Structure–Activity Relationship for the Picolinamide Antibacterials that Selectively Target Clostridioides difficile

Abstract

Clostridioides difficile is a leading health threat. This pathogen initiates intestinal infections during gut microbiota dysbiosis caused by oral administration of antibiotics. C. difficile is difficult to eradicate due to its ability to form spores, which are not susceptible to antibiotics. To address the urgent need for treating recurrent C. difficile infection, antibiotics that selectively target C. difficile over common gut microbiota are needed. We herein describe the class of picolinamide antibacterials which show potent and selective activity against C. difficile. The structure–activity relationship of 108 analogues of isonicotinamide 4, a compound that is equally active against methicillin-resistant Staphylococcus aureus and C. difficile, was investigated. Introduction of the picolinamide core as exemplified by analogue 87 resulted in exquisite potency and selectivity against C. difficile. The ability of the picolinamide class to selectively target C. difficile and to prevent gut dysbiosis holds promise for the treatment of recurrent C. difficile infection.

Clostridioides (formerly known as Clostridium) difficile is a Gram-positive anaerobic bacterium which causes intestinal infections that can lead to diarrhea, sepsis, and life-threatening colitis.¹⁻³ Since the outbreak of the hypervirulent strain NAP1/027 in 2000,^4,5 the incidence of C. difficile infections (CDIs) continues to trend upward.⁶ In 2017 alone, CDI resulted in 223 900 hospitalizations and 12 800 deaths in the United States, with nearly $6.3 billion of estimated healthcare costs.^7,8 A challenge in the treatment of CDI is the formation of spores that are not susceptible to antibiotics, resulting in 24–27% recurrence of CDI.^9,10

CDI is currently treated with metronidazole (MTZ), vancomycin (VAN), and fidaxomicin (FDX), with VAN and FDX recommended for initial CDI and MTZ used only in mild cases where VAN or FDX is unavailable. Furthermore, MTZ is no longer recommended for repeated or prolonged CDI treatment due to neurotoxicity.¹¹ The failure rates of these antibiotics are 22.4%, 14.2%, and 12%, respectively.^9,12 In addition, treatment of CDI with MTZ and VAN promotes emergence of vancomycin-resistant enterococci.¹³ The perturbation of gut microbiota (gut dysbiosis) caused by antibiotics allows for C. difficile colonization.¹⁴ While VAN and FDX have similar clinical cure efficacy, patients treated with FDX had lower recurrence (15.4% for FDX compared to 25.3% for VAN), which was attributed to the higher selectivity of fidaxomicin toward C. difficile than to gut microbiota.¹⁵ New antibacterial agents that selectively target C. difficile are needed to tackle persistent and recurrent CDI.

We discovered the isonicotinamide antibacterial 4 during the structure–activity relationship (SAR) exploration on the cinnamonitrile class of antibiotic adjuvants (compounds 1 and 2, Figure 1) against methicillin-resistant Staphylococcus aureus (MRSA).¹⁶ Cinnamonitrile 1, a protein kinase(s) inhibitor, showed promise as an adjuvant for β-lactam antibiotics and also exhibited modest antibacterial activity.¹⁷ Since azoles are common among protein kinase inhibitors,^17,18 the oxazole was introduced as a favored moiety (compound 3, Figure 1). Upon insertion of the oxazole, the resulting series of molecules showed respectable Gram-positive antibacterial activity (minimum inhibitory concentration (MIC) ≤ 4 μg·mL^–1 against MRSA strain NRS70). Further modifications of 3 (amide linker, pyridine core, and removal of the bridging oxygen) led to the potent isonicotinamide antibacterial 4.

Figure 1.

Evolution of the structural template of the cinnamonitrile class of potentiators into the isonicotinamide antibacterial 4.

