Cholesterol—The Devil You Know; Ceramide—The Devil You Don’t

Abstract

Ectopic lipids play key roles in numerous pathologies including heart disease, stroke, diabetes, etc. Of all the lipids studied, perhaps the most well understood is cholesterol, a widely-used clinical biomarker of cardiovascular disease and target of pharmacological interventions (e.g., statins). Thousands of studies have interrogated the regulation and action of this disease-causing sterol. As a growing body of literature indicates, a new class of lipid-based therapies may be on the horizon. Ceramides are cholesterol-independent biomarkers of heart disease and diabetes in humans. Studies in rodents suggest that they are causative agents of disease, as lowering ceramides through genetic or pharmacological interventions prevents cardiovascular disease and diabetes. Herein we discuss the evidence supporting the potential of therapeutics targeting ceramides to treat cardiometabolic disease, contrasting it to the robust datasets that drove the creation of cholesterol-lowering pharmaceuticals.

Introduction

“Better the devil you know than the devil you don’t,” an Irish proverb from the 1300s.

The long and distinguished history of research relating cholesterol to disease started in the early 1900s, following its discovery in plaque-laden human aortas [1,2]. Subsequent animal studies supported the idea that cholesterol played causal roles in cardiovascular pathologies: rabbits fed diets enriched in cholesterol developed atherosclerotic lesions [3]. With growing interest in this malicious lipid, Carl Müeller described familial hypercholesterolemia [4], leading to the paradigm-shifting discovery that cholesterol levels were potently regulated by genetics. The subsequent discovery that mutations in the APOB, LDLR, and PSCK9 genes influence serum cholesterol and lipoprotein levels and alter susceptibility to atherosclerotic cardiovascular disease revolutionized our understanding of disease mechanisms and revealed highly-efficacious therapeutic strategies for treating cardiovascular disorders [5–7].

By comparison, ceramides are a relative newcomer to the lipid-mediated disease arena. Though the sphingolipid class that includes ceramides was initially described over 200 years ago, an understanding of the biological actions of these enigmatic lipids was slow to materialize. In the years preceding the 21^st century, animal models of diabetes and obesity were found to accumulate ceramides, which were in turn shown in cultured cells to induce numerous stress responses, including cell death. Shortly after the turn of the century, researchers found that pharmacologic or genetic inhibition of ceramide synthesis prevented or reversed many of the pathogenic comorbidities of obesity, including diabetes and heart disease, in rodents. More recently, sizeable human cohort studies revealed strong correlations between serum ceramides—which also traffic in the lipoproteins that carry cholesterol—and the incidence of cardiometabolic disorders. Indeed, blood ceramides are now being measured clinically (see Figure 1). Nonetheless, the body of literature on ceramides pales in comparison to that of cholesterol.

Figure 1.

Timeline depicting major advances in either cholesterol or ceramide research.

Cholesterol as a Driver of Disease

Sometimes The Devil is a gentleman – Percy Bysshe Shelley

Michael Brown and Joseph Goldstein conducted the innovative and elegant winning work that underlies much of our current understanding of hypercholesterolemia and its roles in disease. The team described in remarkable detail how patients with defective low-density lipoprotein receptors (LDLRs) overproduce cholesterol, owing to a compensatory upregulation of the rate-limiting enzyme in the cholesterol biosynthesis pathway: 3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) reductase [8]. Using novel techniques and rigorous molecular analyses, they showed that cholesterol-laden low density lipoprotein (LDL) particles enter the cell through the specific LDLRs [8]. Impairment of LDLR functionality decreases cholesterol import, which induces transcription of HMG-CoA reductase in order to synthesize additional cholesterol for membranes. This intricate system self-regulates, as rising cholesterol levels decrease LDLR and HMG-CoA reductase expression to slow cholesterol import and synthesis [8], thus enabling exquisite control of the cholesterol content of biological membranes (see Figure 3). These foundational ideas led to the development of a novel class of HMG-CoA reductase inhibitors known as statins—one of the most widely prescribed drug classes in the world [9]. Built upon the work of generations of scientists, statins were shown by meta-analyses to lower LDL levels and reduce major cardiac events in people of different races, ages, and risk categories [10]. Exciting findings continue to emerge as we further discern the mechanistic intricacies that control cholesterol synthesis, secretion, and action.

Figure 3.

Schematic depicting cholesterol synthesis and the negative feedback pathway that controls LDL and cholesterol production.

