Visible-Light Controlled Release of a Fluoroquinolone Antibiotic for Antimicrobial Photopharmacology.

Owing to the dwindling arsenal of antibiotics, new methodologies for their effective and localized delivery are necessary. The use of optical control over delivery of drugs, also known as photopharmacology, has emerged as an important option for the spatiotemporally controlled generation of drugs and bioactive molecules. In the field of antimicrobial photopharmacology, most strategies utilize ultraviolet light for triggering release of the antibiotic. The use of such short wavelength light may have limitations such as phototoxicity. Here, a small molecule that is activated by visible light to release a fluoroquinolone, a broad-spectrum antibiotic, is reported. A boron-dipyrromethene, which is sensitive to cleavage at 470 nm, was used, and levofloxacin was used as a model fluoroquinolone. BDP-Levo was found to undergo cleavage in the presence of visible light to release the active antibiotic. Using growth inhibitory studies in Gram-positive as well as Gram-negative bacteria, the efficacy of BDP-Levo is demonstrated. Together, our study demonstrates that visible light can be used for optical control over antibiotic release and lays the foundation for visible-light-mediated antimicrobial photopharmacology.

Introduction

The modern antibiotic era has exacerbated the morbidity and mortality associated with bacterial infections. The overuse and misuse of antibiotics have, in part, contributed to the emergence of drug resistance.[1] Antimicrobial drug resistance, compounded with a limited pipeline of new antibiotics, has now become a major global public health threat.[2,3] The number of antibiotics that continue to remain effective is dwindling at an alarming rate, and the existing classes of antibiotics need to be better used to increase longevity.[4] Because antibiotics are used systemically, a large buildup of the bioactive molecule is invariably unavoidable. This affects not only the pathogen of interest but also gut microbiota that are crucial for our health and well-being. Furthermore, environmental exposure to antibiotics leads to the possible buildup of drug-resistant pathogens. For certain types of infection, if spatiotemporal control over antibiotic generation can be achieved, side effects can be reduced and perhaps patient compliance can be improved. The past several decades have seen the emergence of photopharmacology, which aims to minimize the problem of off-target activity and decrease deleterious side effects by offering an external control over the action of the drug.[5,6] In the field of antimicrobial photopharmacology, an antibiotic is attached to an optical switch or a photocleavable group;[7] upon irradiation, the active form of the antibiotic is produced (Figure a). For example, Feringa and co-workers have reported a diazo functional group linked to a model fluoroquinolone (Figure b).[7−9] This functional group serves as an optical switch. Upon exposure to UV light, isomerization occurs, which results in the active antibiotic being formed. It is proposed that after the antibiotic action is completed, another source of light could be used to inactivate the antibiotic. Fuchter and co-workers have developed a methodology for aminohydrolase-based antibiotics.[10] Using a 2-nitroaryl cleavable group, Forsythe and co-workers have reported a hydrogel that is triggered by UV light (Figure c).[11] Although this liberates ciprofloxacin, the use of UV light is not highly desirable and may have problems associated with phototoxicity. The use of visible light, on the contrary, has considerably lower toxicity and is therefore highly desirable. Here, we report a methodology that is suitable for visible-light-triggered release of a fluoroquinolone.

Figure 1

(a) Generic design of a visible-light-triggered antibiotic. (b) Example of an optically controlled antibiotic: UV light is used to convert an inactive trans isomer to the active cis isomer. (c) UV-triggered release of ciprofloxacin, a clinically used antibiotic. The sphere represents a tether for cross-linking.

Among the various fluorophores that have been previously used as triggerable scaffolds to deliver bioactive molecules,[12] boron-dipyrromethene (BODIPY)-based fluorophores have distinct advantages. They are stable in buffer, have high quantum yields, are relatively nontoxic, and their wavelengths of absorption can be tuned by systematic structural modifications.[13] Recently, Urano and co-workers have reported the cleavage of a B–O bond in aryloxy-BODIPY derivatives.[14,15] In addition, there are two other reports of a polymeric scaffold for drug delivery[16] as well as a small molecule for delivery of the pharmacologically active gaseous species, hydrogen sulfide.[17] We thus designed BDP-Levo, 1, to deliver a fluoroquinolone using visible light as a trigger (Figure ). This compound is expected to undergo cleavage by visible light to produce a self-immolative phenolate that rearranges to generate the active antibiotic (Scheme S1, Supporting Information (SI)).

Figure 2

Design of a BODIPY-based scaffold for visible-light-triggered release of a fluoroquinolone antibiotic.

