Circadian disruption by short light exposure and a high energy diet impairs glucose tolerance and increases cardiac fibrosis in Psammomys obesus

Type 2 diabetes mellitus (T2DM) increases cardiac inflammation which promotes the development of cardiac fibrosis. We sought to determine the impact of circadian disruption on the induction of hyperglycaemia, inflammation and cardiac fibrosis. Methods: Psammomys obesus (P. obesus) were exposed to neutral (12 h light:12 h dark) or short (5 h light:19 h dark) photoperiods and fed a low energy (LE) or high energy (HE) diet for 8 or 20 weeks. To determine daily rhythmicity, P. obesus were euthanised at 2, 8, 14, and 20 h after ‘lights on’. Results: P. obesus exposed to a short photoperiod for 8 and 20 weeks had impaired glucose tolerance following oral glucose tolerance testing, compared to a neutral photoperiod exposure. This occurred with both LE and HE diets but was more pronounced with the HE diet. Short photoperiod exposure also increased myocardial perivascular fibrosis after 20 weeks on LE (51%, P < 0.05) and HE (44%, P < 0.05) diets, when compared to groups with neutral photoperiod exposure. Short photoperiod exposure caused elevations in mRNA levels of hypertrophy gene Nppa (atrial natriuretic peptide) and hypertrophy transcription factors Gata4 and Mef2c in myocardial tissue after 8 weeks. Conclusion: Exposure to a short photoperiod causes impaired glucose tolerance in P. obesus that is exacerbated with HE diet and is accompanied by an induction in myocardial perivascular fibrosis.


Type 2 diabetes mellitus (T2DM) increases cardiac inflammation which promotes the development of cardiac fibrosis. We sought to determine the impact of circadian disruption on the induction of hyperglycaemia, inflammation and cardiac fibrosis. Methods: Psammomys obesus (P. obesus) were exposed to neutral (12 h light:12 h dark) or short (5 h light:19 h dark) photoperiods and fed a low energy (LE) or high energy (HE) diet for 8 or 20 weeks.
To determine daily rhythmicity, P. obesus were euthanised at 2, 8, 14, and 20 h after 'lights on'. Results: P. obesus exposed to a short photoperiod for 8 and 20 weeks had impaired glucose tolerance following oral glucose tolerance testing, compared to a neutral photoperiod exposure. This occurred with both LE and HE diets but was more pronounced with the HE diet. Short photoperiod exposure also increased myocardial perivascular fibrosis after 20 weeks on LE (51%, P < 0.05) and HE (44%, P < 0.05) diets, when compared to groups with neutral photoperiod exposure. Short photoperiod exposure caused elevations in mRNA levels of hypertrophy gene Nppa (atrial natriuretic peptide) and hypertrophy transcription factors Gata4 and Mef2c in myocardial tissue after 8 weeks. Conclusion: Exposure to a short photoperiod causes impaired glucose tolerance in P. obesus that is exacerbated with HE diet and is accompanied by an induction in myocardial perivascular fibrosis.
Cardiovascular disease is the leading cause of death worldwide 1 . Type 2 Diabetes Mellitus (T2DM) is regarded as pan epidemic, with more than 640 million people predicted to have T2DM by 2040 2 . It is well-established that there is a link between cardiovascular disease (CVD) and T2DM with them frequently occurring simultaneously along with an associated elevated risk of adverse outcomes 1,3,4 . Specifically, people with diabetes or pre-diabetes are more likely to develop cardiovascular disease than their non-diabetic counterparts 1,3,4 . Diabetes causes a range of myocardial abnormalities that can lead to heart failure including left ventricular hypertrophy, an impaired tolerance to ischemic injury and pathophysiological changes in the myocardium that can result in diabetic cardiomyopathy [5][6][7][8][9] . The characteristics of diabetic cardiomyopathy include: myocardial fibrosis, dysfunctional remodelling, diastolic and systolic dysfunction, an elevation in collagen-based myocytes and extracellular matrix stiffness and impaired calcium handling. Inflammation is increasingly recognised as a central underlying cause of these changes which can lead to heart failure 5,10 .

