A novel diabetic mouse model for real-time monitoring of clock gene oscillation and blood pressure circadian rhythm

Abstract

Diabetic patients have an increased prevalence of blood pressure (BP) circadian rhythm disruption, which is associated with increased risk of target organ damage and detrimental cardiovascular events. Limited information is available regarding the role of clock genes in the disruption of the BP circadian rhythm in diabetes due to the lack of a diabetic animal model that allows real-time monitoring of clock gene oscillation. Here, we generated a novel diabetic db/db-mPer2^Luc mouse model by crossing the type 2 diabetic db/db mice with the mPer2^Luc knock-in mice. The daily rhythms of BP, heart rate, locomotor activity, and food and water intake were acquired by radiotelemetry or metabolic chambers. The daily oscillation of mPer2 bioluminescence was recorded by LumiCycle in real-time in tissue explants and by IVIS system in vivo. Our results showed that the db/db-mPer2^Luc mice were obese, diabetic and glucose intolerant. The db/db-mPer2^Luc mice displayed a compromised BP daily rhythm, which was associated with the disruption of the daily rhythms in baroreflex sensitivity, locomotor activity, and metabolism, but not heart rate or food and water intake. The phase of the mPer2 daily oscillation was advanced to different extents in the explanted peripheral tissues from the db/db-mPer2^Luc mice relative to that in the control mice. In contrast, no phase shift was detected in the mPer2 daily oscillation in the explanted suprachiasmatic nucleus (SCN). Moreover, the advanced phase shift of the mPer2 daily oscillation was also detected in the liver, kidney and submandibular gland in vivo in the db/db-mPer2^Luc mice. In conclusion, the diabetic db/db-mPer2^Luc mouse is a novel animal model that allows real-time monitoring of mPer2 circadian rhythms ex vivo and in vivo. The results from db/db-mPer2^Luc mice suggest that the desynchrony of mPer2 daily oscillation in the peripheral tissues contributes to the loss of BP daily oscillation in diabetes.

Introduction

Diabetes is one of the most common metabolic disorders, and its prevalence is rising to epidemic levels worldwide (Collaboration 2016). Cardiovascular diseases are the major common causes of mortality and morbidity in diabetic patients (Rask-Madsen and King 2013). Accumulated evidence from human and rodent studies suggests that disruption of normal circadian rhythms is associated with increased incidence of obesity and type 2 diabetes (Bass and Takahashi 2010). Night or rotating shift workers have a higher prevalence of type 2 diabetes (Gan, Yang et al. 2015). Imposing chronic jet-lag or light at night in mice disrupts food intake pattern and is associated with higher body weight (Fonken, Workman et al. 2010, Oike, Sakurai et al. 2015) and leptin resistance (Kettner, Mayo et al. 2015). Food intake at the “wrong” time (i.e., during the rest period) and increased snacking has been demonstrated to be critical for the development of obesity and diabetic metabolic disorder, independent of total caloric intake and macronutrient quality (Hatori, Vollmers et al. 2012). In humans, genetic variations in BMAL1, a core clock gene, are associated with increased incidence of type 2 diabetes and hypertension (Woon, Kaisaki et al. 2007, Corella, Asensio et al. 2016, Uemura, Katsuura-Kamano et al. 2016). In mice, either global (Rudic, McNamara et al. 2004, Kondratov, Kondratova et al. 2006) or tissue-specific Bmal1 deletion in liver (Lamia, Storch et al. 2008, Jacobi, Liu et al. 2015) or pancreas (Marcheva, Ramsey et al. 2010) causes impaired glucose homeostasis and/or insulin resistance. Taken together, a large body of evidence indicates a critical role of the circadian rhythm disruption in the pathogenesis of obesity, diabetes, and metabolic disorders. In contrast, much less research has been conducted on the impact of diabetes on the circadian rhythm in particular related to diabetic cardiovascular complications.

Blood pressure (BP) in humans undergoes a daily oscillation that is characterized by lowest levels at night (i.e., a “nocturnal dip”) and a peak before awakening (morning surge) (Millar-Craig, Bishop et al. 1978). The importance of the BP circadian rhythm is highlighted by a meta-analysis that demonstrated the early morning BP surge is associated with a 40% higher risk of acute myocardial infarction, a 29% higher risk of sudden cardiac death, and a 49% higher risk of stroke (Cohen, Rohtla et al. 1997). The results of a cross-sectional analysis of a 20,000-patient database revealed that up to 70% of diabetic patients are non-dippers (Gorostidi, Sobrino et al. 2007), in which the decrease in BP during the nocturnal sleep period is less than 10% of the daytime values. The non-dipping BP rhythm is the most common form of BP circadian rhythm disruption in diabetic patients and is associated with increased risks of target organ damage and detrimental cardiovascular outcomes (Routledge, McFetridge-Durdle et al. 2007, Eguchi 2011, Yano and Kario 2012, Ayala, Moya et al. 2013). Importantly, the non-dipping BP is emerging as an independent predictor of future cardiovascular events (Friedman and Logan 2009, Cuspidi, Vaccarella et al. 2010). The nocturnal laboratory animals, such as mice, have a BP daily oscillation that is opposite to that in humans, i.e., BP is low during the day (rest phase) and peaks at night (active phase). In agreement with these studies, we and others recently reported that the db/db mice, a widely-used diabetic mouse model (Herberg and Coleman 1977), are not only hypertensive but are also non-dippers (Park, Bivona et al. 2008, Su, Guo et al. 2008, Goncalves, Tank et al. 2009, Senador, Kanakamedala et al. 2009). Moreover, we and others have demonstrated that either global or selective deletion of Bmal1 in smooth muscle or adipocytes abolishes or attenuates BP circadian rhythm (Curtis, Cheng et al. 2007, Xie, Su et al. 2015, Yang, Chen et al. 2016, Chang, Xiong et al. 2018). Furthermore, the daily oscillations of the mRNA encoding many core clock genes and clock target genes are altered in isolated tissues from db/db mice (Kudo, Akiyama et al. 2004, Su, Guo et al. 2008, Caton, Kieswich et al. 2011, Su, Xie et al. 2012, Nernpermpisooth, Qiu et al. 2015, Grosbellet, Dumont et al. 2016). However, most of these studies investigated the core clock genes and clock target genes oscillations by real-time PCR or Western blot quantification of mRNAs or proteins in tissues collected every 4 to 6 hours in only one day. Consequently, the time resolution of circadian rhythm analysis is limited due to the limited sampling intervals and duration. One of the major reasons accounting for these limitations is that increasing the time resolution of clock gene oscillation requires a significant increase in time points of tissue collections which is cost prohibitive.

