Hibernating brown bears are protected against atherogenic dyslipidemia

To investigate mechanisms by which hibernators avoid atherogenic hyperlipidemia during hibernation, we assessed lipoprotein and cholesterol metabolisms of free-ranging Scandinavian brown bears (Ursus arctos). In winter- and summer-captured bears, we measured lipoprotein sizes and sub-classes, triglyceride-related plasma-enzyme activities, and muscle lipid composition along with plasma-levels of antioxidant capacities and inflammatory markers. Although hibernating bears increased nearly all lipid levels, a 36%-higher cholesteryl-ester transfer-protein activity allowed to stabilize lipid composition of high-density lipoproteins (HDL). Levels of inflammatory metabolites, i.e., 7-ketocholesterol and 11ß-prostaglandin F2α, declined in winter and correlated inversely with cardioprotective HDL2b-proportions and HDL-sizes that increased during hibernation. Lower muscle-cholesterol concentrations and lecithin-cholesterol acyltransferase activity in winter suggest that hibernating bears tightly controlled peripheral-cholesterol synthesis and/or release. Finally, greater plasma-antioxidant capacities prevented excessive lipid-specific oxidative damages in plasma and muscles of hibernating bears. Hence, the brown bear manages large lipid fluxes during hibernation, without developing adverse atherogenic effects that occur in humans and non-hibernators.

www.nature.com/scientificreports/ interval: 2,636 to 2,877) individuals in 2018 50 . The hunting period for bears begins on 21 August and ends when the area-specific quota has been filled, usually mid-to late September 51 , and can overlap with the pre-denning period 10 . Most den abandonments occur early in the denning season: a recent study documented that 22% of bears changed dens during winter, but only 4% after mid-December 52 .

Animals and sample collection. All personnel in the Scandinavian Brown Bear Research Project (SBBRP)
have advanced experience and training in capturing and handling free-living brown bears during all seasons. Brown bears are captured annually by the SBBRP and fitted with neck collars, which include a global positioning system ('GPS'), dual-axis motion sensors (to monitor activity), very-high-frequency ('VHF') transmitters, and a global system for mobile mobilization ('GSM') modem (Vectronic Aerospace GmbH, Berlin, Germany). As a backup to relocate bears if the collar malfunctioned, VHF transmitters were implanted into the abdomen (Telonics, Inc, Mesa, Arizona, USA) 53 . GPS positions are recorded every 30 min to 1 h. Bears that were the offspring of marked females were followed from birth; otherwise, age was determined by counting the annuli of a cross-section of the premolar roots 54 . Fourteen bears, including two animals studied during two consecutive years (for more details see Table 1), were included in this study. During winter, bears hibernate at moderate (30-36 °C) hypothermia 9 , and the torpid state hence corresponds to their entire hibernation from November to April 10 . Body temperature measurements of the bears were performed during hibernation in winter and during the active phase in summer (see Table 1 for details). During hibernation, torpid animals were captured in February 2011, 2012, and 2013 by darting them directly in their den, as previously described 55 . The same individuals were re-captured when active in June 2011, 2012, and 2013 by darting from a helicopter 56 . In winter, the anesthesia consisted of a combination of medetomidine-ketamine-tiletamine-zolazepam 10,55 , whereas bears in summer were anesthetized with a mix of medetomidine-tiletamine-zolazepam 53 . Once anesthetized, we took each bear out of the hibernating den in winter or located it on the ground in the case of darted individuals from a helicopter in summer, and placed it on an insulated blanket to conduct the experiments and to collect tissue samples. The same tissues, i.e., blood (~ 20-30 mL) and skeletal muscle (Vastus lateralis) biopsies (~ 250-350 mg), were taken from these bears during both seasons and were used to perform the biochemical analyses. Blood samples were collected, approximately 20-30 min after darting, from the jugular vein in heparinized tubes (BD Vacutainer, Lithium Heparin 34 IU, Plus Blood Collection Tubes 2.0 mL), then plasma was isolated by centrifugation within one hour at 3,500 rpm at 4 °C for 10 min. Plasma and muscle samples were snap-frozen and stored at − 80 °C for subsequent analyses.  Table 1. Summary of the brown bears used in the study and their individual physiological parameters. We applied linear mixed-effects models to test for the effect of season (fixed variable) on the different lipoproteins, lipid groups, enzymatic activities, inflammatory processes, antiradical resistance, and oxidative stress markers (predicted variables) in each tissue, i.e. muscle tissue ('M'), and blood plasma ('P'). Bear's ID was included as random effect for taking repeated measurements among animals into account.  from summer active ('summer') and winter hibernating ('winter') brown bears. In contrast to summer active individuals, winter hibernating bears were all torpid as indicated by their body temperature during hibernation (see Table 1). The lipid main categories are non-esterified cholesterol ('CHNE'), total cholesterol ('CHT'), nonesterified fatty acids ('NEFA'), phospholipids ('PL'), and triacylglycerols ('TAG'). Error bars represent standard errors. Winter levels differing significantly from their respective summer level are denoted by subscripts (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 2.
