In humans, numbers of circulating T cells show a circadian rhythm with peak counts during the night and a steep decline in the morning. Sleep per se appears to counter this rhythm by acutely reducing the total number of T cells. The T-cell population, however, is rather heterogeneous, comprising various subpopulations with different features and functions and also different circadian rhythms. Therefore, we examined here whether sleep likewise differentially affects these subsets. We measured eight different T-cell subsets (naïve, central memory, effector memory, and effector CD4+ and CD8+ T cells) over a 24-h period under conditions of sustained wakefulness compared with a regular sleep-wake cycle in 14 healthy young men. Sleep reduced the number of all T-cell subsets during nighttime with this effect reaching the P < 0.05 level of significance in all but one subpopulation, i.e., effector CD4+ T cells, where it only approached significance. Furthermore, sleep was associated with an increase in growth hormone, prolactin, and aldosterone levels, whereas concentrations of catecholamines tended to be lower than during nocturnal wakefulness. The effect of sleep uniformly decreasing the different T-cell subsets is surprising considering their differential function and circadian rhythms, and even more so, since the sleep-induced decreases in these subsets are probably conveyed by different hormonal mediators. Although the reductions in cell numbers are rather small, they are comparable to changes seen, for example, after vaccination and are, therefore, likely to be of physiological relevance.
- T-cell subsets
t cells constantly migrate through the blood to reach different lymphoid and nonlymphoid tissues, where they are activated if they encounter their cognate antigen (50). As a consequence of this recirculation, human T-cell numbers in the blood are not constant, but follow a strong circadian rhythm with peak numbers during the night and a steep decline in the morning hours (e.g., 9, 35, 41). Studies comparing a regular 24-h sleep-wake cycle with continuous wakefulness revealed that on top of this strong circadian rhythm, sleep per se exerts a subtle, but robust control of circulating T cells, by acutely reducing T-cell numbers (9, 19, 20). This sleep-induced reduction has been shown for CD4+ and CD8+ T cells (9, 20). However, these populations are composed of subpopulations that are quite disparate in terms of functions and migratory routes. Both CD4+ and CD8+ T cells can be further subdivided into naïve (TN), central memory (TCM), effector memory (TEM), and terminally differentiated effector T cells (TTE) [also called CD45RA+ effector memory T cells (TEMRA)], according to their stage of differentiation (29, 45). TN and TCM have a low differentiation status and continuously recirculate through secondary lymphoid organs, where they differentiate and proliferate following activation by their cognate antigen (45). In contrast, the more differentiated TEM and TTE subsets preferentially migrate to inflamed tissues where they can exert immediate effector functions upon antigen recognition (44).
Interestingly, the expression of circadian rhythms of these subsets appears to be related to their differentiation state (16): circulating numbers of TN have the strongest circadian rhythm with a peak during the early night, followed by TCM and then TEM, showing rhythms with gradually lower amplitudes. In contrast, CD8+ TTE show an inverse rhythm with highest numbers during daytime, and CD4+ TTE numbers are generally very low and seem to lack a circadian rhythm (16). Because of the obvious differences in functional properties and circadian rhythms, we wondered whether these subsets are also differentially affected by sleep. Likely mediators of the reduction of circulating T-cell numbers during sleep are hormones, such as growth hormone, prolactin, aldosterone, and catecholamines (7, 8), which are regulated by sleep (e.g., 11, 40, 47, 49) and, in turn, can affect T-cell migration (e.g., 8, 18, 42, 46). As the CD4+ and CD8+ TN and CD8+ TTE subpopulations are most strongly affected by sleep-dependent hormones, we expected the most pronounced reducing effects of sleep on these subsets (5, 8, 18). To test this hypothesis, we conducted a randomized cross-over study in healthy men and analyzed numbers of CD4+ and CD8+ TN, TCM, TEM, and TTE in the blood repeatedly during a regular sleep-wake cycle compared with 24 h of sustained wakefulness.
