Cultured hypothalamic neurons are resistant to inflammation and insulin resistance induced by saturated fatty acids

Abstract

Hypothalamic inflammation induced by high-fat feeding causes insulin and leptin resistance and contributes to the pathogenesis of obesity. Since in vitro exposure to saturated fatty acids causes inflammation and insulin resistance in many cultured cell types, we determined how cultured hypothalamic neurons respond to this stimulus. Two murine hypothalamic neuronal cell cultures, N43/5 and GT1–7, were exposed to escalating concentrations of saturated fatty acids for up to 24 h. Harvested cells were evaluated for activation of inflammation by gene expression and protein content. Insulin-treated cells were evaluated for induction of markers of insulin receptor signaling (p-IRS, p-Akt). In both hypothalamic cell lines, inflammation was induced by prototypical inflammatory mediators LPS and TNFα, as judged by induction of IκBα (3- to 5-fold) and IL-6 (3- to 7-fold) mRNA and p-IκBα protein, and TNFα pretreatment reduced insulin-mediated p-Akt activation by 30% (P < 0.05). By comparison, neither mixed saturated fatty acid (100, 250, or 500 μM for ≤6 h) nor palmitate exposure alone (200 μM for ≤24 h) caused inflammatory activation or insulin resistance in cultured hypothalamic neurons, whereas they did in control muscle and endothelial cell lines. Despite the lack of evidence of inflammatory signaling, saturated fatty acid exposure in cultured hypothalamic neurons causes endoplasmic reticulum stress, induces mitogen-activated protein kinase, and causes apoptotic cell death with prolonged exposure. We conclude that saturated fatty acid exposure does not induce inflammatory signaling or insulin resistance in cultured hypothalamic neurons. Therefore, hypothalamic neuronal inflammation in the setting of DIO may involve an indirect mechanism mediated by saturated fatty acids on nonneuronal cells.

diet-induced obesity (DIO) is caused by high-fat (HF) feeding in rodent models and is associated with low-grade systemic inflammatory responses. In both humans (20) and animal models (13), DIO is associated with increased circulating inflammatory markers such as tumor necrosis factor-α (TNFα) and interleukin-6 (IL-6). In addition, white adipose tissue (WAT) becomes infiltrated by macrophages as DIO progresses, leading to increased expression and secretion of inflammatory mediators in WAT (26). Circulating cytokines (11) as well as nutrient excess from exposure to fatty acids (13, 14) also activate intracellular inflammation in a cell-autonomous manner. In liver, muscle, adipocytes, and endothelial cells, for example, inflammation can arise from intracellular processes ranging from mitochondrial dysfunction and reactive oxygen species formation to endoplasmic reticulum (ER) stress and the associated unfolded protein response. Among the key consequences of this inflammation is cellular insulin resistance (11).

Recent studies have demonstrated that, just as in peripheral tissues, DIO induces inflammation in the hypothalamus of rats and mice (7, 18, 21, 22, 27) and that hypothalamic cytokine gene expression, IL-1β, TNFα, and IL-6, is increased by either DIO or intracerebroventricular (icv) administration of saturated fatty acids (7, 18). Furthermore, experimental interventions that block hypothalamic inflammation [e.g., inhibition of hypothalamic IκB kinase-β (IKKβ), a key intracellular enzyme involved in cellular inflammatory responses, by either a pharmacological or gene therapy approach] reduce food intake and lower body weight in animals exposed to HF feeding but not in controls fed a low-fat diet (18, 22, 27). In these studies, a key mechanism whereby hypothalamic inflammation appears to favor weight gain is by causing hypothalamic resistance to leptin (8, 21, 27) and insulin (7, 22, 27). These observations collectively suggest that, during high HF feeding, inflammation in key hypothalamic neurons (27) is required for excess weight gain, and they implicate exposure to saturated fatty acids as a potential mediator of these effects, as occurs in other cell types (13, 14). This suggestion is supported by a very recent study demonstrating that prolonged high-dose palmitate exposure causes ER stress and insulin resistance in a cultured hypothalamic neuronal cell line (17). In the current work, we sought to determine the effects of saturated fatty acid exposure on inflammatory signaling in cultured hypothalamic neurons and whether inflammatory signals contribute to fatty acid-induced changes of insulin sensitivity in these cells.

