Chronotype-Dependent Changes in Sleep Habits Associated with Dim Light Melatonin Onset in the Antarctic Summer

The biological clock has evolved in all living beings to adapt their functioning to a predictable light–dark cycling environment. Consequently, although one of the most essential characteristics of the biological clock is to keep a circadian time reference even in free-running conditions, the daily adjustment of the clock is very sensitive to light exposure [1,2]. Undoubtedly, light is the most important zeitgeber and entrainer of the clock across evolution [3,4,5]. It is indisputable that modern urban life has affected human light exposure by increasing it during nighttime (due to electric light and screens) and decreasing it during daytime (due to indoor work in enclosed buildings). It is a matter of current debate how the human clock copes with these quite recent changes in environmental light, by which the reliable day–night natural cycle has been distorted [6,7,8].

Human life schedules are more frequently organized to respond to family, work, and social demands than to the environmental light cycle. The pineal secretion of melatonin acts as the main signal for human integrative time keeping as indicator of the beginning of the night [9,10]. Though melatonin always shows a nocturnal increase even in people that remain active at night (for example, in nocturnal-shift workers [11,12]), the secretion of melatonin is highly sensitive to light. Light exposure at night suppresses melatonin secretion [13], but bright light at daytime (especially in the morning) increases nocturnal melatonin secretion [14,15]. In addition, the rhythm of melatonin secretion provides the best available measure of the timing of the internal clock [9]. The dim light melatonin onset (DLMO, the start of the evening rise of melatonin in plasma or saliva) is the best proxy of the functioning of the circadian clock in humans [10,16]. The DLMO itself is affected by light, so it can be either advanced by bright light exposure in the morning [17,18,19,20] or delayed by light exposure during the evening [21,22]. Likewise, it is clear that seasonal changes in day length affect the duration and phase of the nocturnal melatonin pulse [23,24].

Circadian preferences represent individual differences among people in the sleep–wake schedule, which depend on genetic factors, sex, and age, as well as on physical and social environmental cues [25,26,27,28]. The adult population can be classified into three chronotypes: Morning-oriented individuals (early chronotypes) that go to bed and wake up early, evening-oriented individuals (late chronotypes) that tend to remain active at night, and intermediate chronotypes that correspond to about 60% of the population [29,30]. Circadian preferences impact on the daily melatonin profile. Early melatonin onset is associated with early chronotypes, while late chronotypes show later DLMO values [31,32,33]. In addition, light exposure has been reported to influence circadian preferences completing the picture of mutual interactions between light, melatonin, and circadian preferences. In Amazonian rubber-tapper workers, later chronotypes and delayed melatonin onset occur in individuals exposed to electric light with respect to those exposed to the natural photoperiod [34].

Two previous studies evaluated the influence of acute light manipulation on the phase relationship between melatonin onset, sleep schedule, and circadian preferences [35,36]. Wright et al. [35] compared subjects under two conditions (constructed environment with electrical lighting and summer natural light–dark cycle camping) and found that not only do sleep patterns and melatonin profiles change and synchronize to solar time but also that later chronotypes exhibit larger circadian advances when exposed to only natural light. On the other hand, Figueiro et al. [36] found that the circadian phase and sleep patterns change in response to controlled advancing and delaying light exposure patterns regardless of individual chronotype. In the present study, we add evidence on the impact of light changes on the triple interaction between melatonin, sleep, and chronotype using the extreme light exposure of the Antarctic summer as an advantageous scenario. We observe chronotype-dependent changes in the melatonin profile and sleep schedule of Uruguayan university students when comparing their Antarctic summer trip with their normal life situation around the autumn equinox in Montevideo, Uruguay.

The chronobiological characterization of the study population was performed in the peri-equinox in Montevideo with data obtained from two largely validated questionnaires (MEQ and MCTQ [38,39]), whose outputs were significantly correlated (R² = 0.51, p = 0.0091). MSFsc (n = 12) was, on average, 5.95 ± 0.83 and ranged from 4.67 to 7.58, indicating participants’ lateness. The average value of the MEQ scores (47 ± 8.01, n = 12) was used to separate late participants (32 < MEQ scores < 47, n = 5) from early participants (47 < MEQ scores < 55, n = 7) within the studied sample. The MEQ scores of early and late participants were significantly different (Mann–Whitney U test, p = 0.0053).

Though the light modulation of melatonin profile has been largely documented [47,48,49,50], the impact of light on the interactions between sleep patterns, melatonin onset, and circadian preferences is a topic of current debate [35,36]. This study demonstrates the chronotype-dependent effects of changes in light exposure on the melatonin profile and sleep patterns of Uruguayan university students who traveled to Antarctica during the summer.

