Exosomal miR-92a Concentration in the Serum of Shift Workers

: Shift work is associated with alterations in the human biological clock and metabolism. Serum exosomal miR-92a concentration was inversely correlated with brown adipose tissue activity playing a pivotal role in energy balance. In this study, miR-92a was measured in serum exosomes of 30 workers engaged in shift and daytime work. No significant metabolic alterations were shown between daytime and shift workers while a difference in serum exosomal miR-92a levels was found between the two groups. The lower levels of miR-92a in shift workers were suggestive of a higher brown adipose tissue activity compared with daytime workers.However, the possibility that other physiological and pathological processes may influence miR-92a cannot be ruled out. Our results suggest further investigations on brown adipose tissue activity and on miR-92a regulatory mechanisms, such as those related to the estrogen pathway, in shift workers.


Introduction
Working on a shift work schedule implies that some daily habits, such as sleeping or eating, are not conducted with regular occurrence [1] resulting in a possible desynchronization between external and biological times [2]. The perturbation of the biological clock leads to alterations of several metabolic processes including reduced glucose tolerance, dislipydemia, weight gain [2][3][4], and the impairment of some molecules related to fat absorption and metabolism [2,5]. Epidemiological studies associated shift work with an increased incidence of metabolic syndrome [6,7] that, in turn, is an important cardiovascular risk factor [8,9]. Since alterations of metabolic parameters and weight gain in shift workers do not seem to be associated with low physical activity or increased food intake [5,10], it is likely that there are other factors that may influence energy balance.
In the last few decades, brown adipose tissue (BAT) has regained scientific and public interest for its role in metabolism and weight gain [11][12][13][14][15][16]. BAT is a rich vascularized tissue with a high number of mitochondria. It catabolizes lipids to generate heat playing a major role in thermogenesis and energy expenditure [17][18][19][20]. Similar to lipids, BAT also metabolizes carbohydrates, contributing to the homeostasis of these two groups of macromolecules resulting in a beneficial effect on overall metabolism [21][22][23]. Obese patients have less active BAT [24]. In fact, an inverse relationship between active BAT and body mass index (BMI) has been discovered [25].
The most used method for quantitatively measuring BAT volume and activity is the 18 fluor-deoxy-glucose positron emission tomography coupled with computed tomography

Subjetct Enrollment
The concentration of exosomal miR-92a was determined in the serum of 30 workers enrolled among the nursing staff of Area Vasta 2 (Ancona, Italy). All nurses were female, fifteen subjects worked a rotating shift schedule that included night shifts (SW nurses), whereas the other 15 workers were involved exclusively in daytime shifts (DT nurses).
Selection criteria were the following: fertile age; no current use ofmedications; a negative history of psychiatric disorders, insomnia, cancer or chronic systemic diseases, such as degenerative, cardiovascular or autoimmune disorders; absence of artificial light when sleeping at home; no occupational tasks involving exposure to ionizing radiation or antiblastic drugs. SW nurses had to be enrolled in the current shift schedule, including more than 60 night shifts/year, for more than 2 years without schedule breaks in the previous 6 months. DT nurses had to go to bed habitually around 23:00 and to wake up around 06:00 without sleepless nights for more than 3 weeks prior to the study.
Workers provided their consent after receiving information about the purpose and procedures of the study, which was conducted according to the Helsinki Statement of Ethical Standards. Being performed as part of required, routine health surveillance, the study required no formal approval by the local ethics committee. The SW nurses worked a fast clockwise shift schedule: day 1, 07:00-14:00; day 2, 14:00-22:00; day 3, 22:00-07:00; rest for 48 h; resumption of the cycle. The DT nurses worked from 07:00 to 14:00 for 6 days/week. To exclude the acute alterations related to night shift, fasting venous blood samples were taken from both SW and DT nurses at the beginning of the morning shift after a regular night's sleep on a day off. Blood samples were processedimmediately after collection. An aliquot of serum was stored at −80 °C until miR-92a analysis.Blood was analyzed the same day for glycemia, total and HDL cholesterol, and triglycerides.

