Rhythmic expression of the melatonergic biosynthetic pathway and its differential modulation in vitro by LPS and IL10 in bone marrow and spleen

Daily oscillation of the immune system follows the central biological clock outputs control such as melatonin produced by the pineal gland. Despite the literature showing that melatonin is also synthesized by macrophages and T lymphocytes, no information is available regarding the temporal profile of the melatonergic system of immune cells and organs in steady-state. Here, the expression of the enzymes arylalkylamine-N-acetyltransferase (AA-NAT), its phosphorylated form (P-AA-NAT) and acetylserotonin-O-methyltransferase (ASMT) were evaluated in phagocytes and T cells of the bone marrow (BM) and spleen. We also determined how the melatonergic system of these cells is modulated by LPS and the cytokine IL-10. The expression of the melatonergic enzymes showed daily rhythms in BM and spleen cells. Melatonin rhythm in the BM, but not in the spleen, follows P-AA-NAT daily variation. In BM cells, LPS and IL10 induced an increase in melatonin levels associated with the increased expressions of P-AA-NAT and ASMT. In spleen cells, LPS induced an increase in the expression of P-AA-NAT but not of melatonin. Conversely, IL10 induced a significant increase in melatonin production associated with increased AA-NAT/P-AA-NAT expressions. In conclusion, BM and spleen cells present different profiles of circadian production of local melatonin and responses to immune signals.


Results
Melatonin production and expression of melatonergic enzymes in the BM and the spleen. The melatonin levels in the BM and the spleen were evaluated at nine different Zeitgeber times (ZT). The content of melatonin in both organs follow a circadian profile (Supplementary Table 1), and the content in the BM was 1000 fold higher than in the spleen (Fig. 1). In order to evaluate whether the BM and the spleen could synthesize melatonin, we determined the expression of the enzymes AA-NAT, its active form (P-AA-NAT), and ASMT.
In the BM, the percentage of cells expressing AA-NAT and P-AA-NAT had complementary profiles and presented circadian rhythms, with maximal and minimal at ZT03, respectively. The mean fluorescence intensity (MFI) of AA-NAT and P-AA-NAT had non-circadian rhythms, but the MFI of P-AA-NAT peaked at ZT18. Thus, the active form of the enzyme that converts serotonin in N-acetylserotonin is available at nighttime. ASMT positive cells showed rhythmic expressions (Fig. 2, Supplementary Table 2). In the context of the BM, it is also relevant to mention that around 40-60% of the cells express P-AA-NAT and/or ASMT.
In the spleen, less than 20% of the cells expressed the enzymes of synthesis of melatonin. The percentage of cells expressing AA-NAT and P-AA-NAT, as well as the MFI of AA-NAT followed a circadian rhythm, with higher expression at daytime. However, neither the MFI of P-AA-NAT nor ASMT presented a rhythmic variation (Fig. 2, Supplementary Table 2), strongly suggesting that it does not contributes to daily melatonin rhythm in the spleen. characterization of monocytic, lymphocytic and other lineages in the BM and spleen. As we know, several studies have already shown that monocytes and T lymphocytic cells can produce melatonin, and it is possible that these cells are contributing to the levels of melatonin observed in the organs. Therefore, before looking at its expression profile of melatonin enzymes, we wanted to characterize the percentage of monocytes/ macrophages/neutrophils (CD11b + ), T lymphocytes (CD3 + ) and other cells (CD11b − /CD3 − ) in the BM and spleen, since this distribution is specific for each organ.
In the BM most of the cells are CD11b − /CD3 − (85-95%) and their percentages followed a circadian rhythm. On the other hand, CD11b + (5-10%) and CD3 + (0.5-1%) did not follow rhythmic profiles (Fig. 3, Supplementary Table 1). In the spleen, the percentage of CD11b − /CD3 − reached 50-65%, while CD3 + mount up to 30-45%, and CD11b + up to 3-10%. All three categories of cells follow a circadian rhythm (Fig. 3, Supplementary Table 1). Thus, the proportion of CD11b − /CD3 − cells is much higher in the BM than in the spleen, while the proportion of CD3 + is higher in the spleen when compared to the BM, and the daily variation of the CD3 + cells was more prominent in the spleen than in the BM.
