Over 2 billion people worldwide are estimated to have nonalcoholic fatty liver disease (NAFLD), defined by excess hepatic fat. Commonly, progression into nonalcoholic steatohepatitis (NASH) is characterized by the onset of liver inflammation following exacerbated steatosis.1 This suggests that reducing liver inflammation (e.g. through lifestyle interventions involving exercise training) may come secondary to a reduction of steatosis. However, both the metabolic and inflammatory processes involved in NAFLD are under circadian control and could respond differently to exercise at different times of day.2
To investigate the time-of-day–dependent effect of exercise training on NAFLD amelioration in the early disease stages we trained high-fat high-cholesterol (HFHC)-fed APOE*3-Leiden cholesteryl ester transfer protein (CETP) mice during their early or late active period. This mouse model was chosen because of its humanized lipid metabolism and its ability to develop all hallmarks of human NAFLD upon HFHC feeding.3 The animals were treadmill trained five times per week for 8 weeks at either Zeitgeber time (ZT)13 (E-RUN) or ZT22 (L-RUN). Corresponding sedentary animals (E-SED and L-SED) were put into empty cages at the same time to control for experiment-induced stress. After 8 weeks, all mice were killed at the same circadian time (ZT17) on the day after the last exercise training bout to allow for comparisons between all four groups and to reduce the confounding effect of the acute exercise (Supplementary File 1).
At the end of the study, body weight (Fig. 1A) and lean body mass (Supplementary Fig. 1A) were similar in all mice, but trained mice had gained less fat mass than sedentary mice (Fig. 1B), indicating a measurable exercise effect. Fasting plasma glucose, which when elevated is independently positively associated with the risk of developing NAFLD,4 was unchanged among the groups (Fig. 1C). Furthermore, no differences in hepatic steatosis, NAFLD activity score, plasma alanine aminotransferase (ALT) levels, portal inflammation and liver weight were observed between the groups (Fig. 1D–F and Supplementary Fig. 1D, E), likely owing to an overall limited potential to improve these parameters in early stages of steatosis without signs of NASH. Accordingly, liver triglyceride (TG), total cholesterol (TC), and phospholipid (PL) levels (Fig. 1G–I) as well as plasma TG and TC (Supplementary Fig. 1B, C) levels remained unchanged in the exercising and sedentary groups regardless of the time of training.
Surprisingly, exercise training had a time-of-day specific impact on liver inflammation, challenging the notion that hepatic inflammation merely follows the level of steatosis. In livers collected at the same circadian timepoint 1 day after the last training, flow cytometry analysis of isolated hepatic immune cells revealed an unexpected hepatic increase in the total number of leukocytes, neutrophils, and monocytes in response to early training that did not reflect increased blood immunocyte levels (Fig. 2A–C and Supplementary Fig. 2C–F). Notably, blood leukocytes were even significantly decreased in E-RUN compared to E-SED at the same time liver leukocytes were elevated (Supplementary Fig. 2C), consistent with migration of these cells to the liver. Late training, however, had no effect on the immune cell populations in blood or liver. The increase of specific cell populations following early training may indicate disease acceleration, as infiltrating neutrophils are associated with NAFLD development and disease progression.5,6 In line with that, infiltrating monocytes, which are recruited to the liver partly through hepatocyte-derived stress signals such as interleukin (IL)-1β and tumor necrosis factor alpha (TNFα), differentiate into proinflammatory macrophages that contribute to tissue damage and the loss of resident macrophages.7 Interestingly, early training also increased the number of natural killer (NK) cells in the liver (Fig. 2D). Although the contribution of these cells to NAFLD development and progression to NASH remains controversial, they can produce large quantities of proinflammatory cytokines such as interferon gamma.8 Taken together, early training led to an inflammatory response in the liver characterized by an increase of proinflammatory and tissue damage-associated cell populations.
An increase in hepatic inflammation following early training was also confirmed by gene expression analysis in isolated hepatic immune cells. Gene expression of the secreted proinflammatory factors IL-1β (Il1b) and TNF-α (Tnf) was increased after early training but not after late training (Fig. 2E, F). Similarly, the expression of the macrophage marker F4/80 (Adgre1) was increased following early training (Fig. 2G). In line with that, early training also increased the expression of Tnf, Il1b and Adgre1 in whole liver tissue (Fig. 1D–F).