Specific neuronal cell lines were selected for this study on the basis of evidence that GT1–7 produce Agouti-related peptide (AgRP) (15), whereas N43/5 express proopiomelanocortin (POMC) (2), and that studies using these cell lines closely mimic various aspects of hypothalamic neuronal physiology (1, 19, 24), including the capacity for “glucose sensing” (N43/5) (2). Our findings demonstrate that, in these cell lines, saturated fatty acid exposure fails to induce inflammatory responses, unlike what is observed in cultured muscle (C₂C₁₂) (3) and human microvascular endothelial cells (HMECs) (14). Fatty acid exposure did induce ER stress, consistent with a recent report (17), but did not cause insulin resistance in these hypothalamic neuronal cells except with prolonged high-dose exposure, when significant cellular toxicity was also present. These findings suggest that neuronal inflammation, a process thought to contribute to the pathogenesis of DIO, is not induced through a direct action of saturated fatty acids on cultured neurons and may instead involve other cell types. In addition, fatty acid-induced ER stress does not reliably cause insulin resistance in cultured hypothalamic neurons.

MATERIALS AND METHODS

Cell cultures.

The mouse cell line GT1–7 was generously provided by Dr. R. I. Weiner (University of California San Francisco) and was derived originally from an immortalized fetal hypothalamic neuron. The mouse cell line N43/5 was purchased from CELLutions Biosysems (Cedarlane Laboratory, Hornby, ON, Canada) and was derived originally from a POMC-positive fetal hypothalamic neuron generated by Dr. D. Belsham. The neuronal cells were cultured in Dulbecco's modified Eagle medium (DMEM; Gibco-Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (FBS; Hyclone, Logan, UT), 2 mM glutamine, nonessential amino acid, and penicillin-streptomycin and maintained at 37°C in 5% CO₂. Mouse muscle cell line C₂C₁₂ was cultured in 10% FBS containing DMEM. To differentiate the cells, DMEM containing 2% horse serum (Gibco-Invitrogen) was used. HMECs (Invitrogen-Cascade Biological, Carlsbad, CA) were cultured in RPMI 1640 supplemented with 10% FBS and 12 μg/ml of bovine brain extract (Clonetics, Walkersville, MD), l-glutamine (2 mmol/l), sodium pyruvate (1 mmol/l), and nonessential amino acids in the presence of penicillin (100 U/ml) and maintained at 37°C in 5% CO₂. Primary neuronal cultures were generated from hypothalamic and hippocampal tissue derived from newborn rats according to a previously published protocol using density gradient (OptiPrep; Sigma-Aldrich, St. Louis, MO) separation and AraC treatment to remove glial cells. Primary cultured neurons were subjected to experimental conditions after 7 days in culture, when cells were 90% confluent.

Reagents.

LPS (E. coli; Sigma), TNFα, and interferon-γ (IFNγ) (both Calbiochem, La Jolla, CA) were added to culture medium at the indicated concentrations. Palmitic acid (C16:0), stearic acid (C18:0), arachidic acid (C20:0), and behenic acid (C22:0) were obtained from Nu-Chek (Elysian, MN). Fatty acid-free bovine serum albumin was purchased from Roche. The fatty acids were dissolved in 0.1 N NaOH at 70°C and then conjugated with 10% bovine serum albumin using a 3:1 ratio. To make the mixture of saturated fatty acids, we mixed each of the fatty acids (arachidic acid, beheneic acid, palmitic acid, and stearic acid) at equal concentrations.

RNA isolation and real-time PCR.