Results
Effect of high energy diet and shorter photoperiod on weights and glucose levels. P. obesus fed the HE diet and exposed to the short photoperiod (5:19HE group) for 8 weeks had significantly higher heart and body weights and a lower heart:body weight ratio, compared to the 12:12LE control group (Table 1). After 8 weeks, fasting blood glucose levels were elevated in the group exposed to a short photoperiod and HE diet (5:19HE), compared to all other groups (12:12LE 53%; 12:12HE 79%; 5:19LE 85%, Table 1). Glucose tolerance was impaired in groups that were exposed to a HE diet or short photoperiod or both (12:12HE,5:19LE and 5:19HE groups). This was demonstrated by a significant elevation in glucose levels 120 min after receiving a bolus infusion of glucose, when compared to the corresponding 0 min timepoints ( Fig. 1a, P < 0.05 for all).
In the 20-week cohort, there were no significant differences in heart weight across the four groups (Table 2). However, the 5:19HE group gained significantly more body weight compared to both the 12:12LE and 5:19LE groups. No differences in heart:body weight ratios were observed. After 20 weeks, there were no changes in fasting blood glucose levels ( Table 2). Glucose tolerance was, however, impaired in groups that were exposed to a short photoperiod (5:19LE and 5:19HE groups) as glucose levels were significantly elevated 120 min after glucose infusions, compared to corresponding 0 min timepoints (Fig. 1b, P < 0.05 for all). 20 weeks of high energy diet and shorter photoperiod induced perivascular fibrosis in heart tissue. Perivascular fibrosis was detected by staining for the presence of collagen around the heart vessels. At 8 weeks, there were no changes in the amount of collagen deposition (Fig. 2a), either in the animals fed a HE diet, nor exposed to a short photoperiod. Animals on both the HE diet and the short photoperiod (5:19HE) had reduced expression of the collagen gene Col1a1 (46%, P < 0.05), when compared to the 12:12HE group (Fig. 2b). However, no daily rhythm differences were observed between time points across all the groups (Fig. 2c).
In the 20-week cohort, when compared to the 12:12LE group, there was a significant increase perivascular fibrosis with higher levels of collagen deposition around myocardial vessels in both groups of animals exposed to a short photoperiod (Fig. 2d, 5:19LE 51%; 5:19HE 44%). There was a trend for an increase in myocardial Col1a1 expression in the 5:19HE group (54%, P = 0.06), when compared to 12:12HE animals (Fig. 2e). Overall, exposure of a short photoperiod for 20 weeks appears to be the main effector stimulating myocardial perivascular fibrosis. Table 1. Weights and glucose levels of P. obesus exposed to neutral or short photoperiods and a high energy or low energy diet for 8 weeks. P. obesus were exposed to low energy (LE) or high energy (HE) diet and either a neutral photoperiod (12 light:12 dark) or short photoperiod (5 light:19 dark) for 8 weeks. Data presented as Mean ± SEM. † P < 0.05, † † P < 0.01 vs. 12:12LE; # P < 0.05, vs. 12:12HE; ^P < 0.05 vs. 5:19LE. www.nature.com/scientificreports/ Effects of high energy diet and short photoperiod on hypertrophy genes expression and cardiomyocyte size. We next assessed the hypertrophy markers atrial natriuretic peptide (ANP) and myosin heavy chain (MHC)α (adult isoform) and MHCβ (foetal isoform), represented by genes Nppa, Myh6 and Myh7, respectively. We also measured changes mRNA levels of the transcription factors that control these genes, Gata4 and Mef2c 26 . Morphological changes in cardiomyocyte size, representing hypertrophy, were also determined. High energy diet and Short photoperiod cause abnormal glucose tolerance. Oral glucose tolerance tests (OGTT) were performed in P. obesus were exposed to low energy (LE) or high energy (HE) diet and either a neutral photoperiod (12 light:12 dark) or short photoperiod (5 light:19 dark) for 8 (a) or 20 weeks (b). P. obesus were fasted for 4 h. A dose of 2 g glucose/kg body weight was administered via gavage. Blood was collected at time 0 and 120 min later. Mean ± SEM, n = 6-15/group. *P < 0.05 compared to time point 0. Table 2. Weights and glucose levels of P. obesus exposed to neutral or short photoperiods and a high energy or low energy diet for 20 weeks. P. obesus were exposed to low energy (LE) or high energy (HE) diet and either a neutral photoperiod (12 light:12 dark) or short photoperiod (  www.nature.com/scientificreports/ In the 8-week cohort, we found that there was a significant increase in Gata4 in hearts of the 5:19HE group, when compared to 12:12LE and the 12:12HE groups ( Fig. 3a, P < 0.05). Mef2c was also elevated in the 5:19HE group, compared to the 12:12LE (Fig. 3b, P < 0.05). Demonstrating that a combination of a short photoperiod and/or HE diet increases these hypertrophy transcription factors. Significant increases were also noted in Nppa (i.e. ANP) in both groups exposed to a short photoperiod such that Nppa was higher in the 5:19LE and 5:19HE groups, when compared to the 12:12HE group ( Fig. 3c, P < 0.05). This shows that short photoperiod exposure is the main driver of these changes. No changes occurred in the Myh7:Myh6 ratio, a marker of cardiac function, between groups ( www.nature.com/scientificreports/ exposed to HE and short photoperiod (5:19HE), there were no changes in cardiomyocyte size across all groups in the 8-week cohort (Fig. 3e).
