Absorption of light (UV or visible) by the ground state of a molecule (S0) generates electronically excited states, either directly (the singlet states) or after intersystem crossing from the singlet manifold (the triplet states). Alternatively, triplet states may be generated by energy transfer from another excited state (a sensitiser). In both multiplicities, very fast internal conversion leads to the lowest states (Sj and Tj respectively). These states, although still quite short lived (typical lifetime x<10-8 s for Sj and <10-6 s for Tj) live long enough that a chemical reaction competes with decay to the ground state.

Electronically excited states are electronic isomers of the ground state, and not surprisingly show a different chemistry. These, however, can be understood with the same kind of reasoning that is used for ground state chemistry, taking into account that the very large energy accumulated in excited states makes their reactions much faster (in the contrary case, there would be no photochemistry at all, in view of the short lifetime of the key intermediates). As an example, ketones are electrophiles in the ground state due to the partial positive charge on the carbon atom. The reaction with nucleophiles occurs. In the nrc* triplet excited state electrons are differently distributed, and the important thing is now the presence of an unpaired electron on the non-bonding orbital localised on the oxygen atom. This makes atom transfer to that atom so fast a process (k«106s_1, many orders of magnitude faster than any reaction of ground state molecules) that it competes efficiently with the decay of such a state.

On the basis of such principles, the many photochemical reactions now known have been rationalised. This is shown in many fine books of photochemistry,1"5 which demonstrate both the dramatic development of this science in the last decades and the high degree of rationalisation that has been reached. The photoreactions of drugs6 obviously can be discussed in the same way, and G. M. J. Beijersbergen van Henegouwen (p. 74) pointed out some key points that one should take into account. It is therefore generally possible to predict the photochemical behaviour of a new drug, as of any other molecule, or at least to point out the most likely alternatives.

More exactly, as it has been pointed out by Grenhill in a recent review,7 it is possible to indicate some molecular features that are likely to make a molecule liable to photodecomposition, even if it is difficult to predict the exact photochemical behaviour of a specific molecule. This is due to the fact that competition between the chemical reaction(s) and physical decay to the ground state depends in a complex way on the structure (and on conditions). Thus both the efficiency of a photochemical reaction and product distribution may vary significantly even among closely related compounds and further depend on conditions.

At any rate, several chemical functions are expected to introduce photoreactivity (see Scheme 1). These are:

a. The carbonyl group. This behaves as an electrophilic radical in the mi* excited state. Typical reactions are reduction via intermolecular hydrogen abstraction and fragmentation either via a-cleavage ("Norrish Type I") or via intramolecular y-hydrogen abstraction followed by Ca-Cp cleavage ("Norrish Type II").

b. The nitroaromatic group, also behaving as a radical, and undergoing intermolecular hydrogen abstraction or rearrangement to a nitrite ester.

c. The //-oxide function. This rearranges easily to an oxaziridine and the final products often result from further reaction of this intermediate.

d. The C=C double bond, liable to E/Z isomerisation as well as to oxidation (see case g).

e. The aryl chloride, liable to homolytic and/or to heterolytic dechlorination f. Products containing a weak C-H bond, e.g. at a benzylic position or a to an amine nitrogen. These compounds often undergo photoinduced fragmentations via hydrogen transfer or electron-proton transfer.

g. Sulphides, alkenes, polyenes and phenols. These are highly reactive with singlet oxygen, formed through photosensitisation from the relatively harmless ground state oxygen.

Such functions are present in a very large fraction, if not the majority, of commonly used drugs. Thus, many drug substances, and possibly most of them, are expected to react when absorbing light. However, photodegradation of a drug is of practical significance only when the compound absorbs significantly ambient light (»330 nm), and even in that case the photoreaction may be too slow to matter, particularly if concentrated solutions or solids are considered. It is important to notice that most information about photoreactions available in the literature refers to the conditions where such processes are most easily observed and studied, viz. dilute solutions in organic solvents, whereas what matters for drug photostability are (buffered) aqueous solutions or the solid state. Under such different conditions the photoreactivity of a drug may be dramatically different. To give but one example, benzophenone triplet - probably the most thoroughly investigated excited state - is a shortlived species in organic solvents, e. g. x ca 0.3 jis in ethanol, and is quite photoreactive via hydrogen abstraction under such conditions, and in general in an organic solution. However,



Products b)



0 0

Post a comment