LDL receptor mutations resulting in familial hypercholesterolemia are both the first and most well-understood cholesterol handling mutations. While familial hypercholesterolemia is a rare disease, untreated individuals suffer from an up to 20-fold increased risk for CAD and heightened risk of fatal heart attacks due to elevated LDL-cholesterol levels. Further mutations have been described in the LDL particle itself. More than 100 mutations have been characterized in APOB, a key recognition signal of the LDL particle used by LDL receptors (LDLR). Defects in APOB reduce LDLR uptake and impair the production of APOB containing lipoproteins.

More recently, PCKS9 mutations have taken center stage due to their role in LDLR recycling. PCSK9 shares the same domain as LDLR on the plasma membrane. When both an LDLR and PCSK9 protein are taken up in the endosome, the LDL receptor is targeted for lysosomal degradation. However, when PCSK9 is not present, the LDL receptor is returned to the plasma membrane, where it can take up additional LDL particles. Both gain and loss of function PCSK9 mutations exist; gain-of-function mutations increase, while loss-of-function mutations decrease, serum LDL-cholesterol. The ‘cholesterol-sensing’ process involves intricate machinery in the endoplasmic reticulum (ER), including the sterol sensor SCAP and its partners insulin induced gene-1 (INSIG1) and sterol response element-binding protein (SREBP), which translate changes in ER cholesterol to altered gene expression.

Cholesterol and Ceramides as Biomarkers of Disease

You can give the Devil too much or too little attention – C. S. Lewis

The numerous studies relating cholesterol to disease, including the well-known seven countries study by Ancel Keys, established associations between dietary saturated fat intake and serum LDL-cholesterol with indices of cardiac events including heart attack and stroke [11,12]. Serum cholesterol is measured throughout the world and pharmacological and dietary guidelines to lower cholesterol are major components of the clinician’s armamentarium for identifying patients at risk of atherosclerotic cardiovascular disease. Nonetheless, these studies left room for other lipids as likely contributors to disease risk. Ceramides are emerging as strong, cholesterol-independent biomarkers of an overlapping but distinct spectrum of diseases.

In the early 2000s, advances in lipidomics enabled researchers to detect ceramides in large clinical cohorts. For example, Laaksonen et al. profiled lipids in the Corogene study, a Finnish cohort that included stable coronary artery disease (CAD) patients. They found that a ceramide-based score comprising several distinct ceramide species potently predicted major adverse cardiovascular events (MACE), including death due to coronary artery disease and acute coronary syndrome (ACS)[13]. The score included three ceramides (C16:0, C18:0, and C24:1) that positively associated with death, and a fourth (C24:0) that showed an inverse relationship. Adjustment for cholesterol levels did not change the prognostic value of ceramides, indicating that they are independent biomarkers. The same group validated this novel ceramide score in the FINRISK 2002 cohort, which includes individuals who underwent blood collection at a visit and were followed for 14 years [14]. In this study, they found that addition of C18:0-ceramide levels to canonical lipid measures of disease (e.g., HDL and LDL cholesterol) improved one’s ability to predict major adverse cardiac event (MACE) risk. Strikingly, the previously developed ceramide score associated with MACE in the entire FINRISK population [14]. In total, Laaksonen and colleagues published ~15 studies showing that ceramide-based scores predicted or correlated with cardiovascular disease endpoints in a large number of individuals from multiple clinical populations (see Figure 2) [13–23].

Figure 2.

Schematic depicting de novo ceramide synthesis and ceramide risk score strategies. As described in the text, several groups have demonstrated associations between ceramides, coronary artery disease, and/or MACE. In these studies, long-chain ceramides (C16:0 and C18:0) and very long chain C24:1-ceramide positively associated with either mortality or disease. Emerging data are now also showing that certain ceramides associate with other conditions related to cardiovascular health including vascular brain aging and dementia, heart failure with preserved ejection fraction, cerebral microvascular disease, and stroke. As more studies are done, researchers are finding nuanced differences between specific ceramide species and particular disease endpoints. These groundbreaking studies correlating ceramides with MACE are being replicated across the world in different countries and ethnicities, speaking to the robust nature of the association.

With the burgeoning evidence showing that ceramides correlate with MACE and cardiovascular disease, groups have begun to perform meta-analyses of the different studies. For example, Alessandro Mantovani and Clementina Dugo aggregated data from nearly 30,000 participants and seven cohort studies, confirming that distinct ceramide species (e.g., C16:0, C18:0, C24:1 and others) consistently associate with cardiovascular disease [21].