To overcome this barrier and to stimulate the research on the BP circadian rhythm in diabetes, we crossed the db/db mice with the mPer2^Luc knock-in mice and generated a novel db/db-mPer2^Luc mouse model. The db/db mouse is an extensively used monogenic type 2 diabetic mouse model. The syndrome in db/db mice is similar to that in maturity-onset diabetes in humans, characterized by obesity, infertility, hyperphagia and marked hyperglycemia (Ktorza, Bernard et al. 1997). Diabetes in db/db mice is caused by a spontaneous point mutation in the “leptin receptor” gene (lepr), resulting in abnormal splicing of the gene transcript, leading to defective in leptin signaling (Chen, Charlat et al. 1996, Lee, Proenca et al. 1996). The db/db-mPer2^Luc mice are expected to not only be diabetic but to allow quantitatively measuring of mPer2 protein oscillation by mPer2^Luc bioluminescence monitoring in real-time ex vivo and in vivo (Yoo, Yamazaki et al. 2004, Tahara, Kuroda et al. 2012) due to a reporter gene luciferase fused in-frame to the 3’ end of the endogenous mPer2 gene (Yoo, Yamazaki et al. 2004). We reported here for the first time that the db/db-mPer2^Luc mice are obese and diabetic and that they are non-dippers although they are normotensive. The compromised daily rhythm in BP in the db/db-mPer2^Luc mice is associated with the disruption of daily rhythms in baroreflex sensitivity, locomotor activity, and metabolism. Moreover, by monitoring mPer2^Luc bioluminescence in various peripheral and suprachiasmatic nucleus or nuclei (SCN) tissues ex vivo and in vivo, we demonstrated that there is a desynchrony of mPer2^Luc bioluminescence daily oscillations in peripheral tissues but not in the SCN in the db/db-mPer2^Luc mice.

Materials and methods

Generation of the db/db-mPer2^Luc mice

The heterozygous leptin receptor (Lepr^db) mutation db/+ mice on the C57BL/KsJ background (Stock No: 000642; also known as C57BL/KsJ-db/+) and the homozygous mPer2^Luc mice on the C57BL/6J background (Stock No: 006852; also known as C57BL/6J-mPer2^Luc) were purchased from the Jackson Laboratory. Since the homozygous C57BL/KsJ-db/db mice are infertile, the heterozygous male C57BL/KsJ-db/+ mice and homozygous female C57BL/6J-mPer2^Luc mice were used as breeders to generate the homozygous diabetic db/db-mPer2^Luc mice and heterozygous non-diabetic db/+-mPer2^Luc control mice (Figure S1A). Of note, both db/db-mPer2^Luc and db/+-mPer2^Luc control mice have a mixed C57BL/KsJ and C57BL/6J background. The genotyping protocol for the db/db mice is listed in the Jackson Laboratory website. The genotyping protocol for the mPer2^Luc mice was described previously (Yoo, Yamazaki et al. 2004). The representative agarose gels for PCR genotyping of the mPer2^Luc and db/db mice are shown in Figure S1B and S1C. The mice were fed normal chow diet and housed under 12:12 light: dark condition. Only the 4-6 month-old male db/db-mPer2^Luc and age- and gender-matched db/+-mPer2^Luc control mice were used in the current study. All animal procedures were approved by the Institutional Animal Care and Use Committee.

Animal characterization

Body composition (lean mass and fat mass) was assessed by NMR spectroscopy (Echo MRITM-100H, Houston, TX, USA). Non-fasting blood glucose level was measured between Zeitgeber Time (ZT; ZT0 is defined as light on and ZT12 is defined as light off) ZT9 to ZT10 by using StatStrip® XepressTM glucometer (NOVA® biomedical, Waltham, MA, USA). Non-fasted plasma insulin level was determined between ZT10 to ZT11 by an ELISA according to the manufacturer’s instructions (Chrystal Chem, Downers Grove, IL, USA). Intraperitoneal glucose tolerance test (IPGTT) was performed at ZT3 after 6-hour fasting by injecting 1 mg/kg glucose.

Radiotelemetry measurement of BP, heart rate, and locomotor activity

The db/db-mPer2^Luc and control mice were chronically instrumented in the left common carotid artery with a radiotelemetry probe (TA11PA-C10, Data Sciences International, St. Paul, MN, USA) as described previously (Su, Guo et al. 2008, Su, Xie et al. 2013, Xie, Su et al. 2015). After 10-day recovery from the surgery, BP, heart rate, and locomotor activity were recorded for three consecutive days.

Baroreflex sensitivity analysis

Spontaneous baroreflex sensitivity was analyzed by sequence techniques using Hemolab software (http://www.haraldstauss.com/HemoLab/HemoLab.html). At least four consecutive sequences where the systolic arterial pressure and pulse interval were positively correlated (r²>0.80) were counted. Baroreflex sensitivity was calculated as the average slope of the systolic pressure-pulse interval relationships as described previously (Xie, Su et al. 2015).

Metabolic chamber measurement of locomotor activity, food and water intake, respiratory exchange ratio (RER) and energy expenditure (EE)

The locomotor activity, food and water intake, RER, and EE were determined by indirect gas calorimetry LabMaster system (TSE System, Bad Homburg, Germany; also known as metabolic chambers). The mice were exposed to a 12:12 light:dark cycle and were individually housed in the acclimation cages for seven days and then transferred to the metabolic chambers. The oxygen consumption and carbon dioxide production in the metabolic chambers were measured every 30 minutes for three consecutive days. RER and EE were calculated by the accompanied TSE PhenoMaster software.