In the 20-week cohort, no changes were found in Merf2C or the Myh7:Myh6 ratio between groups ( Fig. 4a,b, P < 0.05). However, histological analyses revealed that P. obesus exposed to the high energy diet and short photoperiod (5:19HE) had significantly smaller cardiomyocytes, when compared to all other groups (Fig. 4c, P < 0.05).
Expression of Clock, inflammatory and pro-fibrotic genes after 8 weeks of high energy diet and/or short photoperiod. We next assessed the average expression and daily rhythms of the established clock gene Per2, inflammatory genes Rela and Ccl2, and pro-fibrotic gene Tgfb1 in our 8-week cohort. No significant differences were observed in the average expression of Per2 across all four groups irrespective of diet or photoperiod (Fig. 5a). It has been shown that the expression of Per2 will change over a photoperiod 27 . As such, we measured the gene expression changes across a 24-h period. Daily Per2 rhythms changed in animals exposed to short photoperiod fed a LE diet (5:19LE) (Fig. 5b). In this cohort, Per2 expression was elevated at ZT8 when compared to ZT2 and ZT20.
Assessment of inflammatory genes showed that while the average Rela (p65-NF-κB) mRNA levels did not differ across the four groups (Fig. 5c), the daily Rela rhythm was found to increase at ZT14 compared to ZT2 and ZT8 in the 12:12HE group (Fig. 5d), indicating a cyclic effect over the day with the peak at ZT14 with a HE diet but normal photoperiod. High energy diet alone with a normo-photoperiod (12:12HE) caused a trend for an increase in the average myocardial Ccl2 expression in the 12:12HE group, when compared to the 12:12LE group (Fig. 5e, 34%, P = 0.0530). Daily Ccl2 rhythm was significantly different in the 12:12LE group with elevated Ccl2 levels observed at ZT20 compared to ZT2 and ZT14 (Fig. 5f), indicating that Ccl2 gene expression increases after a longer light exposure in our control group (12:12LE). By contrast, the daily Ccl2 rhythms in the groups exposed to HE diet and/or short photoperiod (12:12HE, 5:19LE or 5:19HE) were not altered. While no differences were a b www.nature.com/scientificreports/ observed with average Tgfb1 mRNA levels (Fig. 5g), variations in daily rhythm were observed in the 12:12HE group (Fig. 5h). Similar to the Rela patterns, Tgfb1 mRNA levels in the 12:12HE group increased at ZT14 compared to ZT2 and ZT8, indicating a cyclic effect over the day with the peak at ZT14. This was stimulated by a HE diet but normal photoperiod.