Total cellular RNA was extracted using Trizol reagent (Sigma), and the amount of RNA was determined using a spectrophotometer. Reverse transcription of 1.5 μg of RNA was performed using AMV reverse transcriptase (Promega, Madison, WI) with oligo(dT)_12–18 as a primer in the presence of 10× buffer, RNase inhibitor (Promega), and dNTP mix. Quantitative real-time PCR was performed with SYBR Green PCR master mix (Applied Biosystems, Foster City, CA), using an ABI 7900. The primers used for real-time PCR included GAPDH: forward 5′-C-GAACATCATCCCTGCATCCA-3′-C, reverse 5′-C-CCAGTGAGCTTCCCGTTCA-3′-C; IL-6: forward 5′-C- GTGGCTAAGGACCAAGACCA-3′-C, reverse 5′-GGTTTGCCGAGTAGACCTCA-3′-C; IkBα: forward 5′-C-TGCCTGGCCAGTGTAGCAGTCTT-3′-C, reverse 5′-C- CAAAGTCACCAAGTGCTCCACGAT-3′; CCAAT/enhancer-binding protein homologous protein (CHOP): forward 5′-CTGGAAGCCTGGTATGAGGAT-3′, reverse 5′-CAGGGTCAAGAGTAGTGAAGGT-3′-C; and Grp78: forward 5′-C-ACTTGGGGACCACCTATTCCT-3′-C, reverse 5′-C-ATCGCCAATCAGACGCTCC-3′-C. Gene expression levels were normalized to GAPDH expression. To determine X-box binding protein-1 (XBP-1) mRNA splicing, PCR was performed using cDNA as a template. The primers used for XBP-1 were forward 5′-TGAGAACCAGGAGTTAAGAACACGC-3′ and reverse 5′-TTCTGGGTAGACCTCTGGGAGTTCC-3′. The PCR product was separated with 4% agarose gel, with a PCR product of 326 bp for unspliced and 300 bp for spliced XBP-1.

Western blotting.

Neuronal cells were stimulated with 200 μM palmitic acid or fatty acid-free albumin for 18 h and then treated with 100 nM insulin (Eli Lilly, Indianapolis, IN) for 15 min before the cell lysates were prepared. The concentration of the protein was determined using the BCA method, and the same amount of proteins were used in each lane for Western blotting. SDS-polyacrylamide gel electrophoresis was performed using 4–20% gradient gels. After transfer, the membrane was incubated with p-IkBα antibody (1:1,000; Cell Signaling Technology, Beverly, MA), total IκBα antibody (1:1,000; Cell Signaling Technology), p-Akt antibody (1:1,000; Cell Signaling Technology), total Akt antibody (1:1,000; Cell Signaling Technology), p-ERK antibody (1:1,000; Cell Signaling Technology), total ERK (1:1,000; Cell Signaling Technology), p-IRS-1 (insulin receptor substrate-1)-Ser³⁰⁷ (1:1,000; Cell Signaling), IRS-1 (1:1,000; Cell Signaling), or GAPDH antibody (Santa Cruz Biotechnology, Santa Cruz, CA). The protein bands were analyzed using densitometry and Image J image analysis-normalizing phosphorylated protein to total protein bands.

Immunoprecipitation of phosphotyrosine-IRS-1.

Cultured neuronal cells were incubated with palmitate for 3 or 6 h and then treated with insulin (100 nM, 10 min). Cell lysates were immunoprecipitated with anti-IRS-1 antibody (Cell Signaling Technology, Beverly, MA) and immunoblotted with anti-phosphotyrosine antibody (Cell Signaling). Total IRS-1 determined by Western blotting was used as a control.

Caspase-3 assay.

Caspase-3 activity was evaluated using caspase-3 fluorometric assay kit from R & D Systems (Minneapolis, MN) according to the manufacturer's instructions. The cells were incubated with 500 μM palmitate containing DMEM for 16 h. The cell lysate was incubated with binding buffer and caspase-3 fluorogenic substrate. The fluorescence was determined by fluorescent microplate reader.

Statistical methods.

Comparisons between group mean values were performed by standard two-sample equal variance, with two-tailed Student's t-test always comparing treatment values with individual control values for every time point and condition. The null hypothesis of no difference between groups was rejected at P < 0.05. All values are presented as means ± SE.

RESULTS

The inflammatory potential of N43/5 and GT1–7 hypothalamic neuronal cell lines.

