Idoxuridine
Since the DNA polymerase of the virus was known to be specifically coded and was known to have different affinities, compared to the cellular DNA polymerase (Herrmann, 1961), it seemed reasonable that effective virus chemotherapy would have to involve a drug that worked specifically on the viral DNA polymerase, either by inhibiting its activity or by being incorporated itself into a false DNA chain, which would no longer code proper genetic material and therefore would not reproduce, so that new virus would not be made. We first tested idoxuridine because it was very similar to thymidine chemically (Kaufman, 1962) (Table 1.1). The van der Waalâs radius of the iodine group was similar to that of the methyl group of thymidine, and it was felt that this similarity would allow idoxuridine to interfere with the uptake of thymidine in the synthetic process. The chloro- and bromo-deoxyuridines appeared to be more toxic than the iodinated form of the molecule.
Table 1.1
Possible sites of inhibition of DNA synthesis by antiviral compounds
Iodo-deoxyuridine was synthesised by William Prusoff, of the Department of Pharmacology at Yale University, approximately three years before its first use as an ocular antiviral agent (Prusoff, 1959). The compound was developed originally as an anticancer drug but was relatively ineffective. From the biochemical studies that were carried out at that time, it was known that idoxuridine directly inhibited DNA polymerase and was incorporated into a false DNA. For this reason, and because of its lesser toxicity, this compound was tested as a topical antiviral, and it was found to be effective. A series of double-blind studies confirmed the activity of idoxuridine, but a similar drug, 5-fluorouracil, was found to inhibit the synthesis of thymidine, rather than its uptake, and was not an effective in vivo antiviral agent.
The testing of antiviral drugs was a new area. With further work, it became clear that in animal and clinical studies, the use of complex scoring systems that included redness of the eye, iritis, etc. (Draize et al, 1944) clouded the precise measurement of drug efficacy and made the prediction of drug potency difficult. The results were more precisely predictive when we concentrated on the size of the corneal ulcer alone and the improvement in corneal healing during drug administration.
One of the problems with idoxuridine in its original form was its relative lack of solubility. In part because of this, it was necessary initially to administer the drug every hour during the day and every two hours at night; later, a longer-lasting ointment was used at night. However, clearly, idoxuridine was effective and clinically useful in the treatment of ocular herpes (Kaufman et al, 1962).
Mechanism of action
For antiviral drugs to be effective and to interact with DNA polymerase, they must be present in the cell in the triphosphate form. The first phosphate is added to the molecule either by cellular thymidine kinase or by the virus-encoded enzyme; the virus-encoded thymidine kinase is much more active than the cellular enzyme. Two additional phosphates are rapidly added by cellular enzymes and the compound is then able to interact with the DNA polymerase.
All of the metabolite inhibitors have multiple sites of activity, and it is not always certain which is the true rate-limiting mechanism of action. In spite of this, we can hypothesise a mechanism of action for these drugs. DNA is a helix with a backbone of sugar molecules hooked to each other with 3-,5-bonds, forming a long chain. Attached to the backbone are the bases: adenine, thymine, guanine, and uracil. These bases are read by chemical processes to produce messenger RNA that provides the code for protein synthesis. The base pattern is also involved in the duplication of DNA. The bases themselves are not essential to the structure of the backbone although theoretically, they could stereochemically hinder the formation of the helix.
It appears that the nucleoside analogues that have normal sugar moieties and abnormal purine or pyrimidine substituents for some of the bases, probably are incorporated in the formation of DNA, but the mistakes in the base pattern code for products do not permit the replication of the virus. There may be some direct inhibition of DNA synthesis through abnormal attachments and binding to the DNA polymerase, but in general it seems that the nucleosides with normal sugars and abnormal bases act by incorporation and false coding. Those nucleosides with abnormal sugarsâespecially abnormalities involving the three or five positions on the carbon chain, which are necessary to link up the sugars to make the DNA chain â probably act largely as chain terminators. They are picked up by the DNA polymerase and one bond forms to a normal sugar molecule of the normal DNA chain, but the chain cannot be continued because the sugar is too abnormal to sustain a 3â5 bonding and permit the synthesis of the backbone of the chain. Thus small pieces of DNA are synthesised and the synthesis of a complete DNA chain is interrupted. This explanation is somewhat oversimplified, and some of the mechanisms are still somewhat uncertain, but it does provide enough of an overview to explain how antiviral drugs exert their effects.
Adenine arabinoside
The next topical antiviral to be developed for ophthalmic use was adenine arabinoside (Underwood, 1962). This arabinose sugar probably acts largely as a chain terminator and is an effective topical antiviral. It is as insoluble as idoxuridine. There has been some suggestion that adenine arabinoside is slightly less toxic than idoxuridine on a chronic basis, but there is no good evidence for this (Dresner & Seamans, 1975). Evidence from double blind studies (Pavan-Langston & Dohlman, 1972) indicate that adenine arabinoside has approximately the same potency as idoxuridine (Table 1.2) and also about the same short term topical toxicity (Dresner & Seamans, 1975). It is also clear that neither adenine arabinoside nor idoxuridine is effective as a primary therapeutic agent in the treatment of stromal disease and iritis.
Table 1.2
Efficacy of idoxuridine and adenine arabinoside
*Generally the severity of the disease and the duration of symptoms were greater in patients entered in the open study
(adapted from Dresner & Seamans, 1975).
Trifluorothymidine
Trifluorothymidine, like idoxuridine, was synthesised originally as an anti-cancer drug by Gottschling & Heidelberger (1963). However, trifluorothymidine has several unique properties. It is extremely potent, perhaps because it has a dual mechanism of action; it inhibits the de novo synthesis of thymidine (as a thymidylic synthetase inhibitor), and it also acts directly as a substitute thymidine analogue on DNA polymerase (Heidelberger et al, 1962). It is extremely soluble, so that high concentrations can be obtained in drop form, and ointments are not necessary (Kaufman & Heidelberger, 1964). The drops are usually given between five and eight times a day in the treatment of dendritic ulcer (Wellings et al, 1972), and approximately 97% of patients are cured within two weeks. Resistance to trifluorothymidine occurs, but it is very rare, possibly because of the multiple sites of activity. In the few cases where resistance occurs, the older, less effective drugs can be used. Trifluorothymidine is the most effective topical antiviral now available for the treatment of ocular epithelial herpes, and because of the low incidence of viral resistance, is probably the drug of choice for ordinary topical use.
Out work, as well as studies by others, suggests that trifluorothymidine is also the most potent drug now available for counteracting the steroid effect, when steroids are used as a primary therapeutic agent. Epithelial toxicity from trifluorothymidine is seen infrequently, unless the drug is misused and continued at high frequency for a long time. The drops should not be applied on a frequent dosage schedule for a period of many weeks; when the early high frequency applications are reduced rapidly after about five to seven days, corneal toxicity is rare.
All of the topical antivirals can produce epithelial toxicity. Idoxuridine may appear to be somewhat more toxic, since allergy may play a part in this reaction. The eye may become red, the epithelium keratinised, and punctal occlusion also occurs with herpes infections alone, so it is not certain that idoxuridine therapy is implicated in this problem. All of these antiviral drugs â idoxuridine, adenine arabinoside, and trifluorothymidineâmay produce typical allergenic reactions, and all of them ma...