It is not clear whether the observed increase of liver inflammation with early training is beneficial or detrimental in NAFLD development. As the number of circulating immune cells as well as their activity exhibits a circadian rhythm in mice and humans, exercise in the early active phase may stimulate cell migration into the liver at the time these cells are most prone to migrate into peripheral organs.9 Consequently, one could speculate that by stimulating liver inflammation in developing steatosis, early training activates a rapid alert system that supports disease resolution. Conversely, it has been shown that early exercise can acutely worsen metabolic diseases as seen in people with obesity and type 2 diabetes where early high intensity cycling elicited unfavorable blood glucose spikes that did not occur with late exercise.10 Accordingly, our findings could indicate that early training accelerated disease progression while late exercise training potentially targets liver steatosis and inflammation at a later disease stage. However, while not affecting liver lipid levels, the hepatic gene expression of Srebp1c (Srebf1), the mediator of insulin-induced fatty acid synthesis, was downregulated with both early and late training (Supplementary Fig. 1G), suggesting that the regulation of metabolic and inflammatory disease drivers may not be synchronized. Future studies need to assess the translatability of our findings to advanced disease stages and to human NAFLD. Notably, we observed distinct inflammatory modulation already at an early disease stage with a low NAFLD score, low grade hepatic steatosis and before the disease becomes inflammation-driven. This may present a previously underappreciated inflammation-targeted treatment opportunity in a large part of the population at risk for NASH. In summary, we showed that early and late exercise training in a mouse model of NAFLD differently influenced liver inflammation in developing steatosis. An unexpected increase in liver inflammation was observed with early exercise training.
Supporting information
Supplementary File 1
Supplementary Materials and Methods.
(DOCX)
Supplementary Fig. 1
Lean mass changes, plasma lipids, plasma ALT levels, portal inflammation, and liver gene expression after 8 weeks of early or late exercise training.
Lean mass was measured before and after 8 weeks of exercise training (A). Plasma triglycerides (B), total plasma cholesterol (C), plasma alanine aminotransferase (ALT) levels (D), portal inflammation after Knodell et al.,7 (E) and liver gene expression of Tnf (D), Il1b (E), Adgre1 (F) and Srebf1 (G) were assessed after 8 weeks. Gene expression is shown relative to E-SED. *p<0.05, **p<0.01 in one-way analysis of variance, n=9–18. E-SED, early sedentary; E-RUN, early running; L-SED, late sedentary; L-RUN, late running; ALT, alanine aminotransferase; Tnf, tumor necrosis factor; Il1b, interleukin-1β; Adgre1, adhesion G Protein-Coupled Receptor E1; Srebf1, sterol regulatory element-binding transcription factor 1.
(TIF)
Supplementary Fig. 2
Gating strategy for the analysis of liver and blood immune cells and circulating immune cell levels after 8 weeks of early or late exercise training.
Isolated hepatic immune cells from livers collected at ZT17 were sorted using spectral flow cytometry and gated as shown (A). Blood immune cells isolated from heart puncture blood collected at the same time were sorted and gated as shown (B). Circulating leukocytes (C), neutrophils (D), monocytes (E) and NK cells (F) were quantified and are shown per mL blood. *p<0.05, in one-way analysis of variance, n=7–9. E-SED, early sedentary; E-RUN, early running; L-SED, late sedentary; L-RUN, late running; NK, natural killers (cells).
(TIF)
Abbreviations
- Adgre1:
adhesion G protein-coupled receptor E1
- ALT:
alanine aminotransferase
- CETP:
cholesteryl ester transfer protein
- E-RUN:
early running
- E-SED:
early sedentary
- HFHC:
high-fat high-cholesterol
- Il1b:
interleukin-1β
- L-RUN:
late running
- L-SED:
late sedentary
- NAFLD:
nonalcoholic fatty liver disease
- NASH:
nonalcoholic steatohepatitis
- NK:
natural killer (cells)
- PL:
phospholipid
- TC:
total cholesterol
- TG:
triglyceride
- Tnf:
tumor necrosis factor
- ZT:
Zeitgeber time
Declarations
Acknowledgement
We thank Trea Streefland and Reshma Lalai (Div. of Endocrinology, Dept. of Medicine, LUMC, Leiden, the Netherlands) for their excellent technical assistance. We also thank Lars Hoeve, Sjahnaaz Bholai and Jack Brouwer for their technical contributions.
Funding
This study was financed by a grant from the Novo Nordisk Foundation to MS (NNF18OC0032394). ZY was supported by a full-time PhD scholarship from the China Scholarship Council.
Conflict of interest
The authors have no conflict of interests related to this publication.
Authors’ contributions
Study concept and design (AK, ZY, PCNR, MS), acquisition of data (AK, ZY, JML, HJPZ, MS), analysis and interpretation of data (AK, ZY, JML, HJPZ, BG, PCNR, MS), drafting of the manuscript (AK, ZY, MS), critical revision of the manuscript for important intellectual content (JML, BG, PCNR), and study supervision (BG, PCNR, MS). All authors have made a significant contribution to this study and have approved the final manuscript.