Muscle levels of neutral lipids according to season. Proportions -% of total neutral lipids -of triacylglycerols ('TAG'), total cholesterol ('CHT'), and non-esterified cholesterol ('CHNE') in skeletal muscle of summer active ('summer') and winter hibernating ('winter') brown bears. In contrast to summer active individuals, winter hibernating bears were all torpid as indicated by their body temperature during hibernation (see Table 1). Error bars represent standard errors. Winter levels differing significantly from their respective summer level are denoted by a subscript (**p < 0.01). www.nature.com/scientificreports/ , and high-density lipoproteins ('HDL'). Black arrows between lipoproteins indicates the formation pathway from VLDL to IDL, and the circular arrow represents the lipid cycling from peripheral tissues leading to HDL formation. Concentrations of the different lipid groups, i.e., cholesteryl esters, free cholesterol, total cholesterol, phospholipids and triacylglycerols, were determined for each of the lipoprotein classes from summer active ('summer') and winter hibernating ('winter') brown bears. The lower panels show the different proportions -% of total lipoprotein class -of (B) high-density lipoproteins 2a ('HDL2a'), 2b ('HDL2b'), 3a ('HDL3a'), 3b ('HDL3b'), and 3c ('HDL3c'), and (C) high-density lipoprotein subunits with a diameter greater than 12.9 nm ('HDL 12-9 nm') that were determined in blood plasma from summer active ('summer') and winter hibernating ('winter') brown bears. (D) The ratio between low-density and high-density lipoproteins ('LDL/HDL') is further indicated. Lipoprotein particles associated with anti-atherogenic effects are indicated in green, whereas pro-atherogenic factors are indicated in red-Bordeaux. In contrast to summer active individuals, winter hibernating bears were all torpid as indicated by their body temperature during hibernation (see Table 1). Error bars represent standard errors. Winter levels differing significantly from their respective summer level are denoted by subscripts (*p < 0.05, **p < 0.01, ***p < 0.001).  57 and electrophoresis using non-denaturing polyacrylamide gradient gels [58][59][60] , respectively (read the sections 'Determination of plasma and lipoprotein lipid composition' and 'Electrophoretic separation of high-density lipoprotein sub-fractions' in supplementary methods for more details). Hence, for each lipoprotein, we identified and quantified the main groups of composing lipids: non-esterified cholesterol ('CHNE'), cholesteryl-esters ('CE'), CHT, free cholesterol ('CHF'), NEFA, phospholipids ('PL'), and TAG. We also quantified the same lipid molecules, along with the concentration of ß-hydroxybutyrate ('ßOHB') using specific kits, in blood plasma. We additionally determined plasma enzymatic activities of cholesteryl-ester transfer protein ('CETP'), lecithin-cholesterol acyltransferase ('LCAT'), and phospholipid transfer protein ('PLTP') via the use of commercial enzymatic kits (see supplementary methods for further details). Using mass spectrometry, we further assessed pro-inflammatory and pro-oxidative properties of the blood by measuring plasma levels of three oxysterols, i.e., 7-ketocholesterol, 7α-hydroxycholesterol In contrast to summer active individuals, winter hibernating bears were all torpid as indicated by their body temperature during hibernation (see Table 1). Error bars represent standard errors. Winter levels differing significantly from their respective summer level are denoted by subscripts (**p < 0.01, ***p < 0.001). . Also, we assessed the antiradical resistance of the blood, i.e., the half-time of the hemolysis of red blood cells ('HT50') exposed to a controlled free radical attack, and various antiradical defense reserves ('RESEDA-1' , 'RESEDA-2' , 'RESEDA-3') within plasma by using the biological KRL/RESEDA™ test from Kirial International/Laboratories Spiral, Couternon, France 61-64 (see supplementary methods for further details). Finally, we measured plasma and muscle levels of oxidative stress markers, malondialdhyde (MDA)-protein adducts and protein-carbonyls via the use of commercial ELISA kits (see supplementary methods for further details).

