Repurposing Eukaryotic Kinase Inhibitors as Colistin Adjuvants in Gram-Negative Bacteria

Abstract

Kinase inhibitors comprise a diverse cohort of chemical scaffolds that are active in multiple biological systems. Currently, thousands of eukaryotic kinase inhibitors are commercially available, have well-characterized targets, and often carry pharmaceutically favorable toxicity profiles. Recently, our group disclosed that derivatives of the natural product meridianin D, a known inhibitor of eukaryotic kinases, modulated behaviors of both Gram-positive and Gram-negative bacteria. Herein, we expand our exploration of kinase inhibitors in Gram-negative bacilli utilizing three commercially available kinase inhibitor libraries and, ultimately, identify two chemical structures that potentiate colistin (polymyxin E) in multiple strains. We report IMD-0354, an inhibitor of IKK-β, as a markedly effective adjuvant in colistin-resistant bacteria and also describe AR-12 (OSU-03012), an inhibitor of pyruvate dehydrogenase kinase-1 (PDK-1), as a potentiator in colistin-sensitive strains. This report comprises the first description of the novel cross-reactivity of these molecules.

Graphical Abstract

The upsurge of antibiotic-resistant bacteria represents an urgent threat to modern healthcare worldwide.¹ To date, bacterial resistance to the entire ensemble of clinically available antibiotics has been reported in some capacity, while many bacterial species have incrementally evolved concurrent resistance to several antibiotic classes, giving rise to multidrug-resistant (MDR) and extensively drug-resistant (XDR) bacteria. Pan-drug-resistant bacteria are no longer hypothetical as, in 2016, a patient in the United States lost their life to a systemic infection caused by a Klebsiella pneumoniae (KP) strain that was resistant to every available antibiotic.² The emergence of these MDR and XDR bacterial infections has highlighted the paucity of new treatment options for patients suffering from Gram-negative (G−) bacterial infections and, in these cases, often forces clinicians to revert to antiquated and unfavorable therapies. Colistin (polymyxin E) is a polycationic, cyclic peptide that is highly effective against G− pathogens; however, it carries significant neurological and nephrotoxic liabilities and is thus considered the drug of last resort for the treatment of G− infections.³ Despite its side effects, clinical use of colistin has risen sharply in recent years.⁴

Before 2015, colistin resistance in G− bacteria was understood as a relatively infrequent phenomenon that typically stemmed from intense selection pressure in critical care settings. In Acinetobacter baumannii and K. pneumoniae specifically, point mutations in genes encoding endogenous two component systems such as the pmrAB operon access high levels of colistin resistance through modification of the lipid A component of lipopolysaccharide (LPS).⁵ Recently, the small, plasmid-borne gene transcript dubbed mobile colistin resistance-1 (mcr-1) was reported, thus providing a mechanism by which colistin resistance could be rapidly disseminated across the human pathogen pool.⁶ Since the first report of mcr-1, seven additional isoforms of mcr (mcr-2–8) have been found across the globe in animals and humans, posing a serious threat to this “last-resort” antibiotic.⁷

Given its significant potency against G− bacteria, a long-standing goal of our research group is to identify small molecules that augment the activity of colistin and either enable reduced dosing to overcome its inherent toxicity or restore efficacy against colistin-resistant bacteria (compounds 1–8, Figure 1).^8-12 Such compounds, generically called antibiotic adjuvants, typically lack inherent antimicrobial activity in monotherapy but, when codosed with specific antibiotics, display marked potentiation of the antibiotic, significantly lowering the amount of drug needed to mitigate infection. Using this approach, we have identified compounds that lower the minimum inhibitory concentration (MIC) of colistin upward of 2048-fold against strains possessing chromosomally encoded colistin resistance determinants, 64-fold against strains that harbor the plasmid-borne colistin-resistant gene mcr-1, and >2048-fold against colistin-sensitive strains.^8,9,11-13

Figure 1.

Structures of compounds 1–8, colistin adjuvants previously disclosed by our group.

