The antibacterial activity of topical retinoids: The case of retinaldehyde

Background. Retinoids such as retinoic acid (RA), retinol (ROL) and retinaldehyde (RAL) are currently used in many formulations and indications ranging form acne to skin aging. Most if not all their pharmacological activities occur through binding to nuclear receptors with subsequent modulation of the activities of several genes. Little attention has been given to the many other potential actions on the surface of the skin. Aim: To analyse the potential anti-infective activities of topical ROL, RAL and RA. Methods: Microbial minimal inhibitory concentrations (MIC) of ROL, RAIL and RA were determined by a microdilution method on reference strains including Staphylococcus aureus, Staphylococcus epidermidis, Micrococcus flavus, Propionibacterium acnes, Micrococcus luteus, Enterococcus faecium, Staphylococcus hominis, Escherichia coli, Pseudomonas aeruginosa, Candida albicans and 133 clinical strains including methicillin-resistant S. aureus, methicillin-sensitive S. aureus, coagulase-negative Staphylococcus, Streptococcus group B, Enterococcus faecalis, vancomycin-resitant E. faecalis, vancomycin-resistant E faecium and Pseudomonas/ Klebsiella. In two clinical trials in healthy human volunteers, skin bacterial densities were evaluated in samples obtained with the cylinder scrub method: (1) 2 and 5 h after a single application of 0.05% RAL or vehicle on the forearm and (2) in a single-blind randomized study where 0.05% RAL or vehicle were applied daily for 2 weeks on the forehead of 22 volunteers. Paired results from treated (or vehicle) and untreated areas were analysed. Results: Of the three retinoids tested, only RAL showed a significant in vitro antibacterial activity; this activity was found against reference strains of gram-positive bacteria like S. aeureus, Micrococcus spp. or P. acnes. No activity was found against gram-negative bacteria. These results on reference strains were confirmed on 133 clinical isolates. MIC50 and MIC90 values for RAL were 8 and 16 mg/l, respectively, for methicillin-sensitive S. aureus and 4 and 8 mg/l for methicillin-resistant S. aureus. The two in vivo studies showed that areas treated with RAL had a significant decrease in the bacterial counts. In the forehead study, the median decrease was 10(2) log/cm(2) for P. acnes and 10(1.8) log/cm(2) for staphylococci. No resistant bacteria were found after 2 weeks of topical use. Preliminary results suggest that the antibacterial effect of RAL is due, in part, to the aldehyde group in the lateral chain, since non-retinoid pseudo-analogues of the chain, like citral and hexenal, showed a similar antibacterial activity. Conclusion: We have shown that RAL differs from parent natural retinoids such as ROL and RA in demonstrating significant antibacterial activities upon topical use. This activity is likely due to the aldehyde group in the isoprenoic lateral chain, which illustrates the potential bifunctional properties of some retinoids. Copyright (C) 2002 S. Karger AG, Basel.