The Role of Fatty Acids in Ceramide Pathways and Their Influence on Hypothalamic Regulation of Energy Balance: A Systematic Review

Obesity is a global health issue for which no major effective treatments have been well established. High-fat diet consumption is closely related to the development of obesity because it negatively modulates the hypothalamic control of food intake due to metaflammation and lipotoxicity. The use of animal models, such as rodents, in conjunction with in vitro models of hypothalamic cells, can enhance the understanding of hypothalamic functions related to the control of energy balance, thereby providing knowledge about the impact of diet on the hypothalamus, in addition to targets for the development of new drugs that can be used in humans to decrease body weight. Recently, sphingolipids were described as having a lipotoxic effect in peripheral tissues and the central nervous system. Specifically, lipid overload, mainly from long-chain saturated fatty acids, such as palmitate, leads to excessive ceramide levels that can be sensed by the hypothalamus, triggering the dysregulation of energy balance control. However, no systematic review has been undertaken regarding studies of sphingolipids, particularly ceramide and sphingosine-1-phosphate (S1P), the hypothalamus, and obesity. This review confirms that ceramides are associated with hypothalamic dysfunction in response to metaflammation, endoplasmic reticulum (ER) stress, and lipotoxicity, leading to insulin/leptin resistance. However, in contrast to ceramide, S1P appears to be a central satiety factor in the hypothalamus. Thus, our work describes current evidence related to sphingolipids and their role in hypothalamic energy balance control. Hypothetically, the manipulation of sphingolipid levels could be useful in enabling clinicians to treat obesity, particularly by decreasing ceramide levels and the inflammation/endoplasmic reticulum stress induced in response to overfeeding with saturated fatty acids.


Background
Obesity has become a major public health issue, and an effective therapy has yet to be well established. Individuals who gain excess weight are at risk of developing chronic diseases, such as type 2 diabetes mellitus, cardiovascular diseases, and cancer [1,2]. Recently, obese individuals were described as being more predisposed to developing severe symptoms of SARS-CoV-2, reinforcing to the scientific community that there is a need to improve the understanding of the pathophysiology of obesity and develop novel treatment strategies [3].
Additionally, the inverse pathway can occur, leading to SM, sphingosine, and glucosylceramide production from ceramide by sphingomyelin synthase (SMS), ceramidase (CDase), and glucosylceramidase (GCase), respectively [24][25][26]. For example, the ceramide made in the endoplasmic reticulum (ER) can be introduced into the Golgi apparatus by its own carriers, subjected to different isoforms of the sphingomyelin synthase enzyme, and converted to sphingomyelin with a similar number of carbons to that of the original ceramide. This happens by converting phosphorylcholine to diacylglycerol [34,35]. The product from this reaction is mainly used for the manufacture of myelin of the cell membrane of neurons [36].
The length of the carbon chain of the added fatty acids for biosynthesis is variable and defines the species of ceramide formed. In mammals, the six existing ceramide synthases are responsible for the diversity of the species and tissue specificity of ceramides synthesized [37,38]. Depending on ceramide length and level, and the CerS isoform expressed, it could be associated with negative biological effects. For example, in peripheral organs/tissues, excessive ceramide has been correlated with lipotoxicity, thereby resulting in insulin resistance and β cell apoptosis [33]. It has been shown in vitro that insulin resistance in myoblasts induced by a saturated fatty acid such as palmitate is associated with increased levels of C16:0 ceramides (presence of a palmitic acid bound to sphingosine), whereas in vivo muscle insulin resistance was associated with ceramide C18:0 [39,40].
Although the nature of the ceramide species involved in insulin resistance due to obesity has not been fully elucidated, many studies have described the interactions of CerS isoforms in biological functions. Decreased CerS6 expression was associated with a protective effect from high-fat diet-induced obesity and glucose intolerance, and also increased energy expenditure and reduced C16:0 ceramide [41]. By comparison, CerS2 deficiency or insufficiency led to susceptibility to diet-induced steatohepatitis and insulin resistance, glucose intolerance, and increased C16 ceramide, suggesting that the altered sphingolipid acyl chain length directly affects insulin signaling, because it was observed as a compensatory effect for the decline in very-long-chain ceramide production by CerS2 [42,43].
Increased ceramide levels are also associated with macrophage infiltration after myocardial infarction, contributing to higher infarct size [48][49][50]. However, specific inhibition of SPT by myriocin leads to an approximately 35% reduction in the level of ceramides in the heart, which was associated with reduced infarct size [50].
More recently, it was reported that enzymes involved in de novo ceramide synthesis are widely expressed in the hypothalamus and hippocampus of rodents [51,52]. In an interesting review, Cruciani et al. (2017) [53] hypothesized that ceramide accumulation in the brain under lipotoxic conditions (triggered by HFD or isolated SFA treatment) might play a role in the downregulation of energy balance control, leading to obesity and associated diseases. S1P is generated from sphingosine phosphorylation by the enzyme sphingosine kinase (Sphk), mostly in the plasma membrane. There are two isozymes of Sphk: sphingosine kinase 1 (Sphk1) and sphingosine kinase 2 (Sphk2) [54,55]. Sphk1 is found in the cytoplasm, whereas Sphk2 can be found in the nucleus [22] (Figure 1). The irreversible breakdown of S1P to phosphoethanolamine and hexadecenal is catalyzed by the enzyme S1P lyase (S1PL).
Interestingly, in contrast to ceramides and other metabolites, S1P has no structural role [20]. S1P mainly acts as a potent signal mediator that modulates multiple cellular functions important for health and diseases, usually with opposite effects of those of ceramide accumulation, and having tissue-or cell-type-specific effects. For example, Hadas et al. (2020) [56] used a new genetic tool called modified mRNA (modRNA) during myocardial infarction to increase the acid ceramidase (AC) enzyme, which increases S1P production, and found that the infarct area was reduced via decreasing cardiomyocytes' cell death and infiltrated neutrophils. Interestingly, it has been found that feeding increases levels of S1P, showing a role in the regulation of energy balance [57,58]. Intracerebroventricular injections (ICV) of the bioactive lipid, S1P, reduce food consumption and increase rat energy expenditure through persistent activation of STAT3 and the melanocortin system.
Thus, sphingolipid metabolism has emerged as being important in energy balance control and is a pathway to be explored to better understand the pathophysiology of obesity.
To date, few studies have addressed the relationship between fatty acids and sphingolipid metabolism in the control of energy homeostasis. Thus, our main objective was to evaluate the role of ceramide/sphingosine-1-phosphate and their influence on hypothalamic energy balance control, particularly in response to fatty acid interventions.
The results summarized here indicate that S1P, as opposed to ceramide, appears to be a central satiety factor in the hypothalamus. Extracted data demonstrated that ceramides are associated with hypothalamic dysfunction in which metaflammation, ER stress, and lipotoxicity lead to insulin/leptin resistance and disturbances in energy balance control, in a vicious cycle. We conclude that manipulation of sphingolipid levels would be an interesting strategy for clinicians to treat obesity, specifically by decreasing ceramide levels and the inflammation/ER-stress induced by overfeeding, particularly with LCSFA.