Fourteen physically and mentally healthy men participated in the study (mean age 25 yr, range 21–30 yr) forming part of a larger parent trial. Results on circadian rhythms of T-cell subsets and on the effect of sleep on other leukocyte subsets have been reported elsewhere (16, 19). Women were not included in the study because of known interactions between sleep and the menstrual cycle (e.g., 4), which would have substantially increased interindividual variance. All participants were nonsmokers, did not suffer from sleep disturbances [as assessed by interviews asking for difficulties falling asleep or maintaining sleep and by visual inspection of the polysomnographic recording from the adaptation night (see below)], and were not taking any medication at the time of the experiments. None had a medical history of chronic disease or mental disorder. Acute illness was excluded by physical examination and routine laboratory investigation. The men had a regular sleep-wake rhythm for at least 6 wk before the experiments and did not nap during the day, as assessed by interviews. Additional inclusion criteria were age between 18 and 30, a body mass index of 20–25 kg/m2, and a habitual bed time between 11 PM and 7 AM ± 1 h. All subjects spent one adaptation night in the laboratory to habituate to the experimental setting and to exclude possible sleep disturbances. The study was approved by the Ethics Committee of the University of Lübeck, and all participants gave written informed consent.
Experiments were performed according to a within-subject cross-over design, as previously described (19). Briefly, each man participated in two experimental conditions, each starting at 8 PM and ending 24 h later. One condition (“sleep”) included a regular sleep-wake cycle (sleep allowed between 11 PM and 7 AM), whereas in the other condition (“wake”), subjects remained awake throughout the 24-h experimental period (Fig. 1). Both experimental sessions for a subject were separated by at least 4 wk and took place between February and April. The order of conditions was balanced across subjects. On the “wake” condition, subjects stayed awake in bed in a half-supine position between 11 PM and 7 AM and were allowed to watch TV, listen to music, and to talk to the experimenter at normal room light (about 300 lux). They were under constant supervision by the experimenter to ensure that they did not fall asleep at any time. On both conditions, blood was sampled first at 8 PM, then every 1.5 h between 11 PM and 8 AM, and every 3 h between 8 AM and 8 PM the next day. Blood was sampled via an intravenous forearm catheter, which was connected to a long thin tube and enabled blood collection from an adjacent room without disturbing the subject's sleep. Standardized meals were provided at appropriate times for breakfast (8 AM), lunch (12 PM), and dinner (6 PM). Standard polysomnographic recordings confirmed that sleep architecture in the sleep condition was normal for laboratory conditions (average time in minutes ± SE spent in the different sleep stages were as follows: total sleep, 430 ± 13; S1, 32 ± 5; S2, 226 ± 11; slow-wave sleep, 72 ± 6; rapid-eye movement sleep, 63 ± 7; wake after sleep onset, 36 ± 10).
Measurement of T-cell subpopulations.
Absolute counts of CD3+ total T cells, CD4+ T cells, and CD8+ T cells were determined by a “lyse no-wash” flow cytometry procedure. Briefly, 50 μl of an undiluted blood sample was immunostained with anti-CD3/FITC (mouse IgG1, clone SK7), anti-CD8/PE (mouse IgG1, clone SK1), and anti-CD4/APC (mouse IgG1, clone SK3) in Trucount tubes (BD Biosciences, San Jose, CA). After 15 min of incubation at room temperature, 0.45 ml of FACS lysing solution (BD Biosciences) was added to lyse erythrocytes for 15 min. Finally, samples were mixed gently, and at least 10,000 CD3+ cells were acquired on a FACSCalibur using CellQuest Software (BD Biosciences). The blood was not washed to ensure the optimal determination of absolute cell counts without losing any cells.