The chronobiological characterization of the study population was performed during their normal daily life in the peri-equinox in Montevideo using two validated questionnaires (MCTQ and MEQ [38,39]), whose results were, as expected, significantly correlated [51]. However, while MCTQ provided very late MSFsc values, as previously reported in a similar population [52], circadian preferences derived from MEQ scores were not markedly late. This discrepancy is not surprising because, while the correlation between MEQ and MCTQ is a good indicator of the coherence and quality of the data, both questionnaires have different aims and are not interchangeable [53]. We selected the MEQ score as a proxy of individual circadian preference. The classification of circadian typologies depends on age, geographic, and cultural differences and, thus, cannot be universal [25,54,55]. However, it is acceptable to define cutoffs points to identify earlier and later participants within homogeneous populations as the participants of this study [56]. The use of the average split criterion to categorize participants into two groups [57,58] was validated by the fact that MEQ scores were significantly different between early and late participants.

Expected but important conclusions emerged from the analysis of both light incidence and participants’ light exposure. A light incidence of high intensity light (>1000 lx) was obviously longer in Antarctica, but, more importantly, participants were actually more exposed to high intensity light in Antarctica than in Montevideo (Figure 1). It is widely recognized that the phase responses of circadian clocks to light could be either advances or delays, depending on the biological time when light stimulation is applied [2,21,59,60,61,62,63,64]. In short, exposure to light during the evening induces phase delays by suppressing nocturnal melatonin secretion [50,63,64,65]. In contrast, exposure to bright light in the morning has been documented to prevent circadian phase shifts and to increase the total amount of nocturnal melatonin secretion [14,15,17,18,20,62]. Therefore, we focused on these sensitive time windows (morning and evening) to evaluate differences in light exposure between Montevideo and Antarctica. Interestingly, we observed that all participants, regardless their chronotype (Table 1), were significantly more exposed to light in Antarctica during the morning (7:00–12:59 h) and at the beginning of the evening (19:00–21:59 h) but not in the late evening time window (22:00–0:59 h). It is interesting to find differences between chronotypes in the light exposure in this late sensitive time window in which 50–100 lx are enough to delay the circadian phase [50]. Both early and late chronotypes were exposed to a similar amount of light from 22:00 to 0:59 h in each location, but a significantly higher light exposure occurred for early participants during the Antarctic trip at this time window (Table 1).

As the most reliable seasonal time cue is the changing photoperiod, Antarctica—with its extreme high latitude geographical distribution and the confinement and shared schedule of Antarctic bases’ crews—has provided a natural laboratory for the observation and manipulation of seasonal changes in human circadian systems [24]. In the long nights of the Antarctic winter, for example, there is no light information to suppress melatonin levels and thus to set circadian time. Therefore, two lines of evidence arise from Antarctic seasonal studies: 1) Daytime melatonin levels are significantly higher in the Antarctic winter than in the summer [66,67], and 2) the melatonin rhythm is delayed in the winter with respect to the summer [66,68,69,70,71,72]. In addition, a long series of studies conducted in the small overwintering community of the British Halley Research Station demonstrated that the winter delay in the melatonin rhythm can be advanced to the summer phase by bright light treatment [73]. Considering that in the Antarctic summer crew members are more exposed to light in both light-sensitive time windows with opposite effects on circadian shifts, these results indicate that, on average, the morning window is more important in resetting the circadian phase. Experimental data have added evidence to support that bright light in the morning, together with a strong contrast in day–night light, are the most relevant determinants of circadian phase advance [20,61,74,75,76,77,78]. In this regard, as light exposure was enormously higher in Antarctica during all morning time (Figure 1; Table 1), we expected all participants to advance their circadian phase in Antarctica with respect to Montevideo.

In both experimental conditions (Montevideo and Antarctica), we found the expected melatonin profile with lower levels in daytime and higher levels in nighttime (Figure 2). In line with the well-known suppressor effect of daytime light [13] and the previous Antarctic summer reports [66,73,79], we found that daytime melatonin levels were extremely low in the Antarctic summer and significantly lower than in the peri-equinox in Montevideo (Figure 2). It is currently accepted that DLMO is the best proxy of the individual functioning of the circadian clock, and it is thus a comprehensive way to measure circadian phase shifts [44,80,81,82]. DLMO correlates with circadian preferences [31,32], shows seasonal changes [9,83,84,85], and can be either advanced or delayed by light [20,22,64,78,86]. However, only few previous studies have tested the effects of light changes on the interaction among DLMO, chronotype, and sleep patterns [35,36]. These studies were consistent in showing that DLMO was advanced when participants were more exposed to light in the morning light-sensitive window and that it was delayed when they were more exposed to light in the evening light-sensitive window [35,36]. In contrast to these previous reports, the participants of this study did not show significant changes in DLMO between Montevideo and Antarctica (Figure 3A). This may be due to life conditions in Antarctica that differed not only in light exposure with respect to Montevideo. The confluence of many social confounding interactions is also known to impact on the synchronization of the circadian clock, such as tight schedules, mealtimes or routine physical activity [87,88,89,90,91,92,93]. Indeed, the organized nature of Antarctic life synchronized the circadian system of the participants as daytime melatonin values, DLMO, and awakening times were less variable in Antarctica for the whole study sample.