Real-time monitoring of mPer2 oscillations in explant tissues by LumiCycle

The procedure for real-time monitoring of mPer2 oscillations in explant tissues by LumiCycle was adapted from the previous report (Yamazaki and Takahashi 2005). Briefly, the aorta, mesenteric artery (MA), kidney, liver, white adipose tissue (WAT), thymus, lung, adrenal gland (AG), and brain were isolated from mice between ZT10 and ZT11. The aorta was cleaned, cut open longitudinally, and denuded of endothelium cells. The MA was dissected to remove fat tissues. The kidney, liver, WAT, thymus, and lung were cut into small pieces, with a diameter varying between approximately 2 and 6 mm depending upon the tissue. The total AG was used. The brain containing the SCN was cut into 250 μm thick sections by using NVSL manual advance vibroslice (World Precision Instruments, Sarasota, FL, USA). Each tissue was cultured in a well-sealed 35-mm Petri dish containing Dulbecco’s Modified Eagle Medium (DMEM) and 0.1 mM D-luciferin (Gold Biotechnology Inc., St. Louis, MO). Details of the medium constituent were described previously (Yamazaki and Takahashi 2005). The light emission from the cultured tissues was measured with photon-counting photomultiplier tubes that count photons for 1 min over a 10 min interval using a LumiCycle 32 system (Actimetrics, Wilmette, IL, USA) as described (Yamazaki and Takahashi 2005). The bioluminescence data obtained from the explanted tissues were analyzed using LumiCycle Analysis software (Actimetrics, Wilmette, IL, USA). To detrend the signal drift over time, the 24-hour moving average was subtracted from the raw data. To eliminate the influence of exposure to environmental lighting before recording, the first 12-hour data collected in the explant culture were excluded. The data collected from 12 hours to 36 hours in the culture were used to determine the oscillation amplitude and acrophase. The data collected from 12 hours to 120 hours in the culture were used to determine the oscillation period by the dampened sine-curve fitting method. The data with a goodness of fit >0.8 were used for analysis in all the tissues except in kidney where data with a goodness of fit > 0.7 were used due to the rapid dampening of the oscillation.

In vivo imaging of mPer2 time-of-day variation in the kidney, liver, and submandibular gland (SG)

The procedure for in vivo imaging of mPer2 time-of-day variation in the kidney, liver, and SG was adapted from the previous report (Tahara, Kuroda et al. 2012). Briefly, at ZT5, 11, 17 and 23, mice were anesthetized with 2.5-4% isoflurane and subcutaneously injected with D-luciferin (15 mg/kg body weight in PBS). The mice were imaged 7 minutes later for dorsal side up and 10 minutes later for later ventral side up for 5 seconds by using the IVIS spectrum (IVIS spectrum in vivo imaging system, PerkinElmer, Waltham, MA, USA). Bioluminescence from the liver was quantified (photon/s/cm²/sr) by setting the region of interest to the same shape and size using Living Image software (IVIS Imaging System). The bioluminescence intensity was expressed as an absolute value or as the percentage of the average value throughout the day as described (Tahara, Kuroda et al. 2012).

Cosinor analysis of circadian rhythm

The daily rhythms of BP, heart rate, locomotor activity, food and water intake, RER and EE were analyzed by using Cosinor analysis as previously reported (Refinetti, Lissen et al. 2007). Briefly, a cosine wave with a known period (24 hours) was fitted by the least squares to the data as an estimate of the pattern of the smooth rhythm. The model equation was written as x_i=M+Acos (θ_i+φ), where M is mesor, A is amplitude, φ is acrophase, and θ_i is trigonometric angles corresponding to the sampling time.

Quantitative analysis of mRNA expression

Mesenteric arteries were isolated from db/db-mPer2^Luc and control mice at ZT5 and ZT17. RNA extraction, cDNA synthesis, and real-time PCR were carried out as described (Guo, Su et al. 2005, Xie, Su et al. 2015). The real-time PCR primers for each gene are described in Table S1.

Statistical analysis

All data were expressed as mean ± SEM. For comparison of 1 parameter between 2 strains of mice, unpaired 2-tail Student’s t-test was used. For comparison of one parameter across a time period between 2 strains of mice, 2-way ANOVA with repeated measures and Bonferroni’s post-test were performed. For comparison of multiple parameters between 2 strains of mice, regular 2-way ANOVA with Bonferroni’s post-test was performed. P < 0.05 was defined as statistically significant.

Results

Db/db-mPer2^Luc mice are obese and diabetic

The db/db mouse is an extensively used monogenic type 2 diabetic mouse model. The syndrome in db/db mice is similar to that in maturity-onset diabetes in humans, characterized by obesity, infertility, hyperphagia and marked hyperglycemia (Ktorza, Bernard et al. 1997). The diabetic phenotype of the db/db mice, however, varies depending on the genetic background. Currently, there are two db/db mouse models: one is on the C57BL/KsJ background with severe hyperglycemia and temporarily elevated plasma insulin; the other one is on the C57BL/6J background with transient hyperglycemia and marked hyperinsulinemia (Hummel, Coleman et al. 1972). To study the disruption of circadian rhythms in type 2 diabetes, we crossed the C57BL/KsJ-db/db mice that have severe diabetes with the mPer2^Luc mice that contain a knock-in luciferase gene fused to mouse Period2 (mPer2) as a clock gene reporter (Yoo, Yamazaki et al. 2004), and generated a novel db/db-mPer2^Luc mice. Since the mPer2^Luc mice are on the C57BL/6J background, the generated db/db-mPer2^Luc mice have a mixed background (C57BL/KsJ and C57BL/6J). It is unclear to what extent the db/db-mPer2^Luc mice retain obesity and diabetes. Therefore we first characterized this novel mouse model with respect to obesity, hyperglycemia, hyperinsulinemia and insulin resistance.

The db/db-mPer2^Luc mice had significantly increased body weight when compared to their littermate db/+-mPer2^Luc control mice (Figure 1a). The body weight increase was mostly attributable to an increased fat mass as the lean mass was comparable between the db/db-mPer2^Luc and control mice (Figure 1b). The non-fasting blood glucose and plasma insulin levels in the db/db-mPer2^Luc mice were also markedly elevated relative to those in the control mice (Figure 1c and 1d). Moreover, the db/db-mPer2^Luc mice exhibited a severely impaired glucose tolerance (Figure 1e). These results indicate that the db/db-mPer2^Luc mice manifest the common characteristics of type 2 diabetes, e.g. obesity, hyperglycemia, hyperinsulinemia, and impaired glucose tolerance.

Fig. 1. The db/db-mPer2^Luc mice are obese and diabetic.

Body weight (a; N = 12), body composition (b; N = 4-6), non-fasting blood glucose (c; N = 12), and plasma insulin (d; N = 4-5) were measured between ZT9 and ZT11 in the db/db-mPer2^Luc and control db/+-mPer2^Luc mice. Glucose tolerance test (e; N = 11-12) was performed at ZT3 after 6-hour fasting. All data were expressed as mean ± SEM. *, P < 0.05; ***, P < 0.001; ns, not significant.