For detection of T-cell subsets, whole blood was incubated with anti-CD4/FITC (mouse IgG2a, clone EDU-2) and anti-CD8/FITC (mouse IgG2a, clone UCHT-4) (Diatec, Oslo, Norway), anti-CD3/PerCP (mouse IgG1, clone SK7) and anti-CD62L/PE (mouse IgG2a, clone SK11) (BD Biosciences), and anti-CD45RA/APC (mouse IgG2b, clone MEM-56) (Invitrogen, Carlsbad, CA). Cells were lysed with FACS lysing solution (BD Biosciences), washed, and resuspended, and at least 10,000 CD4+ or CD8+ T cells were acquired on the FACSCalibur using CellQuest Software (BD Biosciences). The subsets within the CD4+ and CD8+ T-cell populations were defined as follows on the basis of their expression of CD62L and CD45RA: TN, CD62L+CD45RA+; TCM, CD62L+CD45RA−; TEM, CD62L−CD45RA−; and TTE, CD62L−CD45RA+. Absolute counts of T-cell subsets were calculated on the basis of the proportion of the respective CD4+ and CD8+ T-cell subpopulation and on absolute counts obtained by the “lyse no-wash” procedure.
Measurement of hormone levels.
Samples for measuring hormone concentrations were kept frozen at −70°C until assay. Growth hormone (GH), prolactin, and cortisol were measured in serum using commercial assays (Immulite, DPC-Biermann, Bad Nauheim, Germany). Aldosterone was assessed in serum by ELISA (DRG, Marburg, Germany). Epinephrine and norepinephrine were measured in plasma by standard HPLC. Sensitivity and intra-assay and interassay coefficients of variation were as follows: GH: 0.01 ng/ml, <6.6%; prolactin: 0.16 ng/ml, <9.5%; aldosterone: 5.7 pg/ml, <10%; cortisol: 0.2 mg/dl, <10%; epinephrine: 2.0 pg/ml, <5.6%; and norepinephrine: 5.0 pg/ml, <6.1%.
Data are presented as means ± SE. Statistical analyses were based on repeated-measures ANOVA after normal distribution of the data was confirmed by the Kolmogorov-Smirnov test. ANOVA factors were Sleep/Wake to represent the two experimental conditions and Night/Day, to represent the respective intervals of the 24-h measuring period and Time reflecting respective five single time points of the night (12:30-6:30 AM) and of the day period (8 AM-8 PM). Degrees of freedom were corrected using the Greenhouse-Geisser procedure. Two-tailed, paired t-tests were applied to analyze post hoc differences at single time points once ANOVA indicated significant effects. A P < 0.05 was considered significant.
Effect of sleep on T-cell subpopulations.
Compared with nocturnal wakefulness, sleep acutely reduced numbers of total CD4+ and CD8+ T cells (see Table 1 for ANOVA results; see Fig. 2 for pairwise post hoc comparisons). Sleep also reduced counts of all CD4+ and CD8+ T-cell subsets during the night except for CD4+ TTE, which failed to reach significance in the overall ANOVA (Table 1). For CD8+ TTE, the effect became apparent as early as 12:30 AM, whereas for the other affected subsets, numbers were significantly changed only from 2 AM on (see Fig. 3 for pairwise post hoc comparisons). There were no differences between T-cell subsets when comparing the % change from the wake to the sleep condition (F7,91 = 0.57, P = 0.6 for main effect of Subset; F63,819 = 1.02, P = 0.4 for Subset × Time). See Table 1 for % changes from the wake to the sleep condition at 2 AM, the time point at which all subsets were affected by sleep. The significant differences between the sleep and wake conditions persisted when calculations were based on T-cell numbers that were divided by the hematocrit. This shows that the observed changes were not due to an unspecific effect of sleep on the cellular content of the blood.
Effect of sleep on hormone levels.
Sleep increased levels of GH during the early night, of aldosterone during the later night, and of prolactin throughout the entire night period (F1,13 ≥ 9.52, P ≤ 0.009 for Sleep/Wake × Night/Day). In contrast, reducing effects of sleep on the concentrations of norepinephrine and epinephrine failed to reach significance in the overall ANOVA (F1,13 = 4.34, P ≤ 0.057 for Sleep/Wake × Night/Day, and F4,52 = 2.39, P ≤ 0.086 for Sleep/Wake × Night/Day × Time, respectively). Data are not shown because effects of sleep on the release of these hormones are well known (11, 40, 47, 49) and because results of hormone levels from an overlapping study population were already partially published (17). Cortisol levels were not affected by sleep (F1,13 or F4,52 < 1.7, P > 0.19 for main effect of condition or any interaction).