Materials and Methods
This review followed the recommendations of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The review protocol was registered in PROSPERO International Prospective Register of Systematic Reviews (Unique Identifier: CRD42020198608).

Data Sources Search Strategy
Four electronic databases (Embase, PUBMED, Scopus, and Web of Science) were searched until May 2020, with an English language restriction. A new search was conducted during the final stage of the preparation of this article (January 2021). The search strategies were developed to trace published reports from studies with rodents and hypothalamic cells assessing the relationship between sphingolipids and energy balance control during obesity and associated conditions. Thus, the index terms used in our protocol were hypothalamus and sphingolipids.
The search was conducted by two authors, and the articles were selected by three independent reviewers. After the first screening, duplicates were removed from our databases. Initially, titles and summaries were screened. When full texts were available, we assessed them for eligibility according to our prespecified inclusion/exclusion criteria.

Eligibility Criteria
The population of interest was limited to rodents of any species receiving HFD, fatty acid treatment, or treatment with other obesogenic factors (such as inflammatory cytokines, sphingolipids), in addition to genetically obese mice models. All sexes and ages were considered. We also investigated studies that used hypothalamic cells treated with fatty acids, inflammatory cytokines, and/or sphingolipid metabolites (ceramides, S1P, or others). Our primary outcome included all data about the levels of sphingolipids. Our selected secondary outcomes were body weight, food intake, energy expenditure assessment, levels of neuropeptides, and inflammatory markers. We excluded articles that presented no criteria for population eligibility, intervention, comparison, and outcomes (Table 1). Table 1. Eligibility criteria applied in this study.

Inclusion Criteria Exclusion Criteria
Population Rodents (offspring and/or adults); hypothalamic cell lines; or primary hypothalamic cell culture.
Other cells than hypothalamic; non-rodents and studies involving human subjects.

Intervention
High-fat feeding; genetic models of obesity; treatment with fatty acids; treatment with cytokines; treatment with sphingolipids (ceramide or sphingosine-1-phosphate).
Other genetic backgrounds; animals with protein restriction or receiving any other treatment.
Studies related to ganglioside metabolism.

Inclusion Criteria Exclusion Criteria
Outcomes Alterations in sphingolipid metabolism following disturbances in energy balance control. Primary outcomes: sphingolipid level (ceramide and sphingosine -1-phosphate). Secondary outcomes: Body weight, food intake, energy expenditure assessments, neuropeptides levels, and inflammatory markers.
Other outcomes not relevant to energy balance control.
Type of publication Original article Nonoriginal article (reviews, conference abstracts).
Other selection criteria Studies following these main characteristics were included: animal and cell culture studies from 1990 until 2020. The articles needed to be published on these databases: PubMed, Embase, Web of Science, and Scopus. Keywords searched: sphingolipid and hypothalamus.
Non-English language.

Data Extraction
Data extraction was performed by two independent researchers after reading the previously selected articles, according to the eligibility criteria. A third researcher was consulted when there were differences or doubts regarding the article data. We extracted data including the following: names of authors, year of publication, species of animal, composition of diets, intervention period, neuropeptide levels, sphingolipid levels, inflammation levels, body weight, food intake, and adiposity. The data were transcribed into tables according to primary or secondary outcomes, separating animal results from cellular models.

Assessment of Methodology Quality
The evaluation of risk of bias in the studies included in this systematic review was conducted in pairs and individually for each article. For in vivo studies, assessment was based on Syrcle's recommendations [60]. This tool establishes 10 questions related to selection bias (sequence generation, baseline characteristics, and allocation concealment), performance bias (random housing and blinding researchers), detections bias (random outcome data and blinding of outcome assessment), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias), and other biases.
For in vitro studies, we adapted Syrcle's tool, based on the studies of Golbach et al. (2016) [61] and Alshwaimi et al. (2016) [62]. Thus, we applied 10 questions to the analyzed articles, individually. The questions were related to selection bias (baseline characteristics of cell culture models, duration and frequency of the exposure, and cellular viability), performance bias (controlled temperature and humidity, randomized and independent experiments), detection bias (blinding), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias), and other biases. These questions were judged as having high, uncertain, or low risks of bias. In Figures 2 and 3, the green symbol means a high level of reliability, the yellow symbol means not clear, and the red symbol means low reliability.