Unlike cholesterol, ceramides are also showing associations with other metabolic disorders, including insulin resistance and diabetes. For example, ceramides associated with fasting blood glucose in the Australian Diabetes, Obesity and Lifestyle Study [24] and with the homeostatic model of insulin resistance (HOMA-IR) and diabetes risk in Native Americans in the the Strong Heart Study [25–28]. Researchers in Finland also found that increased dietary intake of saturated fats increased serum ceramides and induced insulin resistance [29]. Moreover, Wigger and colleagues found that dihydroceramides, the less abundant precursor to ceramides, were tightly associated with type 2 diabetes in European cohorts [30]. Indeed, we and others have found found that dihydroceramides are particularly good markers of flux through the biosynthetic pathway and frequently show strong associations with disease ([31], unpublished observations).

Some studies have also evaluated relationships between tissue ceramides and disease, though the patient numbers are obviously far lower in studies using biopsies. Nonetheless, the findings are intriguing, as they point to the important sites of ceramide action. In heart tissue biopsies, ceramides were elevated in failing hearts and were decreased in individuals that had received a left ventricular assist device [32]. In liver, ceramide species were elevated in subjects with insulin resistance [33]. In adipose tissues, ceramides were elevated in individuals with insulin resistance [34], obesity [35], and diabetes [36]. And in muscle, ceramides associated with insulin resistance [37–41] and were decreased following successful insulin-sensitizing interventions (e.g., exercise, insulin-sensitizing drugs, and bariatric surgery) [37–40]. Some controversy has surrounded the associations in skeletal muscle and liver, as a few reports have indicated that ceramides were unchanged in biopsies from individuals with differing insulin sensitivities [42]. However, some of these researchers have re-evaluated their data and found that other sphingolipids like the dihydroceramides—which are markers of sphingolipid synthesis—were elevated in relation to insulin resistance [43,44].

Thus far, the data relating ceramides to disease endpoints are small in comparison to the robust data related to cholesterol. Nonetheless, we describe above over 30 papers showing that ceramides associate with cardiometabolic disease endpoints in humans. The associations are strong enough that clinics have started to measure serum ceramides as a means of gauging disease risk [45].

Ceramides in Lipotoxicity

“The Devil has put a penalty on all things we enjoy in life. Either we suffer in health or we suffer in soul or we get fat.” – Albert Einstein

The majority of fatty acids entering cells are either burned for energy, incorporated into glycerophospholipids (e.g., phosphatidylcholine, phosphatidylethanolamine, etc.) or sphingolipids (e.g., sphingomyelin) that comprise the bulk of cellular membranes, or stored as inert triglycerides for later use. As fatty acid availability exceeds the cell’s storage and energetic capacities, ceramides and other intermediates in these biosynthetic pathways start to accrue.

Ceramides emerged as major regulatory signals of cellular stress in the late 20^th century, approximately 100 years after their initial description [46]. Lina Obeid and Yusuf Hannun first discovered that tumor necrosis factor alpha (TNFα) rapidly hydrolyzed sphingomyelin to generate ceramide, which in turn triggered cytochrome c release and programmed cell death [46]. Around the same time, Kolesnick and colleagues found that interleukin 1-beta activated a comparable sphingomyelinase pathway, using cell free systems to discern the direct sphingomyeolinase-receptor interaction [47]. These discoveries revealed that ceramides were ‘second messengers’ and triggered a race to understand the mechanistic underpinnings of ceramide-driven signaling and apoptosis.

One of the first clues that ceramides might be vital in metabolic disease was when Turinksy et al. observed elevated levels of ceramide in genetically obese Zucker rats [48]. Shortly thereafter, Roger Unger coined the word “lipotoxicity,” describing the provocative idea that an accumulation of lipids causes the tissue dysfunction that underlies obesity-related disease [49]. He described ceramide as “the most damaging lipid” owing to its necessity for lipid-induced apoptosis in pancreatic β-cells, an event which drives type 2 diabetes [50]. Concurrently, ceramides were found to blunt insulin signaling and glucose uptake in 3T3L1 adipocytes [51]. In the decades that followed, researchers found that specific pharmacological inhibitors or genetic depletion of the enzmes involved in ceramide synthesis ameliorated diabetes. For example, treating ZDF rats with the SPT inhibitor myriocin prevents the onset of hyperglycemia [52].

