Leptin stimulates autophagy/lysosome-related degradation of long-lived proteins in adipocytes

ABSTRACT Obesity, a condition most commonly associated with hyper-leptinemia, is also characterized by increased expression of autophagy genes and likely autophagic activity in human adipose tissue (AT). Indeed, circulating leptin levels were previously shown to positively associate with the expression levels of autophagy genes such as Autophagy related gene-5 (ATG5). Here we hypothesized that leptin acts in an autocrine-paracrine manner to increase autophagy in two major AT cell populations, adipocytes and macrophages. We followed the dynamics of autophagosomes following acute leptin administration with or without a leptin receptor antagonist (SMLA) using high-throughput live-cell imaging in murine epididymal adipocyte and macrophage (RAW264.7) cell-lines. In macrophages leptin exerted only a mild effect on autophagy dynamics, tending to attenuate autophagosomes growth rate. In contrast, leptin-treated adipocytes exhibited a moderate, ~20% increase in the rate of autophagosome growth, an effect that was blocked by SMLA. This finding corresponded to mild increases in mRNA and protein expression of key autophagy genes. Interestingly, a long-lived proteins degradation assay uncovered a robust, >2-fold leptin-mediated stimulation of the autophagy/lysosome-related (bafilomycin-inhibited) activity, which was entirely blocked by SMLA. Collectively, leptin regulates autophagy in a cell-type specific manner. In adipocytes, autophagosome dynamics is moderately enhanced, but even more pronounced stimulation is seen in autophagy-related long-lived protein degradation. These findings suggest a causal link between obesity-associated hyperleptinemia and elevated adipocyte and AT autophagy-related processes.


Introduction
Dys-regulation of autophagy, the evolutionarily-conserved processes by which cells degrade intracellular components to maintain cell homeostasis, is increasingly considered in the pathophysiology of various diseases. 1 Although mostly dysfunctional (attenuated) autophagy is implicated in pathogenesis as a result of accumulated damaged or unnecessary cell components, over-activated autophagy is likely also pathogenic, and was proposed to contribute to the pathophysiology of several diseases, including chronic pulmonary disease and cancer. 2,3 In obesity, autophagy is dys-regulated in a cell-type/tissuespecific manner. It is considered to be attenuated in the liver, altered in beta-cells, and decreased in skeletal and cardiac muscles. 4,5 In adipose tissue, how autophagy is dys-regulated remains unsettled, with a general agreement that autophagy genes are overexpressed in adipose tissue, particularly in depots and in obesity sub-types that are more associated with metabolic dysfunction. [6][7][8] Yet, although autophagy, particularly in chronic diseases, may be regulated also at the gene expression level, 9,10 much more is known about its more rapid regulation by post-translational modifications and proteinprotein interactions. Indeed, while two independent groups suggested that autophagic activity (i.e., 'flux') is inhibited in adipose tissue, 11,12 at least four groups, including our own, provided evidence that the elevated gene expression of autophagy genes in adipose tissue manifests by increased autophagic flux. 6,[13][14][15][16] The molecular consequence of such elevated activity is not fully-mapped, but hyper-activated autophagy in adipose tissue may not only associate with, 17 but can also be causally linked to, decreased secretion of adiponectin. 17,18 This finding suggests a putative mechanistic link between activated adipose tissue autophagy in obesity, adipose tissue dysfunction, and a dys-metabolic obese phenotype. Some outstanding questions emerge from the above: (1) since adipose tissue is composed of various cell types, what is the cell-specificity of autophagy alterations in obesity? Indeed, adipose tissue macrophages (ATM) may actually exhibit inhibited autophagy in obesity. 19 Moreover, a transcriptional activator of adipose tissue autophagy in obesity, E2F1, is up-regulated in the adipocyte cell fraction but not in the stromalvascular fraction of adipose tissue in obesity. 17 (2) Among the various alterations that occur in the adipose tissue milieu in obesity, which is/are the factor(s) that dysregulate autophagy? Indeed, classically, autophagy is activated under conditions of nutrient deficiency. Inflammatory cytokines are common activators of autophagy in disease states, and although adipose tissue inflammation signifies metabolically-dysfunctional obesity, many of the proinflammatory cytokines are only mildly elevated in obesity compared to overtinflammatory reaction. Circulating levels of leptin are robustly and positively associated with BMI, 20,21 and the effect of leptin on autophagy was studied extensively in various cell types, and may be highly cell-typespecific: Leptin activated autophagy in muscle, heart, liver, kidney, 22 conventional T cells 23 and cancer cells, 24,25 while inhibited autophagy in human lung cell line 26 and recently was shown to inhibit ERstressinduced autophagy in whole adipose tissue. 27 A gap of knowledge remains in understanding whether and how leptin, produced by adipocytes at increased amounts in obesity, can regulate autophagy in adipocytes and other adipose tissue cell types.