Statistical analyses.
Data analyses were performed using R 3.4.4 65 . We used linear mixed-effects models ('lme' in library 'nlme' 66 ) to test for the effect of season (fixed variable) on the various predicted variables, considering repeated measurements among animals by including bear's ID as random effect, with a compound sym- www.nature.com/scientificreports/ metry (CS) as a structure for covariance (equivalent to paired Student's t-test). Standardized residuals from statistical models were tested for normality using Kolmogorov-Smirnov tests, and data were either log-or Box-Cox transformed when necessary. Predicted variables included (i) the different lipoprotein categories, sub-classes or ratios, (ii) the different lipid groups, (iii) the plasma enzymatic activities, (iv) the concentrations of oxysterols and isoprostanes, (v) the antiradical resistance of red blood cells or the various antiradical defense reserves, and (vi) the plasma levels of oxidative damages. Initial inspection of the data gave no evidence for an effect of sex or sampling year on any of predicted variables. Differences of means between seasons were assessed. Values are means ± SE and means of log-or Box-Cox transformed differences ± SE (except for percentages of total neutral lipids, which are actual log-transformed differences). We further performed regression models between levels of 7-Ketocholesterol or 11ß-PGF2α and proportions of HDL 12-9 nm, HDL3c and HDL2b particles by using similar linear mixed-effects with bear's ID as random effect to consider repeated measures. We also applied linear mixed-effects models to test for potential associations between the antiradical resistance of red blood cells and the specific markers of either lipid peroxidation (MDA-protein adducts) or protein oxidation (protein carbonyls) in muscle. Further, we computed (conditional) R-squared values accounting for the random effect in linear mixed-effects models ('r.squaredGLMM' in library 'MuMIn' 67 ). P-values < 0.05 were considered as significant.

Results
Plasma and muscle lipid levels. Plasma concentrations of the main lipid categories and molecules were higher in hibernating brown bears than summer-active individuals. Specifically, we found significant higher plasma levels of CHNEs (53% increase), CHT (42% increase), NEFAs (67% increase), PLs (24% increase), and TAGs (twofold increase) during hibernation (Fig. 1A, Supplementary Table S1). In addition, plasma concentration of ßOHB was tremendously higher, by 120-fold, in brown bears during winter hibernation than the summer active season (Fig. 1B, Supplementary Table S1). Muscle concentrations of neutral lipids were also different in brown bears during winter hibernation compared to the active state in summer. Hibernating bear muscles showed a 77% higher level of TAGs and a 2.1-fold lower level of CHT compared to muscles from summer active individuals (Fig. 2, Supplementary Table S1). No significant seasonal change was detected in the level of CHNEs in muscle (Fig. 2).
Lipoprotein composition, subclasses and sizes. Composition of lipoproteins was substantially different in brown bears during winter hibernation compared to the summer active phenotype. We found significant increases of TAG levels by 9, 1.5, 3.2, and 1.4-fold within all lipoprotein classes, namely VLDL, LDL, IDL, and HDL respectively, from winter hibernating bears compared to summer active animals (Fig. 3A, Supplementary  Table S2). Further, concentrations of CEs, CHF, CHT, and PLs increased by 8.8, 1.5, 2.8-fold on average in VLDL, LDL, and IDL respectively, but no significant changes were detected in the HDL subunit between winter hibernating and summer active bears (Fig. 3A, Supplementary Table S2).