It is known that sleep is important for proper immune functions, and the adaptive immune system seems to be especially sensitive to the effects of sleep (7, 43). T cells are key players in the adaptive immune system, and their constant recirculation between the blood and various tissues is essential to the development and exertion of adaptive immune functions (3). Several studies have already explored the impact of sleep on numbers of circulating T cells. However, results were rather inconsistent with studies reporting reducing, as well as increasing, or no effects of sleep on T-cell counts (9, 21, 23, 43). The discrepancies between these studies might be explained by essential differences in study designs, like the frequency and time point of blood sampling. In addition, these studies distinguished only between CD4+ and CD8+ T cells, but these subpopulations are each composed of several further subsets that might be differentially affected by sleep. In the present study, we, therefore, analyzed the effects of sleep versus nocturnal wakefulness on eight different T-cell subsets, with blood samplings every 1.5 h during the night and every 3 h during daytime. Sleep compared with nocturnal wakefulness acutely reduced the circulating numbers of total CD4+ and CD8+ T cells, as well as of their subsets during the night. The effect of sleep was significant for the TN, TCM, and TEM CD4+ and CD8+ cell subsets, as well as for TTE CD8+ cells, and only approached significance for the TTE CD4+ subpopulation.
So far, only one study has compared different T-cell subsets between conditions of sleep and total sleep deprivation and found no effect on CD4+ and CD8+ TN or non-naïve T cells (encompassing TCM, TEM, and TTE) (1). These diverging results could reflect methodological differences. In particular, our higher sampling rate during the night (every 1.5 h vs. every 3 h in Ref. 1), and our balanced cross-over design likely increased the sensitivity of our study to the acute effects of sleep on T-cell subset numbers.
TN, TCM, TEM, and TTE are very diverse in terms of their functional properties. For instance, they substantially differ in their circadian rhythm characteristics, in their migration patterns, and in their expression of hormone and chemokine receptors (16, 29). As their sensitivity to the effects of sleep-associated hormones is also disparate (8, 18), the similar effect of sleep on these subsets was rather unexpected. However, different temporal dynamics of the sleep-induced changes in cell numbers of TTE vs. TN, TCM, and TEM (sleep effect starting at 12:30 vs. 2:00 AM, respectively) suggest disparate hormonal mediators. Epinephrine rapidly and selectively mobilizes highly differentiated leukocytes like CD8+ TTE from the marginal pool and seems to mediate the circadian rhythm of this subset with a maximum during daytime (16, 18, 51). Therefore, the reduction in epinephrine levels during early sleep compared with nocturnal wakefulness might also be responsible for the parallel drop in circulating CD8+ TTE. Although the effect of sleep on epinephrine levels failed to reach significance in the present study, this was presumably due to the rather low sampling rate, which did not allow to detect transient changes in hormone levels. Moreover, because other studies found a robust sleep-stage-specific regulation of epinephrine concentrations (22, 40), it seems premature to exclude epinephrine as a potential mediator in the present study. Future studies with higher sampling rates are, therefore, important to prove the connection between the sleep-dependent drop in epinephrine levels and CD8+ TTE numbers.
Unlike CD8+ TTE, numbers of circulating TN, TCM, or TEM in humans are not altered by physiological variations in epinephrine (16, 18). In addition, these subsets follow an opposing circadian rhythm to that of TTE, with a strong drop in cell counts in the morning hours (16). This decline is thought to reflect a redistribution of the cells to the bone marrow following the morning increase in cortisol (6). Here, this overall circadian rhythm in cell numbers persisted after one night without sleep, and cortisol was likewise not changed in the present, as well as in other studies, by short-term sleep manipulation (e.g., 23, 27, 31, 32). Therefore, other factors than cortisol likely mediated the acute effect of sleep on these subsets. Candidate mediators are the hormones GH, prolactin, and aldosterone, which all were increased here by sleep, as also shown by others (11, 12, 28, 47, 49), and can affect T-cell migration (8, 30, 33, 42, 46, 48). However, additional immunomodulatory substances not measured in the present study, e.g., neuropeptides or chemokines that are released locally, might have added to this migratory effect of sleep (34). Again, experiments with higher sampling rates that allow for fine-grained correlational analyses between sleep-induced changes in hormones and T-cell subsets are needed to conclusively identify the endocrine mediators of the sleep effect.