Ceramides Are Evolutionarily-Conserved Signals of Lipid Excess that Modify Metabolic Programs to Alter Fuel Choice

“We are our own devils; we drive ourselves out of our Edens.” – Johann Wolfgang von Goethe

Thus far we have discussed ceramides as signals of apoptosis, which undoubtedly contributes to the tissue dysfunction that underlies cardiometabolic disease. Prior to triggering that terminal event, we theorize that ceramides serve as evolutionarily-conserved indices of lipid excess and change the cellular metabolic program [53]. Indeed, we and others have found that ceramides initiate actions to enable storage or utilization of fatty acids, likely as a means of helping the cell adapt to the surplus of fatty acids and protect membrane bilayers from these detergent-like molecules. The following actions of ceramides help fulfill this role (see Figure 4):

1. Sphingomyelins, a sphingolipid produced by the addition of choline to the ceramide scaffold, associates with cholesterol in plasma membranes, forming detergent resistant subdomains that support fatty acid diffusion (see Box 2).

2. Ceramides promote the insertion of fatty acid translocases into the plasma membrane, which facilitates fatty acid passage through the membrane to enable their rapid esterification [54].

3. Ceramides induce SREBP (in the liver) to increase expression of enzymes required for triglyceride storage [53].

4. Ceramides block Akt/PKB phosphorylation, which inhibits the uptake and oxidation of glucose and amino acids (in adipose tissue and skeletal muscled) and allows the cell to alter fuel choice to favor use of fatty acids [53,55,56].

5. Finally, ceramides decrease mitochondrial complex efficiency, thus allowing the cell to burn excess fat without creating unneeded energy [53].

This altered metabolic program—which leads to increased fat storage, decreased glucose utilization, and diminished mitochondrial efficiency—defines key components of the metabolic syndrome (e.g., insulin resistance and hepatic steatosis). We surmise that this altered metabolic program precedes apoptosis, which is a terminal response when the cell yields to the excessive fatty acid intake in order to protect the organism from looming cell lysis.

Figure 4.

Ceramides induce selective insulin resistance in the liver, form detergent resistant membranes, induce mitochondrial dysfunction, and drive glycerolipid synthesis.

Box 2. Cholesterol and Ceramide – Location Location Location.

“Have you ever danced with the Devil in the pale moonlight?” -- Jack Nicholson as the Joker

Though ceramides and cholesterol are independent biomarkers of cardiovascular disease, their biological activities are inextricably linked. Ceramides are not only found in LDL particles alongside cholesterol, but cholesterol-lowering interventions (e.g., statin treatment) decrease circulating ceramides in humans [92]. The dynamic regulation between cholesterol and ceramide is yet to be fully understood. Groups are working diligently to understand how changing one molecule may lead to alterations in the other [93]. The link between cholesterol and ceramide is most obviously demonstrated by the interaction between cholesterol and sphingomyelin. These two lipids are physically associated and colocalized in detergent-resistant domains in the plasma membrane and in low density lipoproteins [94,95].

Both of these lipids are synthesized and accumulate in various subcellular compartments. For example, cholesterol sensors in the ER (e.g., INSIG1) detect changes in cholesterol levels leading to changes in gene expression while the effects of ceramide accumulation in the ER are less understood. Some studies suggest that the propensity of ceramide to form dense microdomains (i.e., lipid rafts) require the ER to tightly regulate ceramide levels to avoid apoptosis [96]. Other groups hypothesize that the hydrophobicity of the ceramide molecule combined with its ability to form hydrogen bonds may alter membrane packing and fluidity in the ER, thus altering potential protein-protein interactions and protein trafficking [97].

Cholesterol and ceramide may also accumulate in mitochondria, where they influence respiration. Increased mitochondrial cholesterol levels leads to decreased membrane fluidity and oxidative stress [98]. Similarly, ceramides inhibit respiration (see Figure 4), in part by inducing mitochondrial fission to create a less-efficient reticular network. Interestingly, the Brüning lab found that C16 ceramides derived from Cers6 bind mitochondrial fission factor (Mff), but C16 ceramides produced from Cers5 do not [99]. This details a level of specificity adding further mechanistic depth to the deleterious actions of specific ceramides.