One group of compounds (6 and 7) that we recently disclosed as highly effective potentiators of colistin in G− bacteria as well as potent bacterial antibiofilm compounds are based upon the marine natural product meridianin D (5).¹¹ Interestingly, meridianin D and related derivatives are also known to inhibit eukaryotic cyclin-dependent and glycogen synthase kinases.¹⁴ This latter activity, coupled with previous reports that kinase inhibitors (KIs) can also inhibit biofilm formation,¹⁵ led us to posit that eukaryotic KIs may serve as adjuvants that potentiate colistin activity. Herein, we report a comprehensive screen of three separate, commercially available KI libraries for synergy with colistin in both colistin-resistant (col^R) and colistin-susceptible (col^S) G− strains. Of the 924 KIs screened, we describe 51 unique hit molecules with activity in one or more G− strains. Further analysis of the five most potent adjuvants identified two highly potent lead molecules: IMD-0354, an inhibitor of the NF-Kβ inhibitor kinase (IKK-β)¹⁶ for use against col^R bacteria, and AR-12 (OSU-03012), an inhibitor of pyruvate dehydrogenase kinase-1 (PDK-1)¹⁷ as a potentiator in col^S strains.

1. RESULTS AND DISCUSSION

1.1. Pilot Screening of Kinase Inhibitor Libraries.

A screening campaign utilizing four G− bacterial strains (two col^S (AB 5075 and KP ATCC 43816) and two col^R (AB 4106 and KP B9)) was initially employed to evaluate and identify active compounds. Each compound was screened against all four strains alone or cotreated with colistin at 1/16th the colistin MIC against that strain (colistin MICs are as follows: AB 5075 = 1 μg/mL; AB 4106 = 1024 μg/mL; KP 43816 = 0.5 μg/mL; KP B9 = 1024 μg/mL). Any compound that inhibited bacterial growth in combination with colistin without displaying stand-alone toxicity was recorded as a hit. The 149-member Kinase Screening Library from Cayman Chemical (Ann Arbor, MI) was first screened at a concentration of 20 μM and returned 26 unique compounds that potentiated colistin in one or more strains (Table S1 records all hits from this library and provides data from a follow-up determination of each hit’s MIC in monotherapy and the colistin MIC shift in the presence of each compound at 5 μM).

From the colistin potentiation data, we noted measurable, reproducible activity of several hit molecules at 5 μM, and thus to conserve compound, further screens utilized a 10 μM concentration of each kinase inhibitor. Given that numerous hits were found in this library, we then screened a larger, 644-member, commercially available KI set (Selleck Chemical, Houston, TX). This library shared an overlap of 57 structures with the Cayman Chemical library. Screening across the same four strains gave 23 additional unique hits (Table S2).

The third and final library screened was the Kinase Chemogenomic Set (UNC Structural Genomics Consortium, Chapel Hill, NC). This library features 188 kinase inhibitors with known, highly specific targets in eukaryotic cells. As most of these compounds are not commercially available, it is dispensed in extremely low quantities, and thus, it was only possible to screen two strains (A. baumannii 5075 and A. baumannii 4106). From this, two unique hits, PFE-PKIS-12 (CAS No. 1421367–67–0) and GW770249A (CAS No. 501693–48–7), were found and displayed activity limited only to A. baumannii 5075.

1.2. Focused Screening of Multiple Gram-Negative Strains.

After collecting all hits and comparing equimolar potency, additional dry solids of five compounds (compounds 9–13, Figure 2) having activity in two or more strains, considerable published eukaryotic toxicity profiles, and well-defined eukaryotic targets were purchased and tested against a panel of 15 bacterial strains, 11 col^R and 4 col^S (Table 1). This group of compounds returned, in some strains, greater than 2048-fold reduction in colistin MIC, all with limited antimicrobial effects of the adjuvant alone. Of these five compounds, 12 (IMD-0354) was both the most consistently active and potent compound against the col^R strains. It showed synergy with colistin in the chromosomally encoded col^R strains (AB 3941, AB 3942, AB 4106, AB 4112, KP A5, KP B9, and KP C3) as well as strains containing the mcr-1 plasmid-borne colistin resistance gene.^11,13 Activity was observed against the three G− ESKAPE pathogens tested, in addition to Escherichia coli, with the greatest activity seen against A. baumannii and K. pneumoniae, and limited efficacy against Pseudomonas aeruginosa was observed. Furthermore, 12 displayed dose-dependent activity with colistin in both AB 4106 and KP B9 as shown in Table 2, and activity was retained in the presence of 0.001% Triton X100 (Table S3).¹⁸ Growth curves for AB 4106 and KP B9 were then constructed in the presence of 12 alone or in combination with colistin and are shown in Figure 3. Here, we see no perturbance of growth in the presence of compound alone and, in dual therapy, antimicrobial activity of colistin at concentrations well below each strain’s respective colistin MIC.