Db/db-mPer2^Luc mice have a compromised BP daily rhythm that is associated with the disruption of the daily rhythms in baroreflex sensitivity but not heart rate

To determine whether the BP daily rhythm is disrupted in the db/db-mPer2^Luc mice, we recorded BP by radiotelemetry under normal 12:12 light-dark cycle for 72 consecutive hours. We found that the daily oscillations of mean arterial pressure (MAP), systolic blood pressure (SBP), and diastolic blood pressure (DBP) were diminished in the db/db-mPer2^Luc mice compared to that in the control mice (Figure 2a; Figure S2a and S2d). The compromised daily rhythms of the MAP, SBP, and DBP were primarily caused by the decreased dipping during the inactive light phase with no change during the active dark phase in the db/db-mPer2^Luc mice relative to the control mice (Figure 2a; Figure S2a and S2d). Quantitative analysis of the daily (24-hour) average of MAP, SBP, and DBP showed no difference between the db/db-mPer2^Luc and control mice (Figure 2b; Figure S2b and S2e), indicating that the db/db-mPer2^Luc mice are normotensive, unlike the C57BL/KsJ-db/db mice (Park, Bivona et al. 2008, Su, Guo et al. 2008, Goncalves, Tank et al. 2009, Senador, Kanakamedala et al. 2009). Further quantitative analysis of the BP during either the light or dark phase (12-hour) revealed a 50% reduction in the difference between the light phase and the dark phase in the MAP, SBP, and DBP in the db/db-mPer2^Luc mice compared with that in the control mice (Figure 2c; Figure S2c and S2f). Cosinor analysis of the oscillations showed that the amplitude (half of the range of oscillation) and robustness of daily rhythms in the MAP, SBP, and DBP were significantly attenuated in the db/db-mPer2^Luc mice compared with that in the control mice (Figure 2d and 2e; Table S2). Interestingly, no differences were found in the acrophase (the time when the cycle peaks) between the db/db-mPer2^Luc and control mice (Figure 2f; Table S2).

Fig. 2. The daily rhythms of blood pressure (BP) and baroreflex sensitivity but not heart rate are disrupted in the db/db-mPer2^Luc mice.

BP and heart rate were recorded by radiotelemetry in the db/db-mPer2^Luc and control db/+-mPer2^Luc mice. a. The 72-hour recording of mean arterial pressure (MAP). The light grey box indicates the dark-phase and the length of the arrowhead lines indicates the BP difference between the light and dark phase in the two mouse strains. b. The 24-hour MAP. c. The 12-hour MAP during the light phase (L) and dark phase (D). d., e. and f.: The amplitude, robustness, and acrophase of the MAP daily oscillation. g. The spontaneous baroreflex sensitivity over the 24-hour day. h. The 72-hour recording of heart rate. i. The 24-hour heart rate. j. The 12-hour heart rate during the light phase (L) and dark phase (D). k., l., and m.: The amplitude, robustness, and acrophase of heart rate daily oscillation. All data were expressed as mean ± SEM (N = 6). *, P < 0.05; **, P < 0.01, ***, P < 0.001; ns, not significant.

Baroreflex is an important rapid negative feedback mechanism for maintaining normal BP. Therefore we investigated whether the compromised BP daily rhythm in the db/db-mPer2^Luc mice is associated with an alteration of the time-of-day variations in baroreflex sensitivity. We analyzed spontaneous baroreflex sensitivity by sequence techniques in the db/db-mPer2^Luc and control mice as previously described (Xie, Su et al. 2015). In the db/+-mPer2^Luc control mice, baroreflex sensitivity was significantly higher during the light phase than during the dark phase (Figure 2g). In contrast, such time-of-day variations of baroreflex sensitivity were abolished in the db/db-mPer2^Luc mice. This result implicates the loss of daily variation in baroreflex sensitivity contributes to the compromised BP daily rhythm.

Because heart rate is an important factor that determines the cardiac output and BP level (Reule and Drawz 2012), we investigated whether the daily heart rate oscillation is also altered in the db/db-mPer2^Luc mice. We found that the daily heart rate, the difference between light phase and dark phase heart rate, and its rhythmicity, including amplitude, robustness, and acrophase, were not significantly altered in the db/db-mPer2^Luc mice compared to that in the control mice (Figure 2h-2m).

The compromised BP daily rhythm is associated with the disruption of daily rhythms in locomotor activity and metabolism but not in food and water intake in the db/db-mPer2^Luc mice

Behavioral factors such as locomotor activity, food and water intake as well as metabolism may affect central and peripheral clock function through the release of neurotransmitters and hormones and thus impinge on BP circadian rhythm (Rudic and Fulton 2009). Therefore, the daily rhythms in locomotor activity, food and water intake, and metabolism were monitored by indirect calorimetry (also known as a metabolic chamber) in the db/db-mPer2^Luc and control mice every 30 minutes over 72 consecutive hours under 12: 12 light: dark condition. We also used radiotelemetry to monitor locomotor activity independently to confirm the indirect calorimetry data. The results from both indirect calorimetry and radiotelemetry data consistently showed that the daily oscillation in locomotor activity was abolished in the db/db-mPer2^Luc mice compared with that in the control mice (Figure 3a and 3b; Figure S3a and S3b). While the absolute counts regarding the daily locomotor activity from indirect calorimetry (Figure 3c) and radiotelemetry (Figure S3c) were not consistent, both methods showed a loss of the locomotor activity daily oscillation in the db/db-mPer2^Luc mice (Figure 3d; Figure S3d). Cosinor analysis revealed that the amplitude and robustness of the locomotor activity daily oscillations were largely diminished in the db/db-mPer2^Luc mice (Figure 3e and 3f; Figure S3e and S3f). Interestingly, in agreement with the compromised BP daily rhythm in the db/db-mPer2^Luc mice (Figure 2f), there were also no differences in the acrophase of the locomotor activity daily oscillation between the db/db-mPer2^Luc and control mice (Figure 3g; Figure S3g).

Fig. 3. The daily rhythm of locomotor activity but not food intake is disrupted in the db/db-mPer2^Luc mice.

Locomotor activity and food intake were recorded by indirect calorimetry. a. The 72-hour recording of locomotor activity in the control mice where the light grey box indicates the dark-phase. b. The 72-hour recording of locomotor activity in the db/db-mPer2^Luc mice where the light grey box indicates the dark-phase. c. The 24-hour locomotor activity. d. The 12-hour locomotor activity during the light phase (L) and dark phase (D). e., f., and g: The amplitude, robustness, and acrophase of locomotor activity daily oscillation. h. The 72-hour recording of food intake in the control mice where the light grey box indicates the dark-phase. i. The 72-hour recording of food intake in the db/db-mPer2^Luc mice where the light grey box indicates the dark-phase. j. The 24-hour food intake. k. The 12-hour food intake during the light phase (L) and dark phase (D). l., m., and n: The amplitude, robustness, and acrophase of the food intake daily rhythm. All data were expressed as mean ± SEM (N = 6). *, P < 0.05; **, P < 0.01, ***, P < 0.001; ns, not significant.