In a previous study, 18 we found that circulating leptin levels associate with visceral AT (VAT) expression of ATG5 mRNA, a putative surrogate of adipose tissue autophagic activity. This led us to hypothesize that leptin, via the leptin receptor, may dys-regulate autophagy in a cell-type specific manner.
Cell culture RAW264.7 murine macrophage cell line (American Type Culture Collection, Manassas, VA, USA) were cultured in DMEM 4.5 g/l glucose containing 10% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin and 4 mM glutamine, as previously described. 19 Epididymal pre-adipocyte cells 28,29 were cultured as described. Briefly, cells were grown in DMEM 4.5 g/l glucose, supplemented with 20% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin and 4 mM glutamine. Twenty-four h after reaching confluence, preadipocytes were induced to differentiate in media containing 0.125 mM indomethacin, 2 g/ml dexamethasone, and 0.5 mM 3-isobutylmethylxanthine for 48 h. For RT-PCR and western blot analysis, cells were seeded in 6-wells plates and experiments were held after 6 d of differentiation.

Live imagining
For macrophages imagining, cells were cultured until they reached 60-80% confluence. Cells were scraped and 3 × 10 4 cells were seeded in 96-well uClear black plate (Greiner Bio One, Kremsmünster, Austria). For adipocytes, on the third day after differentiation induction, cells were detached with trypsin. 4 × 10 4 cells were seeded into 96-well uClear black plate.
3 h before treatments, the media was replaced to DMEM 4.5 g/l glucose containing 0.05% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin and 4 mM glutamine. Cells were treated with or w/o leptin (10 and 100 ng/ml, orders of magnitude larger than endogenous leptin production from adipocytes cell-line (~2 pg/ml/h)), chloroquine (10 μmol/l). 3MA (5 mmol/l) and SLAN (50 μg/ml) were added to the media 30 minutes prior to the treatments mentioned above. For cell imaging and dynamic autophagosomes tracking, Hoechst (final dilution 1:1000) and CYTO-ID (final dilution 1:1000) were added to the media. For lipid droplets staining, BODIPY-C 12 (1 µM) was used. Identification and quantification of images was done semi-automatically by Columbus software, and the mean value derived from each experiment was used to calculate the final comparison between experimental conditions.