Similar to lipoprotein composition, subclass proportions and sizes of lipoproteins were also altered, although in different ways between classes/sizes, in hibernating brown bears during winter compared to active individuals In contrast to summer active individuals, winter hibernating bears were all torpid as indicated by their body temperature during hibernation (see Table 1). Error bars represent standard errors. Winter levels differing significantly from their respective summer level are denoted by a subscript (*p < 0.05, **p < 0.01, ***p < 0.001). www.nature.com/scientificreports/ in summer. Plasma proportions of all subclasses of HDL decreased significantly by 20%, except HDL2b, which is known to be cardioprotective, that increased by 40% in hibernating bears versus summer active animals (Fig. 3B, Supplementary Table S3). Further, we observed a higher plasma proportion of large-size HDL particles, i.e., those with a diameter greater than 12.9 nm, in brown bears during winter hibernation than during the summer active period (Fig. 3C, Supplementary Table S3). No significant seasonal change was detected in the ratio between LDL and HDL (LDL/HDL) (Fig. 3D, Supplementary Table S3).
Enzymatic activity levels. Activities of enzymes involved in lipid and lipoprotein metabolisms were substantially modified in hibernating bears during winter compared to active individuals in summer. We found a significant 36% increase in the activity of plasma CETP and a significant 40% reduction of the activity of plasma LCAT in hibernating brown bears versus summer active animals (Fig. 4A,B, Supplementary Table S4). No significant seasonal change was detected for the activity of PLTP (Fig. 4C,F, Supplementary Table S4).  www.nature.com/scientificreports/ winter samples, and for 7α-hydroxycholesterol in 8 of 16 summer samples and in all winter samples but one. On the other hand, 11ß-PGF2α was nicely detected in all samples. We found significantly lower plasma concentrations of both the oxysterol, 7-ketocholesterol, and the isoprostane, 11ß-PGF2α, in hibernating brown bears during winter compared to active individuals in summer (Supplementary Table S5). Interestingly, plasma levels of 7-ketocholesterol and 11ß-PGF2α correlated negatively with HDL 12-9 nm and positively with proportions of HDL3c small particles across all individual brown bears (Fig. 5, see legend for specific statistics and regression coefficients). Further, the level of 11ß-PGF2α was negatively associated with proportions of HDL2b, known to be cardioprotective, but not with that of 7-ketocholesterol (Fig. 5).

Oxysterols and isoprostanes.
Blood antioxidant defenses and oxidative stress markers. The overall antioxidant capacity of blood was altered according to the seasonal status of the brown bear. We found a significant 20% greater half-time of hemolysis (HT50; KRL test), i.e., a higher antiradical resistance of red blood cells during winter hibernation (Fig. 6A, Supplementary Table S6). Interestingly, RESEDA-2, the second contingent of antiradical defense reserves corresponding to sulfatases, showed a significant increase of 27%, which was paralleled by a 40% reduction of RESEDA-3, composed of glucuronidases, in winter hibernating bears compared to summer individuals (Fig. 6A, Supplementary Table S6). We observed no significant seasonal difference in RESEDA-1, i.e., glucosidases, between seasons.
We found that antiradical resistance of red blood cells (half-time of hemolysis of the KRL) was negatively correlated with a marker of lipid peroxidation, i.e., malondialdehyde (MDA)-protein adducts, but was not significantly associated with a marker of protein oxidation, i.e., protein carbonyls (Fig. 6B,C, read legend for specific statistics and regression coefficients). Muscle level of MDA-protein adducts was significantly lower in brown bears during winter hibernation vs. summer active season (Supplementary Table S7). Further, we observed a trend to lower level of protein carbonyls in hibernating bears compared to summer individuals (Supplementary Table S7). In contrast, plasma level of MDA-protein adducts, but not that of protein carbonyls, was significantly higher by 30% in winter hibernating brown bears compared to summer active individuals (Supplementary Table S7).