Figure 2.

Structures of adjuvants utilized in postscreen follow-up experiments: sorafenib (9); AG 879 (10); 6-bromoindirubin-3-oxime (BIO) (11); IMD-0354 (12); OSU-03012 (AR-12) (13); NVP-BAG956 (14); celecoxib (15).

Table 1.

Colistin MIC (μg/mL (Fold Reduction)) of 11 Gram-Negative Strains of A. baumannii (AB), E. coli (EC), K. pneumoniae (KP), and P. aeruginosa (PA) Codosed with 5 μM Adjuvant and Colistin^a

	colistin MIC (μg/mL)	9	10	11	12	13
AB 3941	512	64 (8)	64 (8)	256 (2)	1 (512)	512 (2)
AB 3942	512	32 (16)	64 (8)	128 (4)	2 (256)	512 (0)
AB 4106	1024	64 (16)	64 (16)	128 (8)	1 (1024)	512 (2)
AB 4112	>1024	64 (>16)	64 (>16)	256 (>4)	1 (>1024)b	1024 (>0)
KP A5	>1024	>1024 (0)	>1024 (0)	256 (>4)	8 (>128)b	>1024 (0)
KP B9	1024	2 (512)	512 (2)	0.5 (2048)	0.25 (4096)b	256 (4)
KP C3	256	4 (64)	256 (0)	1 (256)	1 (256)b	256 (2)
AB ATCC 17978+mcr-1	16	2 (8)	16 (0)	16 (0)	0.0625 (256)c	16 (0)
EC ATCC 25922+mcr-1	8	1 (8)	8 (0)	1 (8)	0.25 (64)c	4 (2)
KP 2210291+mcr-1	16	8 (2)	16 (0)	4 (4)	0.5 (32)	16 (0)
PA ATCC 47085+mcr-1	4	4 (0)	4 (0)	4 (0)	2 (2)	2 (2)
AB ATCC 17978	1	0.5 (2)	0.5 (2)	1 (0)	1 (0)c	0.5 (2)
EC ATCC 25922	0.5	0.5 (0)	0.5 (0)	0.5 (0)	0.5 (0)b	0.5 (0)
KP 2210291	1	0.5 (2)	1 (0)	0.5 (2)	0.5 (2)	0.25 (4)
PA ATCC 47085	2	1 (2)	2 (0)	2 (0)	1 (2)	1 (2)

^aAll adjuvant MICs are >200 μM unless noted.

^bCompound MIC = 200 μM.

^cCompound MIC = 100 μM.

Table 2.

Colistin MIC (μg/mL (Fold Reduction)) of AB 4106 and KP B9 Treated with 12 at Various Concentrations

	AB 4106	KP B9
compound MIC	>200 μM	200 μM
colistin MIC	1024 μg/mL	1024 μg/mL
colistin + 10 μM	0.5 (2048)	0.0625 (>4096)
colistin + 7.5 μM	1 (1024)	0.25 (4096)
colistin + 5 μM	1 (1024)	0.25 (4096)
colistin + 2 μM	4 (256)	1 (1024)
colistin + 1 μM	8 (128)	4 (256)

Figure 3.

Bacterial growth over time for AB 4106 and KP B9 alone and treated with 5 μM 12 only or 5 μM 12 in combination with colistin at the denoted concentrations.

