. Such selective TTR ligands may induce a desired partial reduction in serum RBP4 levels required for inhibition of lipofuscin bisretinoid synthesis in the retina while simultaneously providing adequate TTR tetramer kinetic stabilization. To the best of our knowledge, prior to the work reported herein, TTR had not been considered as a drug target for mechanisms or indications that may be treated via selective RBP4 antagonists. In the reported work, we sought to identify novel TTR tetramer kinetic stabilizers that may also demonstrate robust RBP4-lowering capability in vivo. We utilized a TTR computational docking model derived from reported X-ray crystallographic data (PDB 4HIQ) and SAR for clinically investigated AG10 to enable our structure-based drug design campaign. Our goal was to identify novel and conformationally constrained compounds that could provide favorable TTR tetramer kinetic stabilization. We initially prepared piperazine ring core linker congener 14, which was found to exhibit excellent potency at TTR (TTR FP IC₅₀ = 220 nM). Our computational docking model derived from PDB 4HIQ provided insights into the molecular determinants responsible for the high-affinity 14 exhibits toward the TTR. Follow-up analogues of 14 included a focused sample set containing alternately substituted benzoic acid appendages as well as analogues that explored SAR effects with alternative pyrazole head groups. Compound 14 emerged as a lead upon conclusion of our SAR and ADME profiling campaigns and was found to possess excellent PK characteristics in mouse. Analogue 14 significantly decreases the formation of TTR high-molecular-weight aggregates in a manner comparable to tafamidis in an in vitro TTR aggregation assay, proving its activity as a TTR kinetic stabilizer. Furthermore, the compound and its bis-hydrochloride salt induced robust and sustained lowering of serum RBP4 levels upon oral dosing in mice. As previously stated, it is currently unclear how the TTR kinetic stabilizers studied in this work are inhibiting RBP4–TTR–all-trans-retinol complex formation, and future studies will be conducted to help gain a better understanding of these observations. However, the current data suggests that they may be efficacious in suppressing the formation of cytotoxic lipofuscin bisretinoids in the retina while preventing possible TTR amyloid fibril formation. This justifies evaluation of selective TTR tetramer ligands as a class of potential therapeutics for Stargardt disease, dry AMD, and other conditions characterized by enhanced accumulation of lipofuscin in the retina, especially in patients who are also prone to ATTR comorbidities such as sporadic SSA or hereditary TTR amyloidosis. Studies with 14 in the Abca4^−/− mouse model are ongoing, and the data will be reported in due course.

EXPERIMENTAL SECTION

Fluorescence Polarization TTR Tetramer Binding Assay.

Binding potency (IC₅₀ values) of novel compounds for unliganded TTR tetramers isolated from human plasma (Clabiochem-Millipore, cat. no. 52957) was measured using an established fluorescence polarization assay designed for TTR kinetic stabilizers.⁴¹ The assay measured competitive displacement of a fluorescent probe (fluorescein isothiocyanate (FITC) coupled to a diclofenac analogue via a PEGylated linker) by a novel compound from TTR. The FITC probe was synthesized at LeadGen Labr, LLC, using the reported route by Alhamadsheh and co-workers.⁴¹ Each well contained 200 nM of TTR and 100 nM of FITC probe in the FP buffer (10 mM Tris–HCl pH 7.5, 150 mM NaCl, 0.01% of 3-(3-cholamidopropyl)-dimethylammonio-1-propanesulfonate (CHAPS), 0.01% Prionex) along with test compounds. Nonspecific binding was determined in the presence of 500 μM unlabeled diclofenac (Sigma-Aldrich). Reactions with test compounds were incubated overnight at 4 °C, and FP was measured on a Spectramax M5e plate reader (Molecular Devices).

TTR Aggregation Assay.

The ability of test compounds to prevent TTR aggregation was evaluated under the acidic conditions that favor TTR aggregation and fibril formation.^63,64 A 2 μL solution of 167 μM human TTR (ACRO Biosystems #H5223) was incubated with 7 μL of 50 mM sodium acetate (pH = 4.0) (Sigma-Aldrich # S7545), and 100 mM KCl (Sigma-Aldrich # S5405) in the presence or absence of 1 μL of TTR inhibitor for 72 h at 37 °C. At the end of the incubation, 3.5 μL of a 500 mM sodium phosphate (Sigma-Aldrich #S5136) buffer (pH = 8.0) was added to each sample for neutralization and 0.6 μL of 5% CHAPS (Sigma-Aldrich #C5070) as a detergent to prevent reassociation of protein. The cross-linking was performed by adding 1.5 μL of 5% glutaraldehyde solution (Sigma-Aldrich # G6257). After 4 min, the reaction was stopped by the addition of 2.5 μL of freshly made 5% sodium borohydride (NaBH₄). Samples were subjected to TTR Western blotting with prealbumin antibodies (1:500; Dako #A0002). Band intensity for TTR monomer and TTR aggregates was quantified from scanned images of the blots.

In Vitro Binding of Compounds to RBP4.

Compound binding to RBP4 was assessed in the radiometric scintillation proximity (SPA) assay that was previously described.^4–7 The assay measured competitive displacement of radiolabeled [³H]-all-trans-retinol from native RBP4 purified from human urine (Fitzgerald, 30R-AR022L). The protein was biotinylated using the EZ-link Sulfo-NHS-LC-Biotinylation kit from ThermoFisher (Cat #21335) as recommended by the manufacturer. Binding assays were implemented in a final volume of 100 μL in SPA buffer (1× PBR, pH 7.4, 1 mM ethylenediaminetetraacetic acid (EDTA), 0.1% BRA, 0.5% CHAPS). The assay reaction included a radioligand, 10 nM [³H]-all-trans-retinol (48.7 Ci/mmol; PerkinElmer, Waltham, MA), along with the 0.3 mg/well streptavidin-PVT beads (PerkinElmer, RPNQ0006) and 50 nM biotinylated human RBP4. Unlabeled retinol (Sigma-Aldrich, cat # 95144) at 20 μM was added to control wells to assess a nonspecific binding. Radioactivity counts were measured using CHAMELEON plate reader (Hidex Oy, Turku, Finland) after 16 h of incubation at rt with mild shaking.

Assessment of Antagonistic Activity in the HTRF RBP4–TTR Interaction Assay.

The ability of analogues to act as antagonists of all-trans-retinol-dependent RBP4–TTR interaction was measured in the homogeneous time-resolved fluorescence (HTRF) assay. Untagged TTR (Calbiochem, cat # 529577) and maltose-binding protein-tagged RBP4 expressed in Escherichia coli were used in this assay. HTRF Cryptate labeling kit from CisBio (Cisbio, cat # 62EUSPEA, Bedford, MA) was used to label TTR with Eu³⁺ cryptate. The assay was performed in a final assay volume of 16 μL in the buffer that contained 10 mM Tris–HCl (pH = 7.5), 1 mM dithiothreitol (DTT), 0.05% NP-40, 0.05% Prionex, 6% glycerol, and 400 mM KF. Other components of the reaction mix included 60 nM of MBP-RBP4, 5 nM of TTR-Eu, 26.7 nM of anti-MBP antibody conjugated with d2 (Cisbio, cat # 61MBPDAA), and 1 μM of all-trans-retinol (Sigma-Aldrich, cat # 95144). All of the reactions were performed under dim red light in the dark. The plates were read in the SpectraMax M5e Multimode Plate Reader (Molecular Devices, Sunnyvale, CA) after the overnight incubation at 4 °C. Fluorescence was excited at 337 nm; emission was measured at 668 and 620 nm with a 75 μs counting delay. The HTRF signal was expressed as the ratio of fluorescence intensity: Flu₆₆₈/Flu₆₂₀ × 10 000.

Animal Care and Use Statement.

Information regarding in vivo experimental protocols is located within the Supporting Information section. All of the protocols used for the in vivo experiments disclosed are in compliance with the U.S. Department of Agriculture’s (USDA) Animal Welfare Act (nine CFR parts 1–3); the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources, National Academy Press, Washington, DC, 1996; and the National Institutes of Health, Office of Laboratory Animal Welfare. Whenever possible, procedures in this study are designed to avoid or minimize discomfort, distress, and pain to animals.

Docking Models.