To verify that the hypothalamic cell lines selected for study activate the IKKβ-NF-κB pathway in response to established proinflammatory stimuli, cultured N43/5 cells were treated for 3, 6, or 24 h with either LPS (100 ng/ml) or TNFα (10 ng/ml). As expected, both LPS and TNFα increased IκBα mRNA expression (3- to 5-fold) at 3 h and to a lesser degree at 6 and 24 h (Fig. 1A) in these neuronal cells. In contrast, GT1–7 cells did not demonstrate induction of these inflammatory genes in response to LPS, although these cells express Toll-like receptor 4 (TLR4) at levels comparable with N43/5 cells [by both quantitative (q)PCR and Western blotting analysis; data not shown] and were cultured with serum containing the essential TLR4 cofactor CD14. Despite no response to LPS, GT1–7 cells did exhibit a robust increase of IκBα gene expression in response to TNFα (10 ng/ml) either alone or in combination with IFNγ (10 ng/ml) pretreatment (IFNγ + TNFα) (Fig. 1A).

Fig. 1.

Effect of acute inflammatory mediators on inflammatory signaling in cultured N43/5 and GT1–7 hypothalamic neuronal cell lines. Gene expression of IκBα (A) and IL-6 (B) was determined by quantitative (q)PCR following 3-, 6-, or 24-h exposure to either LPS (100 ng/ml), TNFα (10 ng/ml), or TNFα + IFNγ (10 ng/ml each). Expression of mRNA species is shown relative to untreated hypothalamic neuronal cells [control (cont)]. Data presented are means ± SE from 3 separate studies. C: representative Western blots and bar graphs showing p-IκBα, IκBα protein in untreated cultured hypothalamic neurons, and neurons exposed to LPS or TNFα for 1 h. Data are means ± SE (n = 2–3). *P < 0.05.

In GT1–7 and N43/5 cells treated with LPS, TNFα, or IFNγ + TNFα, we also examined “downstream” signals dependent on NF-κB signaling, including mRNA encoding IL-1β, TNFα, and IL-6. Of these, only IL-6 mRNA expression was increased consistently in both hypothalamic neuronal cell lines. As shown in Fig. 1B, IL-6 mRNA expression was increased by LPS in N43/5 cells at both 6 and 24 h, whereas IFNγ + TNFα increased IL-6 gene expression only in GT1–7 cells. TNFα exposure alone caused a more modest, transient increase in IL-6 expression in N43/5 cells and had no effect in GT1–7 cells. No change in IL-6 mRNA expression was demonstrated at the 3-h time point in either hypothalamic cell line. In addition to changes in gene expression, exposure to either LPS or TNFα increased the quantity of phosphorylated IκBα (p-IκBα) protein in N43/5 cells, whereas only TNFα was effective in GT1–7 cells (Fig. 1C). Thus, cultured hypothalamic cell lines demonstrate evidence of NF-κB activation and downstream immune signaling in response to proinflammatory mediators, although the capacity to respond to LPS is not ubiquitous.

Effect of saturated fatty acid exposure on inflammatory signaling in N43/5 and GT1–7 hypothalamic neuronal cell lines.

To determine whether hypothalamic neurons can mount an inflammatory response to saturated fatty acids, we cultured N43/5 and GT1–7 neuronal cell lines with a mixture of four different saturated fatty acids (MFA; behenate C22:0, arachidate C20:0, stearate C18:0, and palmitate C16:0) complexed with albumin at one of three concentrations: 100 μM (physiological), 250 μM (high physiological), or 500 μM (supraphysiological). Exposure to MFA did not increase IκBα gene expression in N43/5 cells at any concentration (Fig. 2A), whereas in GT1–7 cells IκBα gene expression was significantly increased only by 24-h exposure to the highest dose, 500 μM MFA (Fig. 2A). Expression of IL-6 mRNA also tended to increase in both N43/5 and GT1–7 cells, but only after 24-h MFA exposure at the two highest concentrations, and this effect achieved statistical significance only at the 500-μM MFA concentration in GT1–7 cells (Fig. 2B). Importantly, exposure to MFA at these higher concentrations (250 and 500 μM) for 24 h also resulted in significant cell death (30–50% by trypan blue exclusion), and we were unable to find a dose of MFA that induced cellular inflammation in the absence of serious cytotoxicity.