1.3. IMD-0354 Abrogates Colistin-Resistant Lipid A Modification.

Colistin is a polycationic, cyclic polypeptide that directly interacts with the anionic bacterial cell membrane and disrupts envelope stability, leading to cell death. Bacteria have evolved resistance to colistin by modifying lipid A, a principle component of the outer cell membrane.¹⁹ Colistin resistance is achieved through covalent modification of phosphate moieties on lipid A with 4-aminoarabinose (Ara4N), galactosamine, and phosphoethanolamine (PEtN).⁵

In previous reports from our group, we have demonstrated that small molecules containing a 2-aminoimidazole (2-AI) moiety both restore a col^S MIC in otherwise col^R G− bacteria and reverse col^R modification of lipid A in these isolates.^8,9,13 Chemical modification of lipid A can be detected utilizing matrix-assisted laser desorption/ionization, time-of-flight mass spectrometry (MALDI/TOF MS), as multiple known structures of lipid A have been described via this method.²⁰

With this precedent in mind, we again utilized this technique to explore the possibility of a reversed lipid A phenotype. Two col^R strains, AB 4106 and KP B9, were each cultured in the presence or absence of 20 μM 12 for 6 h, after which lipid A was extracted by Karoff’s method and analyzed by MALDI-TOF. As seen in Figure 4, all m/z associated with the PEtN modification in AB 4106 and the Ara4N modification in KP B9 were either greatly reduced in intensity or completely undetectable in the presence of 12.

Figure 4.

Lipid A mass spectra of AB 4106 and KP B9 both untreated or treated with 20 μM 12. The +PEtN modified species in AB 4106 are highlighted in red on the left, while the +Ara4N species corresponding to colistin resistance are shown in blue on the right.

1.4. AR-12 Displays Unprecedented Activity in Colistin-Sensitive Strains.

In both of the col^S strains used in our initial screen, we noted that 13 was the only compound with measurable and dose-dependent activity. To probe the limits of activity of 13, it was challenged against a 41-member panel of G− col^S isolates. This panel includes 30 primary clinical isolates of A. baumannii that represent nearly all clinically relevant A. baumannii clades²¹ and 11 additional G− col^S isolates (AB ATCC 19606, Acinetobacter baylyi ADP-1, 8 additional strains of K. pneumoniae, and P. aeruginosa PAO1). As seen in Table 3, each strain was dosed with 13 at both 10 and 5 μM, and dose-dependent activity was noted in each strain, returning as much as a 1024-fold reduction of colistin MIC. As with 12, the activity of 13 is retained in the presence of detergent, ruling out activity due to aggregation (Table S4). In an additional assay, the activities of both 12 and 13 were determined in media spiked with fetal bovine serum (FBS). We saw reduced but notable activity with 12 in AB 4106 in as much as 50% FBS (Table S5), while 13 had limited activity against AB 5075 in serum mixtures (Table S6). A growth curves for AB 5075 was constructed in the presence of 13 alone or in combination with colistin Figure 5. As for compound 12, we observe no perturbance of growth effected by 13 alone, while the combination of 13 with colistin causes growth reduction at concentrations below colistin MIC.

Table 3.

Colistin MIC (ng/mL (Fold Reduction)) of Primary Clinical Isolates of A. baumannii (AB) Treated with 5 μM 13^a

	col MIC (ng/mL)	+10 μM 13	+5 μM 13
AB ATCC 19606	1000	15.6 (64)	125 (8)
AB WRAIR 967	1000	15.6 (64)	125 (8)
AB WRAIR 2828	1000	15.6 (64)	62.5 (16)
AB WRAIR 3340	1000	1.9 (512)	15.6 (64)
AB WRAIR 3560	1000	15.6 (64)	0.25 (4)
AB WRAIR 3638	500	1.9 (256)	62.5 (8)
AB WRAIR 3785	1000	31.3 (32)	500 (2)
AB WRAIR 3806	1000	7.8 (128)	31.3 (32)
AB WRAIR 3917	1000	31.3 (32)	125 (8)
AB WRAIR 3927	1000	7.8 (128)	31.3 (32)
AB WRAIR 4025	1000	3.9 (256)	31.3 (32)
AB WRAIR 4026	1000	62.5 (16)	500 (2)
AB WRAIR 4027	1000	3.9 (256)	31.3 (32)
AB WRAIR 4052	1000	15.6 (64)	125 (8)
AB WRAIR 4269	1000	1.9 (512)	62.5 (16)
AB WRAIR 4448	1000	1.9 (256)	3.9 (256)
AB WRAIR 4456	1000	31.3 (32)	250 (4)
AB WRAIR 4490	1000	7.8 (128)	15.6 (64)
AB WRAIR 4498	1000	15.6 (64)	62.5 (16)
AB WRAIR 4795	500	0.5 (1024)	15.6 (32)
AB WRAIR 4857	1000	1.9 (512)	62.5 (16)
AB WRAIR 4878	1000	1.9 (512)	62.5 (16)
AB WRAIR 4932	1000	7.8 (128)	62.5 (16)
AB WRAIR 4957	1000	15.6 (64)	62.5 (16)
AB WRAIR 4991	1000	1.9 (512)	62.5 (16)
AB WRAIR 5001	1000	15.6 (64)	125 (8)
AB WRAIR 5075	1000	3.9 (256)	62.5 (16)
AB WRAIR 5197	1000	3.9 (256)	125 (8)
AB WRAIR 5256	500	1.9 (512)	31.3 (64)
AB WRAIR 5674	1000	15.6 (64)	62.5 (16)
AB WRAIR 5711	1000	1.9 (512)	125 (8)
A. baylyi ADP-1	1000	125 (8)	1000 (0)
KP ATCC 43816	500	7.8 (64)	15.6 (32)
KP BAA 2146	1000	250 (4)	250 (4)
KP B2	500	15.6 (32)	125 (4)
KP B5	250	3.9 (64)	15.6 (16)
KP B8	500	15.6 (32)	31.3 (16)
KP C2	250	15.6 (16)	62.5 (4)
KP C4	500	15.6 (32)	62.5 (8)
KP D4	500	15.6 (32)	125 (4)
PA O1	2000	1000 (2)	1000 (2)