Clinical Parameters Collection
Socio-demographic characteristics were collected as part of the medical history. The chronotype was investigated by the "Morningness-Eveningness Questionnaire" (MEQ) [40], a questionnaire with 19 items and a total score ranging from 16 to 86, widely used in adults and workers [41][42][43][44]. Sleep quality was investigated using the Pittsburgh Sleep Quality Index (PSQI) [45], while daytime sleepiness was assessed by the Epworth Sleepiness Scale (ESS) [46,47]. Measurement of anthropometric parameters was performed during the medical examination. The fat mass was calculated using Durnin and Womersley formula [48]. A food diary was kept by workers for 5 consecutive days and daily average values of total energy intake and fat, carbohydrate, and protein intake were calculated [49]. Daily average values of total energy expenditure and energy expenditure due to motor activity were investigated using a metabolic monitor (SenseWear Armband ® ; BodyMedia Inc., Pittsburgh, PA, USA) [50], secured on each participant's right arm for 5 days. There were no cases of interrupted or invalid recordings.

Determination of Serum Exosomal miR-92a
Serum exosomal miR-92a was measured after a few months in storage at −80 °C. Samples of SW and DT nurses were coded in a mixed order and analyzed by blinded personnel. After all the samples completed the miRNAs extraction, they underwent the reverse transcription and finally the relative qRT-PCR. The determination of serum exosomal miR-92a was completed in about two months.

MiRNAs Extraction
Exosomes were extracted from 1 mL of serum using a "Total Exosome Isolation (from serum)" Kit (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's instructions. Exosomal proteins were then quantified through the Bradford method and 20 µg of exosomal protein were diluted in 250 µL of PBS for the following RNA extraction process. We used the same quantity of exosomal protein to prevent differences in miR-92a related to the total amount of exosomes. Before RNA isolation, and to avoid further technical errors during the subsequent quantification of miRNAs by qRT-PCR, 60 fmol of a synthetic spike-in small RNA from Caenorhabditis elegans (cel-miR-39) were added to the samples as external normalizer [51]. Total RNA was then extracted by adding 750 µL of Tri-Reagent BD (Sigma-Aldrich, Saint Louis, MO, USA) and 200 µL of chloroform. The Phase Lock Gel (Eppendorf, Hamburg, Germany) was used to enhance the phase separation. Finally, miRNAs were isolated using the Purelink miRNA Isolation Kit (Invitrogen; Thermo Fisher Scientific, Inc.) and eluted in a final volume of 40 µL of RNase-free water, centrifuging twice 20 µL to improve the yield. Samples were stored at −80 °C until reverse transcription.
Data was normalized for biological variability, using the small nuclear RNA U6, and for possible technical variations during miRNAs extraction, using the cel-miR-39 as the exogenous normalizer. After the reaction was run, Ct values were determined and the relative quantification of miR-92a was calculated using the comparative 2 −∆Ct method [52]. The expression of miR-92a was calculated by double normalization: Both controls (CtU6 and Ctcel-miR-39) were subtracted from the Ct value of the sample of interest [51][52][53].

Statistical Analysis
Kolmogorov-Smirnov test was used to assess the normality of distribution. Since miR-92a values were not normally distributed, a non-parametric statistical approach was chosen. Continuous variables were expressed as median (25th-75th percentile) and differences between DT and SW nurses were analyzed by Mann-Whitney U test. Dichotomous or categorical variables were expressed as percentages and differences were analyzed by Chi-square test. All the tests were two-tailed considering a p<0.05 as statistically significant. Data analysis was performed with the Statistical Package for Social Sciences version 19.0 (SPSS Inc., Chicago, IL, USA).

Results
All the workers who participated in the study were females. No significant differences were found in age, job seniority, numbers of smokers, or of alcohol drinkers between DT and SW nurses ( Table 1). None of the SW nurses had changed the type of schedule from the beginning of the work as evidenced by the same values of job seniority and shift work seniority. SW nurses worked a median of 6.0 nights per month. No significant difference was found in chronotype (MEQ score), sleep quality (PSQI score), and daytime sleepiness (ESS score) between DT and SW nurses. There were no significant differences in metabolic parameters between DT and SW nurses ( Table 2). DT nurses showed a higher median value of total cholesterol, but the difference was not significant. Other blood parameters, systolic/diastolic pressure, BMI, waist circumference, skin thickness, and body fat were also similar in the two groups.  Data about total energy intake, diet composition, and total energy expenditure were similar in the two groups ( Table 3). The statistical analysis did not show significant differences. Table 3. Energy intake, diet composition, and total energy expenditure of daytime (DT) and shift-working (SW) nurses. Results are expressed as median (25th-75th percentile). Levels of miR-92a in serum exosomes are reported in Figure 1. We found significantly higher levels of miR-92a in DT nurses than in SW nurses [median 14.2 versus 4.3 (2 −ΔCt /10 3 ), p=0.002]. Since miR-92a in serum exosomes inversely correlate with BAT activity, results were indicative of a lower BAT activity in DT nurses compared with SW nurses.

