|Home | About | Journals | Submit | Contact Us | Français|
Obesity, type 2 diabetes mellitus (T2DM), and non-alcoholic steatohepatitis (NASH) are associated with cognitive impairment, brain insulin resistance, and neurodegeneration. Recent studies linked these effects to increased pro-ceramide gene expression in liver and increased ceramide levels in serum. Since ceramides are neurotoxic and cause insulin resistance, we directly examined the role of ceramides as mediators of impaired signaling and central nervous system function using an in vivo model. Long Evans rat pups were administered C2Cer:N-acetylsphinganine or its inactive dihydroceramide analog (C2DCer) by i.p. injection. Rats were subjected to rotarod and Morris water maze tests of motor and cognitive function, and livers and brains were examined for histopathology and integrity of insulin/IGF signaling. C2Cer treatment caused hyperglycemia, hyperlipidemia, and mild steatohepatitis, reduced brain lipid content, and increased ceramide levels in liver, brain, and serum. Quantitative RT-PCR analysis revealed significant alterations in expression of several genes needed for insulin and IGF-I signaling, and multiplex ELISAs demonstrated inhibition of signaling through the insulin or IGF-1 receptors, IRS-1, and Akt in both liver and brain. Ultimately, the toxic ceramides generated in peripheral sources such as liver or adipose tissue caused sustained impairments in neuro-cognitive function and insulin/IGF signaling needed for neuronal survival, plasticity, and myelin maintenance in the brain. These findings support our hypothesis that a liver/peripheral tissue-brain axis of neurodegeneration, effectuated by increased toxic lipid/ceramide production and transport across the blood-brain barrier, could mediate cognitive impairment in T2DM and NASH.
Obesity, type 2 diabetes mellitus (T2DM), nonalcoholic steatohepatitis (NASH), Alzheimer’s disease (AD), and experimental AD-type neurodegeneration produced by intra-cerebral streptozotocin treatment are all associated with insulin resistance, oxidative stress, mitochondrial dysfunction, and pro-inflammatory cytokine activation [1–9]. Increasing evidence suggests that ceramides play a major role in the pathogenesis of obesity, T2DM, and NASH [10–13] because ceramides cause insulin resistance [14–18] and they activate proinflammatory cytokines. On the other hand, ceramide synthesis is stimulated by pro-inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α) , and pro-inflammatory cytokines are highly activated in obesity, T2DM, NASH, and AD [19–25], and previous studies demonstrated that ceramide levels are increased in AD brains [12,26]. We hypothesize that ceramides mediate some aspects of brain insulin resistance associated with both AD and T2DM/obesity mediated neurodegeneration because ceramides: 1) can be generated in brain [10,13,27,28]; 2) are increased in various dementia-associated diseases, including AD [10,12,29, 30]; and 3) are lipid soluble and therefore likely to readily cross the blood-brain barrier, providing a mechanism by which obesity, T2DM, or NASH could lead to brain insulin resistance.
Ceramides represent a family of lipids generated from fatty acid and sphingosine [13,31,32]. Ceramides are distributed in cell membranes, and in addition to structural functions, they regulate signaling pathways that mediate growth, proliferation, motility, adhesion, differentiation, senescence, and apoptosis. Ceramides are generated biosynthetically through ceramide synthase and serine palmitoyltransferase activities [27,28, 33]. Alternatively, ceramides are generated by sphingolipid catabolism through activation of neutral or acidic sphingomyelinases [28,32], or degradation of complex sphingolipids and glycosphingolipids localized in late endosomes and lysosomes . Ceramides are metabolized to sphingosine by ceramidases, and ceramide, sphingosine, and sphingosine-1-phosphate are implicated in the pathogenesis of obesity and insulin resistance . Correspondingly, inhibition of ceramide synthesis or its accumulation prevents obesity-associated insulin resistance [14,17].
