A document has been prepared by Fritz Weber in consultation with Henry B. F. Dixon and other members of our committees. It has already been published  by Fritz Weber (chairman of the former Committee II/1 on Nutritional Terminology of the International Union of Nutritional Sciences) and Athel J. Cornish-Bowden (chairman of our committees). It is reprinted in this newsletter by permission of the chairman of the Editorial Board of the British Journal of Nutrition.
The term 'vitamin A' was defined  as the generic descriptor for all C20--ionone derivatives that exhibit qualitatively the biological activity of all-trans retinol. The term 'provitamin A' for the carotenoids giving rise to vitamin A is retained.
Chemically, vitamin A belongs to the 'retinoids', defined  as a class of compounds consisting of four isoprenoid units joined in a head-to-tail manner. These recommendations also contain the statement: all retinoids may be formally derived from a monocyclic parent compound containing five carbon-carbon double bonds and a functional group at the end of the acyclic portion (Formula I).
Formula I: Structure of the parent compound of retinoids
The two definitions do not contradict each other. There are, however, certain implications in the words 'vitamin A' and 'retinoids' that should be considered when using the terms.
'Vitamin A' means a group of substances (retinol, retinyl esters, and retinal) with defined biological activities. Further, there are certain metabolites of vitamin A, such as all-trans and cis-isomeric retinoic acids, that can perform some, but not all, of the biological functions of vitamin A; they are incapable of being metabolically converted into retinol, retinal, etc. .
Retinoic acid and some of its isomers and derivatives, together with a number of structurally modified retinoids, have been shown to control cell differentiation in many epithelial tissues and to prevent metaplasia [5, 6]. Some of these substances are used in the treatment of various types of keratinization disorders. Such compounds cannot substitute for vitamin A; indeed some of them even act as vitamin A antagonists [7, 8].
The term 'retinoids' is widely employed for this class of compounds. This practice arose from an earlier proposal  to use the name 'retinoids' collectively for both natural forms and synthetic analogues of vitamin A that are capable of preventing the development of cancer. General usage of this term is, however, misleading for two reasons. Firstly, the customary practice gives the name 'retinoids', which has an agreed definition based on chemical structure , to a class of compounds defined by their biological activity. Secondly, many synthetic members of this class of compounds, the so-called 'arotinoids'  or 'retinoidal benzoic acid derivatives'  as well as others, are not chemically retinoids. They contain, e.g., aromatic rings replacing either the basic -ionone type ring structure or unsaturated bonds of the tetraene side chain of the retinoid skeleton (Formula II).
Formula II: Structure of an 'arotinoid'
We now suggest that the compounds that control epithelial differentiation and prevent metaplasia, without possessing the full range of activities of vitamin A, should be termed 'retinoate analogues'. Although they are usually called 'retinoids', we discourage their designation by a term that has a defined, but different, meaning.
A new term for the group of substances with such antimetaplastic activities may be desirable, especially if it is based on their biological activity. It should not imply a chemical structure because of heterogeneity among the compounds. Proposals for such a term are welcome.
Return to Contents page
Go to Reference section