Autophagy-mediated long lived proteins degradation assay
We adopted a well-described protein degradation assay 30,31 for use in our differentiated epididymal preadipocyte cells. Five days after differentiation induction, cells were 'pulsed' for 24 h with 50 µCi/ml L-[ 14 C(U)]-Valine (NEC291EU050UC, PerkinElmer), then washed thoroughly, and incubated for the following 24 h period in medium containing 10 mM nonradioactive valine (V0513, Sigma-Aldrich). After an additional washout of the media with nonradioactive valine, the cells were 'chased' during a 6 h period in the above media, while being treated with: EBSS+0.1% BSA (= positive control for autophagy induction), control media (con), media containing leptin (100 ng/ml) only, or media containing leptin + SMLA (1 μg/ml). All conditions were given in the presence or absence of Baf.A. (0.1 μM), as indicated. SMLA was added 30 minutes prior to and during leptin treatment. A sample of 200 µl was taken at 2 h intervals and stored at −80• (200 µl of media was added back to the well). After 6 h, 700 µl were collected into tubes, the remaining media in the wells were removed, and wells were washed with cold PBS containing 10 mM valine. In order to separate the free amino-acids from the proteins, 20%TCA (T/2950/60, Fisher Scientific) was mixed with 1%PBS, 1:1 ratio, and was added to the media. Radioactivity was measured in the supernatant fractions (which contain mainly the free amino acids) and in the pellet of the cells (which contains mainly cellular proteins). % of [ 14 C]-valine release was calculated as ¼ free animo acids radioactivity in the supernatant Â 100 total radioactivity supernatant þ cells 0 pellet ð Þ

Statistical analysis
Data are reported as mean +/-standard error of the mean (SEM). For two-groups comparison, noneparametric Mann-Whitney test was used. Comparison of> 2 groups was analysed after homogeneity of variances was confirmed using Levene test by one way analysis of variance (ANOVA) and the post-hoc Tukey's method. When no equal variance was present, Kruskal-Wallis and Dunns tests were used. All statistical analyses and graphs were performed using GraphPad Prism 5. p < 0.05 was considered statistically significant.

Results
Immune cell functions are altered by both changes in autophagy, and by leptinan adipokine that signals through a cytokine receptor family member (the leptin receptor) that induces well-described immunemodulatory functions. 32 Thus, we first assessed if obesity-related activation of adipose tissue autophagy could be attributed to the most abundant non-adipocyte cell type composing this tissuei.e., macrophages. 33 RAW 264.7 macrophages were treated with the lysosomal acidification inhibitor choloroquine to prevent degradation of newly-formed autophago(lyso)somes, and the dynamics of autophagosome formation and expansion (maturation) was studied using high-throughput livecell imaging after staining cells with the fluorescent dye (CYTO-ID ©). CYTO-ID stains vesicles from the postphagophore autophagosome stage to autophagolysosomes, irrespective of vesicular pH (but does not stain primary lysosomes), and is being used in a growing number of studies to detect autophagosome dynamics in live cells. 19,34,35 Leptin exhibited only a mild effect on autophagy dynamics, tending to attenuate the initial rate of increase in total autophagosome area per cell (a composite measure of new autophagosome formation and expansion/maturation) ( Figure S1(a,b)). This result was variable and not significant with 10 ng/ml ( Figure S1(c)), but demonstrated a modest, though significant,~15% decrease in the rate of total autophagosome area growth with 100 ng/ml. To determine the likelihood that this might represent a physiologicallyrelevant response, we measured LepR expressionboth the total amount, and the long isoform (ObRb), using specific primers (detailed in Methods). Compared to adipocytes and to whole adipose tissue or liver samples, RAW264.7 cells expressed only minute amount of total LepR, but not its long isoform (Figure 1(a)). Thus, this cellular macrophage system provides only limited support for leptin's effect on macrophage autophagy, and given the modest inhibitory effectcould not explain contribution to the activation of whole adipose tissue increased autophagy in obesity.
To estimate if leptin might activate autophagy in adipocytes, we first assessed LepR expression in the adipocyte cell line. Both pre-adipocytes and differentiated adipocytes expressed a readily-measurable amount of total LepR, but not the ObRb isoform. These levels were markedly lower than in whole adipose tissue (in which ObRb was also expressed, being 15% of the total LepR expression), but comparable to the expression of adipocyte fraction isolated from mouse adipose tissue after collagenase digestion (Figure 1(a)). Furthermore, leptin acutely increased AMPK phosphorylation in these cells (~30% increase at 5 min, an effect completely inhibited by the LepR inhibitor SMLAdata not shown). We next utilized the established live-cell imaging approach (used above in macrophages) to detect autophagosome dynamics in differentiated pre-adipocyte cell line (lipid droplets were stained by the fatty acid fluorescent probe BODIPY-C 12 ). Over a > 4h tracking time, among the BODIPY-C 12 positive cells (i.e., differentiated adipocytes), we observed that more autophagosomes accumulated in the cell when their degradation was prevented with choloroquine (Figure 1(b,c)). Conversely, the pan-PI3kinase inhibitor, 3MA, frequently used to inhibit autophagosome formation, markedly attenuated the process. Jointly, live-cell imaging of autophagy dynamics can be reliably achieved in differentiated adipocytes by tracking CYTO-ID-positive puncta.