^aThe MIC of 13 alone was >200 μM for each strain unless noted.

Figure 5.

Bacterial growth over time of A. baumannii 5075 with 10 μM 13 alone or in dual treatment with colistin at the denoted concentrations.

Previous reports have documented that colistin can serve as a potentiator for other antibiotics against G− bacteria by inducing increased membrane permeability.²² To probe if this was the case for 13, we determined its MIC in the presence of polymyxin B nonapeptide (PMBN), a truncated analog of colistin lacking a lipid moiety that is non-antimicrobial (PMBN MIC against AB 5075 is >32 μg/mL).²³ Despite a lack of antimicrobial activity, PMBN retains the ability to increase membrane permeability analogous to colistin. When 12 was coadministered with equimolar concentrations of PMBN, no change in MIC was observed (Table S7), establishing that, in this synergistic pair, colistin is the causative agent of bacterial death, while 13 is the adjuvant.

We then sought to probe synergy with other antibiotics. Using A. baumannii 5075 again as the test strain, 13 was dosed at 10 μM and codosed with 15 additional antibiotics (Table S8). Here, we see modest synergy with daptomycin, which is typically active against G+ bacteria only.²⁴ Daptomycin alone lacks antimicrobial activity against AB 5075 in concentrations exceeding 1024 μg/mL but, in combination with 10 μM 13, achieves an MIC of 128 μg/mL, while no synergy with other antibiotics was observed. In an additional experiment, we sought to evaluate whether or not the effects of 12 and 13 could be additive in col^R bacteria but found that, in both AB 4106 and KP B9, the combined treatment of 12, 13, and colistin did not return any additional reduction in colistin MIC (Table S9).

Compound 13 is a known inhibitor of human pyruvate dehydrogenase kinase-1 (PDK-1). Across the three KI libraries used in this study, several other PDK-1 inhibitors were present in each library but were not identified as hit molecules under our screening conditions. One known PDK-1 inhibitor, NVP-BAG956 (compound 14, Figure 2), was not represented in any library.²⁵ Additionally, the parent molecule of 13 is celecoxib, a nonsteroidal anti-inflammatory drug (NSAID) that has previously reported adjuvant activity with other antibiotics in Staphylococcus aureus (compound 15, Figure 2).^26,27 Given these relationships, we challenged AB 5075 to both of these small molecules at various concentrations (Table 4) and found insignificant activity with 14. While 15 did have minimal synergy with colistin, compound 13 greatly outperforms the parent molecule (Figure 5).

Table 4.

