Synthesis and Evaluation of Benzylamine Inhibitors of Neuropathogenic Naegleria fowleri “Brain-Eating” Amoeba

Abstract

Current therapy for primary amoebic meningoencephalitis (PAM), a highly lethal brain infection in humans caused by Naegleria fowleri amoeba, is restricted to repurposed drugs with limited efficacy and success. Discovery of an antiamoebic benzylamine scaffold 2 precipitated a medicinal chemistry effort to improve potency, cytotoxicity profile, and drug-like properties. Thirty-four compounds were prepared, leading to compound 28 with significant gains in potency (EC₅₀ = 0.92 μM), solubility, and microsomal stability and a demonstrated absence of cytotoxicity in SH-SY5Y human neuroblastoma cells (CC₅₀ > 20 μM). The compounds demonstrated excellent blood–brain barrier permeability in an in vitro assay, thereby providing a new structural scaffold that inhibits N. fowleri viability and permits the investigation of therapeutic interventions in an understudied neglected disease.

Naegleria fowleri is a thermophilic protozoan that does not require a host to survive; however, the amoeba can cause an almost always fatal brain infection in humans called primary amoebic meningoencephalitis (PAM).¹ The pathogen is typically found in freshwater, soil, and poorly treated recreational and tap water. The amoeba exists in a cyst, trophozoite, or flagellate life stage, and human infection occurs when trophozoites are introduced into the nasal cavity, where the amoeba infiltrates the olfactory mucosa and travels to the frontal lobes of the brain, where it causes tissue damage.^1,2 PAM is usually associated with swimming in freshwater rivers or lakes during warm months when humans encounter the amoeba.¹ The infection progresses rapidly and has a reported fatality rate of greater than 97%.^1,3 Infection often results in death within 1 to 18 days of symptom onset. Out of 154 reported PAM cases in the United States between 1962 and 2021, there have been only four survivors.¹

Current treatment for PAM typically involves the administration of amphotericin B and miltefosine, which can be used in combination with other drugs such as azithromycin, fluconazole, and rifampin.^2,4,5 The efficacy of these drugs is unclear, as very few treated patients have survived. Additionally, the cocktail components have limitations, including toxicity, undesirable side effects, or limited blood–brain barrier (BBB) permeability.^2,4 Successful treatment of PAM is attributed to early detection, which allows for expedient therapeutic intervention.³ However, this infection is not routinely tested for, and the symptoms are often mistaken for viral or bacterial meningitis.⁶

These factors underscore the need for effective and safe PAM treatments. Recent research has described N. fowleri inhibitors, though some discoveries are relatively new^7,8 and progress is slow to yield high-caliber leads with properties that are developmentally suitable.^6,9−14 To identify potential chemical matter that impacts N. fowleri, we surveyed compounds from our lab compound library in a previously described N. fowleri cell viability assay conducted initially at a single 10 μM compound concentration.¹⁵ For compounds demonstrating >50% inhibition, the assay was repeated in a dose–response format, and half-maximal effective inhibitory concentration (EC₅₀) values were resolved. The screen included a series of picolinic acid derivatives, including benzamide 1, that we investigated as Toxoplasma gondii inhibitors (Figure 1).¹⁶

Figure 1.

Structures of T. gondii inhibitor 1 and N. fowleri hit 2.

While benzamide 1 was inactive at 10 μM, the benzylamine analog 2 exhibited promising N. fowleri trophozoite inhibition (EC₅₀ = 6.0 μM). We further examined the compound cytotoxicity in human neuroblastoma cells (SH-SY5Y) as a species-relevant cell model¹⁷ for assessing neurotoxicology in the primary site of PAM infection. Evaluation at two independent concentrations revealed that analog 2 demonstrated a mean 55% cell viability at 20 μM, which was reduced from a mean 86% at a concentration of 1 μM. Given the results, we explored structure–activity relationships (SARs) governing N. fowleri inhibition and sought an improved cytotoxicity profile. Analogs were designed to examine benzylamine linker changes exclusive from or in combination with picolinic core substituent changes or terminal phenyl ring modifications. Methyl picolinates, available commercially or prepared via Fischer esterification or by using SOCl₂, were alkylated on the core-appended amine (analogs 2–6, Scheme 1). Halogenation provided additional core modifications, including those produced through Buchwald–Hartwig or Suzuki couplings (analogs 7–15).

Scheme 1. Synthesis of Picolinic Acid Analogs with Core Substituent or Terminal Phenyl Modifications.