Complex sphingolipids including gangliosides , and long-chain naturally occurring ceramides (i.e., up to 24 carbon atoms in length) ,stimulate cell growth and functions, whereas sphingosine-containing lipids, including shorter ceramides, have inhibitory effects, resulting in increased apoptosis and cytotoxicity, or impaired growth [34,36,37]. Sphingomyelinases are activated by pro-inflammatory cytokines (i.e., TNF-α ), and pro-apoptotic stimuli including ionizing radiation, Fas, and trophic factor withdrawal [31,32]. Ceramides impair cellular functions and cause apoptosis by: 1) modulating the phosphorylation states of various protein, including those that regulate insulin signaling ; 2) activating enzymes such as interleukin-1β converting enzyme (ICE)-like proteases, which promote apoptosis ; or 3) inhibiting Akt phosphorylation and kinase activity  through activation of protein phosphatase 2A .
In obesity, adipose tissue, skeletal muscle, and liver tissue exhibit abnormal sphingolipid metabolism that results in increased ceramide production, inflammation, and activation of pro-inflammatory cytokines, and impairments in glucose homeostasis and insulin responsiveness [13,16,28]. In both humans with NASH , and the C57BL/6 mouse model of diet-induced obesity with T2DM and NASH , ceramide levels in adipose tissue are elevated due to increased activation of sphingomyelin transferase, and acidic and neutral sphingomyelinases . In addition, ceramide synthase and sphingomyelin transferase mRNA levels in liver are increased during the early stages of hepatic steatosis, but with emergence of NASH and neurodegeneration, those mRNA transcripts decline while sphingomyelinase gene expression increases . Since the neurodegeneration in diet-induced obesity was not associated with increased central nervous system (CNS) expression of pro-ceramide genes, we extended the analysis by directly investigating the role of exogenous ceramide exposure in the pathogenesis of neurodegeneration and brain insulin resistance using an in vivo model.
Ceramide analogs, D-erythro-Ceramide (C2Cer:N-acetyl-D-erythro-sphingosine, C6Cer: N-Hexanoyl-Derythro-Sphingosine), dihydroceramide analog (C2D Cer; Dihydro-N-Acetyl-D-erythro-Sphingosine) were purchased from CalBiochem (San Diego, CA). Histochoice fixative was purchased from Amresco, Inc. (Solon, OH). The Amplex UltraRed soluble fluorophore, and the Akt Pathway Total and Phospho 7-Plex Panels were purchased from Invitrogen (Carlsbad,CA). MaxiSorb 96-well plates used for ELISAs were from Nunc (Thermo Fisher Scientific; Rochester, NY). QIAzol Lysis Reagent for RNA extraction and QuantiTect SYBR Green PCR Mix were obtained from Qiagen, Inc (Valencia, CA). The AMV 1st Strand cDNA Synthesis kit and Universal Probe Library and rat β-actin reference gene assay were purchased from Roche Applied Science (Indianapolis, IN). Monoclonal anti-ceramide, polyclonal anti-phospho-Tau (pS199/202-Tau), and Tau, and synthetic oligonucleotides used in quantitative polymerase chain reaction (qPCR) assays were purchased from Sigma-Aldrich Co. (St. Louis, MO). Fine chemicals were purchased from CalBiochem (Carlsbad, CA) or Sigma-Aldrich (St. Louis, MO). The triglyceride assay kit was from Sigma (St. Louis, MO).
Long Evans rat pups were given 7 alternate day intraperitoneal (i.p.) injections of 2.0 mg/kg (50 µl volumes) ceramide analogs, D-erythro-N-Acetyl-Sphingosine (C2Cer) or D-erythro-Dihydro-N-Acetyl-Sphingosine (C2 Dihydroceramide; C2D), beginning on postnatal day 3. C2D is a structurally similar, inactive analog of C2Cer, and was used as a negative control. The doses of C2Cer and C2DCer were within the ranges employed to generate other in vivo models [44–50], and the concentrations we used previously to demonstrate ceramide-mediated neuronal insulin resistance in vitro . In addition, we performed empirical studies to assess the dose range and time course of treatment that were not acutely toxic, yet caused peripheral insulin resistance. Ceramide reagents were dissolved in ethanol and diluted in sterile saline prior to use. All animals survived the procedure, and did not exhibit any aberrant behavior or adverse responses such as failure to thrive, poor grooming, reduced physical activity, or weight loss. Rats were weighed weekly. Rats were subjected to rotarod testing on P15-P16, and Morris water maze testing on P24–P28. On P30, after an overnight fast (14 h), rats were sacrificed by i.p. injection of 120 mg/kg pentobarbital, and blood, liver, and brain were harvested.