Discussion
The present study investigates and highlights (as illustrated in Fig. 7) the unique ability of a large hibernator, the brown bear, which does not excrete excess cholesterol (by defecating and urinating) during hibernation, to challenge the handling of large fluxes of lipids while fasting during several months in winter. Here we report that bears display a plasma lipid profile, i.e., increased plasma levels of NEFA and TAG (Fig. 1), typical of a phase 2 fast with lipolysis providing the main source of energy during winter. Such an increase in lipid mobilization reflects the occurrence of a switch in substrate metabolism toward the sparing of muscle proteins during hibernation (for review, see 68 ). Our results further highlight that hibernating bears handle fluxes of TAGs and CHT via futile cycles and re-esterification through lipoprotein metabolism. Specifically, the lipid composition of HDL particles remains stable, while solely HDL2b subunit, known to be cardioprotective, increases during hibernation. In addition, hibernating bears increase blood antioxidant capacities and selectively mobilize SFAs, which are less prone to peroxidation than unsaturated fatty acids, to limit oxidative damages associated to higher lipid fluxes.
Winter fast does not affect the composition of high-density lipoproteins, of which subclass levels decrease in hibernating brown bears. During fasting, fatty acids are continuously released from white adipose tissue for oxidation or recycled by the liver, leading to the efflux of cellular cholesterol pools into plasma and contributing to the rise in cholesterol levels in lipoproteins, notably in HDL 69 . In contrast to fasted rodents and other hibernators, we observed a stable lipid composition of HDL with only a modest increase in TAG concentrations in hibernating brown bears compared to active individuals in summer (Fig. 3A). Such limited alterations of HDL composition in hibernating brown bears suggest the existence of an active process of re-esterification of lipids and cholesterol through a futile cycle during the hibernation of brown bears. This hypothesis is notably supported by higher levels of enzymatic activity of the CETP, i.e., cholesteryl-ester transfer protein, found in the plasma of brown bears of this study during hibernation compared to the active individuals in summer (Fig. 4A,D). Indeed, CETP, also called plasma lipid transfer protein, facilitates the transport of CEs and TAGs between lipoproteins, by collecting TAGs from VLDL or LDL in exchange to CEs from HDL, and vice versa (for review, see 70 ). This enzyme is then directly involved in regulating the amount of circulating HDL-cholesterol within the plasma, and lower levels of CETP promote HDL formation. In particular, genetic deficiency in CETP leads in humans to a drastic rise of HDL-cholesterol [71][72][73] , and it is the most common cause of hyperalphalipoproteinemia, i.e., elevated HDL, in Japanese people 74 . Hyperalphalipoproteinemia is particularly associated with risks of cardiovascular diseases, including coronary artery disease (CAD). Hence, the increased activity of CETP observed in hibernating brown bears likely contributes to a re-esterification of cholesterol and TAGs via a futile cycle between HDL and VLDL/LDL during the winter fast. However, because of a stable, but not reduced, HLD-cholesterol, one could also argue for an increased hepatic uptake of cholesterol in hibernating brown bears during winter. The occurrence of such a process cannot be excluded and would lead, in parallel to a recycling of CEs, to an accumulation of cholesterol into the liver of hibernating brown bears (for overview, see Fig. 7).
Plasma CETP activity is also closely related to HDL subclass distribution. In particular, CETP enhances HDL remodeling from large (> 10 nm) to small (< 8 nm) HDL subclasses, and human CETP deficiency is associated with a predominance of large HDL particles displaying increased capacity to mediate cellular cholesterol efflux [75][76][77] , a marker of cardiovascular disease risk. Our findings in hibernating brown bears revealed a substantial decrease in proportions of nearly all subclasses of HDL, along with an increase in HDL particle size, as indicated by a higher proportion of HDL subunits of diameter greater than 12.9 nm, i.e., HDL 12-9 nm (Fig. 3B,C). www.nature.com/scientificreports/ Interestingly, plasma proportion of large HDL2b particles was increased and that of small HDL3c subunits was reduced in hibernating brown bears during winter compared to the summer bears phenotype. CAD severity in humans is known to be positively associated with the levels of large VLDL and small HDL particles, such as HDL3c, across 158 men 78 . Further, large HDL2b particles have known specific effects for cardiovascular protection (for review see 79 ). Freedman and colleagues 78 also reported that increased levels of both large (10 to 13 nm) HDL, i.e., HDL2b, and intermediate (8.2 to 10 nm) HDL, i.e., HDL2a, are associated with reduced CAD severity in humans. In hibernating bears, together with the increased proportion of HDL2b particles and decreased proportion of HDL3c, our findings of reduced levels of 7-ketocholesterol (oxysterol) and 11ß-PGF2α (isoprostane), metabolites with inflammatory and pro-oxidative actions promoting atherosclerosis and cardiovascular diseases 