It is an unsolved question as to where the cells are redistributed during sleep since we cannot follow their migratory route in healthy humans. TN and TCM routinely travel to secondary lymphoid organs to become activated by their cognate antigen. There are some hints from previous studies that these cells accumulate in lymph nodes during sleep (15, 53). This assumption fits with data suggesting that the sleep-associated hormones GH and aldosterone can promote the migration of T cells to lymph nodes (8, 46). An increased homing to lymphoid tissues could, indeed, be one mechanism mediating the well-known beneficial effect of sleep on vaccination responses (7, 24, 26, 52). On the other hand, the reduction in CD8+ TTE counts during sleep might reflect a stronger attachment of the cells to the vessel walls following minimal levels of epinephrine (16, 18). Although speculative, it may be of adaptive value for cells with immediate effector potential to remain in the marginal pool during sleep when the probability of antigen encounter is low, but to be mobilized into circulation during wakefulness (at night or during daytime) to enhance immune surveillance in situations with higher antigen exposure. This idea is compatible with the line of argument from the human stress literature that posits that mobilization of cells with immediate effector functions into circulation during acute stress is evolutionarily beneficial, as it facilitates subsequent egress of these cells to sites of injury or inflammation (2, 10, 14, 51). Still, future studies are needed to ultimately determine the specific destination of the cells during sleep.
Perspectives and Significance
An essential feature of T cells is their constant migration between the blood and different tissues, which allows them to detect and combat the plethora of antigens with which the body is confronted every day. Sleep acutely affected this migratory behavior within only 1.5–3 h. This effect was evident for several different T-cell subpopulations, but the implications for each subset may differ substantially. Sleep presumably promotes homing of less differentiated T-cell subsets to lymphoid tissues and in this way fosters the development of adaptive immune responses. On the other hand, sleep seems to facilitate the attachment of more differentiated T cells to vessel walls, thus diminishing immunosurveillance at a time of reduced antigen exposure. All in all, animal studies are clearly warranted here to prove these putative distributions of the different T-cell subsets. In addition, experiments investigating how sleep affects the different T-cell subsets also on a functional level, like, for instance, on their cytokine production or adhesive properties, might help to further unravel the importance of sleep for T-cell immunity. Previous studies have found an association between short habitual sleep duration and an increased susceptibility to infectious illness (13, 36, 38, 39), as well as a reduced response to vaccination (37), highlighting the health implications of good sleep for the long term. Complementing these findings, the present results demonstrate that sleep even acutely impacts basal immune functions like the distribution of several T-cell subsets. Although the sleep-induced reduction in T-cell numbers is much smaller than their circadian variations, it is comparable, for example, to the decline in TN numbers during acute immune activation following vaccination against the yellow fever virus (25), and, thus, is indeed likely to be physiologically relevant.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB654 “Plasticity and Sleep”) and from the German Federal Ministry of Education and Research to the German Center for Diabetes Research (01GI0925).
No conflicts of interest, financial or otherwise, are declared by the authors.
L.B. and T.L. analyzed data; L.B., S.D., J.B., and T.L. interpreted results of experiments; L.B. prepared figures; L.B. and T.L. drafted manuscript; L.B., S.D., J.B., and T.L. approved final version of manuscript; S.D., J.B., and T.L. conception and design of research; S.D. and T.L. performed experiments; S.D. and J.B. edited and revised manuscript.
We are grateful to Anja Otterbein, Christiane Otten, Christian Benedict, Dennis Heutling, Thomas Kriesen, and Eicke Böschen for technical assistance and to all subjects, who participated in this study.
- Copyright © 2016 the American Physiological Society