Blood or serum was used to measure glucose, insulin, neutral lipid, and ceramide levels as previously described [43,52]. Brain glucose levels were was measured in PBS homogenates of temporal lobe tissue using a glucometer and results were normalized to sample protein concentration (µg/mg protein). Cerebella, temporal lobes, and liver were harvested for histopathological, biochemical, and molecular studies. For histopathology, tissue samples were immersion fixed in Histochoice and embedded in paraffin. Histological sections of brain (8-µm thick) were stained with Luxol Fast Blue, Hematoxylin, and Eosin (LHE), while liver sections were stained with H&E. For molecular and biochemical assays, brain and liver tissues were snap-frozen in a dry ice-methanol bath and stored at −80°C. We studied cerebella and temporal lobes because both brain regions: 1) require intact insulin/IGF signaling to maintain their structural and functional integrity [53,54]; and 2) they are targets of neurodegeneration in insulin-resistance diseases [8,55–58]. Our experimental protocol was approved by the Institutional Animal Care and Use Committee at Lifespan-Rhode Island Hospital, and conforms to the guidelines set by the National Institutes of Health.
We used rotarod testing to assess long-term effects on motor function  resulting from the i.p. ceramide treatments. On P15, rats were trained to remain balanced on the rotating Rotamex-5 apparatus (Columbus Instruments) at 1–4 rpm. On P16, rats (n = 8–10 per group) were administered 10 trials at incremental speeds up to 4 rpm, with 10 min rest between each trial. The latency to fall was automatically detected and recorded with photocells placed over the rod. However, trials were stopped after 30 s to avoid exercise fatigue. Data from trials 1–3 (1–2 rpm), 4–7 (2.5–3.5 rpm), and 8–10 (4 rpm) were culled and analyzed using the Mann-Whitney test.
Morris water maze testing  of spatial learning and memory was performed on 4 consecutive days as previously described [3,4]. On the first day of testing, the rats were oriented to the water maze and educated about the location of the platform. On the 3 subsequent days of testing, the platform was submerged just below the surface, and rats were tested for learning and memory by measuring the latency period required to reach and recognize the platform. The rats were placed in the same quadrant of the water maze for every trial on Days 1 and 2, but on days 3 and 4, the start locations were randomized. Data from the 3 trials each day were used to calculate latency area under the curve. Inter-group comparisons were made using the Mann-Whitney test.
Lipid analyses were performed with serum samples and chloroform-methanol (2:1) extracted fresh frozen liver and brain homogenates . Total lipid content was measured using a Nile Red fluorescence-based assay [61–63], and fluorescence intensity (Ex 485/Em 572) was measured in a SpectraMax M5 microplate reader (Molecular Devices Corp., Sunnyvale, CA). Triglyceride levels were measured in lipid extracts using a commercial colorimetric assay kit. Ceramide immunoreactivity was measured by direct-binding ELISA  using 96-well Polysorp black plates (Nunc, Rochester, NY) . In brief, lipids (50 µl in methanol) were adsorbed to well bottoms for 2 h at room temperature, then blocked for 1 h with Superblock-TBS, and incubated with monoclonal anti-ceramide (2 µg/ml) overnight at 4°C. Immunoreactivity was detected with horseradish peroxidase (HRP)-conjugated secondary antibody (1:10000) and enhanced chemiluminescence substrate (ECL) . Luminescence was measured in a TopCount NXT (Packard, Meriden, CT). Positive control reactions included spotting known quantities of C2 or C6 synthetic ceramide into the wells. Immunoreactivity was normalized to sample protein content. Negative control reactions included substitutions with nonrelevant primary or secondary antibodies, and omission of primary or secondary antibody.