Biosynthetic pathway in the BM and the spleen specific cells. The expression of AA-NAT, P-AA-NAT and ASMT was evaluated according to the percentage of cells expressing the enzyme in each cell category as well as the MFI in the positive cells, which reflects the amount of enzyme expressed. Considering that the synthesis of melatonin is directly dependent on the presence of P-AA-NAT and ASMT, it is important to evaluate whether the rhythm of these enzymes is in phase or out of phase.
In the BM, 70% of cells expressing AA-NAT are CD11b + and CD3 + , while only 3% are CD3 − /CD11b − cells. Fourier and Cosinor analyses showed that neither the percentage of cells nor the MFI expressing AA-NAT presented a circadian rhythm. Otherwise, the active enzyme P-AA-NAT and ASMT were highly expressed (at least 40%) in the three categories of cells, with a circadian rhythm observed only in the percentage of CD11b − / CD3 − and CD3 + cells expressing P-AA-NAT and ASMT, respectively. In the three cell types analyzed, the MFI of P-AA-NAT and ASMT were higher compared to AA-NAT. The MFI of P-AA-NAT followed a circadian rhythm in the CD11b + cells and presented a rhythmic profile in the CD11b − /CD3 − cells (Fig. 4, Supplementary Table 2).
A different profile was observed for the spleen cells. In spite of only 1% of CD3 + cells expressed AA-NAT, 20-40% of these cells expressed P-AA-NAT and/or ASMT, strongly suggesting that the cells were instrumented to synthesize melatonin. Regarding the monocytic lineage (CD11b + ), around 30% of the cells expressed AA-NAT and 40-80% expressed P-AA-NAT and/or ASMT, while CD11b − /CD3 − cells almost did not express the melatonergic synthetic enzymes (Fig. 5). The expression of P-AA-NAT in the three cell types evaluated and the MFI of AA-NAT in the CD11b + cells, presented circadian rhythms (Supplementary Table 2). expressed in the cells are the moments where we expected to find higher melatonin levels. Therefore, we created predictive indexes of melatonin synthesis by summing either the MFIs or the percentage of P-AA-NAT and ASMT in each ZTs for total cells and each cell subtype. We then correlated such values with the levels of melatonin detected in the tissues.
In the BM, several of the correlations between melatonin and the MFIs of the enzymes (separated or summed) for total cells and the different cellular populations were positive (Fig. 6, Table 1). Among the enzymes, the correlation of P-AA-NAT expression and melatonin was consistently significant in total cells and in all cell populations. ASMT and AA-NAT expressions were also significantly correlated with melatonin in CD11b + and CD11 − / CD3 − cells, respectively. Accordingly, the variations observed throughout the day in the levels of melatonin and in MFI index display a similar pattern ( Supplementary Fig. 1). Finally, the correlations between melatonin and the frequency of cells expressing the enzymes (separated or summed) were not significant in any case (Fig. 6, Table 1). In the spleen, only one significant positive correlation was seen between the local melatonin and the percentage of cells expressing ASMT in CD3 + cells (Fig. 6, Table 1).

Melatonergic system regulation by immunological modulators. BM and spleen cells, collected at
ZT06 (low levels of melatonin), were stimulated or not with LPS [1 μg/ml] or IL10 [3 and 100 ng/ml] as described in materials and methods.
For BM cells, both LPS and IL10 induced an increase in melatonin levels that can be easily explained by an increase in the expression of P-AA-NAT and ASMT. The effect of IL10 on the production of melatonin appears to be dose dependent (Fig. 7). In the case of spleen cells, LPS induced an increase in the expression of P-AA-NAT but, this effect was not sufficient to alter melatonin levels. On the other hand, IL10, increased melatonin in the medium, possibly due to the increase in the expression of AA-NAT and P-AA-NAT. Again, the effect of IL10 was dose dependent (Fig. 8). Here, we characterized the local expression of the melatonergic biosynthetic pathway and the local melatonin levels throughout the day in the BM and in the spleen, two important organs of the immunological system, comprising a hematopoietic tissue and a lymphoid secondary organ, respectively. Our results show that in the BM, the expression of the enzymes varies rhythmically, especially in the CD11b + and CD11b − /CD3 − cells. In the spleen, the variation of the enzymes is rhythmic only for AA-NAT/P-AA-NAT, these effects were observed in all of the cellular population evaluated. Importantly, the local melatonin levels vary rhythmically in both organs (with higher levels at the night phase).
