9,13-Dicis-retinoic acid as an isomerization product of 9-cis-retinoic acid.

With great interest we have read the article by Marchetti et al. on oxidation and isomerization of retinoic acid (RA) isomers in vitro, which was published in this journal recently (1). However, there are misleading statements in that publication, which require clarification. For instance, more than the indicated “few studies” on the metabolism of 9-cis-RA exist. Among the literature not referred to are the articles describing the identification of the glucuronides of 9-cis-RA and 9-cis-4-oxo-RA in mice, rats, and humans after treatment with 9-cis-RA or 9-cis-retinaldehyde (2–4). In contrast to the view of Marchetti et al., both in vivo and in vitro comparisons of biotransformation of 9-cis-RA, 13-cis-RA and all-trans-RA are available (3, 5–7). For instance, a higher degree of glucuronidation of 9-cis-RA and 13-cis-RA was observed in NMRI mice in vivo as compared with the all-trans-isomer, whereas all-trans-RA was oxidized to all-trans-4oxo-RA to a higher extent than the 9-cis and 13-cis-isomers. We consider a major shortcoming of the publication by Marchetti et al. (1) that isomerization of 9-cis-RA to 9,13-dicis-RA is completely ignored and profound transformation to 13-cis-RA is proposed instead. 9,13-dicis-RA has been identified as a major plasma metabolite of administered 9-cis-RA in mice and rats (3, 8, 9). Horst et al. demonstrated that 9,13-dicis-RA is a physiological constituent of neonatal calf plasma as well as of bovine plasma during the periparturient period (9, 10). In addition, 9,13-dicis-RA was found as a profound plasma retinoid in rabbit and human plasma after administration of retinyl palmitate and liver consumption, respectively (11, 12). 9,13-dicis-RA is also formed from 9-cis-RA in vitro after incubation with rat and bovine microsomes (6, 13). Based on those results, it has to be expected that in the experiments described by Marchetti et al., a significant part of 9-cis-RA is transformed into 9,13-dicis-RA in analogy to the isomerization of 13cis-RA to all-trans-RA, whereas the extent of the formation of 13cis-RA (requiring two steps of isomerization) has probably been overestimated in this study (1). Although most other reversed-phase HPLC methods do not allow separation of 9,13-dicis-RA from 13cis-RA (or 9-cis-RA), this can be achieved with a reversed-phase HPLC method (14), which enables the determination of four RA isomers as well as of three retinoyl-b-glucuronides and has already been applied for analysis of biological samples including microsomal preparations (4, 6, 12, 14). If HPLC methods that do not allow reliable separation of isomers are used, simultaneous detection at at least two wavelengths is essential to detect peak impurities as caused by coeluting isomers. In addition, recording the UV-spectrum of the putative cis-isomer would provide valuable information on the identity of this retinoid, as has been done previously (3, 9, 11, 12). Doing so would certainly have indicated that the so-called “13-cis-RA” peak of Marchetti et al. (1) also contains great amounts of 9,13-dicis-RA. It is questionable whether 9,13-dicis-RA has biological activity on its own or exerts activity via isomerization to 9-cis-RA and/or alltrans-RA only. In vivo formation of 9,13-dicis-RA is especially marked after administration of high doses of 9-cis-RA. This has led to the speculation that 9,13-dicis-RA might represent a detoxification product of 9-cis-RA or excess vitamin A (6). In conclusion, formation of 9,13-dicis-RA should always be considered for a complete and correct description of the metabolism of 9-cis-RA.