We used qRT-PCR to measure mRNA expression as previously described [4,66,67]. In brief, tissues were homogenized in Qiazol reagent (Qiagen Inc., Valencia, CA), and total RNA was isolated using the EZ1 RNA universal tissue kit and the BIO Robot EZ1 (Qiagen, Inc., Valencia, CA). RNA was reverse transcribed using random oligodeoxynucleotide primers and the AMV First Strand cDNA synthesis kit. The resulting cDNA templates were used in probe-based qPCR amplification reactions with gene specific primer pairs as reported previously . Primers were designed using ProbeFinder software (Roche, Indianapolis, IN), and target specificity was verified using NCBI-BLAST (Basic Local Alignment Search Tool). The amplified signals from triplicate reactions were detected and analyzed using the Mastercycler ep realplex instrument and software (Eppendorf AG, Hamburg, Germany). Relative mRNA abundance was calculated from the ng ratios of specific mRNA to β-actin mRNA measured simultaneously in duplex PCR reactions. Inter-group statistical comparisons were made using the calculated mRNA/β-actin ratios.
Tissue homogenates were prepared in NP-40 lysis buffer containing protease and phosphatase inhibitors, as previously described . Protein concentration was measured using the bicinchonic acid assay. To examine signaling through the insulin and IGF-1 receptors and downstream through IRS-1 and Akt in liver tissue, we used a bead-based multiplex ELISA and measured immunoreactivity to the insulin receptor (IR), IGF-1 receptor (IGF-1R), IRS-1, Akt, and glycogen synthase kinase 3β (GSK-3β), and pY pY 1162/1163-IR, pY pY 1135/1136-IGF-1R, pS312-IRS-1, pS473-Akt, and pS9-GSK3β. For these studies, liver tissue was homogenized in NP-40 lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet-P (NP-40)), containing protease and phosphatase inhibitors . Samples containing 100 µg protein were incubated with the beads, and captured antigens were detected with secondary antibodies generated in goat, and phycoerythrin-conjugated anti-goat antibody. Plates were read in a Bio-Plex 200 system (Bio-Rad,Hercules, CA). In addition, the samples were analyzed using direct binding ELISAs to measure myelin-associated glycoprotein-1 (MAG-1), Hu (neuronal marker), tau, and phospho-tau as previously described .
Data depicted in the tables represent the mean ± SEM for each group. The box-and-whisker plots depict the lower quartile (25th percentile)-bottom of box, median (horizontal bar), upper quartile (75th percentile), and range (i.e., the lowest value is represented by the bottom whisker, and the highest value is represented by the top whisker). Eight independent samples were included in each group. Inter-group comparisons were made with the Student T test (gene expression and immunoreactivity) or the Mann-Whitney-U test (rotarod and Morris Water Maze) using the GraphPad Prism 5 software (GraphPad Software, Inc., San Diego, CA). Software generated significant p-values are shown in the graphs, and included in the tables.
Age-associated increases in mean body weight, and the mean body and brain weights measured at the time of sacrifice were similar for C2Cer or C2DCer (negative control) treated rats (Fig. 1A–C). However, the mean fasting blood glucose (Fig. 1D), serum total lipid content (Nile Red assay), and serum ceramide immunore-activity were significantly higher in C2Cer-treated relative to C2DCer-treated control rats. Brain glucose levels (µg/mg protein) were also significantly increased in the C2Cer– (33.6 ± 1.22) relative to C2DCer-treated controls (26.0 ± 1.49) (p = 0.0009). In contrast to serum, total lipid content in brain and liver were significantly lower in the C2Cer relative to control. In addition, in serum, brain, and liver, triglyceride concentrations were all significantly reduced in the C2Cer-treated relative to C2DCer-treated control rats. As observed in serum, ceramide levels in liver and brain were significantly higher in the C2Cer-treated rats. The proportional increases in ceramide levels detected in this model are consistent with the findings in previous studies of diet-induced obesity, chronic alcohol feeding or nitrosamine exposure [43,68–70]. Therefore, intraperitoneal treatment with bioactive C2Cer ceramide in the early postnatal period caused hyperglycemia, hyperlipidemia, and increased circulating and tissue levels of ceramides. The ceramides present in serum could potentially cross the blood-brain barrier and cause CNS injury and insulin resistance, and thereby result in further endogenous ceramide production and accumulation with attendant neurodegeneration.