Results and Discussion

First, the BODIPY derivative, 3, was synthesized using a reported procedure, and this compound was reacted with levofloxacin (Levo) to produce the desired BDP-Levo, 1, with good yield (Scheme ).[18] Because the release of carboxylic acids has hitherto not been studied using this BODIPY protective group, benzoate ester 5 was synthesized (Scheme ). The photophysical properties of these BODIPY derivatives were studied. Fluorescence measurements (excitation 470 nm, emission 509 nm) were conducted (see Figure S1, Supporting Information), and quantum yields were determined using standard protocols. Consistent with previous reports, the esters were weakly fluorescent and diminished quantum yields were recorded (see Table S1, Supporting Information). Next, using the irradiation conditions that were previously reported, 1 was incubated in methanol and exposed to 470 nm light.[14] High performance liquid chromatography (HPLC) analysis of the reaction mixture showed complete decomposition of 1 to produce the methoxy derivative, BDP-OMe (4, Scheme S1; see SI). In the absence of light, we found no evidence for the decomposition of 1 (Figure a). Quantum yield measurements of the irradiated samples showed a significant increase compared to those of the unirradiated samples (see Table S1, SI). Because the product in each case is the BDP-OMe, the quantum yields were identical (see Table S1, SI).

Scheme 1

Syntheses of 1 and 5 from 3

Figure 3

(a) Portions of HPLC traces of 1 incubated in methanol: Ctrl, t = 0; dark, t = 30 min; light, reaction mixture was irradiated for 30 min with 470 nm light (30 mW/cm2), decomposition of 1 produces the methoxy derivative, BDP-OMe (4, Scheme S1; see SI). Here, the detector wavelength is 500 nm. (b) Irradiation for 30 min with 470 nm light was followed by incubation of 1 in pH 7.4 phosphate buffer. A fluorescence detector was used with excitation 330 nm and emission 510 nm. Yield was estimated as 31% using authentic Levo.

The BDP-Levo derivative, 1, was next incubated in pH 7.4 phosphate buffer and was found to be stable for 30 min. After irradiation, the reaction mixture was diluted in pH 7.4 buffer, and thin-layer chromatography (TLC) analysis showed nearly complete decomposition of 1 when exposed to light and the formation of levofloxacin 2. The experiment in dark, as expected, did not generate 2, suggesting the stability of 1 in pH 7.4 buffer (Figure S2; see SI). Next, the formation of Levo was assessed by HPLC analysis of 1 in pH 7.4 buffer that was irradiated for 30 min. A distinct peak that corresponded to the formation of 2 was observed (Figure b). The yield of 2 under these conditions was estimated as 31%. This yield is similar to the yield of histamine that was generated by a previous report of a BODIPY-histamine scaffold.[14,17]

To study the bactericidal effects of the BODIPY scaffold, bacteria were exposed to the BODIPY derivative 3 (Figure S6; see SI). No significant inhibition of growth was observed in dark as well as in light, suggesting that the scaffold does not contribute to the growth inhibitory effects. Under similar conditions, levofloxacin 2 was a potent inhibitor of growth of Escherichia coli (E. coli) in dark as well as in light, suggesting that the efficacy of levofloxacin was not dependent on irradiation (Figure b). A similar result was obtained when this experiment was conducted using the Gram-positive pathogen Staphylococcus aureus (S. aureus) (Figure ). Together, these data support the use of 1 to effectively inhibit the growth of bacteria only when exposed to visible light.

Figure 4

Growth curve analysis of E. coli in broth: (a) Ctrl indicates bacteria in dark; 1, dark indicates bacteria treated with 1 (5 μM) but not irradiated with light; 1, light indicates bacteria treated with 1 (5 μM) and irradiated with light for 5 min at 470 nm; 2, dark indicates bacteria treated with 2 (5 μM) but not irradiated with light; (b) at time point 6 h: Ctrl indicates bacterial growth control in dark as well as in light. *p-Value < 0.001 for comparison of bacteria treated with 1 in dark versus irradiated with 470 nm light.

Figure 5

Growth curve analysis of S. aureus in broth. Ctrl indicates bacteria in dark; 1, dark indicates bacteria treated with 1 (5 μM) but not irradiated with light; 1, light indicates bacteria treated with 1 (5 μM) and irradiated with light for 5 min; 2, dark indicates bacteria treated with 2 (5 μM) but not irradiated with light.