Results
The main aim of this systematic review was to evaluate ceramide/S1P and their influence on hypothalamic energy balance control, particularly in response to fatty acid interventions.

Methodological Approach to Assess Sphingolipids
After article selection, the first step was analysis of the manner in which researchers assessed sphingolipid levels and enzymes involved in their synthesis. First, we observed that in vitro studies (Table 2) used the following methods: commercial kits (n = 1); lipidomics (n = 1); liquid chromatography coupled with mass spectrometry (n = 2); liquid chromatography coupled with electrospray ionization and mass spectrometry (n = 1); mass spectrometry (n = 1). Studies also used Western blot (n = 1) and quantitative polymerase chain reaction (qPCR) (n = 1) to monitor enzymes involved in the pathway. However, it should be noted that these are not suitable methods for ceramide quantification. Finally, some papers had no information on the methodological approach used or did not present sphingolipid levels (n = 3), as described in Table 2.

Uninformed
↑ TNF-α and IL-6 mRNA expression in N43 and BV2 cells exposed to PA for 8 h. ↓ TNF-α and IL-6 in N43 cells transfected with siRNA for PGC-1a, followed by treatment with PA. ↑ TNF-α and IL-6 in N43 cells transfected with siRNA for PGC-1a, followed by infection with AdGFP-Era followed by infection with AdGFP-ERa, with or without exposed to PA;

Mass spectrometry Uninformed
↑ Inflammation in N43 cell line and primary neurons exposed to PA. ↓ Inflammation in N43 cell line and primary neurons were pretreatment with E2. Astrocytes were isolated from P1-P3 postnatal WT and S1P3 KO mice.
↑ COX2 and IL-6 levels in S1P3 KO astrocytes treated with S1P.  ↑ Body weight of males and females after consuming HFD for 4 weeks. In Wistar rats, either C2-ceramide or DH-C2-ceramide or vehicle (25nM) were acutely ICV injected.

3.
Prior to sacrifice, animals received an ICV injection of insulin (2 mUI) or saline.
Central myriocin treatment did not affect body weight in either obese or lean Zucker rats.
Central myriocin treatment did not affect food intake in either obese or lean Zucker rats.
Central myriocin treatment did not affect lean and fat body mass in either obese or lean Zucker rats.  ↓ Locomotor activity in LPL knockout with HFD. ↑ Energy expenditure in LPL knockout with HFD ↑ Energy expenditure in OVX rats after treatment with estradiol. ↓ Respiratory quotient in OVX rats after treatment with estradiol.   ↓ Energy expenditure and locomotion in HDF-fed mice. ↑ Energy expenditure and locomotion in HDF-fed mice with telmisartan treatment, in dark periods. ↑ Total ceramide levels in the hypothalamus of obese Zucker rats.
↓ Total ceramide levels in the hypothalamus of obese Zucker rats after ICV injection of myriocin.
↑ TNF-α and Cxcl12 in HFD-fed mice. ↓ CXCL12 in HFD-fed mice after Telmisartan treatment. ↑ IL-4 and IL-6 levels in plasma of both HFD-fed mice and HFD-fed mice treated with telmisartan. ↑ IL-5 levels in plasma of HFD-fed mice with treatment of telmisartan.

Additional in Vitro Methodological Approach
As mentioned above, we decided to include in vivo research and in vitro findings in our review because the anatomy of the hypothalamus is highly complex, and mechanistic investigations of neuropeptide gene regulation and signal transduction events occurring in neurons are a challenge to perform in vivo [14]. Thus, some of our selected studies used hypothalamic cells (primary or lineage) alone or combined with animal models (n = 10 studies), as shown in Table 2. In general, these cells were exposed to fatty acids (palmitate, oleate, lauric, oleanolic, linoleic, or stearic acid) and had ceramide levels assessed (n = 8 studies) ( Table 2). Other treatments, such as TNF-α, IL1-β, estradiol, leptin, lipopolysaccharide (LPS), or Toll-like receptor 4 (TLR4) inhibitor, were also applied in these studies (n = 10) ( Table 2).