To assess leptin-induced changes in adipocyte autophagosomal "mass" (number and size), cells were pre-treated with chloroquine to prevent autophagosome degradation, and the change in autophagosome number and area was followed during the first 1.5 h of leptin treatment. As seen in individual cells quantified by the image analyses software (Figure 2(a)) and in whole representative fields (Figure 2(b)), Leptin induced a moderate,~20% increase in the rate of autophagosome area/cell (Figure 2(c,d)), suggestive of increased autophagosome biogenesis and/or maturation/expansion. This effect was nearly fully inhibited by the leptin-R blocker, SMLA, confirming the specificity of the response to the leptin-LepR system. Since in adipose tissue in obesity autophagy genes' mRNA and protein levels were increased, we also determined whether leptin could elevate the expression of such gene products in adipocytes. The mRNA levels of ATG5 and of Map1lc3b were mildly but nevertheless significantly elevated (by~15-20%), following stimulation of adipocytes with leptin (Figure 3(a)). A similar change was also detectable in some autophagy genes' protein levels (Figure 3(b)).
Given the mild effect of leptin on adipocyte autophagy detected by the cell imaging system, we aimed to further assess whether leptin can activate autophagyrelated functions, by adopting for use in adipocytes, an autophagy activity assay based on degradation of longlived proteins. 31 For this purpose we metabolically labelled long-lived proteins by pulsing cells with radiolabeled valine (a branched amino acid that does not affect mTOR, a regulator of autophagy), followed by a wash-out period of 24 h with media containing nonlabelled valine, as detailed in methods. Over a 6 hour subsequent follow-up, 14 C-valine release from longlived proteins to the media was linear. More importantly, the autophagy inhibitor bafilomycin markedly inhibited, while serum and nutrientfree medium (EBSS) accelerated the rate of 14 C-valine release (Figure 4(a)), consistent with a significant contribution of autophagy to this process. Non-autophagy long-lived protein degradation (i.e., in the presence of bafilomycin) was similar under all conditions (Figure 4(b), black bars). Nutrient-and serum-free buffer increased total 14 C-valine release by >2-fold, and the rate of autophagy-mediated degradation by >3 fold (Figure 4(b,c)). Leptin treatment increased total and autophagymediated 14 C-valine release (the latter by >2 fold), an effect that was completely blocked by the Leptin-R blocker SMLA (Figure 4(b,c)).

Discussion
Although still debated, autophagy is likely activated in adipose tissue in obesity. 6,[13][14][15]36,37 The regulation of autophagy is highly cell and tissuetype, and biological context dependent, and indeed, multiple inducers of autophagy can be envisioned in the altered adipose tissue milieu in obesity. These include the development of adipose tissue hypoxia, inflammation, oxidative stress, ER stress, insulin resistance, etc., 38 although nutrient overload per-se (like high concentrations of free fatty acids and glucose) would theoretically tend to attenuate autophagy. 30,39 In this study, we demonstrate that leptin can modestly modulate autophagosome dynamics as evident by the acceleration of the rate of growth of autophagosome area per cella composite measure of autophagosome number and size. This effect of leptin is evident in adipocytes, but not in the RAW macrophages. Although we demonstrate very low expression of LepR in these cells, others demonstrated leptin-induced activation of mTOR activity, which would inhibit autophagy. 40 Utilizing autophagy-related degradation of long-lived proteins, leptin had a more pronounced activating effect in adipocytes. Taken together our results suggest that obesity-related hyperleptinemia may contribute to autophagy activation in adipose tissue, acting in an autocrine/paracrine manner via the leptin receptor.