Rotarod tests performed at low speed revealed no significant differences in performance between the C2Cer-and C2DCer-treated controls. However, with increasing speed of rotation, the C2Cer treated rats performed significantly worse than control, as manifested by their shorter latencies to fall (Fig. 2B,C). Morris Water maze testing demonstrated that C2Cer-treated rats exhibited significantly longer latencies for learning how to locate and land on the platform during the acquisition phase (Fig. 2D), but on the two subsequent days, their performance was not significantly different from control, indicating relative preservation of memory once the task had been learned. However, on the 4th and final day of testing, the C2Cer treated rats exhibited significantly longer latencies in locating the hidden platform from randomized quadrants of the maze (Fig. 2D). While the poorer performance on Day 1 might have been due to higher levels of anxiety in the C2Cer-treated rats , this phenomenon would not likely account for the significantly impaired performance on Day 4, or the deficits in motor function observed with increasing speed of the rotarod. In addition, the longer latencies measured in the C2Cer group were associated with to rapid swimming, but in varied (seemingly random) directions, particularly on Day 1, reflecting difficulty learning to locate and land on the platform, and Day 4, when the platform was hidden and the starting points were randomized. These results indicate that the C2Cer treatment caused significant abnormalities in spatial leaning and memory.
Livers from control (C2D injected) rats exhibited regular chord-like architecture with minimal or no inflammation or steatosis, whereas livers from C2Cer-treated rats also exhibited regular chord-like architecture, but had evidence of mild steatohepatitis characterized by the presence of multiple foci of lymphomononuclear cell inflammation, and scattered areas of microvesicular steatosis, apoptosis, and hepatocellular necrosis (Fig. 3A–D). In contrast, there were no consistent histopathological abnormalities detected in the brains, including cerebella, hippocampi, and temporal lobes of C2Cer- compared with C2D-treated rats (data not shown). C2Cer treatments resulted in significantly reduced total neutral lipid and triglyceride content, but significantly increased levels of ceramide in both liver and brain (Table 2).
Immunoreactivity corresponding to Hu (neurons), MAG-1 (oligodendrocyte myelin), Tau, and phospho-Tau (pS199/202-Tau) was measured in temporal lobe and cerebellar tissue by direct binding ELISA with results normalized to ribonuclear protein levels measured in the same samples. These studies demonstrated significantly reduced mean levels of Hu in the temporal lobe (p = 0.0015), and increased levels of phospho-Tau and Tau protein in the cerebellum of C2Cer-treated relative to controls (Table 3). In contrast, no significant inter-group differences were measured with respect to MAG-1 in either temporal lobe or cerebellum, Hu in the cerebellum, or phospho-tau and tau in the temporal lobe.
We used qRT-PCR analysis to quantify long-term effects of ceramide exposure on gene expression corresponding to insulin and IGF polypeptides and receptors, and insulin receptor substrate (IRS) molecules that transmit signals required for growth, survival, energy metabolism, and neuronal plasticity downstream of the insulin and IGF receptors (Fig. 4). Early C2Cer exposure significantly reduced insulin, IGF-2, IRS-1, and IGF-1 receptor gene expression, and increased insulin receptor and IRS-2 expression in liver. In addition, C2Cer treatments resulted in significantly reduced mRNA expression of insulin and IGF-1 receptor, and increased expression of IGF-1 in brain (Fig. 5). In contrast to liver, IRS gene expression in brain was not significantly modulated by C2Cer exposure.