Main Findings Regarding In Vitro Sphingolipid Levels and Their Role in the Hypothalamus
Interestingly, as previously reported in cells from peripheral tissues [48,49,83], the studies selected here identified a relationship between ceramide levels and pro-inflammatory cytokines in hypothalamic cells. Importantly, these levels are tissue-and stimulus-specific and dependent on the type of ceramide species assessed. For example, the authors of [77] found that exogenous C6 ceramide treatment in microglial cells decreased interleukin-6 (IL-6) levels ( Table 2). In another study, Dusaban et al. (2017) [80] found that S1PR3, but not S1PR1, induces an increase in IL-6, vascular endothelial growth factor A (VEGFa), and cyclooxygenase-2 (COX-2) mRNA levels in mice astrocytes (Table 2). Interestingly, the authors also found that astrocyte inflammation in vitro leads to an increase in Sphk1 and S1P3 expression, which they speculated could contribute to astrocyte autocrine inflammatory signaling.
Furthermore, treatment with exogenous TNF-α or IL1-β was able to increase neuronal ceramide production that was associated with neuronal apoptosis [74,75] (Table 2). Sortino et al. (1999) [75] also investigated the effects of decreasing C1P levels using D609 (tricyclodecan-9-xanthogenate), an inhibitor of phosphatidylcholine-specific phospholipase C. This is a key enzyme in the activation of Ac SMase (sphingomyelin synthase) responsible for transferring the phosphocholine group from phosphatidylcholine to ceramidegenerating sphingomyelin and diacylglycerol (DAG) (Figure 3). Using this approach, the authors found that D609 decreased the accumulation of C1P in response to 15 min exposure to 20 ng/mL TNF-α by 50% (Table 2), and partially prevented TNF-α-induced apoptosis [75].
PA is also associated with increased ceramide levels and neuronal inflammation [63,67,79] (Table 2).   [68] showed that PA decreases peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α) and estrogen receptor (ERα) expression in hypothalamic neurons and astrocytes, thereby promoting inflammation associated with increased ceramide levels ( Table 2). Consistently, Sergi et al. (2018) [79] showed that Lcycloserine (a SPT inhibitor, Figure 1) inhibited PA-induced inflammation in the hypothalamic mHypoE-N42 cell line (Table 2). Conversely, oleic acid (OA) and eicosapentaenoic acid (EPA) exerted anti-inflammatory effects by decreasing PA-induced intracellular ceramide build-up ( Table 2). The authors also tested the involvement of the TLR4 receptor on neuronal inflammation in response to fatty acids and unexpectedly found that TLR4 inhibitors failed to inhibit PA-induced upregulation of proinflammatory cytokines.
The in vitro studies, although providing potential mechanistic insights into the role of ceramide species in the regulation of energy balance, lack the ability of in vivo studies to provide a physiological outcome, such as food intake. In this context, Silva et al. (2014) [57] showed that S1P treatment increases the signal transducer and activator of transcription 3 (STAT3)-dependent axis in GT1-7 cells (Table 2). By comparison,   [15], using hypothalamic mHypoA-POMC/GFP neurons, found that PA increases neuronal Pomc expression. This phenomenon is dependent on palmitoyl-CoA synthesis, but not de novo ceramide synthesis, since inhibition of SPT enhanced palmitate-induced Pomc expression (Table 2). McFadden et al. (2014) [70] found that exposure to palmitate promotes lipidomic remodeling in primary hypothalamic neurons (PHNs). The authors observed an increase in lipid species, including ceramides 18:0/16:0, when PHN neurons were exposed to excess C16:0. This outcome was reversed by increasing FA catabolism with C75, a stimulator of carnitine palmitoyltransferase-1 (CPT1C) and fatty acid oxidation. This stimulation of FA oxidation in neurons produced an increase in ATP levels, which the authors suggested led to a decrease in food intake and body weight. Finally, Campana et al. (2017) [63] demonstrated that myriocin treatment, in addition to serine palmitoyltransferase 2 (SPT2) knockdown by small interfering RNA (siRNA) (Figure 1), was able to restore neuronal AKT phosphorylation [63] in neuronal GT1-7 cells, which is possibly associated with improved energy balance control under obesogenic conditions.
In summary, these studies demonstrated that saturated fatty acids and inflammatory cytokines are closely related to the sphingolipid pathways, thereby influencing food intake, which is of potential importance for the understanding and treatment of obesity.

Additional In Vivo Methodological Approach
We found 24 articles in total with n = 18 using animals to answer their main hypotheses, as shown in Tables 3 and 4. The strains used were Ob/Ob mice (n = 1); Db/Db mice (n = 1); Swiss mice (n = 1); C57BL/GJ mice (n = 7); Sprague Dawley rats (n = 4); Zucker rats (n = 2); Wistar rats (n = 4); and models of knockout mice (n = 4) ( Table 3). The male sex was the most used (with n = 12 only using male) ( Table 3). Only two studies (n = 2) used females exclusively (Table 3). Finally, some studies used both (n = 4), as shown in Table 3. In general, the experimental age started at 5 weeks, with some studies analyzing animals up to 10 weeks old (Table 3).
Regarding the stimulus used to examine differences between sphingolipid levels, we mainly focused on studies administering HFD and/or FFA treatment. Other treatments/manipulations were reported (such as hormones, exogenous ceramide, agonists, or antagonists of the sphingolipid pathway). We used these manipulations to compare or improve our conclusions regarding diet exposure. We classified the intervention into three classes: hormones/genetic tools (n = 3); HFD or FFA exclusively (n = 4); and mixed stimulus (HFD or FFA combined with hormones, genetic tools, feeding/refeeding, exogenous ceramide treatment, overnutrition by litter size reduction, and others), as shown in Table 3.