Lessons Learned from Global Knockout Mice

“Man can hardly even recognize the devils of his own creation.” – Albert Schweitzer

Most of the initial discoveries about cholesterol trafficking and its role in disease occurred prior to the development of knockout mouse technologies. Indeed, statins were already in the marketplace well-before modern in vivo gene editing tools were an everyday resource in biomedical research. Moreover, ablation of the enzymes controlling cholesterol biosynthesis frequently produces lethal phenotypes because cholesterol is such a critical component of normal cell function. These severe abnormalities limit the utility of rodent models for discerning the consequences of the modest changes in cholesterol that are associated with atherosclerotic cardiovascular disease phenotypes. For example, mice lacking both copies the Hmgcr gene that encodes HMG-CoA reductase do not develop past the blastocyst stage [57]. Homozygosity for loss-of-function mutations in HMG-CoA reductase have not been described in humans, as they are undoubtedly incompatible with life.

Mutations downstream from Hmgcr have been characterized in humans and mice (e.g., 24-dehydrocholesterol reductase [DHCR24]) [58]. Disruptions in DHCR24, the final step in the cholesterol synthetic process converting desmosterol to cholesterol, lead to congenital malformations and shortened lifespan in humans [59]. Dhcr24 ablation in mice leads to partial lethality in utero and failure to thrive post-birth [60]. Interestingly, the only difference between desmosterol and cholesterol is a steroid side-chain double bond saturated by DHCR24. DHCR24 inhibition by triparanol lowers cholesterol, however, triparanol was pulled from the market due to harmul side effects (e.g., teratogenicity) [61].

Germline knockout of ceramide-synthesizing genes Sptlc2 and Degs1 is also ultimately lethal [52,62]. Moreover, humans with inactivating mutations in both copies of the Degs1 gene develop neurological disorders [63–65]. However, studies on inducible knockouts, either in the whole body or in select tissues, have been instrumental for determining the biological roles of ceramides. In numerous instances, intervening in the ceramide synthetic pathway by either limiting ceramide biosynthesis or expressing ceramide degradation enzymes in metabolically active tissues leads to improved health (e.g., resolution of glucose intolerance, insulin resistance, hepatic steatosis, metabolic syndrome, etc.) [53,55,66]. Recently, we determined that these beneficial effects could be recapitulated by depleting the Degs1 gene that encodes DES1 from adult animals (~six weeks of age). Conditional depletion of the gene from most tissues (i.e. using an inducible cre-recombinase inserted into the ROSA26 locus) removes this double bond, as mice instead produce dihydroceramides that do not drive metabolic dysfunction. Knocking out Degs1 in adult animals leads to no gross defects or mortality. The knockout animals exhibit improved glucose tolerance on a chow or high fat diet while being protected from most comorbidities of obesity [53].

Tissue-Specific Manipulations of Ceramides

“My devil had been long caged, he came out roaring,” Robert Louis Stevenson

Both cholesterol and ceramides are major component of LDLs and atherosclerotic lesions. Moreover, ceramides have been implicated in plaque aggregation [67–69]. Beyond these activities in the bloodstream, investigations using tissue-specific knockout mice or isolated cells have revealed roles for cholesterol and ceramides in the tissue damage that further drives cardiometabolic pathologies. An exhaustive analysis of the mechanisms by which altered cholesterol content impairs the function of cardiomyocytes, endothelial cells, and other cell types is beyond the scope of this review.

Ceramides may prove to be as deleterious as cholesterol, as they elicit a non-overlapping spectrum of tissue defects and ultimately trigger cell death. Herein we limit our discussion to the phenotypes observed in various mouse models of ceramide depletion, contrasting them to the phenotypes seen following tissue-specific removal of HMG-CoA reductase.

Hepatic Ceramide and Cholesterol Interventions

Ablation of hepatic Hmgcr results in death by six weeks of age (all males and 75% of females)[70]. Prior to death, the mice display hypoglycemia, severe hepatic steatosis, and hepatocyte apoptosis, which can be rescued by supplementing with mevalonate, the direct product of HMG-CoA reductase [70].

By contrast, liver-specific ablation of Degs1 leads to no adverse phenotypes. Animals lacking hepatic Degs1 are resistant to high fat diet-induced hepatic steatosis and display improved glucose tolerance and insulin sensitivity [53]. Similarly, liver-specific overexpression of acid ceramidase—which degrades ceramide—resolves hepatic steatosis and enhances insulin action while improving liver (and adipose) health [66]. Furthermore, expressing the adiponectin receptors (e.g., AdipoR1 or AdipoR2), which contain intrinsic ceramidase activity [71], in the liver leads to targeted degradation of ceramide and improvements in whole-body glucose metabolism, including resolution of hyperglycemia and glucose intolerance caused by leptin deficiency [72,73].