Multiplex bead-based ELISAs were used to measure sustained effects of C2Cer treatment on insulin and IGF signaling mechanisms in liver and brain. We measured total and phosphorylated levels of insulin receptor (pYpY1162/1163), IGF-1 receptor (pYpY1135/1136), IRS-1 (pS312), Akt (pS473), and GSK-3β (pS9), and calculated the phospho-/total ratios to assess relative levels of phosphorylation. C2Cer treatments did not significantly alter the mean levels of total or phosphorylated insulin receptor, IGF-1 receptor, IRS-1, or Akt, but reduced the relative levels of IRS-1 and Akt phosphorylation in liver (Fig. 6). In addition, C2Cer-exposed livers had significantly reduced levels of total GSK-3β, and increased mean levels of pS9-GSK-3β and pS9-GSK-3β/GSK-3β relative to C2DCer-treated livers. With respect to the brain, the C2Cer exposures resulted in significantly reduced mean levels of AkT, pY pY 1162/1163-insulin receptor/total insulin receptor, and pS312-IRS-1/total IRS-1, and increased levels of insulin receptor and IRS-1 immunoreactivity in the cerebellum (Fig. 7). In contrast, in the temporal lobe, the sustained effects of C2Cer treatment on insulin/IGF signaling mechanisms were more limited in that pS473-Akt/total Akt was significantly lowered, while total IRS-1 protein immunoreactivity was significantly increased (Fig. 8).
The current study represents an extension of previous work demonstrating that in various disease states of peripheral insulin resistance, including diet-induced obesity and nitrosamine exposure, the expression of several genes regulating ceramide production via de novo biosynthesis or sphingomyelin degradation pathways was increased in liver, and ceramide levels (immunoreactivity) were increased in liver and/or blood [2, 43,51,69]. Importantly, these abnormalities were associated with brain insulin resistance and mild neurodegeneration [43,52]. For example, in experimental diet-induced obesity with T2DM and NASH, ceramide gene expression was shown to be increased in liver, and accompanied by mild neurodegeneration with brain insulin/IGF resistance . With alcoholor nitrosamine-induced steatohepatitis, pro-ceramide gene expression in liver was correlated with hepatic and brain insulin/IGF resistance and tissue injury [2,69]. Finally, in vitro experiments demonstrated that direct exposure to cytotoxic ceramides impairs liver and brain cell viability, mitochondrial function, and insulin/IGF signaling mechanisms , consistent with previous reports [10,12,13,17,18,29,43,51,72].
The main objective of this study was to demonstrate the potential role of cytotoxic ceramides originating from the periphery as mediators of neurodegeneration. We and others have shown that in peripheral insulin resistance diseases associated with brain insulin resistance and neurodegeneration, serum and hepatic levels of ceramides, and other toxic lipids are increased [2, 43,51,69]. Moreover, we demonstrated that in vitro cytotoxic ceramide exposure causes neurodegeneration with impaired viability, energy metabolism, and insulin/IGF signaling in neuronal cells . Although generally longer chain ceramides have been detected in insulin-resistance disease models and humans [27,33, 72–75], in our studies we used relatively short-chain synthetic ceramides because this approach has been validated in a number of experimental models, and the compounds are known to be cell permeable, impair signaling, and promote inflammation, mitochondrial dysfunction, and cell death [39,46,48,49,51,76], all of which are features of insulin resistance diseases. Future studies will employ longer chain synthetic ceramides, once their bio-distributions and cell permeability characteristics have been determined.
In the current study, we did not regenerate models of obesity, T2DM, or NASH, and instead focused on testing the hypothesis that peripherally administered cytotoxic ceramides can cause brain insulin resistance with impairments of cognitive and motor functions. Therefore, we examined the degree to which limited early-life bioactive ceramide exposure leads to both hepatic and CNS insulin/IGF resistance and neurodegeneration. This study is novel because it directly examines the role of extra-CNS ceramides in relation to neurobehavioral deficits and brain insulin resistance in the absence of confounders produced by chronic high dietary fat intake, exposures to alcohol or toxins, and aging. The expectation was that ceramide exposure would impair insulin/IGF-1 signaling mechanisms in liver and brain, irrespective of peripheral blood indices of insulin resistance because the toxic lipids were deemed the culprits rather than hyperglycemia or hyperlipidemia per se. Correspondingly, in this study, C2Cer-induced hyperglycemia and hyperlipidemia were modest, and triglyceride levels were in fact reduced in serum, yet serum, liver, and brain ceramide levels were significantly increased. The significance of this work is that it helps establish mechanistic links among hepatic insulin/IGF resistance, lipotoxicity states, cognitive impairment, and neurodegeneration associated with deficits in brain insulin/IGF signaling. Moreover, the findings suggest that examining peripheral blood levels of ceramides may aid in identifying individuals at risk for developing cognitive impairment in the setting of obesity, irrespective of traditional biomarkers of type 2-diabetes and peripheral insulin resistance.