Colistin MIC Given as μg/mL (Fold Reduction) of AB 5075 (Colistin MIC = 1 μg/mL) Cotreated with 14 and 15 at Denoted Concentrations

	14	15
compound MIC	>200 μM	>200 μM
colistin + 30 μM	0.5 (2)	0.25 (4)
colistin +15 μM	0.5 (2)	0.5 (2)
colistin +10 μM	0.5 (2)	1 (0)

1.5. Combination Treatment with Colistin Does Not Increase Eukaryotic Cell Toxicity.

One major advantage of utilizing known KI compounds is that, generally, a wealth of eukaryotic cytotoxicity profiles have been published. However, our intent is to utilize these compounds in dual therapy with colistin, and thus, we evaluated the potential for cytotoxicity of the therapeutic combination in mouse 4T1 mammary gland tumor cells (ATCC). Cells were brought to confluence in RPMI, and either 12 or 13 was dosed in media alone or simultaneously given with 1 μg/mL colistin. After 18 h, cell viability was then measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT).²⁸ The concentration of adjuvant resulting in 50% cell death was taken as that compound’s respective CT₅₀ (using DMSO as 100% viability control and 1% Triton X100 as 0%). As seen in Figure 6, the respective CT₅₀ values of both 12 (162.8 μM alone, 158.6 μM with colistin) and 13 (11.9 μM alone, 10.5 μM with colistin) were not changed significantly in the presence of colistin.

Figure 6.

Cell viability vs log[adjuvant (μM)] of mouse 4T1 in the presence of 12 or 13 alone or in combination with 1 μg/mL colistin.

In conclusion, we report a comprehensive screening of 942 known KIs against both col^S and col^R G− bacteria. Through follow-up evaluation of 51 unique hits, we narrowed our focus to two highly potent compounds, one having efficacy in col^R bacteria (12) and the other having significant activity in col^S strains (13). We demonstrate that 12 suppresses lipid A modification in col^R strains, giving insight into its mechanism of action (MOA). While completing our studies, we noted that niclosamide, a structurally similar yet significantly more toxic salicylamide,²⁹ was discovered to also sensitize colistin-resistant bacteria to the effects of colistin.³⁰ No mechanistic studies were described; therefore, it is unclear whether niclosamide and 12 have similar mechanism of actions regarding lipid A modification. We also report the near-ubiquitous activity of 13 in a diverse array of col^S G− bacilli. Growth-curve analysis shows no perturbation of bacterial proliferation in 12 or 13 alone at their respective active concentrations, while both clearly elicit cell death upon cotreatment with colistin. Additionally, in vitro eukaryotic cell toxicity does not increase when either compound is codosed with colistin. In all, this report provides a robust platform for further structure–activity relationship (SAR) studies of both 12 and 13, while probes of their respective MOAs and murine infection models are pursued.

2. EXPERIMENTAL SECTION

2.1. Kinase Inhibitor Library Screening.

The three kinase inhibitor libraries utilized in this study were procured as follows: Cayman Kinase Screening Library (Ann Arbor, MI), Selleck Chemical Kinase Inhibitor Library (Houston, TX), and UNC Kinase Chemogenomic Set (Chapel Hill, NC). Each library was diluted to 1 mM in DMSO for dosing. Bacterial strains of interest were cultured for 4 h in cation-adjusted Muller Hinton broth (CAMHB), subcultured to 5 × 10⁵ CFU/mL, and divided in 100 μL aliquots on a 96-well plate. Kinase inhibitors were then spiked in media to a final concentration of 10 or 20 μM and either codosed with colistin to a desired concentration or given as monotherapy. Each respective condition was performed in duplicate, and all wells returning a consensus lack of bacterial growth in dual therapy with colistin without toxicity in monotherapy were recorded as hit molecules.

2.2. Single-Compound Bacterial Susceptibility Assay.

This procedure follows the guidelines set by the Clinical Laboratory Sciences Institute.³¹ Bacteria were cultured for 4 to 6 h in CAMHB and subcultured to 5 × 10⁵ CFU/mL in fresh CAMHB. For each compound to be tested, a 1 mL aliquot of subculture was collected and dosed with a compound of interest to a final concentration of 200 μM. Samples were then dispensed (200 μL) into the first row of a 96-well microtiter plate in which subsequent wells were prefilled with 100 μL of subculture. 100 μL of dosed subculture was then serially diluted a total of 6 times in each subsequent row of the plate, save for the last row for a control. Plates were then sealed and incubated stationary at 37 °C. After 18 h, the plates were removed and MIC values were recorded. All compounds tested had a purity of >95%.