The lack of an overall DLMO shift between Montevideo and Antarctica did not mean that all individual DLMOs remained unchanged (Figure 3B–C). Rather, participants’ DLMO shifted in a chronotype-dependent manner (Figure 4). Interestingly, although we did not find the expected correlation between DLMO and circadian preferences [32], in Antarctica, late participants tended to advance their DLMO while early participants delayed their DLMO. This chronotype-dependent reversion in DLMO shift may be explained if early participants were significantly more exposed to light than late participants during the evening phase-delay time window or if there was an inter-individual difference in the sensitivity to evening light [94,95]. No differences in Antarctic light exposure between early and late participants were observed at any time (Table 1). However, from 22:00–0:59 h, only early participants were significantly more exposed to light in Antarctica than in Montevideo. Therefore, only early participants were subjected to a differential effect of light during the phase-delaying evening window that was probably enough to induce the observed phase delay. Though it has been previously reported that 200 lx light at the peak of the evening phase-sensitive window delays the circadian phase [50], it is outstanding to note that a subtle increase in light exposure during this late phase-sensitive window (from 100 lx in Montevideo to 200 lx in Antarctica) can counteract the expected effect of the enormous morning light exposure (from 1000 lx in Montevideo to 9000 lx in Antarctica).

Only few previous studies reported chronotype dependent shifts in DLMO. For example, Lewy and Sack [10] found that the later the DLMO (late chronotypes), the higher the phase advance by light in the morning. In addition, Wright et al. [35] reported that the phase advance observed in natural light respect to electric light conditions was higher in late chronotypes. As expected from these observations, we can interpret that late participants of this study were affected by the Antarctic increase in morning light and thus advanced their DLMO, disregarding the effect of the increased light exposure they received during the evening. However, early participants did not respond as expected to the increase in light exposure during the phase-advancing time window, but they did delay their DLMO in Antarctica. Early participants seem to be more responsive to the evening phase-delaying time window than to the morning phase-advancing one. Chronotype differential responsiveness may be explained by the observation that early participants (but not late ones) received a subtle but significant increase in light exposure in Antarctica during the late evening phase-delaying time window. Our data suggest that individual circadian preferences set a differential light history that tunes the susceptibility to light-sensitive time windows. In line with these observations, the Antarctic trip also affected participants’ sleep patterns in a chronotype-dependent manner (Figure 5). As previously reported in a similar population [52], the sleep duration of all participants decreased in Antarctica with respect to Montevideo, probably due to the tight agenda of both social and academic schedules of the Antarctic summer school. However, it is interesting to note that early participants (but not late ones) delayed their sleep onset in Antarctica despite their important sleep debt. We interpret that the light-induced delay in the circadian phase observed in early participants also prevented them from going to sleep earlier to ameliorate their sleep deficit.

The summer trip to Antarctica meant, for this group of university students (when compared with their normal lives in the peri-equinox in Montevideo), a significant increase in light exposure in both the morning phase-advancing (7:00–12:59 h) and in the evening phase-delaying (19:00–0:59 h) time windows. Though participants with earlier and later circadian preferences were exposed to a similar amount of light in each site and in both time windows, in Antarctica, late chronotypes advanced their DLMO with no changes in sleep onset time and early chronotypes delayed their DLMO and sleep onset time. The behavior of late chronotypes is in line with previous observations in the Antarctic summer [66,68,69,70] and in light manipulation experiments [35,36,73], which demonstrated the powerful circadian phase advancing effects of increasing light in the morning. In contrast, early chronotypes seem to be more responsive to the Antarctic increase in evening light exposure, which delayed their DLMO and thus prevented them from advancing their sleep start times to mitigate their strong sleep deficit. In this natural experiment, we have shown that there is a different susceptibility between chronotypes to respond to light. Despite the fact that all participants were exposed to an enormous light increase in Antarctica during the morning phase-advancing window, early participants delayed their circadian phase and sleep patterns in response to a subtle increase in light exposure during the evening.

We wish to thank the students of the Second Summer School on Introduction to Antarctic Research, who kindly volunteered as participants of this study, in particular Delfina Castiglioni, Mercedes Paz and Eliana Nicolaisen who helped in data acquisition and preliminary data processing. We are especially thankful to the Instituto Antártico Uruguayo for the logistic support and to GeneActive for the generous actimeter loan.

Funding, project administration, experimental design, and writing—original draft preparation, A.S. and B.T.; statistical analysis and figures design, D.S. and B.P.; melatonin assays, C.C. and A.D. All authors provided critical revisions to the manuscript and approved the final version.

Financial support was provided by Comisión Sectorial de Investigación Científica (CSIC I+D_2016/623) and Facultad de Ciencias, Universidad de la República, Montevideo, Uruguay.

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.