The over-arching hypothesis is that ceramides, which are recognized mediators of insulin resistance with demonstrated inhibitory effects on PI3K-Akt signaling [14,15], may mediate neuro-cognitive deficits, brain insulin resistance, and neurodegeneration in the context of obesity, T2DM, and NASH. In this regard, we propose that ceramides generated from the periphery (i.e., liver or perhaps adipose tissue), cross the blood-brain barrier to mediate these adverse effects on brain structure and function. The lipid-soluble nature of ceramides makes it feasible for this class of lipids to regulate and alter brain function. This phenomenon could explain how obesity and T2DM pose increased risk of cognitive impairment and neurodegeneration in humans [77–80]. In the present study, since it is now known which species of ceramides may be responsible for neurodegeneration in obesity and T2DM, and previous studies demonstrated that specific cell permeable synthetic ceramides impair insulin/IGF signaling and are cytotoxic in vitro [2,13–15,17,18,51], we utilized synthetic bioactive (C2Cer) and inactive (C2DCer) ceramide molecules to test our hypothesis.
Corresponding to previous findings that ceramides promote insulin resistance [2,13–15,17,18,51], we observed that following i.p. treatment of rat pups with synthetic C2Cer, as adolescents, the rats exhibited mild hyperglycemia and hyperlipidemia accompanied by increased serum ceramide levels. In addition, the livers showed evidence of on-going inflammation and injury with reduced lipid content but increased ceramide levels, and the brains exhibited normal histology, but had reduced lipid content and increased ceramide levels as well. The somewhat unexpected finding of reduced triglyceride levels in brain, liver and serum of C2-Ceramide treated versus control rats is consistent with results in another recent study in which it was suggested that ceramide treatments may mediate this effect by inhibiting adipogenesis .
Degradation of sphingolipids promotes ceramide generation, which can have adverse effect on intracellular signaling, cell survival, and inflammatory mediators . Correspondingly, we detected significant reductions in neuronal Hu expression in the temporal lobes of C2Cer-treated rats. This finding corresponds with the impairments in spatial learning and memory detected by Morris water maze testing. In the brain, myelin is the most abundant lipid, and myelin maintenance via oligodendroglial metabolism, is regulated by insulin and IGF signaling . Although we did not detect any reductions in MAG-1 immunoreactivity, conceivably, the impairments in insulin signaling effectuated by the C2Cer treatments, and as demonstrated in previous experiments , led to dysregulated lipid metabolism and increased ceramide generation, which in turn, further impaired insulin/IGF signaling mechanisms, and at least in liver, also impaired cell survival.
Although there were no overt histopathological abnormalities detected in brain, the reductions in neutral lipid and triglyceride, and increased ceramide levels may reflect the early stage of neurodegeneration. Correspondingly in previous studies, it was demonstrated that relatively early abnormalities in AD include white matter atrophy and increased ceramide content in brain [26,84]. Moreover, the impairments in cognitive and motor functions were also associated with reductions in Hu (reflecting neuronal injury or loss), and increased levels of phospho-tau and tau immunoreactivity in brains of C2Cer-treated rats. Therefore, neurobehavioral (functional), biochemical, and molecular abnormalities in brain may provide more sensitive indices of neurodegeneration, and precede many of the structural changes detected by histopathological examination. Even in humans with mild cognitive impairment, structural neurodegenerative lesions are often mild, focal, or absent [85–87]. As ceramides and other toxic sphingolipids are generated by myelin breakdown or altered biosynthesis, virtually any pathophysiological process that leads to their accumulation would also impair CNS function. Therefore, we interpret the CNS functional impairments to be consequential to combined effects of neuronal loss/degeneration precipitated by C2Cer-mediated insulin/IGF resistance, and attendant locally increased ceramide generation. At this point, it is not possible to know the relative contributions of exogenous versus endogenous ceramides mediating brain insulin resistance and neurobehavioral deficits; nonetheless, what is clear is that the neurodegeneration process can be initiated by cytotoxic ceramides generated in the periphery, i.e. outside of the CNS. While ceramide levels are increased in sera, skeletal muscle, adipose tissue, and/or liver in peripheral insulin resistance diseases [13,16,18,41,43,88,89], and in AD brains [90–92], the levels cannot be accurately quantified for comparison with our experimental model due to heterogeneity of the expressed or accumulated ceramides and other sphingolipids and the lack of standardized methods for measuring such compounds.