Main Findings Regarding In Vivo Sphingolipid Levels and Their Role in the Hypothalamus
Growing evidence suggests that hypothalamic lipid sensing plays a key role in controlling energy balance [7,84], and that its dysregulation could lead to the development of obesity and associated comorbidities, such as insulin resistance and type 2 diabetes [7,10,57].
Interestingly, it has been discussed in the literature that there is a sexual dimorphism in brain metabolism in response to HFD. Following HFD consumption, the brain tissue of males, but not females, showed elevated levels of saturated fatty acids, such as PA [67,68]. The authors also found that male mice presented elevations in ceramides, which were associated with increased markers of inflammation when compared to female mice (Table 4). They also found that recovering ERα expression in males was able to reduce markers of inflammation [67,68].
Gonzales-Garcia (2018) [65] consistently found that central steroid hormone 17-β estradiol (E2) treatment decreases hypothalamic ceramide levels (Table 4) and ER stress, and increases brown adipose tissue (BAT) thermogenesis (Table 3). Increased ceramide levels were also associated with obesity and earlier female puberty in a study by Heras et al. (2020) [81] (Tables 3 and 4). However, when the authors blocked CerS, the obese phenotype was delayed [81] (Tables 3 and 4). Thus, Cer metabolism and ER stress pathway components could be potential therapeutic targets for the treatment of obesity associated with irregular estrogen levels. It is also important to mention that ceramide is able to trigger ER stress in the hypothalamus, impacting the energy balance [51].
HFD feeding was associated with increased ceramide levels and inflammation by other studies reported here using male animal models [72,78,82] (Tables 3 and 4). For example, obese male mice also presented increased hypothalamic ceramide levels [51,63,81] (Table 4). Overall, some authors suggested that lipid content in the hypothalamus is directly determined by dietary lipid content and not obesity per se, because exercise training for 6 weeks ameliorated inflammation triggered by diet but did not affect hypothalamic lipid profile [72] (Tables 3 and 4). Notably, Campana et al. (2018) [63], using deuterated palmitic acid, demonstrated that, in the hypothalamus, PA activates several enzymes of the de novo ceramide synthesis pathway, which is associated with the development of insulin resistance. Interestingly, ICV myriocin treatment reduced hypothalamic ceramide levels ( Table 4) and improved insulin sensitivity. However, it is important to note that body weight reduction can also lead to higher ceramide levels in some areas (plasma, liver, and adipose tissue) [78]. Lyu et al. (2020) [82], using transcriptomic analysis from animals receiving HFD, showed upregulation of genes related to sphingolipid metabolism associated with loss of POMC neurons (Table 4). Interestingly, treatment with telmisartan recovers HFD-induced ceramide levels, preventing lipotoxicity and contributing to weight loss (Rawish et al., 2020) (Table 3). Thus, lipid intake appears to be critical to sphingolipid metabolism and neuropeptide expression, confirming our main hypothesis.
Ceramide levels in the brain are also associated with lipoprotein lipase (LPL) activity [66]. LPL deficiency is associated with increased ceramide levels in the hypothalamus, in addition to increased body weight (Table 4) and glucose intolerance. The authors showed that LPL is differentially regulated by saturated and unsaturated fatty acids, and the lack of LPL causes ceramide accumulation that is associated with microglia inflammation, ER stress, and increased AgRP levels (Table 4) [66]. Laperrousaz et al. (2017) [64] also found that the partial LPL in mediobasal hypothalamus (MBH) increases body weight and decreases locomotor activity, without changes in AgRP/NPY levels. However, the authors found that decreased LPL in MBH led to a lack of ceramides. It is possible that the differences between the protocols and brain area analyzed are the main reason for the different results found by Gao et al. (2017) and Laperrousaz et al. (2017) [64,66].
Others have addressed the association between ceramides and food intake. For example, ceramide administration was able to induce food intake and orexigenic neuropeptide expression (NPY and AgRP) in CPT1C knockout (CPT1C KO) mice [73] (Tables 3 and 4). The authors suggest a ghrelin/CPT1C/ceramide axis is important for the understanding and treatment of obesity [73], in which CPT1C KO mice had the orexigenic action of ghrelin totally abolished. The same phenotype was observed with myriocin treatment. Others have noted an important relationship between ceramides and CPT1C. For example, Gao et al. (2011) [71] found that leptin decreases CER through malonyl-Coa and, consequently, CPT1C, which in turn affects NPY expression and food intake (Tables 3 and 4). Moreover, Mera et al. (2014) [69] demonstrated that CPT1C contributes to feeding regulation by modulating neurotransmitter expression and neuronal lipid profiles (Tables 3 and 4). Contreras et al. (2014) [51] also showed that ceramides regulate energy balance through the induction of hypothalamic inflammation, and ER stress associated with reduced BAT thermogenesis and weight gain (Tables 3 and 4). Maldonado et al. (2019) [77] also found increased hypothalamic inflammation (Table 4).
Finally, Silva et al. (2014) [57] found that S1P ICV treatment decreases food intake and increases energy expenditure (Table 3) by increasing STAT3-phosphorylation and POMC expression in lean mice. Importantly, using obese mice, the authors found decreased levels of S1PR1, whereas ICV S1P treatment decreased food intake, body weight, and adiposity (Tables 3 and 4). Vozzella et al. (2019) [58] also consistently found that feeding increases S1P, whereas fasting decreases it. This modulation is related to the de novo pathway.
In summary, the studies here provide evidence that S1P, as opposed to ceramide, appears to be a central satiety factor in the hypothalamus. A critical examination of the use of WB and RTq-PCR must always be undertaken to interpret data from the literature, because these methods are not suitable for ceramide quantification. Finally, sphingolipid assessment could be an important pathway to explore to help the scientific community develop new therapies for patients with obesity and to treat their metabolic disorders.