Reagents and conditions: (a) 1:4 H₂SO₄/CH₃OH, 70 °C, 24 h, 42–69%; (b) BrCH₂R, DIPEA, CH₃CN, 70 °C, 18–24 h, 22–41%; (c) 1.3 M LiOH, CH₃OH or THF, RT to 50 °C, 4–24 h, 20–83%; (d) NBS or NCS, CH₃CN, 70 °C, 18 h, 59%; (e) phenylboronic acid, Pd(PPh₃)₂Cl₂, Na₂CO₃, 10:1 CH₃CN/H₂O, 100 °C MW, 3.5 h, 23%; (f) 4-tert-butylbenzyl bromide, KI, K₂CO₃, CH₃CN, 70 °C, 18 h, 47%; (g) phenylboronic acid, Pd(PPh₃)₄, Cs₂CO₃, 4:1 dioxane/H₂O, 85 °C, 18 h, 83%; (h) morpholine or H₂NR, Pd₂(dba)₃, xantphos, Cs₂CO₃, toluene or dioxane, 85–100 °C, 9–18 h, 22–88%; (i) SOCl₂, CH₃OH, 70 °C, 4 h, 69–92%.

Carboxylic acid and aminoalkyl linker changes were also surveyed (Scheme 2). Linker-modified analogs were prepared, after hydrolysis, through N-acylation to give 16, N-sulfonation to provide 17, and deprotonation of 5-chloro-3-methylpicolinic acid with LDA followed by alkylation to afford 18. Buchwald–Hartwig coupling between an aryl bromide and the corresponding amine, followed by hydrolysis, afforded analogs 19–27. An S_NAr reaction between 3-bromo-5-fluoropicolinonitrile and sodium methoxide provided a coupling partner to afford 28 and 29. Nitrile 30 and amide 31 were derived from the same starting material, and chloropyridine 32 was delivered through the coupling of the corresponding aryl bromide starting material.

Scheme 2. Synthesis of Linker- and Acid-Modified Analogs.

Reagents and conditions: (a) 1:4 H₂SO₄/CH₃OH, 70 °C, 24 h, 42–69%; (b) 2-(4-tert-butylphenyl)acetyl chloride, DIPEA, CH₃CN, RT, 24 h, 18%; (c) 1.3 M LiOH, CH₃OH or THF, RT to 50 °C, 4–24 h, 20–83%; (d) 4-tert-butylbenzenesulfonyl chloride, pyridine, and cat. DMAP, CH₂Cl₂, RT, 2 h, 29%; (e) LDA, TEA, then 4-tert-butylbenzyl bromide, THF, −78 °C to RT, 24 h, 60%; (f) SOCl₂, CH₃OH, 70 °C, 4 h, 69–92%; (g) H₂NR, Pd₂(dba)₃, xantphos, Cs₂CO₃, toluene, 85 °C or dioxane, 100 °C, 9–18 h, 22–88%; (h) NaOMe, CH₃OH, 0 to 40 °C, 24 h, 77%; (i) 2.5:1 EtOH/50% aq NaOH, 100 °C, 2 h, 84%; (j) 2.5 M aq NaOH, 30% H₂O₂, acetone, RT, 2 h, 71%.

Reductive amination was employed to generate analogs bearing other heterocyclic cores to the picolinic acid (Scheme 3). Thiazole acid 33 and nicotinate 34 resulted directly or after hydrolysis of the corresponding ester intermediate.

Scheme 3. Synthesis of Analogs with Other Heterocyclic Cores.

Reagents and conditions: (a) 4-tert-butylbenzaldehyde, CF₃CO₂H, STAB, toluene, RT, 2 h, 60%; (b) 1.3 M LiOH, CH₃OH or THF, RT to 50 °C, 4–24 h, 54–60%; (c) 1:4 H₂SO₄/CH₃OH, 70 °C, 24 h, 42–69%; (d) 4-tert-butylbenzaldehyde, AcOH, STAB, CH₂Cl₂, RT, 24 h, 22%.

Compounds in this chemical series were tested against N. fowleri trophozoites (Nf69 strain) at a single compound concentration of 10 μM.¹⁵ Compounds that inhibited amoeba viability by ≥50% were subsequently tested in a dose–response format to obtain EC₅₀ values for comparison. The compounds were also assessed for cytotoxicity in human neuroblastoma cells (SH-SY5Y) at concentrations of 1 and 20 μM (Table 1). Miltefosine, included as a control, afforded an EC₅₀ value of 37.5 μM in the amoeba assay.

