Forty 6-week-old male C57BL/6J mice, weighting 19~21 g at the beginning of the experiment, were purchased from the laboratory animal center of Chinese PLA general hospital. All mice had ad libitum access to high-fat diet (HFD, 60% of energy from lipid, D12492) to simulate unhealthy diet habit. Two mice were housed in each transparent polypropylene cage (34 cm× 15 cm× 16 cm). To prevent the interferences of other ambient cues (e.g., natural light and noise), every cage was placed in a customized, lightproof, soundproof, and ventilated cabinet (50 cm× 35 cm× 40 cm) with integrated, programmable LED lighting in a standard housing environment [(21± 1) ℃, 50%~60% humidity]. All irradiances of light in this study were 120 μ W· cm^-2 at the level of 4 cm above the cage floor. Irradiance (μ W· cm^-2) was measured with a spectral power densitometer (SPD-370), which also provided light intensity values in units of lux and photons cm^-2· s^-1. See Table 1 for light measurements.

Table 1

Table 1

Table 1 Light measurement of LEDs used in the experiment Light measurements WL RL BL Irradiance/(μ W· cm-2) 120 120 120 Illuminance/lux 380 231 93 Photon flux/(photon cm-2· s-1) 3.41× 1014 3.83× 1014 2.91× 1014 Peak wavelength/nm 593 629 471 FWHM/nm none 13 20 Note: LEDs: Light-emitting diodes; WL: White light; RL: Red light; BL: Blue light; FWHM: Full width at half maximum

Table 1 Light measurement of LEDs used in the experiment

During the accommodation, mice were maintained in a normal 12 h:12 h light-dark cycles for 2 weeks. Then the control mice were still housed in a the normal light-dark cycle with white light (control WL), while the other mice were subjected to “ jetlag” light cycles based on a weekly repeated schedule with an 8-h phase advance in light offset on Monday and an 8-h phase delay in light offset on Thursday to simulate shift work [Fig.1(a)]. Under “ jetlag” light cycles, mice were randomized into three subgroups and exposed to a different colors of LEDs (3535ALN), including white light (jetlag WL, 4 000 K), red light (jetlag RL, peak at 629 nm), and blue light (jetlag BL, peak at 471 nm). Spectral power distributions are shown in Fig.1(b). During the experiment, body weight, food and water intake were measured at least once a week. At 13 weeks of the experiment, all groups’ lights were recovered to the normal WL for 2 weeks to resynchronize the circadian systems. During the resynchronization, glucose (GTT) and insulin tolerance tests (ITT) were performed, and then mice were fasted overnight and sacrificed for the collection of serum, adipose, pancreas, and liver samples. The Laboratory animal center approved all animal procedures of Chinese PLA general hospital (Permit Number: 2020-X16-28).

Fig.1

	Figure Option ViewDownloadNew Window
	Fig.1 Experimental schedule and spectral power distributions of light source (a): Zebra bar represents the lighting schedule in which white and black segments indicate “ lights on” and darkness, respectively. The length of the segment indicates the duration; (b): Spectral power distributions of white, red, and blue LEDs

Mice were fasted for 14 or 6 h for intraperitoneal GTT (ipGTT) or ITT, respectively. The blood glucose was determined with a glucometer (Roche) before and at 15, 30, 60, and 120 min after intraperitoneal injection of D-glucose (1 g· kg^-1 body weight) for ipGTT or after intraperitoneal injection of human insulin (1.0 U· kg^-1 body weight) for ITT. The area under the curve (AUC) was calculated by GraphPad 9.0. Insulin sensitivity was assessed by the homeostatic model assessment of insulin resistance (HOMA-IR). HOMA-IR=fasting glucose (mg· dL^-1)× fasting insulin (mU· L^-1)/450.

Animals were anesthetized, and blood was collected through extirpation of the eyeball at the end of the 15-week experiment. Serum glucose, triglyceride, total cholesterol, alanine aminotransferase (ALT), and aspartate aminotransferase (AST) were measured with an automatic biochemistry analyzer (Chemray 800). Insulin was measured by Mouse Ultrasensitive Insulin ELISA Kit (ALPCO), and corticosterone was measured by Mouse and Rat Corticosterone ELISA (ALPCO) by the manufacturer’ s protocol. Hepatic triglyceride concentration was assayed using the Triglyceride Assay Kit (Nanjing Jiancheng) and normalized to liver protein content.