Compound 4 was evaluated against the ESKAPE panel (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species). These bacteria account for the majority of human problematic infections.^19,20 The compound was active only against the Gram-positive members of the panel (E. faecium and S. aureus). We further evaluated 4 against C. difficile ATCC 43255, also a Gram-positive bacterium, and obtained an MIC of 0.25 μg·mL^–1. These results provided the impetus for an SAR investigation around this structural template. We report herein the preparation and evaluation of 108 analogues of compound 4 (Figure 2), of which the prototype picolinamide 5 was reported recently.²¹ As will be outlined, several of these analogues exhibited exceptional activities against C. difficile, both in potency of the antibiotics and in selectivity toward the organism. The aforementioned selectivity in targeting C. difficile is extremely important, as it would avoid gut dysbiosis and mitigates recurrent infections, which are at the root of the bad outcome in the clinical setting.

Figure 2.

Left, the four sites of structure modification for the SAR study as highlighted by the boxes. Right, the structure of picolinamide 5.

The initial compound of interest, 4, was prepared in three steps following the general Scheme 1A. The first step was a Suzuki-Miyaura coupling reaction²² between the chloro-cyanopyridine and arylboronic acid 6, using Pd(PPh₃)₄ as catalyst and potassium carbonate as a base, yielding the cyano-derivative 7. Alkaline hydrolysis of the nitrile gave the carboxylic acid 8. A coupling reaction between 8 and 1,3-oxazol-2-amine (or other amines) using propanephosphonic anhydride (T3P)²³ as the coupling reagent led to derivatives 9. The preparation of compounds possessing the diarylether moiety followed a similar synthetic route, but using an initial nucleophilic aromatic substitution of a chloro-cyanopyridine with a phenol derivative (Scheme 1B). We used the general synthetic route to derivative 13(21) for syntheses of compounds 112–119.

Scheme 1. Synthetic Routes for Accessing (A) Biaryl and (B) Biarylether Derivatives.

The syntheses produced 108 analogues corresponding to diversification at four structural regions (Figure 2, left structure with four different colored boxes). A representative C. difficile strain (ATCC43255) along with MRSA strain NRS70 were chosen for MIC evaluation. Compounds with MICs of ≤8 μg·mL^–1 against C. difficile were considered active.

SAR¹ involved modifications of the oxazole ring in lead compound 4, a main substructure that defines the antibacterial activity of the isonicotinamides. Compounds 14–31 were prepared. The majority of these compounds retained activity against both bacteria, albeit with poor selectivity (Table S1).

The SAR study then focused on optimizing the phenyl ring (SAR³) with respect to the central pyridine ring (SAR²), while leaving the oxazole group intact. Most of the compounds containing isonicotinamide (32–74, Table S2) were active against C. difficile, however the selectivity between MRSA and C. difficile was <16, except for the nitro-containing analogues (49–50). The 2,6-substituted picolinamides (75–80) were inactive.

The most compelling modification was 2,4-substitution of the picolinamide (87–98). This substitution imparted the desired selectivity toward C. difficile compared to MRSA. Picolinamide 87, a constitutional isomer of compound 4, was inactive against MRSA (MIC = 128 μg·mL^–1) yet active against C. difficile (MIC = 0.125 μg·mL^–1). This exquisite selectivity of 1024-fold was achieved by the mere repositioning of one nitrogen atom. Moreover, compounds 87–98 lacked activity against MSSA and E. faecium (Table S3).

Introduction of an ether linkage between the pyridine core and the phenyl ring resulted in MIC values for C. difficile of ≤1 μg·mL^–1. The picolinamides (106–111, Table 1) had significantly higher selectivity for C. difficile than the isonicotinamides (99–105, Table S2). Unfortunately, the introduction of the ether linkage did not increase the antibacterial activity of the picolinamides. These SAR studies led us to introduce a carboxylate, either as a salt or as an ester (compounds 112-119), which improved water solubility of the family while still maintaining potent activity against C. difficile (0.12 μg·mL^–1 for 5 and 0.50 μg·mL^–1 for 116).

Table 1. Activity of Picolinamides Analogues against MRSA and C. difficile (SAR³ and SAR⁴).

^aNRS70 strain, resistant to erythromycin and spectinomycin.

^bATCC 43255, ribotype 087.

^cSelectivity is defined as MIC_MRSA/MIC_C.difficile.