Discussion
In recent years, numerous studies have noted that obesity and the excessive consumption of SFA are accompanied by hypothalamic dysfunction driven by lipotoxicity, ER stress, metaflammation, oxidative stress, autophagy, and apoptosis. As a result, neurons and glial cells show impaired function, resulting in disorders in food intake and energy expenditure. The hypothalamus integrates the signals from the periphery to transmit signals within the brain and back to the peripheral organs to sustain energy balance. The energy derived from neuronal mitochondria is crucial in this context [85], and dysfunction in this organelle is related to changes in homeostasis observed in obesity. Indeed, neuronal mitochondrial dysfunction has been linked to the higher SFA content found in lard, as the fat source of HFD, because PA treatment of mHypoA-2/23 CLU183 cells leads to mitochondrial fission in addition to altered mitochondrial dynamics, suggesting that SFAs affect mitochondrial activity [86].
Increased evidence has shown that, as in the case of SFA, dietary intake of an increased ω6to ω3-polyunsaturated fatty acid (PUFA) ratio is correlated with metabolic diseases [87]. A Western-type diet containing soybean oil and cholesterol caused obesity in mice with alterations in liver function similar to those observed during metabolic syndrome in humans, including mitochondrial dysfunction [88]. Although linoleic acid (LA) is considered to be beneficial for mitochondrial activity, due to its incorporation in cardiolipin species, the in vitro treatment of hypothalamic cells with LA induces a rise in protein carbonylation, promoting oxidative damage in neurons [86].
Preventing obesity is a major challenge for clinicians and researchers because this disease has diverse causes, including genetic and epigenetic factors associated with lifestyle. Similarly, understanding the effects of not only the amount of fat ingested, but also the type of dietary fatty acids, on energy homeostasis under different conditions represents a challenge.
Here, we highlighted the consequences of hypercaloric intake, including excessive sugar and saturated fatty acid intake, usually from ultraprocessed food (UPF) on the dysmetabolic phenotype. UPF is associated with increased risk of cardiovascular diseases and all causes of mortality [89]. Even more concerning is the fact that children under 3 years of age eating UPF showed increased blood cholesterol levels at 6 years of age [90]. Therefore, it is highly important to understand the alteration of pathways in response to excessive SFA intake. This review aimed to address this important issue by examining the available evidence on the involvement of fatty acids and inflammation in ceramide levels and their consequences for hypothalamic function, particularly related to energy balance control.
It should be noted that sphingolipid species possess substantial structural diversity in contrast to the common ceramide structure, and as a result, their biological effects differ. Ideally, therefore, the methodological approaches used to study sphingolipids must present a high degree of specificity, distinguishing the structural similarity and interconversion between samples, thereby providing an accurate determination of metabolic flux [91].
Here, we found that some studies used common molecular biology tools, such as Western blot and qPCR to monitor enzymes, receptors, or genes involved in the process. These are easier and cheaper methods used in all molecular biology laboratories. However, it should be noted that these methods are not validated for assessing ceramide levels. Gene expression or protein content (for example, CerS and Sphk) cannot be always correlated with sphingolipid levels involved in the process because post-translational regulation may occur. Therefore, the optimal approach to validate gene and protein expression data is via a direct assessment of ceramide levels using analytical techniques such as high-performance liquid chromatography-mass and tandem mass spectrometry (HPLC-MS/MS). Additionally, pharmacological tools such as classical agonists/antagonists, in addition to genetic manipulations such as viral injection or knockout models, are useful tools for helping to interpret the data from Western blots and qPCR.
High-performance liquid chromatography (HPLC) is a relatively inexpensive tool that is able to provide a better resolution of sphingolipid levels. This approach has the capability to trace, quantify, and determine the structures of free sphingoid bases, free sphingoid base phosphates, ceramides, and sphingomyelin, among others, in a single sample, showing molecular mass and structure [91]. This technique provides high sensitivity by detecting low magnitudes, and the signal from the samples can be correlated with their concentration, constituting quantitative data that researchers can use to better understand biological effects of several stimuli [91,92]. However, accurate quantification usually depends on the definitive resolution of sphingolipid species, which HPLC alone cannot guarantee. This disadvantage exemplifies the requirement of additional techniques for structural elucidation [91].
For example, additional approaches found in the selected studies were electron ionization mass spectrometry (EI) and electrospray ionization mass spectrometry (ESI). EI enables the analysis of intact sphingolipid species, in addition to fragmented molecular species, whereas ESI enables sphingolipids in solution to be continuously infused at a low flow rate, providing great sensitivity. Finally, we also observed that HPLC-MS/MS is usually the analytical tool of choice for "sphingolipidomic" studies [92].
Here, our main focus was related to the description of recent studies regarding hypothalamic energy balance control and obesity, particularly triggered by HFD, because it has been widely documented that overnutrition, especially lipid overload, induces a state of hypothalamic dysfunction in which low-grade inflammation (or "metaflammation") through TLR4 receptor and NFK-B pathways, ER stress, and lipotoxicity are tightly associated with insulin/leptin resistance in a vicious cycle.
Thus, obesity has been associated with increased de novo ceramide synthesis in response to a higher bioavailability of FFA, particularly LCSFA, such as palmitate [72,93]. Recent work involving acute lipid infusions in rats suggests that, in situations of acute lipid oversupply, ceramide levels particularly increase if the FA source provided is saturated [94], most likely because this promotes de novo ceramide synthesis through the incorporation of palmitate at an early rate-limiting step related to an increase in palmitoyl-CoA [95]. Thus, palmitoyl-CoA formation and TLR4 stimulation, which has been previously related to the activation of SMases that convert SM to CER (Figure 3), are responsible for increased Cer levels in response to lipid overload [96,97]. As we described in our summarized results, others have found that levels of proinflammatory cytokines, such as IL-6 and TNF-α, can upregulate SMases and SPT (Figure 3), triggering excessive Cer levels [32]. Interestingly, n-6 PUFA consumption was also associated with increased ceramide levels in muscle (C16:0, C18:0; C22:0, and C24:1) [98]. Furthermore, when ceramide synthesis was blocked using L-cycloserine, there was an observed decrease in PA-induced inflammation [79].