All animals were anesthetized and euthanized at 15 weeks of the experiment. Formalin-fixed liver tissues were stained with hematoxylin and eosin (H& E). Epididymal white adipose tissue (eWAT) was precisely dissected and weighed. Sections (3 μ m) of paraffin-embedded eWAT were stained with H& E. The adipocyte area was quantified by averaging counts from three random fields per section. Paraffin-embedded pancreatic sections were stained with H& E and stained for insulin (anti-insulin rabbit pAb, Servicebio). The mass of β -cells was calculated by the insulin-positive area ratio to the pancreatic cross-sectional area multiplied by the wet tissue weight. Image quantification of adipocyte area and β -cells mass was performed using ImageJ software.

Statistics were performed by using GraphPad Prism software version 9.0. Statistical analyses were performed using two-way repeated measured ANOVA with Tukey’ s post hoc tests for time series. For two independent groups, statistical differences were evaluated by Student’ s two-tailed t-test or nonparametric Mann-Whitney U-test. All data are presented as mean ± SE. Statistical significance was considered at p< 0.05.

Under WL, chronically alternating light-dark cycles significantly promoted body weight gain from 9 weeks of the experiment compared with the control WL group (103.3%± 3.9% vs. 86.5± 6.1%, p< 0.05)[Fig.2(a)]. Remarkably, BL prevented jetlag mice from body weight gain as no difference was observed in body weight gain between jetlag BL mice and control WL mice throughout the experiment [Fig.2(a)]. Moreover, jetlag BL mice gain much less body weight than jetlag WL mice (113.3%± 10.3% vs. 138.0%± 3.9%, p< 0.01)[Fig.2(b)]. At 15 weeks of the experiment, we evaluated the visceral fat mass. Although there was no significant difference between the groups in left eWAT mass [Fig.2(c)], jetlag RL mice gained more fat accumulation in right eWAT compared with control WL mice [(1.12± 0.08) g vs. (0.91± 0.04) g, p< 0.05] [Fig.2(d)].

Fig.2

Figure Option ViewDownloadNew Window

Fig.2 Effects of chronically alternating light-dark cycles on energy homeostasis are wavelength-dependent (a): Relative body weight gain during the experiment; (b): Relative body weight gain at the end of the 15-week experiment; (c): Left eWAT mass; (d): Right eWAT mass; (e): Cumulative food consumption; (f): Percentage of nocturnal food intake; (g): Cumulative water consumption; (h): Percentage of nocturnal water intake. Results are means± SE (n=8 per group); * p< 0.05, * * p< 0.01, * * * p< 0.001 vs. control WL group; #p< 0.05, ##p< 0.01, ###p< 0.001 vs. jetlag groups

Besides, we also analyzed the percentages of nocturnal food and water intake in the first three days after shifting the light-dark cycle to evaluate how the eating and drinking patterns were affected by different lighting conditions. Results showed that the cumulative food consumption was indistinguishable in all groups [Fig.2(e)]. However, the eating patterns were significant altered by “ jetlag” light cycles as the percentages of nocturnal food intake were significantly decreased in all jetlag mice compared to control WL mice [Fig.2(f)]. Interestingly, across three tested light cycles, significantly increased nocturnal food intake was only observed in jetlag WL mice rather than jetlag mice under RL or BL, suggesting that WL accelerated eating behaviors [Fig.2(f)]. The cumulative water consumption of jetlag BL mice was significantly higher than that of control WL mice after 9 weeks of the experiment [Fig.2(g)]. In addition, the nocturnal drinking patterns were restored more quickly than eating behaviors. The percentages of nocturnal water intake in the first two tested light cycles were similar in all groups, while jetlag RL mice consumed less at night than control WL and jetlag WL mice on the third day after shifting light-dark cycle[Fig.2(h)].