Since autophagy is a dynamic process in which autophagosomes are generated and degraded, snap-shot analyses of cellular autophagosome number, size, and of autophagy-related proteins' content can lead to misleading conclusions. An increase in autophagosomal 'mass' can result from increased biogenesis, and/or inhibited degradation of autophagosomes. Therefore, dynamic assays have become critical when evaluating autophagic activity, and ideally, this should be performed by more than a single assay. 41 Here, we utilized dynamic, high throughput live-cell imaging to track the increase in autophagosome area when autophagolysosome degradation was inhibited. This allows assessing the effect of leptin on initial rate of autophagosome biogenesis and expansion (maturation). As stated above, leptin had a modest effect using this assay, as well as on the expression of several key autophagy gene mRNA and protein levels. Leptin's capacity to activate autophagy was more robust when autophagic activity was assessed by the degradation rate of long-lived proteins. It is possible that the apparent differences of leptin's efficacy to activate autophagy are due to different sensitivities of the two assays. Yet, it is also plausible that they uncover a true biological difference. Chaperone-mediated autophagy (CMA), and micro-autophagy, are two processes in which intracellular components, in particular proteins, are destined for degradation by the lysosomal compartment, by inserting them into the lysosome, independent of the creation of new autophagosomes that would need to merge with lysosomes to degrade their cargo. 42 Indeed, degradation of the lipid droplet-associated proteins perilipin 2 and 3 was shown to be CMA-dependent. 43 Thus, it is tempting to speculate that leptin may only modestly activate macro-autophagy, while more robustly augment CMA and/or micro-autophagy. Current literature provides little evidence, if any, on direct regulation of these processes by leptin: Chaperone-mediated autophagy was proposed to be activated by calorically restricting high-fat fed mice, 44 a condition that would be predicted to diminish, not increase, leptin levels. Moreover, the effect was demonstrated in liver, a tissue in which macro-autophagy may be regulated opposite to adipose tissue in obesity. 38 No information at all was found in the literature on leptin and micro-autophagy. Thus, it would seem worthwhile to address future studies on the potential capacity of leptin to particularly target one or both of these generally lessstudied autophagic processes. This may be of particular relevance to adipocyte biology. In adipocytes 38% of detected proteome was reported to have an exceedingly long half-life, with no discernible degradation evident even after 96h of follow-up. 45 Moreover, in that study, (macro)autophagy was proposed to contribute only minimally to total adipocyte protein degradation, 45 suggesting that leptin, by particularly regulating other autophagytype processes/lysosomal compartmentrelated process, could be an important regulator of adipocyte protein homeostasis.
It is noteworthy that a recent extensive study concluded that leptin inhibits ER-stressinduced autophagy. 27 The in-vivo experiments of the study cannot exclude an indirect effect of leptin treatment, such as via its actions on hypothalamic centres. Yet, some invitro studies that also support the authors' conclusions cannot be easily reconciled with our studies, though we propose that our highly dynamic approach to detect changes in autophagic flux, and more sothe finding that leptin may particularly augment long-lived protein degradation, possibly through mechanisms distinct from macroautophagy, may underlie the differences between the studies.
Obesity is thought to be associated with leptin resistance, 21 questioning the relevance of our findings in obesity. Yet, whether adipocytes exhibit leptin resistance is not unclear: several papers concluded that obesity induces leptin resistance in whole AT, 46,47 characterized by decreased pSTAT3 levels and elevated SOCS-3 expression. Yet, other studies 27 do show leptin responsiveness of AT in obese (ob/ob) mice. Regardless, the chronicity of obesity may suggest that any leptin regulation of adipocyte autopahgic process as we show here, even if with a small effect size due to LepR expression and/or to obesity-related leptin resistance, may nevertheless bear physiological relevance over time.
Leptin is only one, out of hundreds of peptides, which are altered in obesity. Yet, since adipocytes are its major source, and given that obesity is usually associated with hyperleptinemia, local concentrations of leptin in adipose tissue may be very high. Systemically, leptin's circulating levels are amongst the adipocytokines with the highest dynamic range, particularly differentiating lean from obese persons. Thus, despite being a single factor amongst many, which is studied here in isolation, it is not impossible that leptin contributes significantly to the dysregulated/activated adipose tissue autophagy and to this tissue's dysfunction in obesity.