2.3. MIC Shift Assay.

This procedure was adapted from the guidelines set by the Clinical Laboratory Sciences Institute.³¹ Bacteria were cultured for 4 to 6 h in CAMHB and subcultured to 5 × 10⁵ CFU/mL in fresh CAMHB. For each compound to be tested, a 5 mL aliquot of subculture was taken and dosed with adjuvant from a DMSO stock. A 1 mL aliquot of each dosed subculture was collected and dosed with antibiotic to a set concentration. Codosed aliquots were then dispensed (200 μL) into the first row of a 96-well microtiter plate in which subsequent wells were prefilled with 100 μL of the corresponding dosed subculture. 100 μL of codosed subculture was then serially diluted a total of 6 times in each subsequent row of the plate, save the last row as a control to afford serial dilution of the antibiotic while holding a constant concentration of adjuvant. Plates were then sealed and incubated stationary at 37 °C. After 18 h, the plates were removed and MIC values were recorded. All compounds tested had a purity of >95%.

2.4. Time-Dependent Bacterial Growth and Death Quantification.

Strains were cultured for 18 h in CAMHB and subcultured to 5 × 10⁵ CFU/mL in fresh CAMHB. The subculture was then transferred to culture tubes in 5 mL aliquots, which were dosed with adjuvant to a set concentration or adjuvants coadministered with colistin, save one aliquot as a control. All subcultures were then incubated at 37 °C with shaking. At 2, 4, 6, 8, and 24 h post subculture, 100 μL samples of each condition were serially diluted in 900 μL aliquots of CAMHB for a total of 5 to 7 times. 100 μL of each dilution point was then plated on LB agar and incubated overnight. The total number of bacterial colonies on each plate were recorded.

2.5. Bacterial Membrane Lipid A Mass Spectrometry.

Strains of interest were cultured for 18 h in CAMHB and subcultured to 1 × 10⁶ CFU/mL in fresh CAMHB. Subcultures were split, inoculated with adjuvant or media alone, and incubated for 6 h with shaking. Each subculture was collected and centrifuged (4000 rpm, 5 °C, 20 min). Supernatants were discarded, and cell pellets were resuspended in endotoxin-free water and spun down in the same manner; the resulting supernatant was discarded. Cell pellets were then resuspended in 400 μL of 70% isobutyric acid/1 M ammonium hydroxide (5:3 [vol/vol]) and incubated at 100 °C for 45 min. Samples were centrifuged at 2000g for 15 min, and supernatants were added to endotoxin-free water (1:1 [vol/vol]), snap-frozen on dry ice, and lyophilized overnight. The resultant material was washed twice with 1 mL of methanol. Lipid A was extracted with 100 μL of a chloroform/methanol/water mixture (3:1.5:0.25 [vol/vol/vol]). 1 μL of extracted lipid A was spotted on a steel matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) plate followed by 1 μL of 10 mg/mL norharmane matrix in chloroform–methanol (2:1, [vol/vol]) (Sigma-Aldrich, St. Louis, MO). Samples were air-dried and then analyzed on a Bruker Microflex mass spectrometer (Bruker Daltonics, Billerica, MA) in negative ion mode. Mass calibration was achieved using an electrospray tuning mix (Agilent, Palo Alto, CA). All spectral data were analyzed with Bruker Daltonics FlexAnalysis software. Estimations of lipid A structures in each condition were made by comparing spectra with predicted structures and molecular weights of known species of lipid A. Structural heterogeneity of lipid A within a single bacterial membrane has been well-described.^32,33

2.6. Eukaryotic Cell Toxicity.

4T1 cells (ATCC Manassas, VA) were plated at a density of 1 × 10⁴ cells/well in 96-well plates in Roswell Park Memorial Institute media 1640 (RPMI) (Gibco, Gaithersburg, MD) supplemented with 10% fetal bovine serum (Gibco), 2 mM GlutaMAX (Gibco), and 50 μM 2-mercaptoethanol (Sigma-Aldrich, St. Louis, MO) and incubated at 37 °C under a 5% CO₂ atmosphere in the dark for 18 h. Cell cultures were treated with serial dilutions of the compounds in the presence or absence of 1 μ/mL colistin (3 replicates per condition) and incubated for an additional 18 h. The following control conditions were used: media only (blank), 1% Triton X100 (0% cell viability), and 0.5% DMSO (100% cell viability). Each condition was then treated with a 10% volume of a 5 mg/mL solution of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma-Aldrich) in sterile filtered 1× phosphate buffered saline (PBS), and the samples were incubated for 2 h at 37 °C in 5% CO₂, after which the media was aspirated and the resulting formazan crystals were resuspended in 100 μL of acidified (4 mM HCl) isopropanol. The 96-well plate was then read at 540 nm on a FLUOstar Optima (BMG Labtech Cary, NC) microplate reader. Cell viability was calculated as a percentage using the two previously mentioned controls.