Expression of apoptosis genes and Serca2 after 8 weeks of high energy diet and/or short photoperiod. We then determined whether high energy diet with a short photoperiod affected apoptosis by measuring the expression of Bax, a pro-apoptotic gene in tandem with the anti-apoptotic gene Bcl2. Overall, there were no differences in either the average expression or daily rhythms of Bax and Bcl2 across all four groups (Fig. 6a-d). The average Bax/Bcl2 ratio, a marker of apoptosis, showed no differences between groups (Fig. 6e). However, there was a difference in its daily rhythm in the 12:12LE group (Fig. 6f). The Bax/Bcl2 ratio increased at ZT20 compared to ZT8, indicating a cyclic effect of overall apoptosis that increases over a longer photoperiod.
No differences were observed in either the average expression or daily rhythm of Serca2, a heart contractility gene (Fig. 6g,h).

Effect of 20 weeks of high energy diet and/or shorter photoperiod on gene expression in sand rat hearts.
Whether HE diet and/or a short photoperiod affected apoptotic and contractility genes after 20-weeks of exposure was then assessed. Per2 mRNA levels were significantly higher in the 5:19LE group when compared to the 12:12LE (Fig. 7a, 1.2-fold, P < 0.05), indicating an effect of short light exposure. No significant differences were observed in the expression of inflammatory genes Rela (Fig. 7b) and Ccl2 (Fig. 7c), pro-fibrotic gene Tgfb1 (Fig. 7d), or the pro-apoptotic gene Bax (Fig. 7e). The 12:12HE animals had significantly lower expression of the anti-apoptotic gene Bcl2 (Fig. 7f, 35%, P < 0.05). There were no changes in the Bax/Bcl2 ratio (Fig. 7g) or Serca2 (Fig. 7h) mRNA levels across all four groups.

Discussion
Circadian disruption caused by disordered exposure to light increases a plethora of metabolic-related complications including T2DM and cardiovascular disease 11 . Using the unique rodent model of P. obesus we investigated the effects of circadian disruption on the onset of T2DM and myocardial pathophysiology. After both 8 weeks and 20 weeks of exposure we found that a short photoperiod (5:19) caused abnormal glucose tolerance, when compared to neutral photoperiod (12:12) exposure. This occurred with both LE and HE diets, but was exacerbated with the HE diet which was coupled to a significant increase in body weight. In parallel, we found the short photoperiod induced myocardial perivascular fibrosis after 20 weeks, but not 8 weeks, which occurred on both the LE and HE diets. This corresponded with an increase in Col1a1 collagen gene expression in the 5:19HE group, compared to the 12:12HE group. Furthermore, we found that the hearts of P. obesus in the 5:19HE group displayed increases in the expression of hypertrophy genes Gata4, Merf2c and Nppa after 8 weeks but not 20 weeks. Taken together, by simulating the conditions of human shift work with P. obesus, our data imply that shift workers exposed to circadian disruption may be at increased risk of developing T2DM, which is more pronounced with an HE diet. This induces features of cardiac fibrosis, a factor involved in the development of heart failure. The findings suggest that interventions that correct circadian disruption are likely to show benefit. Hyperglycaemia increases cardiac fibrosis via an induction of pro-inflammatory signals that increase collagen deposition and subsequently become organised into a fibrotic scar 5 . The fibrosis occurs either around myocardial vessels (perivascular fibrosis) or interstitially 28,29 . In the current study, we have shown that when placed on a HE diet for either 8 or 20 weeks P. obesus display features of T2DM development. Circadian disruption via a shortened photoperiod substantially increased both fasting glucose levels after 8, but not 20 weeks, of exposure and impaired glucose tolerance as determined using OGTT at both time points. Of particular interest was that in the 20-week group the HE diet alone did not induce impaired glucose tolerance, but only occurred when combined with a short photoperiod. This demonstrates the important regulatory effects of disrupted circadian cycles on the development of T2DM. These findings suggest that exposure to disrupted circadian cycles causes exacerbated detrimental effects on glucose tolerance and weight gain, particularly when accompanied with a HE diet.