In contrast, unsaturated fatty acids have been reported to improve insulin sensitivity and metaflammation. This may be because LCSFAs are less prone to β-oxidation when compared to medium-chain fatty acids and unsaturated fatty acids [16]. Importantly, when reviewing the literature, we found a lack of in vitro and in vivo studies showing the assessment of ceramide levels in response to unsaturated fatty acids such as palmitoleate or EPA/DHA. For instance, the authors of [99] used GLP-1-secreting GLUTag cells and found that PA (C16:0) but not palmitoleate (PO, C16:1) induces ceramide production, whereas unsaturated fatty acids confer lipoprotection by enhancing cell viability. Additionally, unsaturated fatty acids can ameliorate inflammation triggered by palmitate [79,100,101]; therefore, it is possible that they can decrease ceramide levels or at least reverse the effects of ceramide. However, further studies are needed to clarify this relationship.
Fatty acid metabolism is also an important regulator of cellular Cer levels and hypothalamic energy balance control by the CPT1-depending axis. Lipid overload or an excess of energetic substrate in the cell (such as carbohydrates) increases the ATP:ADP ratio, promoting citrate efflux from mitochondria into the cytoplasm and inducing forma-tion of malonyl-CoA, which classically inhibits CPT-1. The inhibition of FA transport to mitochondria increases the availability of palmitoyl-CoA in the ER for de novo ceramide synthesis, and it appears to be one of the main pathways for ceramide synthesis in obesity. The brain mainly contains a specific CPT1 isoform, named CPT1C, which is localized in the ER of neurons. CPT1C was the last member of the CPT1 family of genes to be described, whereas CPT1A is present in the liver and other tissues capable of high rates of FA synthesis, and CPT1B are expressed mainly in tissues characterized by high rates of fatty acid oxidation, such as muscle and brown adipose tissue [102]. CPT1C has been proposed to be a sensor of malonyl-CoA levels in hypothalamic neurons, therefore contributing to energy homeostasis. Indeed, CPT1C KO mice have reduced food intake and weight gain [103]. Additionally, Ramirez and colleagues (2013) [73] found that ceramide metabolism is involved in the anorectic effects of CPT1C in the Arc. The authors discovered that CPT1C KO mice are unresponsive to orexigenic effects of ghrelin. Interestingly, ghrelin seems to act in an additional pathway, increasing levels of C18:0 ceramide, which is mediated by CPT1C. Accordingly, myriocin completely blunted the orexigenic action of ghrelin. Overall, these data indicate that, in addition to formerly reported mechanisms, ghrelin also induces food intake through the regulation of hypothalamic CPT1C and ceramide metabolism, a finding of potential importance for the understanding and treatment of obesity.
Because CPT1C and SPT are expressed in the endoplasmic reticulum, previous studies suggest that palmitoyl CoA might be available via CPT1C action, which induces de novo ceramide synthesis. Additionally, it is possible that CPT1C can also act as a transporter of palmitoyl-CoA to the endoplasmic reticulum [17]. Moreover, it was shown that the stimulation of FA oxidation in neurons produced an increase in ATP levels, leading to a decreased food intake and body weight [70]. The authors showed that an increase in FA catabolism through CPT1C stimulation reversed the lipidomic remodeling in PHN hypothalamic neurons promoted by palmitate. Thus, FA metabolism by CPT1C seems to constitute an additional mechanism that regulates ceramide levels; however, controversy remains regarding these findings.
Because CPT1C is present in the ER, it does not participate in β-oxidation. Mera et al. (2014) [69] found that CPT1A overexpression in the ventromedial hypothalamus contributes to an obese phenotype, characterized by hyperphagia and, consequently, overweight, insulin resistance, glucose intolerance, and hyperglycemia, and led to higher levels of ceramides and sphingolipids, accompanied by increased mitochondrial reactive oxygen species, demonstrating an association between mitochondrial function and ceramide levels. Contributing to this evidence, Jeng and colleagues (2009) [104] found that higher ceramide levels led to mitochondrial dysfunctions, which alter fatty acid catabolism. They treated C6 glioma cells with C2 ceramide (50uM) and observed an increased cytochrome c release and caspase-3 activation, which could trigger apoptosis. The authors hypothesize that this ceramide-induced cell death may have been due to disruption of mitochondrial membrane potential (analyzed with 5,5 ,6,6 -tetrachloro-1,1 ,3,3 -tetraethylbenzimidazolylcarbocyanine iodide JC-1), ATP production, and mitochondrial biogenesis. Stoica et al. (2002) [105] showed a rapid mitochondrial depolarization and activation of the intrinsic caspase pathway, as determined by cytochrome c release to cytosol and increased caspase-9 content in primary rat cortical neurons with C2 ceramide (50 uM) treatment. As mentioned above, these events may induce FA oxidation disturbances because they occur in mitochondria compartments.
Furthermore, HFD feeding increases the total content of several neutral lipid species, such as phospholipids and triacylglycerol, while also increasing signaling lipids, such as sphingolipids and diacylglycerol, among others. Interestingly, fatty acid profiles appear to have an important role in sphingolipids, because mice subjected to 6 weeks of training showed no reduction in ceramide levels, despite having reduced inflammation. Thus, obesity per se does not seem to be the most important stimulus for altering hypothalamic lipid levels [72]. In addition, recent findings showed the existence of sexual dimorphism related to metabolism and obesity. For example, male mice had elevated levels of saturated fatty acids, such as PA, and elevations in total sphingolipids and ceramides, compared to female mice. Such changes were associated with elevated markers of inflammation in male mice [67,68], emphasizing the importance of sex-based research.