The OD600 of compound 1 under irradiation conditions was negligible during the assay, suggesting no interference in the broth dilution assay by the compound (see SI). In addition, an agar-growth method was used to evaluate the efficacy of 1 (Figure ). Here, bacteria after exposure to 1 (either light or dark) were grown on an agar plate. Images of these plates were recorded periodically (Figure S8; see SI).

Figure 6

Growth inhibition of E. coli on agar plates: E. coli grown on an agar medium for 6 h after exposure to 1. Dark indicates agar plates with 1 (100 μM) were not exposed to light; light indicates agar plates with 1 (100 μM) were exposed to light for 30 min at 470 nm.

After 6 h, a clear inhibition of growth of bacteria was recorded with respect to the control (Figure ). This data was consistent during 24 h incubation as well. No significant inhibition of growth of bacteria incubated with 1 in dark was observed. Similarly, the negative control, 3, showed no inhibition of bacterial growth. In the presence of light, the inhibitory potency of 1 was comparable with the clinically used antibiotic, 2 (Figure S8; see SI).

Taken together, we report for the first time a small-molecule-based methodology for visible-light-triggered release of levofloxacin, a clinically used antibiotic in its pharmacologically active form. This third-generation antibiotic has superior antibacterial properties compared with those of ciprofloxacin that has been previously used.[19,20] Feringa and co-workers have proposed that inactivation of the antibiotic after use is important, and this is not possible in the present methodology and may be a shortcoming that will need to be addressed.[21] A number of dendrimer-/polymer-based scaffolds for phototriggered release of an antibiotic are known.[22,23] A cell-wall-targeted dendrimer nanoconjugate containing ciprofloxacin was recently reported. Using this method, phototriggered release of ciprofloxacin was achieved. Our approach can be adapted to incorporate such targeting ligands in future. Recently, hollow microspheres that can rapidly produce localized heat activated by near-infrared light and control the release of an antibiotic via a “molecular switch” in their polymer shells have been reported.[22,23] This photothermally responsive drug delivery system has distinct advantages and presents an attractive methodology for combination therapy.[23] Because the scaffold reported herein is triggered at 470 nm, further modification is necessary to enhance the wavelength of cleavage as well as for incorporation into polymeric scaffolds. These studies are presently under way in our laboratory.

Although there is an urgent need to develop antibiotics with novel targets and new mechanisms of action,[24−28] adjuvants that can enhance the activity of antibiotics,[29−31] preserving the existing classes of antibiotics, is a major public health priority. Improving patient compliance and monitoring inappropriate use are among the long-term solutions. Directed delivery of antibiotics to the site of infection,[32] perhaps, will result in minimizing the systemic exposure to the antibiotic[6] and may minimize certain unwanted side effects of fluoroquinolones.[33,34] Although this strategy is in its infancy, our small molecule lays the foundation for visible-light-activated antibiotic delivery. Lastly, the mechanism of antibiotic action is being investigated and the role of redox mechanisms in lethality is under scrutiny.[35] Development of new tools, such as the compound reported herein that can facilitate a better understanding of antibiotic mechanisms, will be useful. These experiments are presently under way in our lab, and results will be reported in due course.

Methods

Synthesis of 4-((5-Fluoro-1,3,7,9-tetramethyl-10-phenyl-5H-4λ4,5λ4-dipyrrolo[1,2-c:2′,1′-f][1,3,2]diazaborinin-5-yl)oxy)benzyl 9-fluoro-3-methyl-10-(4-methylpiperazin-1-yl)-7-oxo-2,3-dihydro-7H-[1,4]oxazino[2,3,4-ij]quinoline-6-carboxylate (1)