Discussion
Shift work may disrupt sleep and meal rhythms. The perturbation of biological rhythms can alter metabolic homeostasis that may lead to weight gain and/or metabolic disturbances [54][55][56][57]. Shift work was associated with higher waist circumference [58][59][60], and a higher risk of metabolic syndrome and cardiovascular events [7,[61][62][63]. In our sample, SW are similar to DT nurses concerning the waist circumference and the other metabolic parameters analyzed. The lack of differences is likely due to the small size effect that shift work has on metabolic parameters.A larger sampling would help this to be more visible. Moreover, our sample constituted of young nurses (median age of 39.0 years) with a median job seniority of 13.0 years. These parameters likely differ from the metabolic alterations shown in older shift workers [64]. Several authors have obtained increased BMI values in SW compared to DT workers [60,65,66], which, however, were not found to be associated to differences in diet style [65,67], nor in physical activity between the two groups of workers [10,68,69]. Other factors,such as a low level of active BAT, may also contribute to weight gain in SW.
This study measured the levels of miR-92a that a study of Chen and colleagues inversely correlated with BAT activity [39]. Contrary to the hypothesis that SW may have a low level of active BAT, we found lower levels of miR-92a in SW nurses compared to DT nurses, indicative of higher BAT activity [39]. Several argumentations may be possible: (1) The increased BAT activity detected in SW nurses may be a physiologic mechanism to counteract metabolic alterations induced by circadian disruption linked to shift work. This counteracting mechanism may be impaired after a long duration of shift work resulting in weight gain and metabolic syndrome. (2) In a study on mice, prolonged daily light exposure (similar to the experiences of shift workers during a night shift) decreases BAT activity [70]. However, mice differ from humans as they usually sleep during the day and are active and eat during the night, resulting in a possible opposite effect of light on BAT. (3) Shiftwork can induce a sympathodominant state [71]. Since the sympathetic nervous system is a stimulator of BAT [72], the higher BAT activity in SW nurses may be caused by a sympathetic stimulation. (4) Since BAT activity has a circadian rhythm [73,74], our finding may be related to a phase misalignment of a circadian cycle of BAT activity in SW compared to DT nurses. (5) Since the activity of BAT produces heat, its increased activity may be the cause of the higher peripheral skin temperature previously demonstrated in a similar sample of SW nurses [75]. (6) The lower levels of miR-92a found in the serum of shift workers may not only be related to BAT activity, but also to other biological processes [76,77].
This last hypothesis must be carefully considered. Further investigations are necessary due to the inverse correlation between serum exosomal miR-92a and BAT activity, which is based only on the study of Chen and colleagues [39]. Moreover, the miR-92a is involved in the regulation of several genes. MiRTargetLink [78], using information from miRTarBase [79,80], found 1419 genes that are targeted by miR-92a: 30 interactions with strong support, 1301 interactions with weak support, and 88 predicted interactions (Table S1). MiR-92a regulates several processes including cell proliferation and the development of cancer [77,81]. Night shift work has been suggested to be associated with the onset of cancer, specifically breast cancer [82][83][84][85]. Interestingly, in breast cancer the miR-92a is deregulated [86][87][88][89][90][91][92]. The miR-92a targets and downregulates estrogen receptor beta (ERβ), and is controlled by estrogen [93]. Since higher levels of estrogen were found in shift workers [94][95][96][97], miR-92a may contribute to the deregulation of estrogen levels induced by shift work. Hence, it would be interesting to perform additional investigations regarding the possible role of miR-92a related to a higher risk of breast cancer in shift workers.

Conclusions
The significant lower levels of serum exosomal miR-92a suggest a higher BAT activity in SW nurses compared to DT nurses. Additional studies using gold standard tests for BAT activity in shift workers are needed. The possibility that the lower levels of miR-92a found in the serum of shift workers may not only be related to BAT activity, but also to other factors cannot be ruled out. Considering that miR-92a takes part in several biological processes, including breast cancer, our results suggest further investigations on miR-92a in shift workers.