Overall, the literature reveals an inverse relationship between ceramide levels and insulin sensitivity in both rodents and humans [106]. For example, excessive ceramide levels were found in both the plasma and skeletal muscle of diabetic patients, and in the liver, plasma, and muscle of obese mice [107][108][109][110]. Studies in rodents show that inhibiting ceramide biosynthesis also warded off several pathologies associated with insulin resistance, including diabetes, atherosclerosis, hepatic steatosis, and cardiomyopathy. Mechanistically, in peripheral tissues, in addition to the relationship with metaflammation, the additional effects of ceramides on insulin resistance appear to result from their ability to block activation of Akt/PKB, a serine/-threonine kinase that is necessary for growth-factor activation of anabolism, the metabolic effects of insulin, and cell survival [106]. Additionally, ceramides have a detrimental effect on pancreatic β cells, by activating the stress-induced apoptotic pathway (i.e., cytochrome C release and free radical production) [111].
In the hypothalamus of rodents, lipid infusions or high-fat feeding are able to increase ceramide levels [97]. Obese Zucker rats also have higher levels of ceramide in the hypothalamus [63]. Mechanistically, PA treatment impairs insulin signaling while increasing ceramide levels in hypothalamic neuronal GT1-7 cells, whereas myriocin or SPT2 siRNA counteracts PA-induced insulin resistance in these cells [63]. Using an in vivo model (obese Zucker rats), myriocin treatment partially restored glucose-homeostasis.
Ceramides stimulate activation of atypical PKC isoforms (PKCζ/λ), which favors their combination with PKB/Akt and further prevents the activation of PKB/Akt in response to insulin [83], contributing to the insulin resistance state. In fact, PKCζ is a highly expressed kinase in the arcuate and paraventricular nuclei from the hypothalamus [112,113]. Interestingly, Campana and colleagues (2018) demonstrated a positive impact of the PKC inhibitor on insulin resistance induced by obesity. They found that the dominant negative form of PKCζ completely prevented hypothalamic insulin resistance induced by ceramide or palmitate. In addition to the central nervous system, PKCζ was also related to peripheral insulin resistance [83], suggesting a unique role of this PKC isoform in the regulation of insulin sensitivity by ceramides.
Thus, de novo ceramide synthesis has a key role in hypothalamic insulin resistance development and glucose homeostasis dysregulation associated with obesity. Here, we also summarized that obesity and associated conditions, such as metaflammation and ER stress, are able to increase hypothalamic ceramide levels, which in turn also have a role in triggering these pathways and inducing insulin resistance.
For example, exogenous ceramides can induce hypothalamic dysfunction by triggering ER stress. C6 ceramide treatment leads to weight gain and increased C16 levels in the mediobasal hypothalamus, which is reversed by GRP78 overexpression [51]. GPR78 is a chaperone that improves protein folding and, when administrated in obese Zucker rats, leads to reduced body weight, and increases thermogenesis and insulin/leptin signaling. However, the levels of ceramide remained elevated when compared to the control group [114], perhaps indicating that ER stress is downstream of the effects of ceramide [114].
This evidence confirmed our initial hypothesis that recent studies have provided information regarding the negative effects of ceramides on hypothalamic function [63]. Other examples reinforcing this idea have used transgenic mice. For example, LPL knockout mice exhibit excessive ceramide levels, which in turn trigger microglial activation and increased AgRP expression levels [66]. This phenomenon was associated with increased feeding, weight gain, and adiposity [51,81]. Recently, Lyu et al. (2020) [82] also noted that irregular sphingolipid metabolism pathways may be associated with the loss of POMC neurons under HFD feeding.
In contrast to the increasing evidence supporting the role of ceramides in obesity and associated disorders, S1P is a recently discovered molecule that appears to have positive effects on metabolism. Specifically, the ICV administration of S1P can decrease food intake and increase energy expenditure with promising results, reversing obese phenotypes in mice. The same effect was observed when a S1P agonist (SEW2871) was administered; mice had an induced anorexigenic effect, preventing obesity and associated metabolic diseases in a STAT3/POMC-dependent axis [57].
Interestingly, once outside the cell, S1P can either bind to albumin [115] or apolipoprotein M (ApoM) [116]. The ApoM/HDL-bound S1P has been found to be a contributor to the vasoprotective effects of HDLs [117], such as antiatherogenic properties [118]. However, the relationship of S1P with insulin resistance and type 2 diabetes (T2D) remains controversial in some tissues. For example, S1P inhibited insulin signaling in the liver, both in vitro and in vivo [119], and PA increases ceramide and S1P in beta-cells [120].
Thus, we highlight that the understanding of sphingolipid metabolism is a prospective area of interest because ceramides/S1P can be targeted to monitor obesity (as a biomarker) or to treat it using pharmacological nutritional approaches. However, precise determinations of different ceramide subspecies require further elucidation, thus warranting the characterization of their actions in obesity and related comorbidities, particularly regarding tissue-specific effects. In addition, exploring the status of sphingolipid pathways in models of metabolic reprogramming by HFD can be a useful tool to understand the obese phenotype, particularly in young individuals.

Conclusions
We conclude that our work contributes to summarizing the evidence related to sphingolipid metabolism and its role in hypothalamic energy balance control. The studies analyzed here monitored sphingolipid levels using different methodological approaches; however, it should be noted that RTq-PCR and WB are not suitable methods for ceramide quantification, and the results need additional interpretation.
Regarding energy homeostasis, ceramide accumulation in the hypothalamus causes inflammation, ER stress, and insulin/leptin resistance, thus interrupting the energy balance associated with an obese phenotype. Conversely, the S1P axis shows the opposite effects on body weight. Dietary approaches appear to be an interesting management strategy because recent evidence has shown that lipids, particularly saturated fatty acids such as PA, can interfere with de novo ceramide synthesis. It should be noted that other models have found an exacerbated hypothalamic injury in response to excessive consumption of SFA associated with dietary simple sugars. Thus, manipulation of sphingolipid levels with drugs or nutrients can be useful approaches for the treatment of obesity by clinicians, particularly by aiming to decrease ceramide levels to have a positive effect on insulin/leptin resistance. Further studies to investigate specific ceramide chains and the impact of unsaturated fatty acids are needed to clarify their specific action on metabolism.