PDB structures were prepared for molecular docking experiments using the protein structural preparation tool followed by Protonate3D, as implemented in the Molecular Operating Environment (MOE) 2020.09 (Chemical Computing Group, Inc., Montreal, Quebec, Canada, http://www.chemcomp.com). Ligand structures were built with MOE and minimized using the MM-FF94x force field until a root-mean-square deviation (RMSD) gradient of 0.05 kcal/(mol Å) was reached. In the first step for protein preparation, we preprocessed the structure using the standard protocol, which included the assigning of bond orders, the adding of hydrogens, the creating of disulfide bonds, and the prediction of the structural protonation state at a physiological pH of 7.4. The structure was subjected to a short energy minimization routine to relax it using the Amber 99 force field as implemented in MOE. The following standard parameters were selected: receptor van der Waals scaling, 0.50; ligand van der Waals scaling, 0.50; and a maximum of 20 poses per ligand. The best docking pose for 14 was selected based on the lowest RMSD value, which was 1.7 Å.

General Chemistry.

All reactions were performed under a dry atmosphere of nitrogen unless otherwise specified. Indicated reaction temperatures refer to the reaction bath, while room temperature (rt) is noted as 25 °C. Commercial-grade reagents and anhydrous solvents were used as received from vendors, and no attempts were made to purify or dry these components further. Removal of solvents under reduced pressure was accomplished with a Buchi rotary evaporator at approximately 28 mmHg pressure using a Teflon-linked KNF vacuum pump. The measurement of pH for neutralizations or acidifications was measured with Hydrion pH paper (MicroEssential Lab). Thin-layer chromatography (TLC) was performed using 1 in. × 3 in. Analtech no. 02521 silica gel plates with a fluorescent indicator. Visualization of TLC plates was made by observation with either short-wave UV light (254 nm lamp), 10% phosphomolybdic acid in ethanol, or in iodine vapors. Preparative thin-layer chromatography was performed using Analtech, 20 × 20 cm², 1000 μm preparative TLC plates. Flash column chromatography was carried out using a Teledyne Isco CombiFlash Companion Unit and a Biotage Selekt System with Teledyne Isco RediSep Rf and Biotage Sfȧr silica gel columns. If needed, products were purified by reverse-phase chromatography, using a Teledyne Isco CombiFlash Companion Unit and a Biotage Selekt System with a RediSep Gold C18 reverse-phase column. Proton NMR spectra were obtained on a 400 MHz Varian nuclear magnetic resonance spectrometer. Chemical shifts (δ) are reported in parts per million (ppm), and coupling constant (J) values are given in hertz, with the following spectral pattern designations: s, singlet; d, doublet; t, triplet, q, quartet; quint, quintet; m, multiplet; dd, doublet of doublets; dt, doublet of triplets; dq; doublet of quartets; br, broad signal. Tetramethylsilane was used as an internal reference. Peak listing, multiplicity designations, and coupling constant calculations were conducted using Mnova v.14 software (Mestrelab Research). Carbon NMR spectra were obtained on a 500 MHz Bruker AV III nuclear magnetic resonance spectrometer, and tetramethylsilane was used as an internal reference. Fluorine NMR spectra were obtained on a 400 MHz Bruker AV III nuclear magnetic resonance spectrometer. Any melting points provided are uncorrected and were obtained using a Stanford Research Systems OptiMelt melting point apparatus (MPA100) with an automated melting point system. Mass spectroscopic analyses were performed using electrospray ionization (ESI) on a Waters AQUITY UPLC MS triple quadrapole mass spectrometer. High-pressure liquid chromatography (HPLC) purity analysis was performed using a Waters Breeze2 HPLC system with binary solvent systems A and B using a gradient elusion [A, H₂O with 0.1% formic acid; B, CH₃CN with 0.1% formic acid] and flow rate = 0.5 mL/min, with UV detection at 254 nm (system equipped with a photodiode array (PDA) detector). An ACQUITY UPLC BEH C18 column, 130 Å, 1.7 μm, 2.1 mm × 50 mm, was used. High-resolution mass spectrometry (HRMS) analysis was performed using an Agilent 6530 Accurate-Mass Q-TOF. All final compounds tested for in vitro and in vivo biological testing were purified to ≥95% purity, and these purity levels were measured by both ¹H NMR and HPLC.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic Acid (14).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 2.00 g, 10.7 mmol) and methyl 3-bromo-4-fluorobenzoate (2.25 g, 9.65 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (10.0 g, 30.6 mmol), X-Phos (0.600 g, 1.25 mmol), and Pd₂(dba)₃·CHCl₃ (0.555 g, 0.536 mmol) were then added, and the mixture was stirred reflux for 16 h under an atmosphere of N₂. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% ethyl acetate (EtOAc) in hexanes) to give tert-butyl 4-(2-fluoro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16a) as a brown oil (3.0 g, 83%). The material was used as is in the next step: ESI MS m/z 339 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(2-fluoro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16a, 3.00 g, 8.87 mmol) in CH₂Cl₂ (30 mL) was added TFA (6.7 mL, 87.5 mmol), and the resulting solution was stirred at rt for 16 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-fluoro-3-(piperazin-1-yl)benzoate (17a) as a white solid (1.20 g, 57%): ¹H NMR (400 MHz, CDCl₃) δ 9.85 (br, 1H), 7.72–7.69 (m, 1H), 7.63–7.60 (m, 1H), 7.21–7.04 (m, 1H), 3.87 (s, 3H), 3.35 (s, 8H); ESI MS m/z 239 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-fluoro-3-(piperazin-1-yl)benzoate (17a, 1.20 g, 5.03 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (0.9 mL, 5.16 mmol) and 3-chloropentane-2,4-dione (0.67 mL, 5.96 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–50% EtOAc in hexanes) to give methyl 3-(4-(2,4-dioxopentan-3-yl)-piperazin-1-yl)-4-fluorobenzoate (18a) as a brown oil (0.600 g, 35%): ¹H NMR (400 MHz, CDCl₃) δ 7.64–7.62 (m, 2H), 7.06– 7.01 (m, 1H), 3.86 (s, 1H), 3.31–3.09 (m, 4H), 3.09–3.05 (m, 4H), 2.27 (s, 1H), 2.24 (s, 6H); ESI MS m/z 337 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(2,4-dioxopentan-3-yl)-piperazin-1-yl)-4-fluorobenzoate (18a, 0.500 g, 1.48 mmol) in CH₃OH (10 mL) was added N₂H₄·H₂O (0.2 mL, 2.67 mmol, 64–65% solution in H₂O), and the resulting mixture was stirred at rt for 1 h. The mixture was then concentrated under reduced pressure, and the resulting residue was chromatographed over silica gel (0–50% EtOAc in hexanes) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoate (19a) as a brown solid (0.420 g, 85%): ¹H NMR (400 MHz, CDCl₃) δ 7.67–7.63 (m, 2H), 7.04–7.01 (m, 1H), 3.86 (s, 1H), 3.16–3.15 (m, 4H), 3.15–3.11 (m, 4H), 2.24 (s, 6H); ESI MS m/z 333 [M + H]⁺.

Step E.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoate (19a, 0.420 g, 1.26 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (91 mg, 3.79 mmol). The reaction mixture was stirred at rt for 16 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid (14) as a white solid (0.390 g, 97%): melting point = 220–222 °C; ¹H NMR (400 MHz, DMSO-d₆) δ 7.53–7.55 (m, 2H, H₁ and H₂), 7.23–7.18 (dt, J = 12, 3.2 Hz, 1H, H₃), 3.04 (m, 4H, H₄), 3.01 (m, 4H, H₅), 2.10 (s, 6H, H₆); ¹³C NMR (500 MHz, DMSO-d₆) δ 128.7, 121.7, 124.1, 124.0, 120.2, 120.2, 116.3, 116.2; ¹⁹F NMR (400 MHz, DMSO-d₆) δ −116.0 (s, F); ESI MS m/z 319 [M + H]⁺; HRMS (ESI⁺) C₁₆H₁₉FN₄O₂ calcd [M + H]⁺ = 319.1565, observed [M + H]⁺ = 319.1562; combustion analysis (%C, H, N): calcd for C₁₆H₁₉FN₄O₂· 0.5H₂O·0.5HCl: %C = 55.61; %H = 5.98; %N = 16.21; found: %C = 55.88; %H = 5.74; %N = 15.97; HPLC >99% (AUC), t_R = 11.5 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic Acid Bis-hydrochloride (14a).

Step A.