In contrast to the locomotor activity, the food and water intake daily oscillations appeared to be preserved in the db/db-mPer2^Luc mice (Figure 3h and 3i; Figure S4a and S4b), although the db/db-mPer2^Luc mice consumed more food and water than the control mice (Figure 3j; Figure S4c). Since the db/db-mPer2^Luc mice consumed more food and water proportionally during both the light and dark phase than the control mice (Figure 3k; Figure S4d), the percentages of daily food and water intake during the light and dark phase were similar between two strains of mice (Figure 3k; Figure S4d). In accordance with these findings, there were also no differences in robustness and acrophase in food and water intake daily oscillations (Figure 3m and 3n; Figure S4f and S4g). Interestingly, there was a trend towards an increased daily oscillation amplitude in food intake (Figure 3l) and a significant increase in water intake (Figure S4e) in the db/db-mPer2^Luc mice.

The respiratory exchange ratio (RER) and energy expenditure (EE) daily oscillations were acquired by the metabolic chamber. The RER is calculated as the ratio between the volume of carbon dioxide (VCO₂) produced and the volume of oxygen (VO₂) used in metabolism. It is an indicator of fuel sources (Even and Nadkarni 2012). The EE is calculated as the total daily energy expenditure (calories) in the metabolic chamber, including basal and physical activity expenditure, thermoregulation, and the thermic effects of food (Even and Nadkarni 2012). The RER daily oscillation was disrupted in the db/db-mPer2^Luc mice compared with the control mice (Figure 4a vs. 4b). Although both strains of mice had a similar average RER (Figure 4c), the db/db-mPer2^Luc mice lost the RER daily oscillation compared to the control mice (Figure 4d). In agreement with these findings, the amplitude and robustness of the RER daily oscillation were suppressed (Figure 4e and 4f), and the acrophase was delayed in the db/db-mPer2^Luc mice (Figure 4g). In contrast, the EE daily oscillation was preserved in both strains of mice (Figure 4h and 4i), although the daily EE level was higher in the db/db-mPer2^Luc than the control mice (Figure 4j). Both strains of mice exhibited a similar EE daily oscillation pattern (Figure 4k). In agreement with these findings, there was no difference in amplitude and acrophase between the db/db-mPer2^Luc and control mice (Figure 4l and 4n). However, the robustness was suppressed in the db/db-mPer2^Luc mice (Figure 4m).

Fig. 4. The daily rhythms of respiratory exchange ratio (RER) but not energy expenditure (EE) is disrupted in the db/db-mPer2^Luc mice.

RER and EE were recorded by indirect calorimetry. a. The 72-hour recording of the RER in the control db/+-mPer2^Luc mice. The light grey box indicates the dark-phase. b. The 72-hour recording of the RER in the db/db-mPer2^Luc mice. The light grey box indicates the dark-phase. c. The 24-hour RER. d. The 12-hour RER during the light phase (L) and dark phase (D). e., f., and g.: The amplitude, robustness, and acrophase of the RER daily rhythm. h. The 72-hour recording of the EE in the control db/+-mPer2^Luc mice. The light grey box indicates the dark-phase. i. The 72-hour recording of the EE in the diabetic db/db-mPer2^Luc mice. The light grey box indicates the dark-phase. j. The 24-hour EE. k. The 12-hour EE during the light phase (L) and dark phase (D). l., m., and n: The amplitude, robustness, and acrophase of the EE daily rhythm. All data were expressed as mean ± SEM (N = 6). *, P < 0.05; **, P < 0.01, ***, P < 0.001; ns, not significant.

Ex vivo LumiCycle recording reveals that the phases of mPer2 daily oscillation are shifted to different extents in various peripheral tissues but not the SCN from the db/db-mPer2^Luc mice

Multiple systems coordinate to maintain the normal physiological BP circadian rhythm (Coffman 2011). To investigate in which tissue the clock genes are altered in the db/db-mPer2^Luc mice that may contribute to the compromised BP circadian rhythm, we monitored mPer2 bioluminescence in real-time in peripheral and central SCN tissues in explant organ culture in the db/db-mPer2^Luc and control mice. In the various tissues from the control mice, the acrophases of mPer2 oscillation varied but were orchestrated in a specific order (Figure 5a through 5j), with the earliest peak shown by the SCN (10.47 ± 0.82 hours) and later peaks shown by the lung (12.08 ± 0.24 hours), kidney (14.23 ± 0.11 hours), liver (14.39 ± 0.77 hours), adrenal gland (15.59 ± 0.20 hours), white adipose tissue (WAT; 15.59 ± 0.39 hours), aorta (16.17 ± 0.24 hours), thymus (19.61 ± 0.77 hours), and mesenteric arteries (MA; 19.69 ± 0.29 hours).

Fig. 5. The phases of mPer2 protein daily oscillation are desynchronized in various explanted peripheral tissues from the db/db-mPer2^Luc mice.

The bioluminescence of mPer2 protein daily oscillation was recorded by LumiCycle in explanted central SCN and peripheral tissues from the db/db-mPer2^Luc and control db/+-mPer2^Luc mice. The mPer2 oscillation acrophase of the tissues was calculated using the LumiCycle analysis software. In the representative mPer2 bioluminescence real-time recording (left panel), the solid vertical line indicates the acrophase of the non-diabetic db/+-mPer2^Luc control mice, whereas the dotted vertical line indicates the acrophase of the diabetic db/db-mPer2^Luc mice. In the acrophase (right panel), the number above the symbol indicates the difference of the acrophase between two strains of mice. All data were expressed as mean ± SEM from the aorta (a; N = 7-11), mesentery artery (MA; b; N = 8-12), kidney (c; N = 4-5), liver (d; N = 6-12), white fat tissue (WAT; e; N = 3-4), thymus (f; N = 3-5), lung (g; N = 4-6), adrenal gland (h; N = 3-6), and suprachiasmatic nucleus (I; SCN; N = 6-11). *, P < 0.05; **, P < 0.01, ***, P < 0.001; ns, not significant.