Although the limited amount of data regarding the daily expression of the melatonergic biosynthetic pathway in extra-pineal tissues, researches have shown that with age, the mRNA expressions and the enzymatic activity of AA-NAT and ASMT changes differentially in the spleen, the spleen, the liver and the heart 24,39 , as well as, the existence of day/night or daily variations in melatonin levels 39,40 . Therefore, the melatonergic biosynthetic pathway is also being regulated in extra-pineal tissues.
In the pineal gland, melatonin production is regulated by the sympathetic nervous system (SNS), where an adrenergic stimulus is necessary to induce the AA-NAT activity. In case of the immune system, it is known that the SNS regulates different rhythmic functions of immunological cells, like specific cellular responses, activation and migration [41][42][43] . Therefore, it is possible to speculate that the expression of the melatonergic biosynthetic enzymes in immune cells could also be controlled by adrenergic stimulation. Moreover, the existence of feedback between melatonin and clock genes has also been discussed 44,45 . This hypothesis is supported by the presence of sympathetic innervation in www.nature.com/scientificreports www.nature.com/scientificreports/ BM and spleen 46 , and because adrenergic stimulation induces melatonin production in macrophage cell lines and in BM-derived dendritic cells 19 . In line with that, the noradrenaline-induced TNF peak in the BM, is pivotal for the local rhythmic profile of melatonin 25 . Moreover, considering the dual effects of corticosterona 37 and the potentiation induced by interferon-gamma on pineal noradrenaline-induced melatonin synthesis 47 , it will be interesting to evaluate the interplay of these immunoregulatory molecules on the daily production of melatonin in the BM and spleen.
In the BM, we found positive and significant correlations between the local melatonin levels and its enzymes, principally marked by the expression of P-AA-NAT in all cellular types evaluated (Fig. 6, Table 1). This data is quite interesting, since the control of the AA-NAT phosphorylation is how the melatonin production is regulated in the pineal. On the other hand, the capacity of the BM to produce melatonin was already showed 23 , and confirmed in pinealectomized animals. In this case, although the circulating melatonin levels significantly decreased, the melatonin content in the BM remained considerably high, without altering the enzymatic activity of local AA-NAT and ASMT 22 , leaving open the possibility that the BM is producing part of that melatonin. In the present work, the profile of the local melatonin levels in the BM is very similar to the profile obtained from the MFIs index (P-AA-NAT + ASMT; with positive correlations), which was significant for the three cellular populations. These data provide further evidences in favor to the idea that the BM could be rhythmically producing its own melatonin, supplementing the one that comes from the pineal gland.
In the case of the spleen, the variation of the melatonergic biosynthetic pathway shows a greater expression during the diurnal phase, at which point a melatonin peak would be expected; but the melatonin peak occurs only in the dark phase, without correlation between melatonin levels and the expression of its enzymes (Fig. 6, Table 1). Therefore, it is most likely that nocturnal melatonin in the spleen is derived from the pineal gland. Even so, one cannot exclude that the spleen is producing melatonin, since studies have shown that, in experiments realized at ZT06, AA-NAT and ASMT present enzymatic activity 24 . Nevertheless, given the low percentage of cells with potential of melatonin production, the low median intensity of fluorescence of those cells (compared with those of the BM) and the enzymatic rhythmic profiles, it is probable that during the light phase, the spleen is producing melatonin in very low concentrations.
To confirm the hypothesis that BM and spleen cells are capable of producing melatonin and that this production is being regulated differently in each organ; we treated cells from both organs in vitro with LPS, an immunological activator capable of inducing melatonin synthesis in macrophages 26 ; and IL10, an important immunomodulatory cytokine. Melatonin levels were again higher in BM cells, and although IL10 modulated www.nature.com/scientificreports www.nature.com/scientificreports/ melatonin synthesis in both spleen and BM cells, LPS only had an effect on BM cells. Interestingly, for both organs, all increases of melatonin were related to an increase in the enzyme expression. Showing that stimuli of different nature, such as LPS and IL10, have different effects on the melatonergic system in spleen cells. In the case of IL10, the effects were dose-dependent in both spleen and BM cells.
Different functions of the immune cells like cytokines production 48 , cellular responses, proliferation and migration 49,50 are controlled by melatonin in a rhythmic way. In this sense, it was recently shown that melatonin is important to synchronize the mature blood cell production and the hematopoietic stem cell repopulation in the BM 25 . Considering the spleen, limited amount of data are available about the effect of the endogen melatonin in specific functions of this organ, but it is know that melatonin affect the activation and differentiation of T cells 51 and increases the lymphocytic proliferation in different animal models 15,52,53 . Interestingly, melatonin levels and the enzymes, as well as the lymphocytes proliferation, decrease with age in the spleen 52,53 ; showing a direct correlation between melatonin and the immune response of the spleen cells. Additionally, melatonin plays a central role in surveillance against infection and inflammatory and recovery phase of acute defense responses 26,54,55 . Taking into account all the above mentioned, this information strengthens the idea that the modulation of local melatonin is tissue and organ specific and these differences are associated not only with the circadian profile of the system but also with the cell responsiveness to immune-related signals.
In conclusion, as many immune cellular functions vary rhythmically, the cellular functions that melatonin exerts are being differentially regulated in each organ and in each cell type in chronobiologic-dependent manners. Therefore, the importance of taking into account the rhythmic profiles of the immune cells in terms of different profiles of the melatonergic biosynthetic pathway expression will provide a better understanding of the physiological role of extra-pineal melatonin production in tissues and cells. In addition, the pioneering data of this work will allow to propose more refined experiments to future researches focused on the modulation of this pathway in the treatment of immunological diseases like hematopoietic tumors and uncontrolled inflammatory conditions, processes that we are increasingly seeing that the time of the day that treatments occurs is crucial to their efficiency.

Methods
Animals. Male Wistar rats (8-12 weeks old, 250-300 g), receiving water and food ad libitum, were kept at 22.0 ± 2.0 °C under a 12:12 h light/dark cycle (lights on at 06:00 h = Zeitgeber time zero or ZT0). Animals were killed by decapitation at nine different ZTs (12.25, 15, 18, 21, 23.75, 3, 6, 9, and 11.75) and the pineal gland, BM   Table 1. Cells were collected at ZT6, plated at 6.67 × 10 6 cells/mL and stimulated with LPS 1 μg/ml, IL10 3 ng/ml (IL10 3) or IL10 100 ng/ml (IL10 100) for 6 h. After stimulation, cells and supernatant were collected for measurement of the enzyme expression by flow cytometry and for melatonin quantification by ELISA, respectively. Enzyme expression: values are normalized by the enzyme expression of the control. Results are expressed as mean ± SEM, n = 3-7 animals from 2 independent experiments. Control vs LPS: data were analyzed by Student "t" test. Control vs IL10 3 and 100: data were analyzed by one-way ANOVA with Tukey's post hoc. *P < 0.05 vs control.
Statistical analysis. Data were expressed as mean ± SEM. Time series were analyzed using the Fourier and Cosinor analysis (Chronobiology Software El Temps, ©Antoni Díez-Noguera, Barcelona, CA, Spain). Comparisons were performed using One-way Analysis of Variance (ANOVA) followed by Tukey's post hoc test to see the time of maximal and/or minimal expressions and for the in vitro treatments with IL10. Student "t" test  Figure 8. Effect of LPS and IL-10 on the AA-NAT, P-AA-NAT and ASMT expression and melatonin levels in spleen cells. Cells were collected at ZT6, plated at 6.67 × 10 6 cells/mL and stimulated with LPS 1 μg/ml, IL10 3 ng/ml (IL10 3) or IL10 100 ng/ml (IL10 100) for 6 h. After stimulation, cells and supernatants were collected for measurement of the enzyme expression by flow cytometry and for melatonin quantification by ELISA, respectively. Enzyme expression: values are normalized by the enzyme expression of the control. Results are expressed as mean ± SEM, n = 3-7 animals from 2 independent experiments. Control vs LPS: data were analyzed by Student "t" test. Control vs IL10 3 and 100: data were analyzed by one-way ANOVA with Tukey's post hoc. *P < 0.05 vs control; # P < 0.05 vs cells treated with IL10 3 ng/ml.