To a solution of 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid (14, 0.700 g, 2.20 mmol) in anhydrous Et₂O (5 mL) at 0 °C was slowly added a 2 N HCl solution in Et₂O dropwise (11.0 mL, 22.0 mmol), and the pH of the mixture was observed at approximately pH = 2 (monitored with Hydrion pH paper). The mixture was then stirred at rt for 3 h, and the Et₂O was removed under reduced pressure. The resulting material was triturated with Et₂O and dried under reduced pressure to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid bis-hydrochloride (14a) as a white solid (0.800 g, 93%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.63 (m, 2H), 7.29 (dt, J = 8.8 Hz, 3.2 Hz, 1H), 3.26 (m, 8H), 2.39 (s, 6H); ¹³C NMR (500 MHz, DMSO-d₆) δ 167.0, 159.1, 157.1, 140.0, 139.2, 128.1, 124.8, 120.8, 116.9, 52.2, 50.0, 10.8; ESI MS m/z 319 [M + H]⁺; HRMS (ESI⁺) C₁₆H₁₉FN₄O₂ calcd [M + H]⁺ = 319.1565, observed [M + H]⁺ = 319.1561; combustion analysis (%C, H, N, Cl): calcd for C₁₆H₁₉FN₄O₂·1.2H₂O·2.0HCl: %C = 46.54; %H = 5.71; %N = 13.57; %Cl = 17.17; found: %C = 46.57; %H = 5.52; %N = 13.39; %Cl = 16.97; HPLC >99% (AUC), t_R = 11.4 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoic Acid (20a).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 3.46 g, 18.5 mmol) and methyl 3-bromobenzoate (2.0 g, 9.30 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (1.04 g, 3.19 mmol), X-Phos (0.600 g, 1.25 mmol), and Pd₂(dba)₃·CHCl₃ (0.621 g, 0.600 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–20% EtOAc in hexanes) to give tert-butyl 4-(3-(methoxycarbonyl)phenyl)-piperazine-1-carboxylate (16b) as a brown solid (0.440 g, 14%). The material was used as is in the next step: ESI MS m/z 321 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(3-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16b, 0.430 g, 1.34 mmol) in CH₂Cl₂ (30 mL) was added TFA (3.0 mL, 39.2 mmol), and the resulting solution was stirred at rt for 16 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure to give crude methyl 3-(piperazin-1-yl)benzoate (17b) as a white solid (0.292 g, >99%, crude): ESI MS m/z 221 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 3-(piperazin-1-yl)benzoate (17b, 0.290 g, 1.32 mmol) in anhydrous THF (5 mL) were added i-Pr₂NEt (1.2 mL, 6.88 mmol) and 3-chloropentane-2,4-dione (0.44 mL, 3.90 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of methyl 3-(4-(2,4-dioxopentan-3-yl)-piperazin-1-yl)benzoate (18b) (monitored by liquid chromatography-mass spectrometry (LC-MS); ESI MS m/z 319 [M + H]⁺), N₂H₄·H₂O (0.26 mL, 3.15 mmol, 64–65% solution in H₂O) was then added, and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure.The resulting crude residue was chromatographed over silica gel (0– 10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoate as a brown solid (19b) along with an inseparable and uncharacterized impurity (0.280 g, 67%, impure): ESI MS m/z 315 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoate (19b, 0.130 g, 0.413 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (49.5 mg, 2.06 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoic acid (20a) as a white solid (7.8 mg, 6.2%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.48 (s, 1H), 7.36 (m, 2H), 7.23 (m, 1H), 3.23 (m, 4H), 3.04 (m, 4H), 2.14 (s, 6H); ESI MS m/z 301 [M + H]⁺; HPLC >99% (AUC), t_R = 10.7 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methoxybenzoic Acid (20b).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.911 g, 4.89 mmol) and methyl 3-bromo-4-methoxybenzoate (1.0 g, 4.08 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (2.6 g, 7.97 mmol), X-Phos (0.230 g, 0.482 mmol), and Pd₂(dba)₃·CHCl₃ (0.124 g, 0.120 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(2-methoxy-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16c) as a light yellow solid (1.40 g, 98%): ¹H NMR (400 MHz, acetone-d₆) δ 7.68 (d, J = 8.0 Hz, 1H), 7.51 (s, 1H), 7.04 (d, J = 8.4 Hz, 1H), 3.93 (s, 3H), 3.82 (s, 3H), 3.53 (brs, 4 h), 2.98 (brs, 4 h), 1.41 (s, 9H): ESI MS m/z 351 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(2-methoxy-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16c, 1.4 g, 3.99 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure to give crude methyl 4-methoxy-3-(piperazin-1-yl)benzoate (17c) as a brown oil (0.650 g, 65%, crude): ESI MS m/z 251 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-methoxy-3-(piperazin-1-yl)benzoate (17c, 0.650 g, 2.59 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (1.4 mL, 8.03 mmol) and 3-chloropentane-2,4-dione (0.6 mL, 5.32 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18c (monitored by LC-MS; ESI MS m/z 349 [M + H]⁺), N₂H₄·H₂O (0.26 mL, 3.15 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methoxybenzoate (19c) along with an inseparable and uncharacterized impurity (0.100 g, 11%, crude): ESI MS m/z 345 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methoxybenzoate (19c, 0.100 g, 0.290 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (34.0 mg, 1.41 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methoxybenzoic acid (20b) as a white solid (38.0 mg, 39%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.58 (dd, J = 8.4 Hz, 2.0 Hz, 1H), 7.44 (d, J = 2.0 Hz, 1H), 7.01 (d, J = 8.4 Hz, 1H), 3.83 (s, 3H), 2.99 (s, 8H), 2.08 (s, 6H); ESI MS m/z 331 [M + H]⁺; HPLC 94.6% (AUC), t_R = 10.2 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methylbenzoic Acid (20c).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.978 g, 5.26 mmol) and methyl 3-bromo-4-methylbenzoate (1.0 g, 4.36 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (2.8 g, 8.59 mmol), X-Phos (0.248 g, 0.520 mmol), and Pd₂(dba)₃·CHCl₃ (0.136 g, 0.131 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(5-(methoxycarbonyl)-2-methylphenyl)piperazine-1-carboxylate (16d) as a brown solid (0.600 g, 41%). The material was used as is in the next step: ESI MS m/z 335 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(5-(methoxycarbonyl)-2-methylphenyl)piperazine-1-carboxylate (16d, 0.600 g, 1.79 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.3 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-methyl-3-(piperazin-1-yl)benzoate (17d) as a white solid (0.300 g, 71%, crude): ESI MS m/z 235 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-methyl-3-(piperazin-1-yl)benzoate (17d, 0.300 g, 1.28 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (0.74 mL, 4.24 mmol) and 3-chloropentane-2,4-dione (0.30 mL, 2.66 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18d (monitored by LC-MS; ESI MS m/z 333 [M + H]⁺), N₂H₄·H₂O (0.14 mL, 1.68 mmol, 64–65% solution in H₂O) was then added, and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methylbenzoate (19d) as a brown oil along with an inseparable and uncharacterized impurity (0.100 g, 23%, crude): ESI MS m/z 329 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methylbenzoate (19d, 0.100 g, 0.304 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (24.0 mg, 1.00 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-methylbenzoic acid (20c) as a white solid (31.0 mg, 32%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.57 (d, J = 1.2 Hz, 1H), 7.53 (d, J = 8.0 Hz, 1H), 7.27 (d, J = 8.0 Hz, 1H), 3.01 (m, 4H), 2.90 (m, 4H), 2.46 (s, 3H), 2.12 (s, 6H); ESI MS m/z 315 [M + H]⁺; HPLC >99% (AUC), t_R = 11.2 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-(trifluoromethyl)benzoic Acid (20d).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.700 g, 3.75 mmol) and methyl 3-bromo-4-(trifluoromethyl)benzoate (0.957 g, 3.38 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.6 g, 11.04 mmol), X-Phos (0.212 g, 0.444 mmol), and Pd₂(dba)₃· CHCl₃ (0.192 g, 0.186 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(5-(methoxycarbonyl)-2-(trifluoromethyl)phenyl)piperazine-1-carboxylate (16e) as a brown oil (0.450 g, 31%). The material was used as is in the next step: ESI MS m/z 389 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(5-(methoxycarbonyl)-2-(trifluoromethyl)phenyl)piperazine-1-carboxy-late (16e, 0.450 g, 1.15 mmol) in CH₂Cl₂ (30 mL) was added TFA (3.0 mL, 39.2 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 3-(piperazin-1-yl)-4-(trifluoromethyl)-benzoate (17e) as a white solid (0.350 g, >99%, crude): ESI MS m/z 289 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 3-(piperazin-1-yl)-4-(trifluoromethyl)benzoate (17e, 0.350 g, 1.21 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (1.0 mL, 5.74 mmol) and 3-chloropentane-2,4-dione (0.40 mL, 3.54 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18e (monitored by LC-MS; ESI MS m/z 387 [M + H]⁺), N₂H₄·H₂O (0.20 mL, 1.82 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-(trifluoromethyl)benzoate (19e) as a brown oil (0.150 g, 32%): ESI MS m/z 383 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-(trifluoromethyl)benzoate (19e, 0.130 g, 0.34 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (81 mg, 3.38 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-(trifluoromethyl)benzoic acid (20d) as an amorphous white solid (79.0 mg, 63%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.99 (s, 1H), 7.84 (d, J = 8.4 Hz, 1H), 7.78 (d, J = 8.4 Hz, 1H), 2.97 (m, 4H), 2.94 (m, 4H), 2.11 (s, 6H); ESI MS m/z 369 [M + H]⁺; HPLC 98.7% (AUC), t_R = 11.9 min.