Another factor that reflected the exacerbation of abnormal glucose tolerance in the 20-week short photoperiod groups, was the induction of myocardial perivascular fibrosis. This was not present in the 8-week cohorts suggesting that it takes a longer period of time for the fibrotic deposits to develop. Furthermore, in the 20-week group the development of perivascular fibrosis only became evident in the P. obesus exposed to the short photoperiod but was independent of diet. Consistent with the increase in myocardial perivascular fibrosis, there was an increase in the expression of myocardial expression of collagen type 1 gene Col1a1. These findings support other studies in P. obesus 24 that have reported increases in myocardial perivascular fibrosis and Col1a1 and has been noted in alternate rodent models of Type 1 and Type 2 diabetes 5,30-32 . However, our study is the first to demonstrate that circadian disruption can drive these pathophysiological changes in the myocardial tissue, independent of diet.
Increases in myocardial perivascular fibrosis and Col1a1 were not accompanied by increases in inflammation in the 20-week groups with a short photoperiod. Proinflammatory genes CCl2 and Rela remained unchanged but are reported to be the underlying mechanistic drivers of hyperglycaemic-driven myocardial fibrosis 6 . Furthermore, pro-fibrotic gene TGF-B1 that is also anti-inflammatory did not change in the 20-week group. A likely explanation for the lack of change is that after 20-weeks the earlier inflammatory phase had already resolved and the subsequent proliferative phase of repair, involving TGF-B1 (which proceeds collagen deposition 33 ), was also complete. At the 8-week time point some of our data support this concept. In the earlier 8-week 12:12HE group, myocardial CCL2 was significantly elevated when compared to the 12:12LE group. Furthermore, TGF-B1 mRNA levels are reported to peak at 2 and 12 weeks post-myocardial damage in rats 34 . Similarly, a study of P. obesus which were also exposed to short photoperiods and HE diets showed an increase in TGF-B1 levels in fat www.nature.com/scientificreports/ tissue at 10 weeks 25 . This suggests the time points of this study were unable to capture changes in inflammatory and pro-fibrotic markers. Previous studies in P. obesus have found exposure to HE diet and/or circadian disruption increase myocyte hypertrophy 21,22,24 . Studies in rats and mice fed high fat diets have reported similar findings 30,32 . Consistent with this, the current study found that after 8-weeks there were increases in the hypertrophy genes Gata4, Merf2C and Nppa in the P. obesus exposed to the HE diet and short photoperiod. Despite this, we did not witness evidence of cardiomyocyte hypertrophy using histological analyses, suggesting that elevation in hypertrophy genes precedes physical increases cardiomyocyte size. In the 20-week cohort of P. obesus there was a reduction in cardiomyocyte size in the 5:19HE group, when compared to LE diet and/or normo-photoperiod (12:12). This was unable to be explained by reductions in hypertrophy gene expression as they remained unchanged. The most significant variation from previous studies is the length of time on HE diet and short photoperiod (16 weeks vs. 20 weeks), which in turn prolongs the overall period of abnormal glucose control in these animals. Stem cell recruitment to the myocardium, their survival and proliferation as part of the myocardial repair process are impaired by hyperglycaemia 35,36 . Extended abnormal glucose tolerance, as is the case for the 20-week group in this study, may tip the balance from an inflammatory-driven hypertrophy response into myocardial atrophy due to diminished stem cell recruitment/repair.
Per2 is one of the core clock genes that is used to determine circadian rhythmicity of the clock molecular machinery that is present in every cell. In the 20-week cohort, Per2 levels were significantly lower in the 5:19HE group, when compared to the 12:12HE group. Interestingly, the 5:19HE group also had pronounced myocardial perivascular fibrosis. Our findings of a decline in Per2 are consistent with previous studies that found deletion of Per2 in rodent models induces deleterious cardiovascular consequences 5,37,38 . The daily average of Per2 gene expression did not change after 8 weeks, although there were alterations in the daily rhythm in Per2 gene expression that were diminished in the 5:19HE group. This response was mimicked by the inflammatory genes Rela and CCl2 and fibrotic gene Tgfb1, in which the daily circadian rhythm of their expression was abolished in the 5:19HE group. CCl2 is controlled by the core clock protein Bmal1 39 , therefore diminished rhythms in Per2 will result in diminished rhythms in all clock-controlled genes, such as CCl2, consistent with the observations of this study.