45,46 , and their negative associations with HDL particles size including the large HLD2b subunit (Fig. 5), suggest a cardiovascular protective state. This is in accordance with the fact that hibernators are particularly resistant to cardiovascular malfunctions and sudden cardiac arrest during the torpor state (for reviews, see [80][81][82], which is especially true for brown bears during winter hibernation [83][84][85][86] . Adaptations of lipoprotein metabolism, including the maintenance of 'healthy' large HDL-cholesterol, would likely contribute to the outstanding ability of bears to avoid cardiovascular dysfunctions during the hibernating state in winter. Taken together, our results suggest the existence of processes of re-esterification via a futile cycle or of reuptake of lipids, especially of cholesterol, by the liver in brown bears during their hibernation in winter. In particular, the conversion of cholesterol to bile acid by the liver is nearly fully arrested, as supported by transcriptional suppression of genes involved in cholesterol metabolism in liver of hibernating American black bears 87 . Hence, the constant supply of fatty acids to fuel hibernation and the inability of hibernating bears to excrete excess cholesterol during winter imply additional specific mechanisms at the synthesis level and/or release rate from peripheral tissues.

Tight control of circulating amounts of cholesterol during the hibernation in brown bears.
Within the organism, plasma concentration of cholesterol depends on the balance between fluxes of cholesterol input, via dietary absorption and synthesis/release by tissues, into and excretion from the body. Because fat-storing hibernators, including brown bears, do not eat during winter, the hibernation eliminates dietary cholesterol intake, making de novo synthesis the only way for new cholesterol to be added to body tissues. Because hibernating bears do not defecate nor urinate 88,89 , excess cholesterol cannot be excreted during the entire winter. Our findings of a stable concentration of cholesterol in HDL of hibernating brown bears (Fig. 3A) suggest the occurrence of (i) a tight control of cholesterol synthesis at the cellular level, (ii) a reduced rate of release of cholesterol from peripheral tissues, and/or (iii) an increased reuptake of cholesterol by the liver. However, the lower concentrations of CHT in bear muscles (Fig. 2) during hibernation strongly suggest that the synthesis and/or release rates of cholesterol are lowered at peripheral tissues in hibernating bears. Fasting lowers the hepatic and intestinal synthesis of cholesterol in rodents 90,91 , and the cholesterol synthesis rate of fat cells is suppressed by 90% during fasting in rats, Rattus norvegicus 92 . Further, the synthesis rate of cholesterol has been found to be slightly inhibited by the presence of serum VLDL or LDL particles in the medium, suggesting the implication of a regulatory feedback mechanism on the cholesterol synthesis rate 93 . Because the synthesis of cholesterol requires the mobilization of energy, a reduction of the cholesterol synthesis rate during hibernation might contribute to the energy saving of hibernating brown bears in winter.
Along with decreased muscle CHT levels, we also found a significantly lower enzymatic activity of LCAT, i.e., lecithin-cholesterol acyltransferase, in brown bears during winter hibernation compared to the summer active state (Fig. 4B,E). This result points toward a reduction of the accumulation rate of cholesterol into HDL of hibernating brown bears during winter. The enzyme LCAT converts plasma CHF into CEs, which are then sequestered into the core of a lipoprotein particle, making new synthetized HDL spherical (for review, see 94 ). A decrease of LCAT activity is fully in line with the reduced synthesis or release rates of CEs and would contribute to the reduction of energy needs during the hibernating state. Comparative studies of LCAT activity in the plasma of several mammalian and ectothermic species, known to have lower daily and yearly metabolic rates, have revealed considerable inter-and intra-species variations 95 . In particular, the activity of LCAT was quantitatively and substantially lower in ectothermic species, such as snakes, amphisbaenids, and amphibians, compared to mammalian species 95 . In lizards during prolonged fasting, LCAT activity and lipid levels were reduced in plasma, and plasma lipoproteins had a lower proportion of lecithin to CHNE 95 . These studies clearly suggest significant relationships between LCAT activity level, lipoprotein metabolism, and metabolic rates of individuals. This would particularly be true for hibernating bears that can reduce their metabolism to 25% of basal metabolic rates, despite regulating their T b between 30° and 36 °C during winter hibernation 9 . In humans, deficiency of LCAT is associated with some pathologies, including eye disease or atherosclerosis 96 . In addition, HDL abnormalities associated to chronic kidney failure (CKF) are largely due to marked downregulation of LCAT activity [97][98][99] . In our study, we found that hibernating brown bears appear to be able to manage large lipids fluxes via modification of lipoprotein metabolism, including substantial reductions of LCAT activity, without developing any pathological state, such as CKF 100 , during winter.