Jetlag BL mice had significantly higher serum triglyceride levels compared with jetlag WL mice [(1.41± 0.12) mmol· L^-1vs. (1.10± 0.08) mmol· L^-1, p< 0.05][Fig.3(a)]. In contrast, jetlag BL mice had significantly lower serum cholesterol levels compared with jetlag RL mice [(5.2± 0.3) mmol· L^-1vs. (6.2± 0.2) mmol· L^-1, p< 0.05] [Fig.3(b)]. Normal liver function is essential for lipid metabolism. The “ jetlag” light significantly increased the serum ALT and AST levels under WL and RL, but not under BL, suggesting that BL prevented jetlag mice from impaired liver function [Fig.3(c, d)]. Although there was no difference in hepatic triglyceride levels between control WL mice and jetlag WL mice, RL significantly promoted fatty liver development in jetlag mice, while jetlag BL mice had the lowest levels of hepatic triglyceride [Fig.3(e, f)]. Histological analysis of eWATs revealed that chronically alternating light-dark cycles significantly promoted fat accumulation in adipose depots and jetlag mice under WL and RL had larger adipocyte size [Fig.3(g, h)].

Fig.3

Figure Option ViewDownloadNew Window

Fig.3 Effects of chronically alternating light-dark cycles on lipid metabolism are wavelength-dependent (a): Serum triglyceride levels; (b): Serum cholesterol levels; (c): Serum ALT levels; (d): Serum AST levels; (e): Representative images of H& E staining of sections of livers. Scale bars, 100 μ m; (f): Hepatic triglyceride levels; (g): Representative images of H& E staining of sections of eWAT. Scale bars, 100 μ m; (h): Average adipocyte area. Results are means ± SE (n=8 per group). * p< 0.05 vs. control WL group; #p< 0.05 vs. jetlag groups

Jetlag WL mice were much more glucose intolerant than control WL mice, while RL surprisingly protected jetlag mice from glucose intolerant as jetlag RL mice had lower blood glucose levels than jetlag WL mice [Fig.4(a)]. ITT results showed no significant difference in insulin sensitivity among all the groups, despite a trend that jetlag BL mice had a better insulin sensitivity [Fig.4(b)].

Fig.4

Figure Option ViewDownloadNew Window

Fig.4 Effects of chronically alternating light-dark cycles on glucose metabolism are wavelength-dependent (a): Glucose tolerance tests. Inset represents AUC for glucose; (b): Insulin tolerance tests. Inset represents AUC for glucose; (c): Fasting serum glucose levels; (d): Fasting serum insulin levels; (e): HOMA-IR; (F): Serum corticosterone levels; (g): Representative images of insulin staining of sections of pancreases. Scale bars, 200 μ m; (h): Quantitation of β -cell mass. Results are means ± SE (n=8 per group). * p< 0.05, * * p< 0.01, * * * p< 0.001 vs. control WL group; #p< 0.05, ##p< 0.01 vs. jetlag groups

As expected, jetlag WL mice had significantly higher fasting serum glucose and insulin levels than control WL mice [Fig.4(c, d)]. In addition, the HOMA-IR index, a model based on a single combination of fasting glucose and insulin levels, was dramatically increased in jetlag WL mice, indicating that WL significantly exacerbated the insulin resistance in jetlag mice [Fig.4(e)]. However, such adverse effects of “ jetlag” light cycles on glucose metabolism were attenuated in RL and BL. Jetlag BL mice had significantly lower fasting serum glucose levels compared with jetlag WL mice [(8.0± 0.8) mmol· L^-1vs. (11.2± 0.9) mmol· L^-1, p< 0.05] [Fig.4(c)], while jetlag mice subjected to RL and BL had significantly decreased fasting serum insulin levels and lower HOMA-IR index than jetlag WL mice [Fig.4(d, e)]. Corticosterone, a hormone that regulates multiple aspects of glucose and lipid homeostasis, was remarkably increased in jetlag RL mice [Fig.4(f)].

Chronically alternating light-dark cycles provoked a 33.6% increase in β -cell mass in jetlag WL mice compared with control WL mice [(2.73± 0.28) mg vs. (2.04± 0.15) mg, p< 0.05] [Fig.4(g, h)]. The compensatory increments of β -cell mass combined with elevated fasting serum insulin levels in jetlag WL mice provide an estimate on the contribution of “ jetlag” light cycle to insulin resistance. However, BL significantly prevented jetlag mice from enlargement of islet morphology, while jetlag RL mice had a trend of β -cell mass expansion compared with control WL mice [Fig.4(h)].