Table 1. N. fowleri and Mammalian Cell Viability Data for Key Substituent Changes at R¹, R², and R³.

	structure			N. fowleri dataa		percent cell viabilityb
cmpd	R1	R2	R3	% growth inhibition	EC50 (μM)	1 μM		20 μM
2	Cl	H	CO2H	98.8 ± 0.3	6.0 ± 0.3	87	85	57	53
3	H	H	CO2H	99.2 ± 0.35	3.0 ± 0.7	100	100	89	91
7	Cl	Cl	CO2H	12.3 ± 4.9	>10	100	100	100	100
8	Cl	Br	CO2H	17.6 ± 4.6	>10	100	100	100	100
9	Cl	Ph	CO2H	20.4 ± 4.9	>10	100	100	86	82
10	F	H	CO2H	99.5 ± 0.3	5–10c	91	93	75	73
11	Br	H	CO2H	84.4 ± 2.0	7.2 ± 0.3	99	99	85	85
12	Ph	H	CO2H	99.3 ± 0.4	5–10c	100	100	81	85
13	morpholinyl	H	CO2H	52.8 ± 8.1	11.5 ± 1.9	89	85	75	77
14	CH3	H	CO2H	100.4 ± 0.02	1.4 ± 0.2	96	98	98	98
28	OCH3	H	CO2H	100.3 ± 0.2	0.92 ± 0.03	98	98	99	100
30	Cl	H	CN	22.2 ± 1.6	>10	96	98	57	64
31	Cl	H	CONH2	27.7 ± 6.3	>10	100	100	77	81
32	Cl	H	H	17.6 ± 7.4	>10	100	100	88	88

^aData are reported as averages of triplicate runs (n = 3) with ±SEM at 10 μM and with half-maximal effective concentration (EC₅₀) against N. fowleri (Nf69 strain, ATCC 30215) trophozoites; miltefosine (positive control) N. fowleri EC₅₀ = 35.7 ± 9.1 μM.

^bPercent viability (n = 2), 72 h, human neuronal SH-SY5Y cells, 1 μM or 20 μM compound concentration.

^cThe Hill slope of the dose–response assay was steep, leading to a lack of resolution of EC₅₀ but a clear range as indicated.

Compared to the hit benzylamine 2, the addition of an adjacent substituent at R² resulted in weaker antiamoebic activity (7–9, Table 1). Exchange of the R¹ chlorine atom for a hydrogen atom (analog 3) gave a similar potency to 2, but an improved cytotoxicity profile was observed at both low and high compound concentrations. Analogously, substitution of the chlorine atom for a phenyl ring in 12, a fluorine or bromine atom (analogs 10 and 11, respectively), or a morpholino substituent in 13 resulted in similar potency to hit 2; however, comparable erosion of mammalian cell viability was also observed. Also, fluorine analog 10 and phenyl analog 12 uniquely and consistently demonstrated steep Hill slopes in the amoeba dose–response assay, thereby limiting a sharp determination of EC₅₀ beyond that of 5–10 μM. Better results were obtained when the chlorine atom of 2 was replaced with a methyl or methoxy group (analogs 14 and 28, respectively), as a 4- to 7-fold improvement in potency was observed without cytotoxicity at the highest tested concentration of 20 μM. Alterations to the carboxylic acid moiety, such as replacement with a nitrile (30), amide (31), or hydrogen atom (32), were not tolerated in terms of antiamoebic activity.

Examination of the 4-tert-butylphenyl moiety of hit 2 revealed poor tolerance for changes (Table 2). Exchange of the tert-butyl group for methyl (3), methoxy (19), fluorine (20), trifluoromethyl (21), or a hydrogen atom (22) all resulted in inferior amoeba inhibition and EC₅₀ values >10 μM. Thiazole (4) or thiophene (5) rings inserted in place of the phenyl ring led to a complete loss of activity. As such, linker region changes were evaluated while preserving the tert-butylphenyl component. Analogs with an extended amide (16), a sulfonamide (17), or the aniline-like nitrogen atom replaced with a methylene (18) were not active. Methylation of the linker methylene group or excision of it altogether also led to inactive analogs (23 and 24). Lengthening the benzyl linker by one methylene unit afforded compound 25 and restored activity to what was observed for compound 2 with an EC₅₀ of 3.3 μM and no notable cytotoxicity. Assessment showed that tert-butyl group changes in combination with the extended linker were not tolerated, as observed with previous analogs. Evaluation of compound 28 had revealed that the incorporation of a C5 methoxy group was beneficial, so this change was integrated into an analog with the extended chain, resulting in compound 29 with an EC₅₀ of 1.1 μM but with little improvement compared to compound 2 with regard to cytotoxicity. Picolinic acid core changes were investigated; however, pyrimidine 15, thiazole 33, and nicotinic acid 34 all afforded N. fowleri EC₅₀ values >10 μM (Figure 2).

Table 2. N. fowleri and Mammalian Cell Viability Data for Analogs with Linker and tert-Butylphenyl Group Changes.