^dReported previously²¹ and included for comparison.

The XTT (2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-carboxanilide) assay²⁴ with HeLa cells was performed on representative compounds (Table 2). Most of the compounds tested had IC₅₀ values >64 μg·mL^–1 (except for picolinamides 94 and 114, and isonicotinamide 99). Compound 87 had an IC₅₀ of 95.2 ± 1.0 μg·mL^–1, 760-fold higher than its MIC (0.125 μg·mL^–1).

Table 2. XTT IC₅₀ of Selected Picolinamides Analogues with HeLa Cells.

	IC50 (μg·mL–1)
4	79.9 ± 1.0
5	179.3 ± 3.6a
87	95.2 ± 1.0
89	170.2 ± 2.5
90	NT at 64b
92	NT at 256b
93	NT at 64b
94	40.1 ± 1.6
97	135 ± 8.6
98	114.8 ± 2.2
99	37.0 ± 1.0
106	105.9 ± 2.1
107	88.0 ± 2.3
108	93.3 ± 1.7
109	77.4 ± 2.3
110	91.6 ± 1.3
114	53.3 ± 3.6
116	NT at 128b
118	79.9 ± 6.3
119	123.6 ± 3.3

^aReported previously²¹ and included here for comparison.

^bNT, no detectable toxicity at the specified concentration.

We evaluated the activity of the picolinamides against the common gut bacteria: Bacteroides fragilis, Bifidobacterium longum, Corynebacterium spp., Fusobacterium nucleatum, Lactobacillus reuteri, Lactobacillus gasseri, Veillonella sp., and Eubacterium sp.²⁵⁻³⁰ An optimum antibacterial will target selectively the pathogenic bacterium and not the natural gut flora. The selectivity (MIC_gutbacteria/MIC_{C.difficileATCC43255}) ranged from 4 to 128 for MTZ, 0.5 to 64 for VAN, 1 to 512 for FDX, 2 to 64 for 4, 32 to 128 for 5, 32 to 128 for 87, and 8 to 16 for 108 (Table 3). Picolinamides 5 and 87 were more selective than 4 and more selective than the clinically used antibiotics VAN, MTZ, and FDX. The selectivity for C. difficile ATCC43255 with respect to Bifidobacterium longum, a major gut bacterium that is reported to help repress CDI,^27,31 was 128 for 87 compared to 4 for MTZ, 0.5 for VAN, and 0.16 for FDX.

Table 3. MIC Values (μg·mL^–1) for MTZ, VAN, FDX, and Compounds 4, 5, 87, and 108 against Major Gut Bacteria.

	MTZ	VAN	FDX	4	5a	87	108
C. difficile ATCC43255	0.25	0.5	0.0625	0.25	0.125	0.125	1
Bacteroides fragilisa	1	16	>32	1	4	4	8
Bifidobacterium longumb	1	0.25	<0.01	>16	16	>16	16
Corynebacterium spp.c	>32	0.5	<0.0625	0.5	8	16	>16
Fusobacterium nucleatumid	2	0.25	<0.0625	8	4	8	8
Lactobacillus reuterie	>32	>32	>32	8	4	4	4
Lactobacillus gasserif	>32	1	2	>16	>16	>16	>16
Veillonella sp.g	2	>32	8	4	16	16	16
Eubacterium sp.h	1	2	16	>16	16	8	8

^aStrain HM-709, a Gram-negative, anaerobic bacterium that is commensal and critical to host immunity and a minor component of the human gut microflora (<1%).

^bStrain HM-846, an anaerobic, nonsporulating Gram-positive bacterium commonly found in the normal human intestinal microflora isolated from human feces.

^cStrain HM-784, an aerobic or facultatively anaerobic Gram-positive bacterium that occurs in the mucosa and normal skin flora of humans and animals.

^dStrain HM-992, an anaerobic, nonsporulating Gram-negative bacterium commonly found in the gastrointestinal tract.

^eStrain HM-102, an anaerobic Gram-positive bacterium commonly found in the normal human gastrointestinal tract, and used frequently as a probiotic to maintain the balance of gut microbial flora.