In conclusion, using the well-characterised Psammomys obesus model of T2DM we simulated circadian disruption, similar to a shift worker, combined with a HE diet to assess the onset of features of T2DM and myocardial pathophysiology. We show that disrupted circadian cycles cause impaired glucose tolerance. Furthermore, fasting glucose levels were significantly increased with HE diet exposure and/or circadian disruption after 8 weeks. This onset in the features of T2DM with circadian disruption is found to coincide with the presence of myocardial perivascular fibrosis and an increase in Col1a1 expression. There were limited changes in inflammatory (Rela, CCl2) and pro-fibrotic (Tgfb1) genes, although their circadian rhythmicity was diminished with circadian disruption and HE diet. These findings provide insights into the potential adverse effects on the heart that may be experienced by shift workers due to T2DM that is induced by circadian dysfunction and amplified by a HE diet. Our studies may provide guidance for the prevention of diabetes-related cardiac fibrosis in shift workers in which a healthy diet and reduced time periods of circadian disruption should be promoted as a strategy for preventing cardiac dysfunction.

Methods
Animal experiments. This study was carried out in compliance with the ARRIVE guidelines. HsdHu diabetes-prone male sand rats (P. obesus, 6-7 months old) were initially maintained on a low energy diet (Koffolk Ltd, Israel) before assignment to experimental groups. All experimental procedures conformed to the NIH guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee (IACUC) of Tel Aviv University (#L15055). Adult male sand rats (n = 11-15/group) were exposed to the following conditions: (1) normal photoperiod, low energy diet (12:12LE); (2) normal photoperiod, high energy diet (12:12HE); (3) short photoperiod, low energy diet (5:19LE) and (4) short photoperiod, high energy diet (5:19HE). Animals in the LE groups were fed a special low-energy diet (14% protein, 1.7% fat, 15.4% Crude fibre; Product No.: 1078; Koffolk Ltd., Israel), which is the caloric equivalent to their native diet. Animals in the HE groups were fed the standard rodent diet (21% protein, 4% fat, 4% Crude fibre; Product No.: 2018; Koffolk Ltd.), which contains a higher caloric density and known to drive T2DM development in P. obesus 20 . In week 10, sand rats were weighed and then euthanized at four different time points post-'lights on' (ZT0), followed by ZT2 and then at 6 h intervals of ZT8, ZT14, ZT20. When sacrificing the animals during the dark hours (ZT14 and ZT20), the head of the sand rat was covered with aluminium foil until the brain was excised to prevent exposure to light. Hearts were excised, weighed and snap frozen for histological and RT-qPCR analyses.
In a parallel cohort (n = 12/group), P. obesus were maintained with the same diet and photoperiod exposure conditions for a longer time period of 20 weeks. Euthanasia was conducted two hours following 'lights on' (ZT 2) and their hearts were excised, weighed and snap frozen for histological and RT-qPCR analyses.
All animals had fasting glucose concentrations assessed using a glucometer. Oral glucose tolerance tests (OGTT) were performed in animals fasted for 4 h. Blood was collected from the tip of the tail at time 0 and 120 min 21 . Tests were performed at ZT 2 by administering 2 g glucose/kg body weight dissolved in 1 ml water using gastric gavages, inserted through the mouth into the stomach.
Histological analysis. The apex of snap-frozen hearts were formalin fixed and paraffin embedded then sectioned on a microtome at a thickness of 5 µm. Two sections were taken ~ 50 µm apart and stained with Masson's Trichrome (Sigma). Heart vessels were chosen randomly within the heart sections. An individual vessel was then photographed under × 400 magnification, with five photos per section selected on two sections (n = 10 images per animal). Perivascular cardiac fibrosis was determined by calculating a ratio for the area of collagen (blue) www.nature.com/scientificreports/ surrounding the vessels divided by the vessel area. Similarly, two sections were taken ~ 50 µm apart and stained with Haemotoxylin and Eosin (Fronine, Thermofisher) then imaged at two separate locations within the heart (total 4 images per animal) to define cell morphology. To determine cardiomyocyte size, the area was calculated by dividing the area of the Eosin stain (Cytoplasm) by the number of haematoxylin stained nuclei. We focussed our analyses on regions of cardiomyocytes that had centralized nuclei. All images were taken using an AxioLab microscope attached to a camera (Zeiss) then analysed with Image-Pro Premier 9.2 (64 bit).