2.7. 4-(4-Hexylphenyl)-1H-imidazol-2-amine (1).

Compound 1 is displayed for reference and context. The synthesis of 1 is described by Harris et al.⁸

2.8. 1-Ethyl-4-(4-hexylphenyl)-1H-imidazol-2-amine (2).

Compound 2 is displayed for reference and context. The synthesis of 2 is described by Brackett et al.⁹

2.9. 4-(4-Hexylphenyl)-1-isopropyl-1H-imidazol-2-amine (3).

Compound 3 is displayed for reference and context. The synthesis of 3 is described by Brackett et al.⁹

2.10. N-(2-(4-(5-(2-Amino-1-(4-butylbenzyl)-1H-imidazol-5-yl)pentyl)-1H-1,2,3-triazol-1-yl)ethyl)-4-pentyl-benzamide (4).

Compound 4 is displayed for reference and context. The synthesis of 4 is described by Harris et al.10

2.11. 4-(6-Bromo-1H-indol-3-yl)pyrimidin-2-amine (5).

Compound 5 is displayed for reference and context. The synthesis of 5 is described by Huggins.¹¹

2.12. 4-(5-Bromo-1H-indol-3-yl)-N-ethylpyrimidin-2-amine (6).

Compound 6 is displayed for reference and context. The synthesis of 6 is described by Huggins et al.¹¹

2.13. N-Benzyl-4-(5-bromo-1H-indol-3-yl)pyrimidin-2-amine (7).

Compound 7 is displayed for reference and context. The synthesis of 7 is described by Huggins et al.¹¹

2.14. N-((1H-Indol-3-yl)methyl)-4,5-dibromo-1H-pyr-role-2-carboxamide (8).

Compound 8 is displayed for reference and context. The synthesis of 8 is described in a previous report.¹²

2.15. 4-(4-(3-(4-Chloro-3-(trifluoromethyl)phenyl)-ureido)phenoxy)-N-methylpicolinamide p-toluenesul-fonate (9).

Compound 9 was purchased from BioTang (Lexington, MA) and used without further purification.

2.16. (E)-2-Cyano-3-(3,5-di-tert-butyl-4-hydroxyphenyl)prop-2-enethioamide (10).

Compound 10 was purchased from Selleck Chemical (Houston, TX) and used without further purification.

2.17. (2Z,3E)-6′-Bromo-3-(hydroxyimino)-[2,3′-biin-dolinylidene]-2′-one (11).

Compound 11 was purchased from Selleck Chemical (Houston, TX) and used without further purification.

2.18. N-(3,5-Bis(trifluoromethyl)phenyl)-5-chloro-2-hydroxybenzamide (12).

Compound 12 was synthesized according to the procedure described by Kang et al.³⁴ Spectral data were consistent with previous reports.³⁴

2.19. 2-Amino-N-(4-(5-(phenanthren-2-yl)-3-(trifluoromethyl)-1H-pyrazol-1-yl)phenyl) acetamide (13).

Compound 13 was purchased from Selleck Chemical (Houston, TX) and used without further purification.

2.20. 2-Methyl-2-(4-(2-methyl-8-(pyridin-3-ylethyn-yl)-1H-imidazo[4,5-c]quinolin-1-yl)phenyl)-propanenitrile (14).

Compound 14 was purchased from Cayman Chemical (Ann Arbor, MI) and used without further purification.

2.21. 4-(5-(p-Tolyl)-3-(trifluoromethyl)-1H-pyrazol-1-yl)benzenesulfonamide (15).

Compound 15 was purchased from VWR (Radnor, PA) and used without further purification.