All together, these results support the existence of a tight control of cholesterol synthesis and/or release from peripheral tissues, such as muscles, and a reduced cholesterol back transportation into HDL, to limit excessive accumulation of cholesterol within the body, with a possible retention by the liver, in hibernating brown bears during winter. Because lipids are prone to autoxidation, these strong adaptions of lipid fluxes and lipoprotein metabolism are expected to increase lipid-derived atherogenic components within the organism. Nevertheless, the unique adaptations of bears to long-term fasting suggest a strong resistance of the latter to oxidative stress associated with lipid handling during winter. www.nature.com/scientificreports/ Increased blood antioxidant capacities dampen lipid oxidative damages in hibernating brown bears. Along with a larger mobilization of lipids associated with the winter fast, our results revealed limited increases of oxidative damage related to lipid peroxidation in the plasma of hibernating brown bears compared to that of summer active individuals (Supplementary Table S7, Fig. 7 for overview). In particular, we observed an unaltered level of protein carbonyls along with a ~ 30% increase of MDA-protein adducts levels in hibernating brown bears compared to summer active individuals. This suggests the existence of significant reserves for antioxidant capacities in the blood of hibernating brown bears during winter (read below for further discussion). Like small heterotherms, hibernating brown bears preferentially mobilize SFAs, less prone to auto-oxidation than unsaturated fatty acids, to fuel energy needs during winter, and retain unsaturated fatty acids in their tissues 18 . Such a selective mobilization of lipids would contribute to limit the increase of oxidative stress damage observed in the plasma of hibernating brown bears during winter. Similarly, the heterothermic grey mouse lemur (Microcebus murinus) selectively increases the oxidation of SFAs in winter, while increasing torpor expression in response to food limitation 101 . This increase of SFAs mobilization was associated with no change of oxidative stress levels specific to unsaturated lipids, which were retained within the body tissues. Also, the relict South American marsupial, Dromiciops gliroides, shows differential expression of genes orchestrating different metabolic changes, including shift in lipid metabolism, in liver and muscle cells during hibernation 20 . Interestingly, hepatic and muscle transcripts of thioredoxin-interacting protein, a potent antioxidant, were found to be overexpressed in D. gliroides during torpor 20 .
Our findings further revealed a greater antiradical blood resistance (KRL test) in hibernating brown bears compared to summer active individuals (Fig. 6A). This increased resistance of the blood to oxidative threats was primarily mediated via the third component of antiradical defense reserves (RESEDA-3) composed of glucuronidases, which decreased substantially in hibernating brown bears during winter (Fig. 6A). This reduced level of RESEDA-3 indicates the mobilization and specific use of antioxidant glucuronides, which are dietary components found notably in berries 102 , on which the brown bears we studied heavily rely upon to increase their body energy reserves during the late summer and autumn prior to hibernation [103][104][105] . Interestingly, glucuronides can also be converted into sulfates, corresponding to the second component of antiradical defense reserves (RESEDA-2), which increased in hibernating brown bears during winter (Fig. 6A). These variations suggest an indirect contribution of glucuronides, via modulations of the sulfate pool, to the overall increased antioxidant capacities of brown bears during winter hibernation. Prior to and during winter, hibernators are known to increase their antioxidant capacities and defenses (for review, see 106 ). For instance, hibernating ground squirrels increase the protein levels of antioxidant enzymes, such as superoxide dismutase ('SOD') -1 and SOD2, as a potential mechanism to prevent ROS-mediated damage in white adipose tissue 107 . Moreover, the activity of antioxidant enzymes including SOD2 was reported to be increased in the brown adipose tissue of golden-mantled ground squirrels fed a high linoleic acid (18:2 n-6) diet prior to hibernation 108 . Altogether, these findings indicate that the increased blood antioxidant capacities likely contribute to dampen the oxidative damage associated with the large lipid mobilization in hibernating brown bears during winter.