In the tissues from the db/db-mPer2^Luc mice, the acrophases of mPer2 oscillations were significantly advanced to different extents relative to the corresponding control in a tissue-specific manner (Figure 5j). The aorta, MA, and kidney, which are crucial for BP and cardiovascular homeostasis, had a 0.98 ± 0.40, 1.70 ± 0.42, and 2.21 ± 0.56 hour phase advance, respectively (Figure 5a to 5c). The liver and WAT, two tissues that are crucial for energy metabolism, had a 3.28 ± 0.77 and 4.65 ± 1.21 hour phase advance (Figure 5d and 5e). The thymus, a primary lymphoid organ, had a 4.24 ± 1.59 hour phase advance (Figure 5f). In contrast, the lung and adrenal gland had no significant phase shift (Figure 5g and 5h). Interestingly, the SCN that has long been believed to be a major regulator of BP circadian rhythm, had also no significant phase shift (Figure 5i). In contrast to the shift in the acrophase in tissues from the db/db-mPer2^Luc mice, no consistent change was detected in period and amplitude of mPer2 luciferase oscillations in most peripheral tissues from the db/db-mPer2^Luc mice (Table S3).

In vivo imaging verifies that the phase of mPer2 oscillation is also advanced in the kidney, liver, and submandibular gland (SG) in the db/db-mPer2^Luc mice

To investigate whether the phase advance of the mPer2 oscillation observed in the explant tissue culture represents in vivo tissue oscillation, we used IVIS spectrum and monitored the mPer2 oscillations of the kidney, liver, and SG in the intact db/db-mPer2^Luc and control mice. The in vivo mPer2 bioluminescence images were obtained with 6 hours interval at ZT5, ZT11, ZT17, and ZT23, respectively. In accordance with the result from the ex vivo LumiCycle recording (Figure 5c and 5d), the in vivo mPer2 bioluminescence of the kidney, liver, and SG exhibited apparent time-of-day variations. The lowest absolute bioluminescence intensity was detected at ZT5 and the highest absolute bioluminescence intensity was detected at ZT17 in all three tissues (Figure 6a and 6b). The absolute bioluminescence intensities were significantly higher in the db/db-mPer2^Luc mice as compared with the control mice at ZT11 and ZT17 in the kidney (Figure 6c), at ZT11, ZT17, and ZT23 in the liver (Figure 6f) and at ZT17 in the SG (Figure 6i).

Fig. 6. The in vivo imaging shows a phase shift in mPer2 oscillation in the kidney, liver, and submandibular gland (SG) in the db/db-mPer2^Luc mice.

The in vivo imaging of mPer2 bioluminescence by the IVIS spectrum show a time-of-day variation in the kidney, liver, and SG. a. Representative in vivo imaging of the mPer2 bioluminescence in the kidney in the db/db-mPer2^Luc (upper panel) and control mice (lower panel). b. Representative in vivo imaging of the mPer2 bioluminescence in the SG and liver in the db/db-mPer2^Luc (upper panel) and control mice (lower panel). The absolute bioluminescence intensity detected in the kidney (c), liver (f), and SG (i). The relative bioluminescence intensity obtained by normalizing to the average of the four-time points’ data in the kidney (d), liver (g), and SG (j). The brown color solid vertical line indicates the acrophase of the control db/+-mPer2^Luc mice, whereas the blue dotted vertical line indicates the acrophase of the db/db-mPer2^Luc mice. The acrophase of the two strains of mice in the kidney (e), liver (h), and SG (k) where the number above the symbol indicates the difference of the acrophase between the two strains of mice. All data were expressed as mean ± SEM (N = 4-5). *, P < 0.05; ***, P < 0.001.

To better quantify the mPer2 oscillation in all three tissues between the two mouse strains, we normalized the absolute mPer2 bioluminescence intensities to the average of the four ZT time points absolute mPer2 bioluminescence intensities, in accordance with a previous report (Tahara, Kuroda et al. 2012). The resulting analysis revealed that the relative mPer2 bioluminescence signal from the db/db-mPer2^Luc mice peaked earlier in all three tissues than those of the control mice (Figure 6d, 6g, and 6j). Moreover, cosinor analysis further illustrated that the phase of the mPer2 oscillation was significantly advanced in all three tissues in the db/db-mPer2^Luc mice compared with that in control mice, with 2.60 ± 0.82, 1.54 ± 0.59, and 1.571 ± 0.61 hour advance in the kidney, liver, and SG (Figure 6e, 6h, and 6k), respectively.

The time-of-day variations in gene expressions are altered in the mesenteric arteries from the db/db-mPer2^Luc mice

Db/db mice exhibit alterations in the daily mRNA expressions of clock genes and BP regulatory genes as we have previously shown (Su, Xie et al. 2012). In addition, we have demonstrated that smooth muscle BMAL1 participates in the control of the BP daily rhythm by regulating one of the contraction regulatory proteins Rho-kinase 2 (ROCK2) in wild-type mice (Xie, Su et al. 2015). To test whether any putative clock-controlled blood pressure-associated genes are dysregulated in db/db-mPer2^Luc mice, we determined mRNA expressions of Bmal1 and several contractile regulatory genes in the MA at ZT5 and ZT17. As shown in Figures S5a through S5e, Bmal1, ROCK1, calponin-1, tropomyosin-2, and smooth muscle protein-22α (SM22α) mRNA expression exhibited a significant time-of-day variation. Importantly, an attenuation or loss of the time-of-day variations was found in the db/db-mPer2^Luc mice compared with the control mice. In contrast, no time-of-day variations were detected in ROCK2, calponin-2, calponin-3, and tropomyosin-1 mRNA in either genotype (Figures S5c, e, f, and g).

Discussion

The current study describes a novel type 2 diabetic db/db-mPer2^Luc mouse model. The major new findings are: 1) the db/db-mPer2^Luc mice are obese, hyperglycemic, and glucose-intolerant and thus resemble type 2 diabetic patients; 2) the db/db-mPer2^Luc mice are normotensive but exhibit a compromised BP daily rhythm, which is associated with the disruption of daily rhythms in baroreflex sensitivity, locomotor activity, and metabolism, but not heart rate or food and water intake; 3) a desynchrony of peripheral tissue oscillation is caused by the various extents of phase advances of the mPer2 oscillation ex vivo of many tissues except the central SCN pacemaker; 4) the similar desynchrony of mPer2 phase is also observed in vivo in the kidney, liver, and SG.