Additional genetic manipulations indicate that the acyl-chain length of ceramide determines their function in the liver. Generally speaking, long-chain ceramides (C16–18) are harmful in the liver, while very long-chain ceramides (C20–26) are either benign or beneficial. Global knockout of Cers2, the predominant ceramide synthase in the liver and the enzyme that produces very long chain ceramides, induces fatty liver disease and insulin resistance [74]. This is attributable to a compensatory upregulation of Cers6 and an increase in long-chain C16 ceramides. Indeed, deletion of Cers6 from the liver resolves hepatic steatosis and insulin resistance caused by HFD or leptin deficiency [35].

Adipose Ceramide and Cholesterol Interventions

Adipose specific disruption of Hmgcr also impairs tissue function, leading to lipodystrophy and diabetes. Interestingly, mevalonate supplementation reversed these effects, but administration of other down-stream products (e.g., squalene) had no effect [75]. We conducted analogous studies by depleting Degs1 from adipocytes and found reduced adipocyte size, increased glucose uptake into the adipose tissue, and enhanced mitochondrial respiration [53]. These effects were strong enough to improve glucose tolerance and insulin sensitivity and resolve hepatic steatosis. Adipocyte-specific depletion of Sptlc2 [36] or overexpression of Asah1 [66] lowers ceramides and produces a comparable spectrum of benefits.

Analogous studies have been conducted in brown adipocytes. For example, use of a Ucp-1 driven cre allows for conditional depletion of genes from thermogenic adipocytes. Removing Cers6 [35] or Sptlc2 [76] increases, while depleting Asah1 inhibits [76], glucose metabolism and uncoupling in thermogenic adipocytes. In these instances, manipulating ceramides led to reciprocal effects on mitochondrial metabolism and adipocyte morphology, resulting in whole animal changes in glucose tolerance and liver fat accumulation [76].

Notably, two other groups found that depleting Sptlc isoforms from adipose tissue led to tissue dysfunction and lipodystrophy [77,78], and thus recapitulated the phenotype seen with the HMG-CoA reductase knockout animals. We previously speculated that this was a consequence of an earlier gene depletion, as pharmacological SPT inhibitors block adipocyte differentiation in vitro [36].

Cardiac Ceramide and Cholesterol Interventions

Particularly relevant to this discussion are the effects of these lipids in the heart, as both ceramides and cholesterol have been reported to have profound actions on cultured cardiomyocytes. No cardiomyocyte-specific Hmgcr ablation models appear to have been generated, and cardiomyocyte-specific constitutive depletion of Sptlc2 induces cardiac dysfunction [79]. Nonetheless, a fair bit of data suggests that ceramides may contribute to the cardiomyocyte dysfunction that underlies heart failure. For example, Ira Goldberg’s group elegantly showed that ceramide is a cardiotoxin by employing a cardiomyocyte-specific GPI-anchored lipoprotein lipase to induce ceramides and other lipid metabolites leading to lipotoxic cardiomyopathy. They demonstrated that treatment with myriocin (SPT inhibitor) or whole-body ablation of a single copy of the Sptlc1 gene preserved fractional shortening and prolongs life in this extreme model of lipotoxic heart failure [80]. In subsequent studies, Schwartz and colleagues observed that myriocin reversed heart failure in a mouse model of myocardial infarction (MI) [32]. Moreover, Drosatos and colleagues found that cardiomyocyte Krüppel-like factor-5 mediates MI-induced cardiac dysfunction by upregulating ceramide biosynthesis [81].

Lessons Learned from Human Genetics

“The Devil is in the Detail” attributed to Gustave Flaubert

Studies on the gene variants that influence the enzymes that produce cholesterol and ceramides have been less informative than the ones above that influence general LDL trafficking. For example, the search for disrupting mutations in HMGCR has been largely unsuccessful, likely owing to the critical importance of this pathway for life. Since more than 50% of individuals fail to achieve treatment goals with statins, researchers have conducted extensive analyses of the impact of genetic diversity to responses to this drug family. Single nucleotide polymorphisms have been identified that appear to influence interactions between statins and HMG-CoA reductase, thus altering the efficacy of this class of drugs [82]. Curiously, studies on individuals with these polymorphisms confirmed that the increased diabetes risk associated with statin use is likely due to on-target actions of the drug on HMG-CoA reductase [83]. Nonetheless, additional studies are needed.