The molecular and biochemical studies demonstrated that the early postnatal treatment with C2Cer impaired insulin and IGF signaling mechanisms in both livers and brains of the adolescent rats. The major impact was on the expression and/or phosphorylation state of the insulin receptor, IGF-1 receptor, IRS-1 or Akt. These findings are consistent with previous reports demonstrating that ceramides impair insulin signaling through the Akt pathway [14,15,39,93], and also with our previous findings that bioactive synthetic ceramides (C2Cer or C6Cer) impair insulin/IGF signaling through inhibition of receptor and IRS expression and function.
The insulin/IGF-1-IRS-1-Akt signaling pathway mediates cell survival, energy metabolism, neuronal plasticity, and neurotransmitter function . Therefore, the observed C2Cer-mediated impairments of this pathway could account for the observed ongoing hepatocellular injury and cognitive-motor deficits. It is noteworthy that we detected increased levels of IRS-1 and/or insulin receptor protein vis-à-vis reduced relative levels of phosphorylated receptor and IRS-1. Conceivably, the increased protein levels reflect a compensatory protective response to the impairments in signaling that limited the degree and rate of cellular injury and death. The same argument could be made for the seemingly paradoxical increases in pS9-GSK-3β in C2Cer-exposed livers.
Although the magnitudes of these effects were variable, the aggregate effects of C2Cer treatment were to reduce insulin ± increase insulin receptor gene expression, and reduce IGF-1 receptor gene expression in liver and brain, inhibit signaling downstream through IRS-1 and Akt with increasing GSK-3β activity in liver, and constitutively impairing insulin signaling at the level of the receptor, IRS-1, or Akt in the brain. These results, together with the reduced levels of insulin gene expression in brain, are reminiscent of the findings of both insulin resistance and insulin deficiency in brains with AD [7,8]. On the other hand, this model clearly does not replicate the abnormalities in AD, and instead seems more closely aligned with the effects of obesity and peripheral insulin resistance [43,52,68,70]. However, in future studies, it will be of interest to examine the effects of aging in relation to ceramide-mediated neurodegeneration.
In conclusion, this study demonstrates that limited in vivo exposure to bioactive toxic ceramides causes mild diabetes mellitus with hyperlipidemia, hepatocellular injury, deficits in cognitive and motor functions, and impairments in insulin/IGF signaling though IRS-1 and Akt. The importance of this work is that it demonstrates that peripherally generated ceramides, such as occurs in obesity, T2DM, alcoholic liver disease, and nitrosamine exposure [1,2,68,69] can mediate cognitive impairment with deficits in brain insulin/IGF signaling that promote neurodegeneration. The results support our hypothesis that in peripheral insulin resistance disease states, cognitive impairment can be mediated via a liver/peripheral-brain axis of neurodegeneration due to increased ceramide production and trafficking across the blood-brain barrier. The consequential brain insulin resistance establishes a reverberating loop of neurodegeneration whereby inhibition of signaling through insulin and IGF receptors, IRS, and Akt, perturbs energy metabolism, lipid and cholinergic homeostasis, and neuronal plasticity. The findings suggest that individuals with peripheral insulin resistance diseases who are at risk for developing cognitive impairment and neurodegeneration may be identified by examining peripheral blood and possibly cerebrospinal fluid levels of ceramides and other toxic sphingolipids, and that preventive/treatment measures could include the use of agents that reduce or block the synthesis and accumulation of such compounds.
Supported by AA-11431, AA-12908, and K24-AA16126 from the National Institutes of Health.
Authors’ disclosures available online (http://www.j-alz.com/disclosures/view.php?id=443).