4-Chloro-3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-benzoic Acid (20e).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.700 g, 3.75 mmol) and methyl 3-bromo-4-chlorobenzoate (0.844 g, 3.38 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.6 g, 11.04 mmol), X-Phos (0.212 g, 0.44 mmol), and Pd₂(dba)₃·CHCl₃ (0.192 g, 0.186 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(2-chloro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16f) as a yellow oil (0.800 g, 60%): ¹H NMR (400 MHz, acetone-d₆) δ 7.71 (s, 1H), 7.68–7.66 (m, 1H), 7.54 (d, J = 8.0 Hz, 1H), 3.88 (s, 3H), 3.58 (brs, 4H), 3.02 (brs, 4H), 1.46 (s, 9H); ESI MS m/z 355 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(2-chloro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16f, 0.700 g, 1.97 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol), and the resulting solution was stirred at rt for 16 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-chloro-3-(piperazin-1-yl)benzoate (17f) as a white solid (0.450 g, 89%, crude): ESI MS m/z 255 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-chloro-3-(piperazin-1-yl)benzoate (17f, 0.200 g, 0.787 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (0.7 mL, 4.01 mmol) and 3-chloropentane-2,4-dione (0.26 mL, 2.38 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18f (monitored by LC-MS; ESI MS m/z 353 [M + H]⁺), N₂H₄·H₂O (0.1 mL, 1.18 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 4-chloro-3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoate (19f) as a brown oil (80.0 mg, 29%): ESI MS m/z 349 [M + H]⁺.

Step D.

To a solution of methyl 4-chloro-3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoate (19f, 80.0 mg, 0.23 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (55.0 mg, 2.29 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 4-chloro-3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)benzoic acid (20e) as an amorphous white solid (30.0 mg, 39%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.66 (s, 1H), 7.58 (d, J = 8.4 Hz, 1H), 7.51 (d, J = 10.8 Hz, 1H), 3.04 (brs, 8H), 2.12 (s, 6H); ESI MS m/z 335 [M + H]⁺; HPLC 98.0% (AUC), t_R = 11.5 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoic Acid (20f).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.700 g, 3.75 mmol) and methyl 3-bromo-2-fluorobenzoate (0.787 g, 3.38 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.6 g, 11.04 mmol), X-Phos (0.212 g, 0.444 mmol), and Pd₂(dba)₃·CHCl₃ (0.192 g, 0.186 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(2-fluoro-3-(methoxycarbonyl)-phenyl)piperazine-1-carboxylate (16g) as a yellow oil (1.0 g, 78%): ¹H NMR (400 MHz, acetone-d₆) δ 7.47–7.45 (m, 1H), 7.31–7.27 (m, 1H), 7.19 (t, J = 8.0 Hz, 1H), 3.87 (s, 3H), 3.56 (brs, 4H), 3.02 (brs, 4H), 1.45 (s, 9H); ESI MS m/z 339 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(2-fluoro-3-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16g, 1.0 g, 2.95 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 2-fluoro-3-(piperazin-1-yl)benzoate (17g) as a brown oil (0.600 g, 85%, crude): ESI MS m/z 239 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 2-fluoro-3-(piperazin-1-yl)benzoate (17g, 300 mg, 1.26 mmol) in anhydrous THF (3 mL) were added i-Pr₂NEt (0.5 mL, 2.52 mmol) and 3-chloropentane-2,4-dione (0.3 mL, 2.52 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18g (monitored by LC-MS; ESI MS m/z 337 [M + H]⁺), N₂H₄·H₂O (0.16 mL, 1.89 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoate (19g) as a brown oil (0.160 g, 38%): ¹H NMR (400 MHz, DMSO-d₆) δ 11.79 (brs, 1H), 7.40–7.36 (m, 1H), 7.32–7.28 (m, 1H), 7.21–7.17 (m, 1H), 3.81 (s, 3H), 3.05–3.00 (m, 8H), 2.11 (s, 6H); ESI MS m/z 333 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoate (19g, 0.150 g, 0.451 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (0.108 g, 4.50 mmol). The reaction mixture was stirred at rt for 16 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoic acid (20f) as an amorphous white solid (90.0 mg, 63%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.29 (m, 1H), 7.19 (m, 1H), 7.12 (t, J = 9.6 Hz, 1H), 3.03 (br, 8H), 2.12 (s, 6H); ESI MS m/z 319 [M + H]⁺; HPLC 98.3% (AUC), t_R = 10.3 min.

5-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoic Acid (20g).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.700 g, 3.75 mmol) and methyl 5-bromo-2-fluorobenzoate (0.787 g, 3.38 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.6 g, 11.04 mmol), X-Phos (0.212 g, 0.44 mmol), and Pd₂(dba)₃·CHCl₃ (0.192 g, 0.186 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(4-fluoro-3-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16h) as a yellow oil (1.05 g, 83%). The material was used as is in the next step: ESI MS m/z 339 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(4-fluoro-3-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16h, 1.2 g, 3.55 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 2-fluoro-5-(piperazin-1-yl)benzoate (17h) as a white solid (0.540 g, 64%, crude): ESI MS m/z 239 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 2-fluoro-5-(piperazin-1-yl)benzoate (17h, 0.300 g, 1.26 mmol) in anhydrous THF (3 mL) were added i-Pr₂NEt (0.5 mL, 2.87 mmol) and 3-chloropentane-2,4-dione (0.30 mL, 2.66 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18h (monitored by LC-MS; ESI MS m/z 337 [M + H]⁺), N₂H₄·H₂O (0.16 mL, 1.89 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoate (19h) as a brown oil along with an inseparable and uncharacterized impurity (0.140 g, 33%, crude): ESI MS m/z 333 [M + H]⁺.

Step D.

To a solution of methyl 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoate (19h, 0.140 g, 0.421 mmol) in CH₃OH (4 mL), THF (4 mL) and H₂O (2 mL) was added anhydrous LiOH (99.4 mg, 4.15 mmol). The reaction mixture was stirred at rt for 16 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-2-fluorobenzoic acid (20g) as an amorphous white solid (75.0 mg, 56%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.29 (m, 1H), 7.19 (m, 1H), 7.12 (t, J = 9.6 Hz, 1H), 3.13 (br, 4H), 2.99 (br, 4H), 2.10 (s, 6H); ESI MS m/z 319 [M + H]⁺; HPLC >99% (AUC), t_R = 10.6 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-5-fluorobenzoic Acid (20h).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.700 g, 3.75 mmol) and methyl 3-bromo-5-fluorobenzoate (0.787 g, 3.38 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.6 g, 11.04 mmol), X-Phos (0.212 g, 0.44 mmol), and Pd₂(dba)₃·CHCl₃ (0.192 g, 0.186 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(3-fluoro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16i) as a brown oil (0.800 g, 70%): ¹H NMR (400 MHz, acetone-d₆) δ 7.39 (s, 1H), 7.08 (d, J = 8.4 Hz, 1H), 6.97 (d, J = 12.0 Hz, 1H), 3.87 (s, 3H), 3.56 (brs, 4 h), 3.26 (brs, 4 h), 1.45 (s, 9H): ESI MS m/z 339 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(3-fluoro-5-(methoxycarbonyl)phenyl)piperazine-1-carboxylate (16i, 0.800 g, 2.37 mmol) in CH₂Cl₂ (30 mL) was added TFA (3.0 mL, 39.20 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 3-fluoro-5-(piperazin-1-yl)benzoate (17i) as an off-white solid (0.540 g, 96%, crude): ESI MS m/z 239 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 3-fluoro-5-(piperazin-1-yl)benzoate (17i, 0.500 g, 2.09 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (1.0 mL, 5.74 mmol) and 3-chloropentane-2,4-dione (0.545 mL, 4.83 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18i (monitored by LC-MS; ESI MS m/z 337 [M + H]⁺), N₂H₄·H₂O (0.26 mL, 3.15 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-5-fluorobenzoate (19i) as a brown oil along with an inseparable and uncharacterized impurity (0.230 g, 33%, crude): ESI MS m/z 337 [M + H]⁺.