The first mutations in the sphingolipid pathway were described in lysosomal storage disorders. For example, inactivating mutations in acid ceramidase lead to Farber’s disease, a condition characterized by accumulation of ceramides and a broad swath of tissue defects, leading to death in infancy. Other inborn errors of metabolism related to sphingolipids include Niemann-Pick disease, Gaucher’s disease, etc. The pathophysiology of sphingolipidoses varies, but each is a direct consequence of deleterious genetic mutations in the sphingolipid pathway. Enzyme replacement or lentivirus mediated gene therapy are emerging as potentially beneficial, but expensive, means of treating these conditions [84,85].

With the expansion of genome wide association studies, more subtle gene variants in sphingolipid-modifying enzymes are starting to emerge as markers of cardiometabolic disease risk. One of the most prevalent is a common missense mutation in the CERS2 gene (rs267738). The CERS2 mutation is found in approximately 20% of non-Finnish Europeans and is tightly associated with elevated HbA1c levels [86] and diabetic kidney disease [87,88]. Our group generated rs267738 knock-in mice and found that the animals exhibited decreased CERS2 activity and slightly impaired metabolic parameters, including worsening glucose tolerance and hepatic steatosis [89]. However, we were unable to detect differences in serum sphingolipids in patients with rs267738 from the Utah Coronary Artery Disease Cohort

Variants have also been found in the DEGS1 gene. Joanne Curran and John Blangero described a partial loss of function DEGS1 variant found in individuals with Hispanic ancestry. Heterozygosity for a L175Q DES1 mutation is associated with increased dihydroceramides, decreased ceramides, and no deleterious comorbidities [90]. Indeed, early studies revealed potential protection from cardiometabolic disease including decreased 2-hour glucose, cholesterol esters, and diabetes risk score [91]. By contrast, homozygous loss of function mutations in DEGS1 have been associated with neurological disorders (e.g., movement disorder, cerebellar atrophy, etc.) and increased dihydroceramide levels [63–65], recapitulating the phenotype seen in the germline Degs1 knockout mice.

Conclusions and Future Perspectives

“If you dance with the devil, the devil don’t change. The devil changes you,” Max California, 8MM

The efficacy and global health benefits of cholesterol-based therapies (e.g., statins, PCSK9 inhibitors) have changed human health for the better. The rationale for cholesterol-lowering strategies emerged in the early 20^th century, when sphingolipids such as ceramide research was limited to a small number of esoteric laboratory projects. Despite rigorous work of the thriving sterol research community, it took approximately 50 years for cholesterol-lowering drugs (e.g., statins) to be approved as atherosclerotic cardiovascular disease treatment. With years of data derived from robust experiments, we now know that lowering cholesterol levels leads to improved cardiovascular health outcomes [2]. Indeed, meta-analysis reveals that a 40 mg/dL drop in cholesterol produces a 20% decrease in CVD risk [2]. In a testament to the indefatigable researchers working on improving our knowledge of cholesterol-based interventions, we are discovering new methods of improving the current standards of care—therapies and clinical protocols made possible by these remarkable discoveries.

One wonders whether someday ceramide-lowering therapeutics could have the same value as statins (see Outstanding Questions Box). In humans, large profiling studies suggest that “hyperceramidemia” is affiliated with cardiometabolic disease progression (i.e. heart disease and diabetes) and mortality. Moreover, gene variants that alter ceramide profiles are starting to emerge in association with diabetes and its comorbidities. In rodents, ceramide-lowering interventions unequivocally mitigate cardiometabolic disease severity. The compilation of evidence suggests that ceramides, occasionally in partnership with cholesterol, contribute to human cardiometabolic pathologies.

Outstanding Questions.

The discoveries surrounding cholesterol were inspired by the discovery of gene mutations that drove hypercholesterolemia and increased susceptibility to cardiovascular disease. Do similar genetic abnormalities drive “hyperceramidemia” and influence susceptibility to heart disease, diabetes, and obesity-related disorders?

The studies surrounding cholesterol-sensing in the cell were elegant and impactful. They revolutionized our understanding of cell biology and had enormous relevance to disease. By contrast, the mechanisms by which cells translate small changes in ceramides into robust alterations in metabolism and survival remain enigmatic. The low abundance of these molecules has made such careful dissection of intracellular pathways exceptionally challenging. How do modest elevations in ceramides elicit such a broad spectrum of deleterious actions in cells and tissues?