The db/db mice have been used extensively for studying the pathogenesis of obesity and diabetes. Interestingly, the diabetic phenotype of db/db mice varies depending on the genetic background. The hyperglycemia is more severe when the leptin receptor mutation is expressed on a C57BL/KsJ background than on a C57BL/6J background (Leiter, Coleman et al. 1981). Probably because of its severe diabetic phenotype, the C57BL/KsJ-db/db mice are most commonly used. Interestingly, the db/db-mPer2^Luc mice have a significantly higher body weight than the age-matched C57BL/KsJ-db/db mice (65.72 ± 1.38 g vs. 47.07 ± 1.05 g; N=12; P<0.001). However, the hyperglycemia in the db/db-mPer2^Luc mice is much less severe than that in the C57BL/KsJ-db/db mice (320.3 ± 18.46 mg/dl vs. 585.9 ± 9.163 mg/dl; n=12; P<0.001). These results suggest that the db/db-mPer2^Luc mice more closely resemble the C57BL/6J-db/db mice (Hummel, Coleman et al. 1972) and mimic diabetic patients with obesity, moderate hyperglycemia, and glucose intolerance.

In agreement with their moderate diabetic phenotypes, the db/db-mPer2^Luc mice are normotensive, which contrasts with the hypertensive C57BL/KsJ-db/db mice (Park, Bivona et al. 2008, Su, Guo et al. 2008, Goncalves, Tank et al. 2009, Senador, Kanakamedala et al. 2009). Despite this difference, the db/db-mPer2^Luc mice also exhibit non-dipping BP, similar to the C57BL/KsJ-db/db mice (Park, Bivona et al. 2008, Su, Guo et al. 2008, Goncalves, Tank et al. 2009, Senador, Kanakamedala et al. 2009), which is typified by a lack of BP fall during the inactive light phase. Although leptin signaling is implicated in obesity-associated hypertension (Simonds, Pryor et al. 2014), such non-dipping BP in the db/db-mPer2^Luc and C57BL/KsJ-db/db mice is unlikely to be mediated directly by the loss of function mutation in the leptin receptor since the disruption of BP circadian rhythm was only detectable in mice older than 11-weeks (Senador, Kanakamedala et al. 2009). The mechanism by which diabetes induces non-dipping BP is unclear. In particular, it is unclear whether hyperglycemia, insulin resistance, or both are responsible for the disrupted BP circadian rhythm. While this important mechanistic issue remains to be elucidated, the current study demonstrates for the first time that the disrupted BP daily rhythm in the db/db-mPer2^Luc mice is associated with the loss of the daily rhythm in spontaneous baroreflex sensitivity but not heart rate. Baroreflex is a critical mechanism for maintaining the BP homeostasis, and baroreflex sensitivity exhibit daily variations in humans (Hossmann, Fitzgerald et al. 1980, Di Rienzo, Parati et al. 2001). Interestingly, the observed loss of baroreflex sensitivity daily variation resembles the loss of baroreflex sensitivity daily variation we reported in the smooth muscle Bmal1 knockout mice (Xie, Su et al. 2015), indicating that dysfunction of clock genes in the db/db-mPer2^Luc mice may cause loss of baroreflex sensitivity daily variation thus contributes to the decreased nocturnal BP decline phenotype.

We have previously reported that the daily locomotor activity rhythm is lost in the C57BL/KsJ-db/db mice (Su, Guo et al. 2008). In accordance with this finding, the current study illustrated that this locomotor rhythm was similarly abolished in the db/db-mPer2^Luc mice (Su, Guo et al. 2008). We speculate that the loss of the locomotor activity rhythm in both strains of db/db mice results from their severe obesity, i.e., that they are too heavy to move around. Although the loss of locomotor activity rhythm may potentially contribute to the loss of the BP daily rhythm [47], the loss of locomotor activity mainly occurred during the night in the db/db-mPer2^Luc mice, whereas the loss of BP dipping occurred during the day. Therefore, it is unlikely that the loss of locomotor activity accounts for the disrupted BP daily rhythm in the db/db-mPer2^Luc mice.

In humans (van Moorsel, Hansen et al. 2016) and rodents (Oosterman, Foppen et al. 2015, Sun, Wang et al. 2015), RER displays time-of-day variations, with higher values during the active phase indicating the preferential use of carbohydrates and lower values during the inactive phase indicating the preferential use of fats. In the C57BL-KsJ-db/db mice, RER was decreased at one specific time of the day (Osborn, Sanchez-Alavez et al. 2010, Choi, Kim et al. 2015). However, it is surprising that it has not been reported whether the daily rhythm of RER is disrupted in the db/db mice. One of the very intriguing findings from the current study is that the db/db-mPer2^Luc mice lost the RER time-of-day variations, mainly due to an increased RER during the inactive light phase as compared to control mice. These results suggest that the flexibility to use different sources of fuel is compromised in the diabetic db/db-mPer2^Luc mice. Moreover, there is a temporal correlation between increased RER and decreased BP decline as both occurred during the inactive light phase. However, it is unclear whether the increased RER during the light phase caused the compromised BP dipping in the db/db-mPer2^Luc mice.

Accumulated evidence from the animal and human studies during the last decade suggests that the BP circadian rhythm is regulated by multiple organs and systems, including the neuroendocrine system, kidneys, and vasculature (resistance arteries) (Rudic and Fulton 2009). It is long believed that the BP circadian rhythm, just like other physiological and behavioral circadian rhythms, is mostly controlled by the master pacemaker in the SCN. However, the current study demonstrated that the phase of mPer2 protein daily oscillation was not significantly altered in the SCN tissue from the db/db-mPer2^Luc mice compared to controls. These results confirm previous reports that there is a little or no change of the SCN mPer2 mRNA daily oscillation in the C57BL-KsJ-db/db mice (Kudo, Akiyama et al. 2004, Nernpermpisooth, Qiu et al. 2015, Grosbellet, Dumont et al. 2016). These results are also consistent with previous reports that peripheral clock gene oscillations are altered in some tissues from diabetic patients (Ando, Takamura et al. 2009, Pappa, Gazouli et al. 2013) and db/db mice (Kudo, Akiyama et al. 2004, Caton, Kieswich et al. 2011, Su, Xie et al. 2012, Nernpermpisooth, Qiu et al. 2015). In addition, in db/db mice, the alternations of peripheral clock expression occur as early as 6-8 weeks of age- (Kudo, Akiyama et al. 2004, Caton, Kieswich et al. 2011) whereas the disruption of BP circadian rhythm is not detectable in db/db mice until 11-weeks or older, indicating that peripheral clock impairment precedes the disruption of the BP circadian rhythm. Taken together, these results suggest that the peripheral oscillators, in contrast to the master SCN pacemaker, are strongly affected by diabetes and may be responsible for the disruption of BP circadian rhythm.