Fig. 2.

Effect of saturated fatty acid exposure on inflammatory signaling in cultured N43/5 and GT1–7 hypothalamic neuronal cell lines. Gene expression of IκBα (A) and IL-6 (B) was determined by qPCR following 3- or 24-h exposure to 1 of 3 concentrations of a saturated fatty acid mixture: 100, 250, or 500 μM. Expression of mRNA species is shown relative to untreated hypothalamic neuronal cells (cont). Data presented are means ± SE from 3 separate studies. *P < 0.05. MFA, mixture of 4 different fatty acids.

We subsequently incubated each of the two neuronal cell lines with individual fatty acids for 24 h and found that the majority of the cell death induced by MFA was caused by the two longest fatty acids (behenate and arachidate), whereas 200 μM palmitate exposure caused only minimal cell death (<5% by trypan blue exclusion). In conclusion, prolonged exposure to MFA at concentrations of ≥250 μM, despite being bound to albumin, causes toxicity in cultured neurons, and MFA-induced inflammation is observed only in this setting.

The effect of palmitate exposure in N43/5 and GT1–7 hypothalamic neuronal cell lines.

Because palmitate reliably induces inflammation in multiple cell types (3, 14), we performed additional studies to more fully delineate responses to palmitate exposure in cultured hypothalamic neurons. Neuronal cell lines were exposed to palmitate (200 μM) for either 1, 3, 6, or 24 h. Even 24-h exposure failed to increase IκBα mRNA expression in either N43/5 or GT1–7 cells (Fig. 3A). Palmitate exposure for both 6 and 24 h induced a nonsignificant increase of IL-6 mRNA expression in N43/5 but not GT1–7 cells (Fig. 3B), and cellular content of p-IκBα was not increased by palmitate exposure (3 or 6 h) in either hypothalamic neuronal cell line (Fig. 3C). By comparison, p-IκBα was increased in both hypothalamic cell lines following exposure to acute inflammatory mediators (LPS or TNFα).

Fig. 3.

Effect of palmitate exposure on inflammatory signaling in cultured N43/5 and GT1–7 hypothalamic neuronal cell lines. Gene expression of IκBα (A) and IL-6 (B) was determined by qPCR following 1-, 3-, 6-, or 24-h exposure to 200 μM palmitate. Expression of mRNA species is shown relative to untreated hypothalamic neuronal cells (cont). Data presented are means ± SE from 3 separate studies. C: representative Western blots and bar graphs demonstrating p-IκBα and IκBα protein in untreated cultured hypothalamic neurons and neurons exposed to palmitate for 3 or 6 h (GT1–7; right) or 24-h palmitate (N43/5; left). LPS (100 ng/ml) or TNFα (10 ng/ml) exposure was used as positive control for stimulation of p-IκBα protein.

Since previous studies have demonstrated that ER stress can be induced by palmitate in cultured adipocytes (9), hepatocytes (16), and neurons (17), we evaluated whether ER stress also occurs in N43/5 and GT1–7 cells. Alternative XBP-1 mRNA splicing (Fig. 4A), a marker of ER stress, began to increase after 6 h of palmitate exposure (200 μM) and increased further after 24 h, as did CHOP mRNA expression (2-fold at 6 h, 5-fold at 24 h) (Fig. 4B) and Grp78 mRNA expression (1.6 ± 0.2-fold at 6 h and 2.6 ± 0.1-fold at 24 h, both P < 0.05), all consistent with induction of ER stress in N43/5 cells. We also examined the effect of palmitate exposure on the ER stress-related MAPK ERK1/2. Exposure to palmitate (200 μM) increased phospho-ERK1/2 in N43/5 cells within 3 h (Fig. 4C). In addition, caspase-3 activity increased (6-fold) in N43/5 cells after 18-h palmitate exposure (500 μM) (data not shown), consistent with an effect of palmitate to cause apoptotic cell death. GT1–7 cells were less sensitive to palmitate-induced toxicity, requiring a higher concentration and longer fatty acid exposure to demonstrate ER stress and apoptosis.