Newly-synthesized or absorbed fatty acids undergo one of three metabolic fates. They can be coupled to carnitine and carried into mitochondria, where they are substrates for fatty acid oxidation. They can be coupled to glycerol, becoming the glycerophospholipids that comprise the bulk of cell membranes or lipid droplets. Or they can be coupled to amino acids and become sphingolipids. What regulatory signals and downstream effectors control this lipid channeling?

Ceramide scores have emerged as clinical tools to gauge disease risk. But what evidence-based recommendations do we give someone with a high ceramide score?

Ultimately, the most obvious and important question is whether one can safely lower ceramide levels to ameliorate disease. Here the great efficacy of cholesterol-lowering drugs gives us hope. What enzymatic targets in the ceramide pathway will prove to be the most efficacious and therapeutically tractable?

Addressing these areas of inquiry could have important implications for our understanding and treatment of cardiometabolic disease.

Like cholesterol, the major test of therapeutic relevance will require the development and testing of ceramide biosynthesis inhibitors. Can such compounds be both safe and effective? Additional studies on this exciting counterpart to cholesterol continue to present exciting potential as a novel treatment paradigm for combating a wide range of prevalent cardiometabolic pathologies.

Box 1. Discovery of the Less Abundant Ceramides.

“The greatest trick the Devil ever pulled was convincing the world he didn’t exist”—Charles Baudelaire

Cholesterol-lowering interventions significantly reduce heart disease burden, but they do not prevent all metabolic pathologies. Statins actually increase risk of diabetes, for example [83]. Might other bioactive lipids contribute to cardiometabolic diseases (e.g., heart disease, stroke, diabetes, etc.)?

Of all the lipid species identified, sphingolipids are attractive as potential contributors to cardiometabolic pathologies. They were discovered in the 19^th century by JLW Thudichum, who was characterizing the various macromolecular components of the brain. He found a unique class of lipids that lacked the glycerol moiety that is present in most phospholipid species. He named them after the Sphinx, owing to their unusual dichotomous structure and enigmatic nature. They are minor components of the lipidome; circulating sphingolipids are approximately 1/1000^th the level of cholesterol, and rigorous quantification requires mass spectrometry. As a result, large clinical studies of the class of lipids didn’t occur until the 21^st century.

The sphingoid base that replaces glycerol in this lipid class is made by the condensation of a saturated fatty acid and an amino acid. The enzyme that catalyzes this reaction, serine palmitoyltransferase (SPT), favors palmitoyl-CoA and serine as substrates, but can occasionally use other species. The condensation of palmitoyl-CoA and serine produces 3-ketosphinganine, a transient intermediate that is rapidly reduced, which removes a highly reactive ketone to produce sphinganine. Sphinganine is then acylated by one of six distinct (dihydro)ceramide synthases (CERS) to create the dihydroceramides. The CERS enzymes are the first branch point of sphingolipid synthesis, and they account for much of the diversity in the sphingolipid class. Each CERS acylates the sphingoid backbone with a subset of specific fatty acyl-CoAs, generally ranging from 14 to 26 carbons. The dihydroceramides are further modified by the introduction of an important trans-4,5 enol double bond by the enzyme dihydroceramide desaturase (DES1), producing ceramides. Ceramides can then be further modified by the addition of hexoses, choline, or a phosphate group to produce complex sphingolipid species such as glucosylceramides, sphingomyelin, or ceramide-1-phosphate. Ceramides can also be deacylated by a family of ceramidases, which produces sphingosine. This latter intermediate can be phosphorylated to produce the bioactive lipid, sphingosine 1-phosphate, which is a ligand for different G-protein coupled receptors and a requisite intermediate in the pathway leading to ultimate destruction of the sphingoid base.

Highlights.

• Pioneering studies on cholesterol have revolutionized our understanding of the role of lipids in cardiovascular disease

• The widely-prescribed statins inhibit cholesterol biosynthesis to prevent coronary artery disease.

• By lowering liver cholesterol, statins reduce levels of circulating lipoproteins and decrease levels of numerous serum lipids (including ceramides).

• Like cholesterol, ceramides contribute to some of the pathologies associated with obesity, but they have a distinct spectrum of actions; they alter tissue metabolism and induce apoptosis.

• In rodents, blocking ceramide accumulation ameliorates a spectrum of disorders including diabetes, fatty liver disease, and heart disease.

• Serum ceramides correlate strongly with markers of diabetes and heart disease

• Therapeutic approaches that lower ceramides could have efficacy in a broad spectrum of metabolic disorders