J. Arendt,     Division of Clinical Biochemistry, Department of Biochemistry, University of Surrey, Guildford, Surrey, GU2 5XH, UK

ABSTRACT

Current available methodology for the assay of melatonin is briefly reviewed. Areas of controversy, due probably to technical problems, are identified, together with observations confirmed by reports from different laboratories. Some technical improvements are suggested.

KEYWORDS

Melatonin assay

TEXT

In a rapidly moving field such as the study of pineal indoleamines, it is likely that an assessment of the “current status” of assay methodology will be out of date long before it appear in print. Nevertheless trends can be recognised and areas of controversy identified.

Wurtman and Axelrod’s (1965) formulation of the pineal gland as a neuroendocrine transducer, and the consequent interest in pineal products as possible humoral effectors of photic signals, led to intensified efforts to measure the methoxyindoleamines in tissues and body fluids. Circulating melatonin in particular was, and is, considered to be largely, if not exclusively (Lewy and coworkers, 1980; Arendt, Forbes, Brown and Marston, 1980) of pineal origin. While several reports demonstrate its presence and synthesis in the retina and the Harderian gland (Cardinali and Wurtman, 1972; Pang and coworkers, 1976), it is possible that from these sites of origin it does not reach the general circulation.

From 1970 to 1975, it was possible to measure melatonin using the tadpole bioassay of Ralph and Lynch (1970). This assay, whilst considered to be specific, was particularly tedious and necessitated a complicated tadpole breeding programme. In addition, the sensitivity was insufficient for determinations in sequential blood samples : of great importance in the study of a rhythmic variable.

During the years 1975-1978 an expolsion of publications concerning the measurement of melatonin occurred, the vast majority using radioimmunoassay (RIA) followed by gas-chromatography-mass-spectrometry (GC-MS) and gas-chromatography (GC). A review of this methodology has appeared (Arendt, 1978). From 1978 until the time of writing there has been little basic change in methodological approach – as assays have been refined, reported values have generally speaking decreased, and areas of controversy have appeared which may or may not be of methodological origin.

Clearly an assay has to be useful within the confines of the system it is proposed to investigate, and the potential user must choose a method according to his or her own requirements. For example clinical screening for pathological disturbance should be rapid and preferably cheap. Where gross differences in measured values are diagnostic, then some degree of noise and assay variation is tolerable. Very few literature reports of gross clinical melatonin variations exist, and these are unconfirmed (Barber, Smith and Hughes, 1978; Silman and coworkers, 1979). Where subtle differences are important, for example phase-changes in circadian rhythm, then a high degree of assay precision and accuracy is required. In the case of one application of melatonin assay – assessment of central nervous syst em rhythmic function – a precise measurement of pineal β-receptor stimulation is required, and interference by other methoxy-indoles, whose synthetic control is differently mediated, is not tolerable.

In the case of urinary melatonin assay, it is likely that a system recognising either 6-hydroxymelatonin (Lewy and coworkers, 1980) or both 6-hydroxymelatonin and melatonin, may well provide more information on pineal output than measurement of melatonin separately (Lynch and coworkers, 1975), in view of the small percentage of melatonin thought to be excreted unchanged (Kopin and coworkers, 1961).

Criteria for melatonin RIA acceptability (which, with slight changes in terminology, can also be applied to other assay systems) are generally considered to be as follows:

Recent reports of low to undetectable melatonin levels following short-term pinealectomy (Lewy and coworkers, 1980; Arendt and coworkers, 1980), or suppression of pineal indoleamine synthesis (Arendt and coworkers, 1980) may provide a more physiologically meaningful assessment of assay performance.

Most published RIA’s fulfil the criteria 1) − 6) (for references see Arendt, 1978).

In view of the prevailing opinion that GC-MS provides the most specific available technique, several authors have made successful, if limited GC-MS/RIA comparisons, however this is clearly a difficult undertaking for the majority of laboratories. It is to be hoped that inter-laboratory comparative studies will include at least one laboratory performing GC-MS assay so that the maximum number of workers may compare their own results with this highly specific but expensive technique.

Clearly, with RIA, every sample cannot be validated : with a good GC-MS system every sample is validated. A compromise solution, both from the point of view of assay specificity, capital investment, running expenses and throughput, may be found in high-performance liquid chromatography (HPLC) separation of melatonin from extracted material, followed by RIA or electrochemical detection (Goldman, Hamm and Erickson, 1980). Whether the sensitivity required for plasma determinations will be achieved by electrochemical detection remains to be seen.

Comparison of results from different assay systems in order to pinpoint the ‘best’ method for a given situation is clearly difficult. In general, for a small molecule obtainable in a pure state, such as melatonin, provided that recovery of added compound from the (so far inevitable) extraction procedures is good, then the lower the reported values, the more specific the assay is likely to be. Amongst the published RIA methods, that of Rollag and Niswender (1976) using iodinated melatonin analogue as tracer has the greatest sensitivity, although not necessarily the lowest values. Unless every sample is validated, it is more correct to refer to measured immunoreactive material as melatonin-like.

It cannot be emphasised too strongly that the establishment of physiological melatonin variations in an adequately large population is an essential prerequisite to pathological studies. However, the complete absence of 24-hour variations is, according to the vast majority of authors an infrequent observation in normal mammals, and would appear to be good evidence of disturbed pineal and/or central nervous rhythmic function. Indeed the areas of least controversy in the measurement of melatonin include its almost uniform dark-phase rise in all species studied to date (Lynch and coworkers, 1975; Kennaway and coworkers, 1976; Rollag and Niswender, 1976; Arendt and coworkers, 1977; Wilkinson and coworkers, 1977; Per low and coworkers, 1980; Vivien-Roëls and coworkers, 1979; Owens, Gern and Ralph, 1980). In addition, several authors have shown a lack of immediate response of human circulating melatonin to artificial light in the dark-phase (Jimerson and coworkers, 1977; Arendt, 1978; Lynch and coworkers, 1978) compared to other species (Rollag and Niswender, 1976; Illnerova and Vanacek, 1979; Perlow and coworkers, 1980) and disturbed plasma melatonin variations in psychiatric disease (Wirz-Justice and Arendt, 1979; Lewy and coworkers, 1979; Wetterberg and coworkers, 1979; Mendlewicz and coworkers, 1979). Most other observations are not yet repeated in competing laboratories. Good general agreement is found amongst different laboratories with respect to pineal and plasma melatonin levels in rats and sheep measured by RIA. Recent results in adult human values show very large (up to 50-fold) variations from the lowest reported daytime levels (Lewy and Markey, 1978 – GC-MS) to the highest (Mendlewicz and coworkers, 1979 – RIA), probably due in part to differences in assay specificity. In addition reports of high “spikes” of circulating melatonin in man during day and night (Weitzman and coworkers, 1978; Mullen and coworkers, 1980), are unconfirmed. In the authors laboratory daytime melatonin in normal volunteers...