Step D.

To a solution of methyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-5-fluorobenzoate (19i, 0.230 g, 0.692 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (0.166 g, 6.93 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-5-fluorobenzoic acid (20h) as an amorphous off-white solid (90.0 mg, 41%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.28 (s, 1H), 6.99 (m, 2H), 3.24 (m, 4H), 2,98 (m, 4H), 2.10 (s, 6H); ESI MS m/z 319 [M + H]⁺; HPLC 97.6% (AUC), t_R = 11.3 min.

5-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluoro-2-methylbenzoic Acid (20i).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 0.771 g, 4.13 mmol) and methyl 5-bromo-4-fluoro-2-methylbenzoate (0.930 g, 3.76 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (3.70 g, 11.35 mmol), X-Phos (0.213 g, 0.446 mmol), and Pd₂(dba)₃·CHCl₃ (0.218 g, 0.211 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 4-(2-fluoro-5-(methoxycarbonyl)-4-methylphenyl)-piperazine-1-carboxylate (16j) as a brown oil (0.260 g, 18%). The material was used as is in the next step: ESI MS m/z 353 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 4-(2-fluoro-5-(methoxycarbonyl)-4-methylphenyl)piperazine-1-carboxylate (16j, 0.250 g, 1.34 mmol) in CH₂Cl₂ (30 mL) was added TFA (3.0 mL, 39.20 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-fluoro-2-methyl-5-(piperazin-1-yl)-benzoate (17j) as a brown oil (0.160 g, 47%, crude): ESI MS m/z 253 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-fluoro-2-methyl-5-(piperazin-1-yl)benzoate (17j, 0.150 g, 0.594 mmol) in anhydrous THF (5 mL) were added i-Pr₂NEt (0.54 mL, 3.10 mmol) and 3-chloropentane-2,4-dione (0.20 mL, 1.77 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of 18j (monitored by LC-MS; ESI MS m/z 351 [M + H]⁺), N₂H₄·H₂O (0.5 mL, 5.95 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluoro-2-methylbenzoate (19j) as a brown oil along with an inseparable and uncharacterized impurity (0.140 g, 68%, crude): ESI MS m/z 347 [M + H]⁺.

Step D.

To a solution of methyl 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluoro-2-methylbenzoate (19j, 0.140 g, 0.404 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (48.3 mg, 2.02 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 5-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluoro-2-methylbenzoic acid (20i) as an amorphous off-white solid (14.6 mg, 11%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.52 (d, J = 8.4 Hz, 1H), 7.13 (d, J = 13.2 Hz, 1H), 3.04 (br, 8H), 2.45 (s, 1H), 2.13 (s, 6H); ESI MS m/z 333 [M + H]⁺; HPLC 98.7% (AUC), t_R = 11.6 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-iodobenzoic Acid (26).

Step A.

A mixture of tert-butyl piperazine-1-carboxylate (15, 1.05 g, 5.63 mmol), ethyl 3-fluoro-4-nitrobenzoate (1.0 g, 4.69 mmol), and K₂CO₃ (1.3 g, 9.38 mmol) in anhydrous CH₃CN (20 mL) was heated at 60 °C for 16 h under an atmosphere of N₂. The mixture was then allowed to cool to rt and concentrated under reduced pressure. The resulting residue was diluted with H₂O (50 mL), and the aqueous mixture was extracted with EtOAc (3 × 75 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–15% EtOAc in hexanes) to give tert-butyl 4-(5-(ethoxycarbonyl)-2-nitrophenyl)piperazine-1-carboxy-late (21) as an orange oil (1.6 g, 90%): ¹H NMR (400 MHz, acetone-d₆) δ 7.89–7.87 (m, 2H), 7.76 (d, J = 8.0 Hz, 1H), 4.38 (q, 2H), 3.54 (brs, 4H), 3.08 (brs, 4H), 1.42 (s, 9H), 1.37 (t, J = 6.8 Hz, 3H); ESI MS m/z 380 [M + H]⁺.

Step B.

A mixture of tert-butyl 4-(5-(ethoxycarbonyl)-2-nitrophenyl)piperazine-1-carboxylate (21, 1.6 g, 4.21 mmol) and Pd/C (10% w/w, 0.160 g) in anhydrous EtOH (30 mL) was stirred at rt under an atmosphere of H₂ gas via balloon (1 atm of pressure) for 24 h. The mixture was filtered through a pad of CELITE pad with repeat rinsing using anhydrous EtOH. The filtrate was concentrated under reduced pressure, and the resulting crude residue was chromatographed over silica gel (0–25% EtOAc in hexanes) to give tert-butyl 4-(2-amino-5-(ethoxycarbonyl)phenyl)piperazine-1-carboxylate (22) as a white solid (1.3 g, 88%): ¹H NMR (400 MHz, acetone-d₆) δ 7.60 (s, 1H), 7.57 (d, J = 8.8 Hz, 1H), 6.77 (d, J = 8.0 Hz, 1H), 5.30 (brs, 2H), 4.24 (q, 2H), 3.57 (brs, 4H), 2.89 (brs, 4H), 1.43 (s, 9H), 1.30 (t, J = 6.8 Hz, 3H): ESI MS m/z 350 [M + H]⁺.

Step C.

A mixture of tert-butyl 4-(2-amino-5-(ethoxycarbonyl)-phenyl)piperazine-1-carboxylate (22, 0.500 g, 1.43 mmol) dissolved in 3 N aqueous H₂SO₄ (5.7 mL, 0.25 M) was cooled to 0 °C. A solution of NaNO₂ (0.108 g, 1.57 mmol) in H₂O (10 mL) was added dropwise slowly at 0 °C over a period of 30 min. A solution of KI (0.356 mg, 2.15 mmol) and urea (17.2 mg, 0.29 mmol) in H₂O (5 mL) was subsequently added, and the resulting mixture was continued to stir at 0 °C for an additional 1 h. The mixture was then quenched via addition of 10% NaHCO₃ (50 mL) and extracted with ethyl acetate (3 × 50 mL). The organic extracts were dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was dissolved in CH₂Cl₂ (10 mL) and cooled to 0 °C. TFA (3 mL, 39.2 mmol) was added, and the resulting solution was stirred for 3 h, gradually warming to rt. The mixture was concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated NaHCO₃ solution, and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure to give crude ethyl 4-iodo-3-(piperazin-1-yl)benzoate (23) as a brown oil, which was carried forward as is without further purification (0.300 g, 58%, crude): ESI MS m/z 361[M + H]⁺.

Step D.

To a 0 °C cooled solution of crude ethyl 4-iodo-3-(piperazin-1-yl)benzoate (23, 0.300 g, 0.83 mmol) in THF (6 mL) were added i-Pr₂NEt (0.75 mL, 4.30 mmol) and 3-chloropentane-2,4-dione (0.28 mL, 2.48 mmol) simultaneously, and the resulting solution was stirred at rt for 16 h. Upon in situ formation of 24 (monitored by LC-MS; ESI MS m/z 459 [M + H]⁺), N₂H₄·H₂O (0.5 mL, 5.95 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give ethyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-iodobenzoate (25) as a brown oil, which was carried forward as is without further purification (0.200 g, 53%, crude).

Step E.

To a solution of ethyl 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-iodobenzoate (25, 0.200 g, 0.44 mmol) in EtOH (4 mL), THF (4 mL) and H₂O (2 mL) was added anhydrous LiOH (0.105 g, 4.38 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The resulting aqueous mixture was diluted with H₂O (10 mL) and acidified with 2 N HCl to pH 3. The mixture was extracted with EtOAc (3 × 50 mL), and the combined organic extracts were washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-iodobenzoic acid (26) as an off-white solid (80.0 mg, 43%): ¹H NMR (acetone-d₆, 400 MHz) δ 7.95 (d, J = 8.0 Hz, 1H), 7.58 (s, 1H), 7.34 (d, J = 8.4 Hz, 2H), 3.12–3.10 (m, 4H), 3.03–2.98 (m, 4H), 2.11 (s, 6H); ESI MS m/z 427 [M + H]⁺.

3-((3aR,6aS)-5-(3,5-Dimethyl-1H-pyrazol-4-yl)hexahydropyrrolo-[3,4-c]pyrrol-2(1H)-yl)-4-fluorobenzoic Acid (32).

Step A.