Perhaps one of the most important findings from the current study is that the phase of the mPer2 protein daily oscillation was advanced into various extents in a tissue-specific manner in peripheral tissues, in the absence of any change in the phase in the SCN. This finding was revealed by monitoring mPer2 protein oscillation in real-time in our novel db/db-mPer2^Luc mice. , In agreement with the important role of BMAL1 in the renal, smooth muscle, and fat tissues in regulation of BP rhythm under physiological conditions (Tokonami, Mordasini et al. 2014, Xie, Su et al. 2015, Chang, Xiong et al. 2018), we found that the phase of mPer2 protein oscillation was advanced in the WAT, kidney, MA, and aorta from the db/db-mPer2^Luc mice to 4.6, 2.21, 1.71, and 0.99 hours, respectively. These results are also consistent with the previous studies reporting mPer2 mRNA daily oscillation was altered in these tissues from the C57BL-KsJ-db/db mice (Su, Guo et al. 2008, Caton, Kieswich et al. 2011, Su, Xie et al. 2012, Nernpermpisooth, Qiu et al. 2015). In contrast, it was surprising that the phase of mPer2 protein daily oscillation in the adrenal gland, an important source of hormones that regulate the BP circadian rhythm, was not significantly changed in the C57BL-KsJ-db/db mice relative to control mice. It was also surprising that the phase of mPer2 protein daily oscillation in the thymus, an important organ that produces T lymphocytes, was advanced up to 4.23 hours. This result is consistent with the recent report that T lymphocytes play a critical role in angiotensin II-induced hypertension (Guzik, Hoch et al. 2007), and suggests that clock genes in T lymphocytes may be crucially involved in the disruption of the BP circadian rhythm in diabetes.

Obesity and diabetes in mice can be induced by a high fat (HF) diet, which also altered activity, feeding, and molecular circadian rhythms (Kohsaka, Laposky et al. 2007, Hatori, Vollmers et al. 2012). Although the causes of obesity and diabetes in HF diet-fed mice and db/db-mPer2^Luc mice are different, it is interesting to note that there are some similarities in respect to the effects of HF diet and leptin receptor mutation (db/db mice) on the mPer2 rhythm. For example, using ex vivo bioluminescent analyses, Pendergast et al. demonstrated that HF diet-fed mice exhibit a 1-4 hour phase advances in the mPer2 rhythms of two organs but not the SCN (Pendergast, Branecky et al. 2013). Using the same ex vivo assay, the current study obtained the similar results in the db/db-mPer2^Luc mice, suggesting that the disruption of peripheral clocks in the absence of a change in the central pacemaker by obesity or diabetes is not model-specific. In addition, a ~ 4-hour phase advance of the mPer2 oscillation in liver explants without a phase shift in explants of lung is observed in HF-fed mice (Pendergast, Branecky et al. 2013, Branecky, Niswender et al. 2015). The current study also obtained the similar results in the db/db-mPer2^Luc mice, suggesting that different peripheral tissues have different sensitivity to obesity and diabetes. Despite these similarities, it should be pointed out that there were some differences in respect to the effects of HF diet and leptin receptor mutation (db/db mice) on the mPer2 rhythm. For example, the mPer2 rhythm in the aorta explants showed a ~ 1-hour phase advance in the db/db-mPer2^Luc mice, but no change in the HF diet-fed mice (Pendergast, Branecky et al. 2013).

Another interesting finding of the current study is that the mPer2 phase shifts observed in vitro from tissues explanted from the db/db-mPer2^Luc mouse reflected phase shifts observed in vivo. This raises the question whether the mPer2 phase shifts in the db/db-mPer2^Luc mice might be caused directly by leptin receptor mutation within cells or indirectly by hyperphagia, obesity, and diabetes. Although it is currently uncertain, there is some evidence to support all possibilities. First, leptin is an adipocyte-derived hormone that binds to the leptin receptor and promotes weight loss by reducing appetite and food intake and by increasing energy expenditure (Kelesidis, Kelesidis et al. 2010). Serum leptin levels display diurnal variations in both humans and rodents. There is also evidence that leptin can directly regulate clock gene oscillations. For example, leptin can phase advance the electrical activity rhythm in the rat SCN in vitro (Prosser and Bergeron 2003). Moreover, leptin is implicated in the regulation of hypertension in obesity (Simonds, Pryor et al. 2014). Thus, leptin receptor mutation in various tissues may directly advance the mPer2 oscillations in vivo. Second, consistent with a previous report (Ktorza, Bernard et al. 1997), the current study demonstrated that the db/db-mPer2^Luc mice consumed more food and water than the control mice. Although food intake pattern was not altered in the db/db-mPer2^Luc mice, it is possible that the increased food intake due to impaired leptin signaling in the db/db-mPer2^Luc mice alters circadian rhythms. In fact, evidence that increased food intake affects behavioral, metabolism, and molecular circadian rhythms has been demonstrated in HF diet-fed mice (Kohsaka, Laposky et al. 2007, Hatori, Vollmers et al. 2012, Pendergast, Branecky et al. 2013, Branecky, Niswender et al. 2015) and db/db mice (Kennedy, Ellacott et al. 2010). Thus, hyperphagia may mediate leptin receptor mutation-associated phase advance of the mPer2 oscillation in vivo. Third, the current study demonstrated hyperglycemia, hyperinsulinemia, and glucose intolerance in the db/db-mPer2^Luc mice. Since both glucose and insulin have been shown to alter clock gene expression rhythms in vitro and in vivo (Hirota, Okano et al. 2002, Dang, Sun et al. 2016), it is likely that obesity and diabetes resulting from leptin receptor mutation may also have an indirect effect on mPer2 phase advances in these tissues in vivo. Nevertheless, future studies are required to distinguish these potential mechanisms.

In conclusion, the current study described a novel diabetic db/db-mPer2^Luc mouse model that allows real-time measurement of diabetes-induced dysregulation of clock genes and disruption of the BP circadian rhythm. Using this novel db/db-mPer2^Luc mouse model, we have revealed that disruption of the BP circadian rhythm in diabetes is associated with the loss of daily rhythms in baroreflex sensitivity, locomotor activity, metabolism, and a tissue-specific phase advance of the peripheral oscillators but not the central SCN pacemaker. These findings provide preclinical evidence for a potentially significant role of peripheral clock gene desynchrony in linking diabetes to compromised BP and metabolism circadian rhythms. The results from the current study may lead to the identification of synchronizing various oscillators as a novel therapeutic strategy against diabetic cardiovascular complications and thus improve the prognosis of diabetic patients.