^aData are reported as averages of triplicate runs (n = 3) with ±SEM at 10 μM and with half-maximal effective concentration (EC₅₀) against N. fowleri (ATCC 30215) trophozoites; miltefosine (positive control) N. fowleri EC₅₀ = 35.7 ± 9.1 μM.

^bPercent viability (n = 2), 72 h, human neuronal SH-SY5Y cells, 1 μM or 20 μM compound concentration.

Figure 2.

Alternative cores to the picolinic acid hit structure.

Given the improvement in antiamoebic potency and cytotoxicity profile, analogs 14 and 28 were selected for comparison to hit 2 with respect to several physiochemical and drug-like properties (Table 3). Cytotoxicity for these analogs was determined in dose–response format, revealing CC₅₀ values of >20 μM, though inspection of the concentration curves showed clear advantages for compounds 14 and 28, which showed no attenuation of cell viability in neuronal cells from 1 to 20 μM. All three compounds landed in the desirable logD range. Compared to hit 2, significant gains were made in solubility for both analogs 14 and 28, and mouse microsomal stability increased, with the greatest improvement observed for methoxy derivative 28. Subsequent evaluations compared 2 and 28 only, revealing excellent stability in mouse plasma, high protein binding, and excellent brain tissue permeability as assessed by a BBB parallel artificial membrane permeability assay (PAMPA),¹⁸ which showed that 2 and 28 are CNS-penetrant, an important consideration for treating PAM infection.

Table 3. Comparison of Physiochemical and Drug-like Parameters for Selected Compounds 2, 14, and 28.

^aHalf-maximal effective concentration in inhibiting N. fowleri trophozoite viability.

^bHalf-maximal cytotoxic concentration after 72 h in human neuronal SH-SY5Y cells.

^cKinetic solubility, PBS buffer.

^dSourced from CD-1 mouse.

^e2 μM protein concentration

^f10 μM compound concentration; porcine brain lipid extract; 4 h incubation; negative control (CNS−): nadalol, Pe < 0.1 nm/s; positive control (CNS+): carbamazepine, Pe = 66.3 nm/s. ND = not determined.

Naegleria fowleri causes a rare, devastating brain infection in patients who have acquired the pathogen through contaminated water entering the nasal cavity. With a lethality of >97%, improved treatments are needed that specifically and efficaciously address this infection. A screening effort directed at identifying new chemical matter with activity against N. fowleri revealed that benzylamines such as 2 reduced the amoeba viability. Optimization efforts focused on improvement of potency and cytotoxicity, understanding of pharmacophoric structural elements of the scaffold, and assessment of physiochemical and preliminary absorption, distribution, metabolism, and excretion (ADME) parameters to determine feasibility for further analysis of pharmacokinetics and efficacy in animal models. Through synthesis, 34 compounds were prepared, resulting in a 6.5-fold potency enhancement for compound 28 with a concomitant advancement at higher concentrations such that human neuronal cells showed no indication of cytotoxic effects. Analog 28 also featured a boost in aqueous solubility and metabolic stability compared to compound 2. The SAR efforts for the series revealed a strong reliance on the position of the picolinic acid nitrogen atom and the need for the carboxylic acid and tert-butylphenyl moieties, as no tolerance was observed for changes in these regions. The aminoalkyl linker was more amenable to alteration, and substituents on the picolinic acid core permitted tuning of physiochemical and ADME parameters. Importantly, the compounds were found to be CNS-penetrant as determined by a BBB permeability assay, which is a significant parameter in the consideration of any scaffold development plan for PAM. Further studies will focus on pharmacokinetic analysis in mice, sensitivity of other patient isolates besides Nf69, and a study of mechanism of action; however, this work demonstrates that benzylamines inhibit N. fowleri amoeba and provides useful chemical probes for additional studies, which may reveal potential therapeutic opportunity.

Glossary

Abbreviations

ADME

absorption, distribution, metabolism, and excretion

BBB

blood–brain barrier

CNS

central nervous system

PAM

primary amoebic meningoencephalitis

SAR

structure–activity relationship

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.3c00440.

• Assay, experimental details, synthetic procedures, and NMR spectra for final compounds 2–34 reported in this work (PDF)

Author Contributions

J.M.P. and M.M.K. synthesized compounds. J.E.G. guided the medicinal chemistry and synthesis efforts. J.E.M. and C.M.P. assessed compounds in the parasite and enzymatic assays. J.C.M. formulated the Naegleria viability assay and assisted with data analysis. The manuscript was written through contributions of all authors. All of the authors approved the final version of the manuscript.

The authors declare no competing financial interest.

Special Issue

Published as part of ACS Medicinal Chemistry Lettersvirtual special issue “Celebrating the 60th Anniversary of the MIKIW Meeting-in-Miniature”.