A mixture of tert-butyl (3aR,6aS)-hexahydropyrrolo[3,4-c]pyrrole-2(1H)-carboxylate (27, 1.0 g, 4.71 mmol) and methyl 3-bromo-4-fluorobenzoate (1.0 g, 4.29 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (4.17 g, 12.8 mmol), X-Phos (0.240 g, 0.503 mmol), and Pd₂(dba)₃·CHCl₃ (0.248 g, 0.240 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl (3aR,6aS)-5-(2-fluoro-5-(methoxycarbonyl)phenyl)hexahydropyrrolo[3,4-c]pyrrole-2(1H)-carboxylate (28) as a brown oil (0.400 g, 26%): ¹H NMR (400 MHz, acetone-d₆) δ 7.37–7.35 (m, 2H), 7.14–7.08 (m, 1H), 3.88 (s, 3H), 3.61 (brs, 4H), 3.36–3.29 (m, 4H), 3.03 (brs, 2H), 1.38 (s, 9H); ESI MS m/z 365 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl (3aR,6aS)-5-(2-fluoro-5-(methoxycarbonyl)phenyl)hexahydropyrrolo[3,4-c]pyrrole-2(1H)-carboxylate (28, 0.390 g, 1.07 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol), and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-fluoro-3-((3aR,6aS)-hexahydropyrrolo[3,4-c]pyrrol-2(1H)-yl)benzoate (29) as a brown solid (0.220 mg, 78%, crude): ESI MS m/z 265 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-fluoro-3-((3aR,6aS)-hexahydropyrrolo[3,4-c]pyrrol-2(1H)-yl)benzoate (29, 0.210 g, 0.79 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (0.73 mL, 4.19 mmol) and 3-chloropentane-2,4-dione (0.27 mL, 2.39 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of diketone 30 (monitored by LC-MS; ESI MS m/z 337 [M + H]⁺), N₂H₄·H₂O (0.26 mL, 3.15 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-((3aR,6aS)-5-(3,5-dimethyl-1H-pyrazol-4-yl)hexahydropyrrolo[3,4-c]pyrrol-2(1H)-yl)-4-fluorobenzoate (31) as a brown oil along with an inseparable and uncharacterized impurity (0.100 g, 35%, crude): ESI MS m/z 359 [M + H]⁺.

Step D.

To a solution of methyl 3-((3aR,6aS)-5-(3,5-dimethyl-1H-pyrazol-4-yl)hexahydropyrrolo[3,4-c]pyrrol-2(1H)-yl)-4-fluorobenzoate (31, 60.0 mg, 0.16 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (20.0 mg, 0.84 mmol). The reaction mixture was stirred at rt for 3 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-((3aR,6aS)-5-(3,5-dimethyl-1H-pyrazol-4-yl)hexahydropyrrolo[3,4-c]pyrrol-2(1H)-yl)-4-fluorobenzoic acid (32) as an amorphous off-white solid (4.8 mg, 8.7%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.40 (m, 2H), 7.16 (m, 1H), 3.60 (m, 2H), 3.20 (m, 2H), 3.01 (m, 2H), 2.88 (m, 4H), 2.11 (s, 6H); ESI MS m/z 345 [M + H]⁺; HPLC 95.4% (AUC), t_R = 11.0 min.

(±)-3-(7-(3,5-Dimethyl-1H-pyrazol-4-yl)-2,7-diazaspiro[4.4]-nonan-2-yl)-4-fluorobenzoic Acid ((±)-38).

Step A.

A mixture of (±)-tert-butyl 2,7-diazaspiro[4.4]nonane-2-carboxylate ((±)-33, 1.07 g, 4.72 mmol) and methyl 3-bromo-4-fluorobenzoate (1.0 g, 4.29 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (4.16 g, 12.84 mmol), X-Phos (240 mg, 0.51 mmol), and Pd₂(dba)₃·CHCl₃ (0.217 g, 0.21 mmol) were then added, and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give (±)-tert-butyl 7-(2-fluoro-5-(methoxycarbonyl)phenyl)-2,7-diazaspiro[4.4]nonane-2-carboxylate ((±)-34) as a yellow solid (0.480 g, 30%): ¹H NMR (400 MHz, acetone-d₆) δ 7.33–7.31 (m, 2H), 7.12–7.07 (m, 1H), 3.84 (s, 3H), 3.56–3.44 (m, 2H), 3.42–3.27 (m, 6H), 1.99–1.91 (m, 4H), 1.42 (s, 9H); ESI MS m/z 379 [M + H]⁺.

Step B.

To a 0 °C cooled solution of (±)-tert-butyl 7-(2-fluoro-5-(methoxycarbonyl)phenyl)-2,7-diazaspiro[4.4]nonane-2-carboxylate ((±)-34, 0.470 g, 1.24 mmol) in CH₂Cl₂ (30 mL) was added TFA (3.0 mL, 39.20 mmol), and the resulting solution was stirred at rt for 16 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure (±)-methyl 4-fluoro-3-(2,7-diazaspiro[4.4]nonan-2-yl)benzoate ((±)-35) as a white solid (0.340 g, 99%, crude): ESI MS m/z 279 [M + H]⁺.

Step C.

To a 0 °C cooled solution of (±)-methyl 4-fluoro-3-(2,7-diazaspiro[4.4]nonan-2-yl)benzoate ((±)-35, 0.340 g, 1.22 mmol) in anhydrous THF (10 mL) were added i-Pr₂NEt (1.1 mL, 6.31 mmol) and 3-chloropentane-2,4-dione (0.40 mL, 3.66 mmol) simultaneously, and the resulting solution was stirred for 16 h under a N₂ atmosphere while gradually warming to rt. Upon in situ formation of (±)-36 (monitored by LC-MS; ESI MS m/z 377 [M + H]⁺), N₂H₄·H₂O (0.62 mL, 12.23 mmol, 64–65% solution in H₂O) was then added and the mixture was continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give (±)-methyl 3-(7-(3,5-dimethyl-1H-pyrazol-4-yl)-2,7-diazaspiro[4.4]nonan-2-yl)-4-fluorobenzoate ((±)-37) as a brown oil (65.0 mg, 14%, crude): ESI MS m/z 373 [M + H]⁺.

Step D.

To a solution of (±)-methyl 3-(7-(3,5-dimethyl-1H-pyrazol-4-yl)-2,7-diazaspiro[4.4]nonan-2-yl)-4-fluorobenzoate ((±)-37, 60.0 mg, 0.16 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (19.2 mg, 0.81 mmol). The reaction mixture was stirred at rt for 16 h and concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL) and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give (±)-3-(7-(3,5-dimethyl-1H-pyrazol-4-yl)-2,7-diazaspiro[4.4]nonan-2-yl)-4-fluorobenzoic acid ((±)-38) as an amorphous white solid (10.5 mg, 18%): 1H NMR (400 MHz, DMSO-d₆) δ 7.25 (m, 2H), 7.14 (m, 1H), 3.46 (m, 4H), 3.02 (m, 2H), 2.99 (m, 2H), 2.11 (s, 6H), 1.94–1.87 (m, 4); ESI MS m/z 359 [M + H]⁺; HPLC 97.3% (AUC), t_R = 11.3 min.

3-(6-(3,5-Dimethyl-1H-pyrazol-4-yl)-2,6-diazaspiro[3.3]heptan-2-yl)-4-fluorobenzoic Acid (44).

Step A.

A mixture of tert-butyl 2,6-diazaspiro[3.3]heptane-2-carboxylate (39, 1.0 g, 5.04 mmol) and methyl 3-bromo-4-fluorobenzoate (2.25 g, 9.65 mmol) in anhydrous 1,4-dioxane (50 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (4.17 g, 12.79 mmol), X-Phos (0.240 g, 0.50 mmol) and Pd₂(dba)₃·CHCl₃ (0.248 g, 0.24 mmol) were then added and the mixture was stirred in a sealed tube at 110 °C for 16 h. The mixture was allowed to cool to rt and then concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give tert-butyl 6-(2-fluoro-5-(methoxycarbonyl)phenyl)-2,6-diazaspiro-[3.3]heptane-2-carboxylate (40) as a brown oil (1.34 g, 76%): ¹H NMR (acetone-d₆, 400 MHz) δ 7.39–7.35 (m, 1 H), 7.11–7.05 (m, 2 H), 4.13 (s, 3 H), 4.12–4.09 (m, 4 H), 3.85–3.78 (m, 4 H), 1.37 (s, 9 H); ESI MS m/z 351 [M + H]⁺.

Step B.

