Dissociation of antibacterial activity and aminoglycoside ototoxicity in the 4-monosubstituted 2-deoxystreptamine apramycin

Abstract

Aminoglycosides are potent antibacterials, but therapy is compromised by substantial toxicity causing, in particular, irreversible hearing loss. Aminoglycoside ototoxicity occurs both in a sporadic dose-dependent and in a genetically predisposed fashion. We recently have developed a mechanistic concept that postulates a key role for the mitochondrial ribosome (mitoribosome) in aminoglycoside ototoxicity. We now report on the surprising finding that apramycin, a structurally unique aminoglycoside licensed for veterinary use, shows little activity toward eukaryotic ribosomes, including hybrid ribosomes which were genetically engineered to carry the mitoribosomal aminoglycoside-susceptibility A1555G allele. In ex vivo cultures of cochlear explants and in the in vivo guinea pig model of chronic ototoxicity, apramycin causes only little hair cell damage and hearing loss but it is a potent antibacterial with good activity against a range of clinical pathogens, including multidrug-resistant Mycobacterium tuberculosis. These data provide proof of concept that antibacterial activity can be dissected from aminoglycoside ototoxicity. Together with 3D structures of apramycin-ribosome complexes at 3.5-Å resolution, our results provide a conceptual framework for further development of less toxic aminoglycosides by hypothesis-driven chemical synthesis.

Fig. 1.

Mitochondrial in organello translation: aminoglycoside effects on mitochondrial protein synthesis. Isolated mitochondria were incubated with various concentrations of aminoglycosides as indicated. Proteins were resolved by SDS/PAGE, and COX1 protein levels were quantified by densitometry. Images of the relevant part of the SDS gel used for quantitation are shown. Quantitation is given in percent. IC₅₀ values indicate the drug concentration in micromolars required to inhibit COX1 synthesis by 50%.

Fig. 2.

Aminoglycoside-induced misreading. Dose–response curves of aminoglycoside-induced misincorporation of amino acids using 245 near-cognate CGC-mutant firefire luciferase (F-luc) mRNA as template. Shown is luciferase activity upon translation of mutant template relative to wild-type F-luc mRNA (mean ± SD; n = 3). Mitochondrial wild-type hybrid (Mit13) ribosomes (A) and mitochondrial A1555G-mutant hybrid ribosomes (B) treated with apramycin (filled green circles), gentamicin (blue open triangles), tobramycin (open black rhombi), or kanamycin (open red squares). Mycobacterium smegmatis bacterial ribosomes (C) and E. coli bacterial ribosomes (D) treated with apramycin (filled green circles and solid green line) or aprosamine (dashes and solid black line). Neamine (crosses and dotted black line) was used as control. Green crosses indicate misreading values below background.

Fig. 3.

Loss of hair cells in cochlear explants and loss of auditory function in vivo. (A) Hair cells in mouse organ of Corti explants were stained for actin with rhodamine-phalloidin. (a) Control. (b) Treatment with 0.2 mM apramycin. (c) Treatment with 0.2 mM gentamicin. Panels a and b show normal morphology with one row of inner hair cells (IHC) and three rows of outer hair cells (OHC). (OHC row 1 is adjacent to IHC.) In c loss of OHC is essentially complete, but some IHC (arrowheads) remain. Sections are from the upper basal turn of the cochlea. (Scale bar: 50 μm.) (B) Hair cell loss was quantified along the entire length of cochlear explants and plotted against drug concentration of gentamicin (circles and solid line) and apramycin (squares and dotted line). Data represent means ± SEM, n = 3–12 per point. (C) Explants of the organ of Corti were stained with antibody to 3-nitrotyrosine (green); blue staining (Hoechst 33342) shows nuclei, and red staining is myosin VIIa antibody outlining hair cells. (a) Control. (b) Treatment with 0.2 mM gentamicin. (c) Treatment with 2 mM apramycin. (D) Effect of chronic aminoglycoside treatment (gentamicin, circles and solid line; apramycin, squares and dotted line) in vivo on ABR. Threshold shift is the difference in auditory threshold before and 3 wk after treatment. Data represent means ± SEM, n = 3–11 per point (except for 160 mg gentamicin, in which only one animal survived). The threshold shift was determined in dB, corresponding to a logarithmic scale (i.e., a 10-dB difference indicates a 1 log₁₀ difference in energy).