^fStrain HM-644, a facultative, anaerobic Gram-positive bacterium commonly found in the normal human gastrointestinal tract, used frequently in yogurt production as a probiotic to suppress Helicobacter pylori infections.

^gStrain HM-49, a nonsporulating Gram-negative bacterium commonly found in the intestinal tract of humans and animals.

^hStrain HM-178, an anaerobic, nonsporulating Gram-positive bacterium commonly found in the gastrointestinal flora of humans and animals.

Recurrence of CDI for the less selective VAN is 25%, while that for the somewhat more selective FDX is 15%.¹⁵ Picolinamide 87 is >1000-fold selective. As gut dysbiosis contributes to CDI recurrence, the importance of this high selectivity is self-evident. Selectivity in targeting C. difficile emerges as an important attribute as mere potency (low MIC) might become less significant given that a compound such as 87 has already a remarkably very low/potent MIC of 0.125 μg·mL^–1 (nanomolar range).

For the treatment of CDI, compounds that are poorly absorbed or not absorbed and achieve high concentrations in the gut are desirable. We selected compounds 4, 87, 107, 108, and 116 for in vivo PK studies in mice based on their potency, safety profile, solubility, and selectivity against C. difficile. Both plasma and feces were collected and analyzed for levels of the compounds and compared to those of compound 5. While all the compounds showed nonquantifiable concentrations in plasma, levels in feces were higher (Table 4). The selected compounds showed 2- to 50-fold higher concentrations in the feces over the MIC values.

Table 4. Drug Concentrations in Fecal Samples of Selected Compounds.

	MIC, μg·mL–1a	selectivityb	dose, mg·kg–1	μg g–1 feces	conc. feces/MIC
4	0.25	1	20	2.8	11
5c	0.125	128	20	13	108
87	0.125	1024	10	0.79	6.6
107	1	≥256	20	2.4	2.4
108	1	64	20	49	49
116	0.50	128	20	3.7	7.4

^aC. difficile strain, ATCC 43255, ribotype 087.

^bSelectivity is defined as MIC_MRSA/MIC_{C. difficile}.

^cReported previously; included for ease of comparison.

The lack of selective antibiotics for the treatment of CDI contributes to gut dysbiosis and recurrence of CDI. The picolinamide family of antibacterials shows exquisite potency and selectivity in targeting C. difficile. Starting with isonicotinamide 4, a compound that is equally active against MRSA and C. difficile, structure optimization gave the new picolinamide 87 with >1000-fold selectivity for C. difficile compared to MRSA (strain NRS70). Compound 87 shows exceptional selectivity against C. difficile while lacking activity against other, normal Gram-positive and Gram-negative gut microbiota. As gut dysbiosis contributes to CDI recurrence, the picolinamides have the potential for treatment of recurrent CDI.

Glossary

Abbreviations

C. difficile

Clostridioides difficile

CDI

C. difficile infection;

ClogP

calculated logP value

FDX

fidaxomicin

MDZ

metronidazole

MRSA

methicillin-resistant Staphylococcus aureus

MSSA

methicillin-sensitive S. aureus

Pd(PPh₃)₄

tetrakis(triphenylphosphine)palladium(0)

PK

pharmacokinetics

SAR

structure–activity relationship

T3P

propylphosphonic anhydride

TLC

thin layer chromatography

VAN

vancomycin

XTT

(2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-carboxanilide).

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00135.

• General synthetic procedures; synthetic experimental procedures and characterization data; X-ray data collection method; crystal data and structure refinement for 87; general biological methods; activity of picolinamide analogues against C. difficile and MRSA, MSSA (ATCC 29213), and E. faecium (NCTC 7171); PK studies; NMR spectra for representative compounds; and molecular formula strings for the synthesized compounds (PDF)

Author Present Address

^§ S.D.B.: Institute of Pharmaceutical Microbiology, University of Bonn, Bonn, Germany.

Author Contributions

The manuscript was written through contributions of all authors. All authors approve the manuscript.

The authors declare no competing financial interest.