To a 0 °C cooled solution of tert-butyl 6-(2-fluoro-5-(methoxycarbonyl)phenyl)-2,6-diazaspiro[3.3]heptane-2-carboxylate (40, 1.30 g, 3.71 mmol) in CH₂Cl₂ (30 mL) was added TFA (5.0 mL, 65.33 mmol) and the resulting solution was stirred at rt for 3 h while gradually warming to rt. The mixture was then concentrated under reduced pressure and diluted with H₂O (30 mL), basified with saturated aqueous NaHCO₃ solution (50 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The crude material was triturated with Et₂O and filtered to give pure methyl 4-fluoro-3-(2,6-diazaspiro[3.3]heptan-2-yl)benzoate (41) as a brown oil (0.600 g, 65%, crude): ESI MS m/z 251 [M + H]⁺.

Step C.

To a 0 °C cooled solution of methyl 4-fluoro-3-(2,6-diazaspiro[3.3]heptan-2-yl)benzoate (41, 0.600 g, 1.65 mmol) in anhydrous THF (5 mL) were added i-Pr₂NEt (1.37 mL, 7.89 mmol) and 3-chloropentane-2,4-dione (0.55 mL, 4.91 mmol) simultaneously and the resulting solution was stirred for 16 h under N₂ atmosphere while gradually warming to rt. Upon in situ formation of 42 (monitored by LC-MS; ESI MS m/z 349 [M + H]⁺), N₂H₄·H₂O (0.82 g, 16.5 mmol, 64–65% solution in H₂O) was then added and the mixture continued to stir at rt for an additional 1 h. The mixture was then diluted with H₂O (50 mL) and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(6-(3,5-dimethyl-1H-pyrazol-4-yl)-2,6-diazaspiro-[3.3]heptan-2-yl)-4-fluorobenzoate (43) as a brown oil (0.100 g, 18%): ESI MS m/z 345 [M + H]⁺.

Step D.

To a solution of methyl 3-(6-(3,5-dimethyl-1H-pyrazol-4-yl)-2,6-diazaspiro[3.3]heptan-2-yl)-4-fluorobenzoate (43, 51.0 mg, 0.14 mmol) in CH₃OH (4 mL), THF (4 mL) and H₂O (2 mL) was added anhydrous LiOH (17.0 mg, 0.70 mmol). The reaction mixture was stirred at rt for 16 h and was concentrated under reduced pressure. The aqueous layer was then diluted with H₂O (30 mL) and neutralized to approximately pH = 7 with 2 N aqueous HCl (monitored with Hydrion pH paper). The aqueous mixture was extracted with EtOAc (3 × 50 mL), and the combined organic solution was washed with brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(6-(3,5-dimethyl-1H-pyrazol-4-yl)-2,6-diazaspiro-[3.3]heptan-2-yl)-4-fluorobenzoic acid (44) as an amorphous white solid (13.0 mg, 26%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.31 (m, 1H), 7.17 (t, J = 9.2 Hz, 1H), 7.7 (d, J = 8.4 Hz, 1H), 4.09 (m, 4H), 3.96 (m, 4H), 2.10 (s, 6H); ESI MS m/z 331 [M + H]⁺; HPLC 98.8% (AUC), t_R = 10.7 min.

3-(4-(3,5-Dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoic Acid (46).

Step A.

To a solution of methyl 3-(4-(2,4-dioxopentan-3-yl)piperazin-1-yl)-4-fluorobenzoate (18a, 80.0 mg, 0.23 mmol) in CH₃OH (2 mL) was added NH₂OH·HCl (32.0 mg, 0.47 mmol), and the resulting solution was stirred at rt for 16 h. The mixture was concentrated under reduced pressure, and the resulting residue was chromatographed over silica gel (0–50% EtOAc in hexanes) to give methyl 3-(4-(3,5-dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoate (45) as a brown solid (20.1 mg, 25%); ESI MS m/z 334 [M + H]⁺.

Step B.

To a solution of methyl 3-(4-(3,5-dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoate (45, 8.1 mg, 0.023 mmol) in CH₃OH (1 mL), THF (1 mL), and H₂O (0.5 mL) was added anhydrous LiOH (2.7 mg, 0.11 mmol). The reaction mixture was stirred at rt for 1 h and then concentrated under reduced pressure. The aqueous layer was diluted with H₂O (15 mL) and neutralized with 2 N aqueous HCl. The aqueous mixture was extracted with EtOAc (3 × 10 mL), and the combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(3,5-dimethylisoxazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid (46) as a white solid (2.5 mg, 34%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.71–7.59 (m, 1H), 7.19–7.10 (m, 2H), 3.18–3.3.17 (m, 4H), 3.12–3.11 (m, 4H), 2.38 (s, 3H), 2.25 (s, 3H); ESI MS m/z 320 [M + H]⁺; HPLC 96.8% (AUC), t_R = 13.7 min.

3-(4-(2H-Tetrazol-5-yl)piperazin-1-yl)-4-fluorobenzoic Acid (48).

Step A.

To a 0 °C cooled solution of methyl 4-fluoro-3-(piperazin-1-yl)benzoate (17a, 0.200 g, 0.84 mmol) and i-Pr₂NEt (0.5 mL, 2.52 mmol) in THF (3 mL) was added CNBr (0.106 g, 1.01 mmol), and the mixture was stirred for 1 h while gradually warming to rt. The mixture was concentrated under reduced pressure, and the residue was dissolved in DMF (3 mL). To this solution were added NH₄Cl (0.449 g, 8.4 mmol) and NaN₃ (0.546 g, 8.4 mmol), and the resulting mixture was heated at 120 °C for 12 h. The mixture was then allowed to cool to rt, diluted with H₂O (10 mL), and extracted with EtOAc (3 × 50 mL). The combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure to give methyl 3-(4-(2H-tetrazol-5-yl)piperazin-1-yl)-4-fluorobenzoate (47) (0.160 g, 62%), which was taken into the next step as is without purification: ESI MS m/z 307 [M + H]⁺.

Step B.

To a solution of methyl 3-(4-(2H-tetrazol-5-yl)piperazin-1-yl)-4-fluorobenzoate (47, 0.160 g, 0.52 mmol) in CH₃OH (4 mL),THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (0.125 g, 5.2 mmol), and the resulting mixture was stirred at rt for 1 h. The mixture was then concentrated under reduced pressure, and the aqueous layer was diluted with H₂O (15 mL) and acidified to pH = 3 with 2 N aqueous HCl. The aqueous mixture was then extracted with EtOAc (3 × 10 mL), and the combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 3-(4-(2H-tetrazol-5-yl)piperazin-1-yl)-4-fluorobenzoic acid (48) as a white solid (80.2 mg, 53%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.59 (m, 2H), 7.25 (dd, J = 7.6, 3.6 Hz, 1H), 3.52 (br, 4H), 3.14 (br, 4H); ESI MS m/z 293 [M + H]⁺; HPLC 98.3% (AUC), t_R = 11.3 min.

4-Fluoro-3-(4-(2-(trifluoromethyl)phenyl)piperazin-1-yl)benzoic Acid (50).

Step A.

A mixture of methyl 4-fluoro-3-(piperazin-1-yl)benzoate (17a, 0.200 g, 0.84 mmol) and 1-bromo-2-(trifluoromethyl)benzene (0.263 g, 1.17 mmol) in anhydrous 1,4-dioxane (10 mL) was degassed with N₂ for 5 min. Cs₂CO₃ (0.816 g, 2.50 mmol), X-Phos (79.1 mg, 0.16 mmol), and Pd₂(dba)₃·CHCl₃ (73.2 mg, 0.08 mmol) were then added, and the reaction mixture was stirred at reflux for 16 h under an atmosphere of N₂. The mixture was allowed to cool to rt and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–30% EtOAc in hexanes) to give methyl 4-fluoro-3-(4-(2-(trifluoromethyl)phenyl)-piperazin-1-yl)benzoate (49) (30.2 mg, 9%) as a brown solid: ¹H NMR (400 MHz, CDCl₃) δ 7.69–7.63 (m, 3H), 7.61–7.60 (m, 1H), 7.52–7.41 (m, 1H), 7.08–7.03 (m, 2H), 3.88 (m, 3H), 3.25–3.23 (m, 4H), 3.10–3.08 (m, 4H); ESI MS m/z 383 [M + H]⁺.

Step B.