In this study, we further found a significant negative relationship between blood antiradical resistance and a marker of lipid peroxidation (MDA-protein adducts) in muscles of hibernating brown bears (Fig. 6B). This association suggests a substantial contribution of greater antioxidant capacities of the blood in limiting oxidative damages in peripheral tissues, such as skeletal muscles. We have previously reported upregulations of cytosolic antioxidant molecules and the maintenance of the GSH/GSSG ratio in skeletal muscles of hibernating brown bears 109 . Along with these upregulations of antioxidant mechanisms in bear muscles, the increased antiradical defenses of the blood could also help to dampen the levels of oxidative threats to muscles during hibernation. Interestingly, neither tissue nor systemic indices of oxidative damage or inflammation increased with fasting in post-weaned pups of northern elephant seal (Mirounga angustirostris), a species adapted to cope with chronic periods of stress 110 . In this species, an increase in blood antioxidant capacities seems to be sufficient to suppress both systemic and tissue indices of oxidative damage during periods of prolonged fasting. Moreover, the innate immune system, associated with inflammatory processes, was found to be downregulated in brown bears during hibernation 111 . Further, we have recently demonstrated that brown bears also reduced levels of eicosanoids (molecules derived from polyunsaturated fatty acids), including those with pro-inflammatory properties, during hibernation 112 . Taken together, these findings likely help explain the limited increase of oxidative damage specific to lipid auto-oxidation in the blood and muscles of hibernating brown bears, despite the large increase of lipid fluxes during winter.

Limitations of the study
One limitation of the study resides in the descriptive aspect of this work that includes only two timepoints (i.e., summer and winter) within the year. Hence, 'season' is not necessarily the same treatment as 'hibernation' or 'fasting' , and their respective effects are then confounded due to the seasonal aspect of this study. Further, the amount of tissue that is possible to collect on bears in the field is limited; hence we could not perform more than a thorough analysis of lipoprotein dynamics, lipid profile, enzyme activities, and oxidative status, which already corresponds to a significant achievement. This collaborative work constitutes, however, a unique study, because it provides a rather full picture of the metabolism of cholesterol and other lipids in brown bears under free-living conditions. For this reason, no analyses of additional targets of interest, such as transporters of cholesterol or macrophages, could have been carried out. Nevertheless, this work is of major relevance because diets in laboratories or zoos fail to reflect the natural diet selection of free-living animals that constrain hibernation physiology and adaptation. In particular, diet is seasonally variable in bears in Scandinavia, as elsewhere 104,105 . In autumn, brown bears in Scandinavia build up fat reserves by mainly feeding on berries, such as from the Vaccinum family, that contribute most (49-81%) of the dietary energy content of the bears [103][104][105] , and also provide them with a www.nature.com/scientificreports/ major source of antioxidant molecules that directly impact on their hibernation physiology and performances, notably in regard to lipid metabolism and protein/muscle maintenance during hibernation (for review, see 68 ).

Conclusion
After approximately four months of hibernation, brown bears display a typical plasma lipid profile of a phase 2 fast with lipolysis providing the main source of energy during winter. As described in studies of fasting, lipolysis provides many times more energy than required to sustain life, leading to high contents of all of the lipid categories in low-and medium-density lipoproteins. Our results suggest that hibernating bears handle fluxes of TAGs and cholesterol via futile cycles and re-esterification through lipoprotein metabolism (for overview, see Fig. 7), keeping the lipid composition of HDL particles stable, while increasing the HDL2b subunit, which is known to be cardioprotective. Despite higher lipid fluxes, oxidative damages seem to be limited, if not reduced, due to greater antioxidant capacities of the blood and the selective mobilization of SFAs, which are less prone to peroxidation than unsaturated fatty acids. Hence, hibernating bears appear able to manage large fluxes of TAG and cholesterol without the classical adverse effects associated with prolonged fasting in humans and non-hibernators.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.