Fig. 4.

Effect of palmitate exposure on endoplasmic reticulum stress signals in cultured hypothalamic neuronal cell lines. A: X-box binding protein-1 (XBP-1) mRNA splicing by PCR demonstrating 326- and 300-bp splice variants following exposure to 200 μM palmitate for either 0, 3, 6, or 24 h in N43/5 cells. B: CCAAT/enhancer-binding protein homologous protein (CHOP) gene expression in N43/5 cells exposed to palmitate for 0, 3, 6, or 24 h. Data are means ± SE. *P < 0.05. C: representative Western blots showing p-ERK1/2 and total ERK1/2 protein in N43/5 cells exposed to palmitate for 3 and 6 h.

To control for the effect of immortalization on the neuronal response to palmitate, we also exposed primary cultured hypothalamic neurons to 200 μM palmitate for ≤24 h. As in immortalized neurons, fatty acid exposure did not significantly increase IκBα or IL-6 mRNA expression in neurons derived from either hypothalamic tissue or cortical/hippocampal tissue (data not shown).

In summary, prolonged palmitate exposure does not detectably activate inflammatory signaling in cultured hypothalamic neurons but does induce ER stress, ERK activation, and, at high doses or with prolonged exposure, apoptosis.

Effect of TNFα to cause insulin resistance in N43/5 and GT1–7 hypothalamic neuronal cell lines.

To determine whether cellular inflammation induces insulin resistance in hypothalamic neurons, cultured neuronal cells were incubated in either the presence or absence of TNFα (10 ng/ml) for 1 h, followed by 100 μM insulin for 15 min. Protein from harvested cells was used to determine p-Akt protein concentration by Western blotting. As shown in Fig. 5, TNFα reduced insulin-stimulated p-Akt production by ∼30–40% in cultured hypothalamic neurons (N43/5 and GT1–7), similar to the ability of inflammatory activation to cause insulin resistance in peripheral cell types (3, 14). Thus, cultured neurons are capable of developing inflammation-induced insulin resistance. Importantly, the standard culture condition for hypothalamic neurons involves exposure to high glucose (20 mM). Therefore, we evaluated whether glucose concentration in the medium affects the ability to evaluate insulin signaling in these cultured cells but found no differences in insulin-stimulated p-Akt/total Akt ratios between cells cultured in high glucose (20 mM) or low glucose (5 mM) (data not shown).

Fig. 5.

Effect of TNFα on insulin signaling in cultured N43/5 and GT1–7 hypothalamic neurons. Representative Western blots and bar graphs demonstrating p-Akt and Akt protein in untreated cultured hypothalamic neurons and neurons exposed to TNFα (10 ng/ml) for 1 h followed by insulin (100 nM) for 15 min. Densitometry analysis (n = 2; means ± SE) revealed an ∼30% reduction in insulin-stimulated p-Akt in TNFα-treated cells. *P < 0.05.

Effect of saturated fatty acid exposure to cause insulin resistance in N43/5 and GT1–7 hypothalamic neuronal cell lines.

Because cellular insulin resistance can result from either inflammatory or noninflammatory mechanisms, we evaluated the ability of saturated fatty acid exposure to cause insulin resistance in hypothalamic cell lines (Fig. 6). A 3-h exposure to arachidate, behenate, palmitate, or stearate, individually (250 μM) or to mixtures of all four saturated fatty acids (100 or 250 μM), did not impair the ability of insulin (100 nM for 15 min) to stimulate p-Akt in either N43/5 or GT1–7 cells relative to cells exposed to a control solution (fatty acid-free albumin) (data not shown). Longer exposure to 200 μM palmitate (6 or 24 h) also did not reduce insulin-stimulated p-Akt production in GT1–7 cells. Although prolonged palmitate exposure (6 or 24 h) reduced insulin-stimulated p-Akt content in N43/5 cells, this decrease was not significant when normalized to total Akt level, which was also reduced by palmitate exposure (Fig. 6A). To further evaluate the effect of palmitate exposure on insulin signaling in N43/5 cells, tyrosine phosphorylation of IRS-1 was determined by immunoprecipitation. Exposure to 200 μM palmitate for 3 or 6 h did not affect insulin-stimulated phosphotyrosine-IRS-1 levels in N43/5 cells (Fig. 6B), nor was phosphoserine-IRS-1 (a marker of cellular insulin resistance) induced by exposure to palmitate in N43/5 cells (data not shown). Thus, fatty acid exposure does not cause insulin resistance in N43/5 or GT1–7 hypothalamic neuronal cell lines.