To a solution of methyl 4-fluoro-3-(4-(2-(trifluoromethyl)-phenyl)piperazin-1-yl)benzoate (49, 30.0 mg, 0.78 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (9 mg, 0.39 mmol). The mixture was stirred at rt for 1 h and then concentrated under reduced pressure. The aqueous layer was diluted with H₂O (10 mL) and acidified to pH = 3 with 2 N aqueous HCl. The aqueous mixture was extracted with EtOAc (3 × 10 mL), and the combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 4-fluoro-3-(4-(2-(trifluoromethyl)phenyl)piperazin-1-yl)benzoic acid (50) as a white solid (15.2 mg, 53%): ¹H NMR (400 MHz, CDCl₃) δ 7.76–7.72 (m, 2H), 7.64 (d, J = 8.0 Hz, 1H), 7.52–7.50 (m, 1H), 7.42 (d, J = 8.0 Hz, 1H), 7.23 (m, 1H), 7.10 (dd, J = 8.0, 4.0 Hz, 1H), 3.26 (m, 4H), 3.11 (m, 4H); ESI MS m/z 369 [M + H]⁺; HPLC 96.7% (AUC), t_R = 16.0 min.

4-Fluoro-3-(4-((3-methyl-1H-pyrazol-5-yl)methyl)piperazin-1-yl)-benzoic Acid (52).

Step A.

A solution of methyl 4-fluoro-3-(piperazin-1-yl)benzoate (17a, 0.200 g, 0.84 mmol) and 3-methyl-1H-pyrazole-5-carbaldehyde (0.111 g, 1.0 mmol) in 1,2-dichloroethane (5 mL) and HOAc (0.01 mL, 0.17 mmol) was stirred at rt for 1 h. NaBH(OAc)₃ (0.267 g, 1.26 mmol) was then added, and the mixture was stirred for 12 h at 60 °C. The mixture was then allowed to cool to rt and diluted with CH₂Cl₂ (10 mL). The organic mixture was washed with saturated aqueous NaHCO₃ (10 mL), H₂O (10 mL), and brine. The organic layer was then dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give methyl 4-fluoro-3-(4-((3-methyl-1H-pyrazol-5-yl)methyl)piperazin-1-yl)benzoate (51) as a white solid (0.210 g, 76%): ESI MS m/z 333 [M + H]⁺.

Step B.

To a solution of methyl 4-fluoro-3-(4-((3-methyl-1H-pyrazol-5-yl)methyl)piperazin-1-yl)benzoate (51, 0.200 g, 0.60 mmol) in CH₃OH (4 mL), THF (4 mL), and H₂O (2 mL) was added anhydrous LiOH (144.5 mg, 6.0 mmol), and the resulting mixture was stirred at rt for 3 h. The mixture was concentrated under reduced pressure, and the resulting aqueous layer was diluted with H₂O (10 mL) and neutralized with 2 N aqueous HCl. The aqueous mixture was then extracted with EtOAc (13 × 10 mL), and the combined organic extracts were washed with brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 4-fluoro-3-(4-((3-methyl-1H-pyrazol-5-yl)methyl)piperazin-1-yl)benzoic acid (52) as a white solid (0.120 g, 63%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.50–7.46 (m, 2H), 7.33 (s, 1H), 7.13 (dd, J = 8.8, 4.0 Hz, 1H), 3.30 (s, 2H),2.95 (br, 4H), 2.45 (br, 4H), 2.12 (s, 3H); ESI MS m/z 319 [M + H]⁺; HPLC 98.7% (AUC), t_R = 10.2 min.

3-(4-(3,5-Dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzamide (53).

Step A.

To a mixture of 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzoic acid (14, 0.100 g, 0.314 mmol), HBTU (0.178 g, 0.471 mmol), and i-Pr₂NEt (0.218 mL, 1.26 mmol) in DMF (4 mL) was added NH₄Cl (16.7 mg, 0.314 mmol). The resulting solution was stirred at rt for 18 h under an atmosphere of N₂. The mixture was diluted with H₂O (10 mL) and extracted with EtOAc (3 × 20 mL). The combined organic extracts were washed with H₂O (3 × 20 mL) and brine, dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting crude residue was chromatographed over silica gel (0–80% EtOAc in hexanes) to give 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzamide (53) as a white solid (71.7 mg, 72%): ¹H NMR (400 MHz, DMSO-d₆) δ 11.84 (br, 1H), 7.94 (bs, 1H), 7.54–7,48 (m, 2H), 7.31 (s, 1H), 7.14 (m, 1H), 3.06–3.01 (m, 8H), 2.10 (s, 6H); ESI MS m/z 318 [M + H]⁺; HPLC >99% (AUC), t_R = 10.5 min.

1-(3,5-Dimethyl-1H-pyrazol-4-yl)-4-(2-fluoro-5-(2H-tetrazol-5-yl)phenyl)piperazine (54).

Step A.

A mixture of 3-(4-(3,5-dimethyl-1H-pyrazol-4-yl)piperazin-1-yl)-4-fluorobenzamide (53, 0.180 g, 0.526 mmol), NaN₃ (0.142 g, 0.375 mmol), and tetrachlorosilane (98.5 mg, 0.579 mmol) in CH₃CN (4 mL) was stirred at 80 °C for 18 h in a sealed vessel. The reaction mixture was allowed to cool to rt and diluted with saturated NaHCO₃ (5 mL). The aqueous mixture was extracted with CHCl₃ (3 × 50 mL), and the combined organic extracts were washed with brine (50 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure. The resulting residue was chromatographed over silica gel (0–10% CH₃OH in CH₂Cl₂) to give 1-(3,5-dimethyl-1H-pyrazol-4-yl)-4-(2-fluoro-5-(2H-tetrazol-5-yl)phenyl)piperazine (54) as a white solid (60.3 mg, 30%): ¹H NMR (400 MHz, DMSO-d₆) δ 7.69 (m, 1H), 7.67 (m, 1H), 7.36–7. (dd, J = 4, 8.4 Hz, 1H); 3.12 (m, 4H), 3.09 (m, 4H), 2.13 (s, 6H); ESI MS m/z 343 [M + H]⁺; HPLC >99% (AUC), t_R = 11.1 min.

ABBREVIATIONS

Aβ

β-amyloid

Abca4

ATP-binding cassette, subfamily A (ABC), member 4

ADME

absorption, distribution, metabolism, elimination

AMD

age-related macular degeneration

Arg

arginine

Asn

asparagine

ATTR

transthyretin amyloidosis

ATTR-CM

transthyretin amyloidosis cardiomyopathy

ATTR-PN

transthyretin amyloidosis polyneuropathy

AUC

area under the curve

CH₂Cl₂

dichloromethane

CH₃CN

acetonitrile

CH₃OH

methyl alcohol

CL

clearance

CL_int

intrinsic clearance

Cs₂CO₃

cesium carbonate

CSF

cerebrospinal fluid

CYP

cytochrome P450

DMF

N,N-dimethylformamide

Et₂O

diethyl ether

EtOAc

ethyl acetate

%F

% oral bioavailability

FAC

familial amyloid cardiomyopathy

FAP

familial amyloid neuropathy

FITC

fluorescein isothiocyanate

FP

fluorescence polarization

Glu

glutamic acid

Gly

glycine

h

hour(s)

HBP

halogen binding pocket

HBTU

(2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate

N₂H₄·H₂O

hydrazine hydrate

HCl

hydrochloric acid

hERG

human ether-a-go-go channel

HLM

human liver microsomes

HOAc

acetic acid

HRMS

high-resolution mass spectrometry

H₂SO₄

sulfuric acid

i-Pr₂NEt

N,N-diisopropylethylamine

Ile

isoleucine

IV

intravenous

K₂CO₃

potassium carbonate

Leu

leucine

LiOH

lithium hydroxide

Lys

lysine

min

minutes

MLM

mouse liver microsomes

NaBH₄

sodium borohydride

NaBH(OAc)₃

sodium triacetoxyborohydride

NaN₃

sodium azide

NaNO₂

sodium nitrite

NH₄Cl

ammonium chloride

PBS

phosphate-buffered saline

PD

pharmacodynamics

PDB

Protein Data Bank

Pd₂ (dba)₃·CHCl₃

Tris-(dibenzylideneacetone)dipalladium(0)-chloroform adduct

Phe

phenylalanine

PK

pharmacokinetics

PO

oral

PPARγ

nuclear peroxisome proliferator-activated receptor-γ

%PPB

% plasma protein binding

RBP4

retinol binding protein-4

RLM

rat liver microsomes

RPE

retinal pigment epithelium

rt

room temperature

SAR

structure–activity relationship

Ser

serine

SPA

scintillation proximity assay

SSA

senile systemic amyloidosis

STD NMR

saturation transfer difference nuclear magnetic resonance

TBG

thyroxine binding globulin

TFA

trifluoroacetic acid

THF

tetrahydrofuran

Thr

threonine

TTR

transthyretin

Trp

tryptophan

Tyr

tyrosine

T4

thyroxine

Val

valine

V _ss

volume of distribution at steady state

X-Phos

2-dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl