Melatonin deficiency decreases brown adipose tissue acute thermogenic capacity of in rats measured by ¹⁸F-FDG PET

Melatonin is a pineal hormone, produced at night, which has a critical role in the synchronization of circadian rhythms, with known metabolic effects in many animal species [1, 2]. One of its metabolic effects in rodents is a reduction in body weight with a minimal decrease in food intake, suggesting an action on energy expenditure, which possibly relates to activation of brown adipose tissue (BAT) [1‐5]. Indeed, many experimental models have shown the role of melatonin on BAT recruitment [3, 6].

BAT is a thermogenic tissue, whose primary function is thermoregulation via non-shivering thermogenesis [7]. Thermogenesis occurs due to a unique and specific enzyme called uncoupling protein-1 (UCP-1), which uncouples ATP energy production in the mitochondria, generating heat [8]. Brown adipose tissue is activated by sympathetic noradrenergic stimuli, and cold is the most important physiological stimulus known [7, 8].

Recently, BAT research has increased after positron emission tomography with 2-deoxy-2-[fluorine-18] fluoro-d-glucose (¹⁸F-FDG PET) has shown that humans still possess active BAT, especially after cold exposure [9, 10]. In this context, a lack of BAT activation could have a potential role in weight gain in humans. Therefore, targeting BAT activation could be promising for the treatment of obesity and/or type 2 diabetes [11].

Many compounds have been studied to determine the ability of BAT recruitment and activation, including melatonin [12, 13]. However, in the literature, several melatonin studies conducted in animals measured BAT mass, in order to analyze BAT recruitment and activation, and some of them focused on BAT thermogenesis, measured by calorimetry or cytochrome activity [3], yet none of them used 18F-FDG PET, which is the gold standard for BAT measurement in humans. More recently, some studies have evaluated a possible increase in BAT and in thermoregulatory responses, using infrared thermography [5, 14, 15].

It is worth highlighting that an increase in BAT mass (recruitment) does not necessarily lead to an increased thermogenic response (activation), if there is no physiological need for heat production [7, 8, 16]. In this regard, an increase in BAT mass or UCP-1 expression would not necessarily mean an increase in BAT uptake in ¹⁸F-FDG PET [16, 17]. Further concerning this point, different protocols of cold exposure or other adrenergic stimuli could lead to different patterns of BAT uptake by ¹⁸F-FDG PET that do not necessarily reflect the total BAT mass of an individual [16].

As ¹⁸F-FDG PET is the gold-standard method for the detection of BAT in humans [18], it would be relevant to use this methodology in experimental studies to understand the role of melatonin deficiency and its supplementation on BAT thermogenic responses. This would be important, not only for melatonin, but also for any other compound that could potentially increase BAT mass (since an increase in BAT mass in humans, by any compound, if not increase thermogenesis as well, could only be detected by tissue biopsy and would probably have minimal therapeutic implications).

In the present study, we ought to determine if melatonin deficiency in experimental models of pinealectomized Wistar rats decreases BAT activation, evaluated by Positron emission tomography-computed tomography (PET-CT), at room temperature and after cold exposure, and if melatonin replacement could revert these changes.

The cold challenge experiment was useful and increased BAT Max SUV in all three animal models with statistical significance (Fig. 1, Additional file 1). In the PET-CT test performed at room temperature, there were no significant differences between the groups. Surprisingly, though, there was an absolute increase in Max SUV in the P group (Fig. 2a and Table 1).

On the contrary, when the cold exposure protocol (which has a higher sensitivity) was analyzed, the pattern was reversed, and the P group had the lowest mean values. However, once again, no statistical significance was observed (Fig. 2b and Table 1).

When the ratio of increase between the after cold challenge was compared with room temperature (what we called “acute thermogenic capacity”—ATC), we found that the P group had a statistically significant reduction in this parameter (Fig. 3 and Table 1), compared with both of the other groups. This suggests, as expected, that melatonin deficiency could impair BAT thermogenic responses (ANOVA P = 0.03), which are restored with melatonin replacement therapy.

We analyzed UCP1 mRNA in a different group of animals that received the exact same protocol.

UCP1 mRNA was analyzed after sacrifice in a similar group of animals, in which the protocol conditions of C, P, and PM were identical to the animals that had the PET-CT performed. ANOVA testing revealed a P < 0.001, confirming differences between the three groups, as expected; the P group had a minimal expression of UCP1, compared with C and PM (Fig. 4). The results confirmed that melatonin deficiency reduces UCP-1 gene expression and that exogenous melatonin supplementation can reestablish its expression.

We aimed to test if melatonin deficiency (due to surgical pinealectomy) decreased BAT activation, and if melatonin replacement could revert this pattern to normal in Wistar rats, using the gold standard for BAT detection in humans—¹⁸F-FDG PET.

This is the first investigation, to our knowledge, to study BAT activation by ¹⁸F-FDG PET scan before and after melatonin treatment in animals. Despite data of increased BAT mass, UCP1 expression, and thermogenesis after melatonin replacement in many animal models and species already being demonstrated [3, 6, 14, 15], we found it relevant to investigate if ¹⁸F-FDG PET was able to detect these differences, facilitating future research in animals and humans, in which indirect measures of BAT are more feasible to be detected than through direct sample tissues. We believe that testing potential BAT activating compounds by performing an ¹⁸F-FDG PET in animals could help the translation of animal to human data. If animals increase their BAT SUV response to a specific compound, future studies can be done in humans to evaluate the same compound using the same method of detection.

In both isolated models (room temperature and after cold challenge), there were no differences between the groups, although some trends could be observed, including an unexpected higher maximal SUV in pinealectomized animals at room temperature and, on the contrary, a reduced maximal SUV in the P group after cold exposure. Since no statistical significance was achieved in any of the isolated methods, we prefer not to over interpret the reasons of this discrepancy. We should bear in mind that performing an ¹⁸F-FDG PET at room temperature is less sensitive to detecting differences; since the animals are chronically adapted to that temperature and have a small biological need to activate a recruited BAT. Perhaps in 23 °C room temperature chronic exposure, which is not thermoneutral for rats, but clearly not as challenging as an acute 4 °C exposure, other mechanisms of temperature regulation could be enough to maintain the normal thermoregulatory responses of the animal [3]. Therefore, the absence of differences between the groups in this specific model was not unexpected, at first. There has been a lot of discussion about what the ideal temperature is to perform BAT and other metabolic studies in rodents, and this remains an open question [24, 25]. The initial goal of performing the test at room temperature was exactly to observe if different patterns occurred when compared to the cold challenge and to calculate the ATC. A larger n could probably bring us more answers in future studies. As very few studies have analyzed BAT by PET in rats, our sample size calculation was based on a pilot study, and due to interindividual variances in Maximal SUV, we believe a larger sample size could have led us to results that were more conclusive.

However, comparing the actual ratio of BAT Maximal SUV increase, the acute thermogenic capacity (ATC), was also a primary goal, since it reveals if BAT is recruited to rapidly exert its effects on thermogenesis. In this context, the melatonin deficient group had a clearly reduced ATC, as expected. Since the PM and C groups had similar responses, this further confirms the hypothesis that melatonin proficiency is critical for thermoregulation in settings of acute cold challenging.

Indeed, the findings presented here encouraged our group to conduct an already published human study that demonstrated an increase in BAT volume and activity after melatonin replacement in a group of pinealectomized individuals [26]. Interestingly, in this study, some individuals have a reasonably high baseline BAT activity before melatonin replacement; suggesting that even if melatonin is important for thermoregulation, other mechanisms still exist.

The UCP-1 RNA data in the present study was performed as an additional tool to analyze and interpret our image data. Importantly, other published manuscripts, some of them by our own group, have already shown a decrease in UCP-1 RNA and UCP-1 protein expression after pinealectomy and a reversal of this reduction by melatonin replacement in slightly different experimental models [5, 6, 14, 15]. Therefore, regardless that our UCP-1 results are not original, they reinforce and validate the previous results and help in the interpretation of the image data [5, 6, 14, 15].

The observed pattern of UCP-1 expression in our model is in line with ¹⁸F-FDG PET results. The groups were statistically different; the P group has the lowest UCP-1 expression, compared with the other two groups, which in some ways could explain the lower ATC.

However, as already pointed out, an increase in UCP-1 RNA does not necessarily mean more heat production, as this recruited tissue can be inactive, if there is no need to increase thermogenesis [3, 16, 17]. Even differences in RNA and protein expression could arise, as post-transcriptional factors may influence protein synthesis. This could explain why the pinealectomized group showed a normal BAT response in ¹⁸F-FDG PET at room temperature, even with having lower UCP-1 expression. We did not expect differences in UCP1 expression between the C and PM groups, since our intent was to reproduce, immediately after pinealectomy, in the PM group, the physiological production of melatonin. Nevertheless, clear differences were observed, so the main hypothesis is that the replacement protocol did not exactly mimic melatonin production. As the animals drink water at specific times of the night, they would probably have several small melatonin peaks, instead of the physiological curve of night melatonin production. In addition, since the melatonin solution was only available from the start of the dark period, the duration of its plasmatic profile might reflect the winter type of melatonin secretion in some animals, changing the physiological pattern of BAT recruitment. These different kinetic curves could explain the differences observed in Fig. 4.

Our main focus in this particular study was to evaluate BAT responses in ¹⁸F-FDG PET. As many other different studies have demonstrated the physiological role of melatonin in body weight regulation, food intake, energy expenditure, metabolic risk factors, and many other parameters, we did not include these data as they would be redundant. Our group has previously shown that pinealectomy leads to a metabolic syndrome phenotype in rats and melatonin replacement reverts it [1]. Melatonin has been shown to decrease body weight with a minimal decrease in food intake, suggesting an effect in energy expenditure, proposed to be mediated by BAT [2, 3]. For a more comprehensive review of metabolic consequences of melatonin deficiency and the physiological role of melatonin in several animal models, see references [1, 2].

The research presented here aggregate data on the role of melatonin on thermoregulation, but the model has several limitations. Many questions still persist. Is there a role of melatonin in maintaining thermoneutrality in animals at room temperature, or is it only important if the temperature acutely decreases? Would the same results appear if the cold challenge was less intense and had a longer duration, probably a more physiological challenge? Would ATC be a reliable way to detect recruited BAT in vivo, as our study suggests? We believe so, but to detect clear effects of different compounds, we need to compare differences between the room temperature and cold challenge, since many adaptive mechanisms should exist, as thermoneutrality is vital for survival. Performing exams in only one of these conditions could lead to misinterpretations of the role of BAT and possible recruiters on thermogenesis and there is clear evidence that small differences in laboratory temperatures could lead to very different results in different animal models [24, 25, 27‐29]. For example, if the tests were performed only after the cold challenge, we could conclude that no differences were seen between the groups, despite large differences in UCP-1mRNA expression. This could have led to an interpretation that UCP-1 is expressed but not necessarily translated, or that FDG-PET is not a reliable way to detect differences in BAT activation in animals.

BAT physiology is very complex, and potential compounds aiming to increase BAT can act in recruitment, activation, or both [3, 7, 11, 12]. Experimental models can help us distinguish between the action of potential BAT recruiters, such as melatonin, and the detection of BAT in vivo, by imaging techniques. Different models have the potential to investigate the role of several recruiters in different laboratory conditions, paving the way for human studies with potential compounds capable of activating or recruiting BAT, besides melatonin [11, 30]. Our finding that melatonin seems critical for acute BAT activation, induced by cold and measured by ¹⁸F-FDG PET, is novel. It can lead to future imaging studies in both animals and humans that will help understand the physiological role of melatonin, concerning BAT, and if melatonin could be a potential target for increasing melatonin recruitment and activation in humans, with potential therapeutic use in metabolic diseases [26]. In the same way, the finding of reduced melatonin production, leading to reduced BAT thermogenic responses, can help the understanding of increased light-at-night exposure as a potential risk factor for obesity and metabolic diseases, as already suggested [31, 32].

This is the first study, to our knowledge, evaluating BAT responses to melatonin by 18F-FDG PET, which is an important essential for future melatonin and circadian studies, in animals and humans, to evaluate BAT responses through non-invasive techniques.

We demonstrated that pinealectomy impairs BAT acute thermogenic capacity, measured by ¹⁸F-FDG PET in Wistar rats, after an acute cold challenge, and that melatonin replacement reverses this impairment, in line with the results obtained by UCP1 RNA expression. However, when the results at room temperature and after the cold challenge were analyzed separately, no significant differences were seen, highlighting the complexity of thermoregulation and the possible limitations of ¹⁸F-FDG PET to detect differences in BAT, depending of the temperature conditions.

These results corroborate many experimental models that suggest that melatonin has a critical role in BAT mass and activity. Studying different compounds that potentially increase BAT by an imaging technique considered to be the gold standard in humans could facilitate future studies in our species.