Following the reported procedure,[18] to a stirred solution of 2 (42 mg, 0.116 mmol), 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (44 mg, 0.116 mmol), and 4-dimethylaminopyridine (5.7 mg, 0.046 mmol) in anhydrous CH2Cl2 (5 mL) were added N,N-diisopropylethylamine (104 μL, 0.583 mmol) and 3 (52 mg, 0.116 mmol) at room temperature (rt). The reaction[20] mixture was stirred for 12 h at rt. After completion of the reaction as monitored by TLC, water (5 mL) and CH2Cl2 (15 mL) were added to the reaction mixture. The organic components were extracted with CH2Cl2, and collected organic phases were dried over anhydrous Na2SO4, concentrated under reduced pressure to obtain the crude product as red residue. The residue was purified by column chromatography using neutral alumina as stationary phase and CHCl3: MeOH (100:0 to 93:7) as mobile phase to afford 1 (62 mg, 69%) as a red-orange solid. FT-IR (νmax, cm–1); 2925, 2849, 2800, 1718, 1617, 1547, 1470, 1300; 1H NMR (400 MHz, CDCl3), δ 8.09 (s, 1H), 7.57–7.47 (m, 3H), 7.42 (d, J = 12.6 Hz, 1H), 7.34–7.29 (m, 2H), 7.23 (d, J = 8.6 Hz, 2H), 6.58 (d, J = 8.6 Hz, 2H), 5.89 (d, J = 15.4 Hz, 2H), 5.16 (d, J = 1.8 Hz, 2H), 4.35 (dd, J = 11.4, 2.3 Hz, 1H), 4.19 (dd, J = 11.3, 2.0 Hz, 1H), 4.06–4.02 (m, 1H), 3.32 (d, J = 4.4 Hz, 4H), 2.53 (s, 4H), 2.48 (d, J = 10.9 Hz, 6H), 2.35 (s, 3H), 1.39 (s, 6H), 1.37 (s, 3H); 13C NMR (100 MHz, CDCl3), δ 172.6, 165.5, 156.8, 156.5, 156.4, 155.8, 154.4, 145.2, 143.3, 141.8, 139.5, 134.8, 131.7, 130.2, 129.3, 129.1, 129.0, 128.1, 127.9, 127.0, 123.6, 121.7, 117.8, 109.6, 105.1, 68.0, 66.6, 55.7, 54.6, 50.5, 46.4, 18.2, 14.9, 14.5; HRMS (ESI) for [C44H44BF2N5O5 + H]+: calcd, 772.3482; found, 772.3498.

Bacterial Strains and Growth Conditions

Both the strains E. coli (ATCC 25922) and S. aureus (ATCC 29213) were obtained from ATCC. All of the bacterial strains were routinely grown in Luria Bertani (LB) broth, Mueller Hinton broth II (MHB) cation supplemented medium, tryptic soy agar and tryptic soy broth, purchased from Himedia. For every experiment, a starting culture was produced by inoculating a single colony picked from the agar plate into liquid medium and incubated overnight at 37 °C with continuous shaking. Levofloxacin was purchased from TCI in the purified powdered form.

E. coli ATCC 25922 was grown overnight in Luria Bertani (LB) broth. Bacterial density was adjusted to 108 colony-forming unit (CFU)/mL corresponding to an optical density (600 nm, OD600) of 0.1. Next, 200 μL of this bacterial suspension was taken in a 96-well microtiter plate. Different lanes were chosen for different conditions such as bacterial control, bacteria with compound 1, bacteria with compound 3 (negative control), bacteria with levofloxacin, and so on. Then, 100× stock solutions were used for all of the compounds. One plate without irradiation was incubated at 37 °C in the dark by covering with an aluminum foil. Another plate was irradiated at 470 nm (30 mW/cm2) by blue light-emitting diode (LED) at room temperature in a closed chamber for 5 min and then incubated at 37 °C in the dark by covering with an aluminum foil. OD600 was measured using a Thermo Scientific Varioskan Flash microwell plate reader for both the plates at an interval of 1 h for 6 h. Values reported are average of six replicates. Errors have been calculated from standard deviation between the values.

Methicillin-sensitive S. aureus ATCC 29213 was grown overnight in Mueller Hinton broth II (MHB) cation supplemented medium. A similar procedure was followed further as was followed for E. coli.

Growth Inhibition of E. coli on Agar Plates

Agar plates were prepared using soyabean casein digest medium (tryptone soya broth) and agar powder. E. coli ATCC 25922 was grown overnight in Luria Bertani (LB) broth. Bacterial density was adjusted to 108 colony-forming unit (CFU)/mL corresponding to an optical density (600 nm, OD600) of 0.1. Next, 100 μL of this bacterial suspension was taken for streaking on the agar plate. Then, 100× stock solutions were used for all of the compounds. Bacteria were streaked on the entire surface of the agar plate using a Hi-Flexiloop 4 (Himedia). For samples with irradiation, bacterial suspensions (with or without compound) were taken in a quartz cuvette and irradiated at 470 nm (30 mW/cm2) by blue LED at room temperature in a closed chamber for 30 min. These irradiated samples were then streaked on the agar plates. All of the plates were then incubated at 37 °C in the dark by covering with an aluminum foil. Images were taken from 0 to 24 h using Nikon D3300 DSLR.