Fig. 6.

Effect of saturated fatty acid on insulin signaling in cultured N43/5 and GT1–7 hypothalamic neurons. A: representative Western blots and corresponding bar graphs demonstrating p-Akt, Akt, and GAPDH protein in untreated cultured hypothalamic neurons (N43/5 and GT1–7) and neurons exposed to 200 μM palmitate (Pal; 3, 6, or 24 h) followed by insulin (100 nM) for 15 min. Densitometry analysis (n = 3; means ± SE) for p-Akt/Akt ratio shown for N43/5 cells. B: immunoblots (IB) demonstrating insulin (Ins)-stimulated (100 nM, 10 min) phospho-Tyr-insulin receptor substrate-1 (IRS-1) protein from immunoprecipitated (IP) IRS-1 protein following 3- and 6-h exposure to Pal in N43/5 cells. Densities were normalized to total IRS-1 protein evaluated by Western blotting (bottom lane).

Effect of palmitate exposure to cause insulin resistance in C₂C₁₂ myotubes and HMECs.

Previous studies have shown that palmitate exposure causes insulin resistance in cultured muscle (3) and endothelial cells (14). As a positive control for lack of insulin resistance in hypothalamic neuronal cell lines, we evaluated the ability of palmitate to cause insulin resistance in C₂C₁₂ myotubes and HMECs. A 3-h exposure to 100 μM palmitate reduced insulin-stimulated p-Akt production in HMECs by ∼30% (P = 0.01, n = 3) relative to either untreated or albumin-treated controls (Fig. 7A). Similarly, in differentiated C₂C₁₂ myotubes, 16-h exposure to 200 μM palmitate reduced insulin-stimulated p-Akt levels by ∼25%, although this change just failed to reach statistical significance (P = 0.06, n = 2; Fig. 7B). Thus, exposure to palmitate causes insulin resistance in muscle and endothelial cell cultures, as has been shown previously, in contrast to the lack of effect of palmitate in the hypothalamic neuronal cultures we studied.

Fig. 7.

Effect of palmitate on insulin signaling in cultured peripheral cells. Representative Western blots and corresponding bar graphs showing p-Akt and Akt protein in human microvascular endothelial cells (HMECs) (n = 3; A) and C₂C₁₂ myotubes (n = 2; B) exposed to palmitate followed by insulin (100 nM) for 15 min. HMECs were incubated with 100 μM palmitate for 3 h, and C₂C₁₂ myotubes were incubated with 200 μM palmitate for 16 h. Bar graphs show means ± SE. *P < 0.05; &P = 0.06. FAFA, fatty acid-free albumin.

DISCUSSION

Saturated fatty acids cause inflammation and insulin resistance in multiple peripheral tissues, including adipocytes, pancreatic β-cells, and hepatocytes, as well as differentiated myotube and endothelial cell cultures (3, 6, 12, 14, 23). Several recent studies have implicated hypothalamic neuronal inflammation and associated neuronal leptin and insulin resistance as a key mechanism promoting weight gain in DIO (7, 18, 21, 22, 27). The observations that DIO increases the amount of saturated fatty acyl-CoA in the hypothalamus, as determined by mass spectrometry (22), and that direct administration of saturated fatty acids into the third ventricle causes hypothalamic inflammation (18) raise the possibility that saturated fatty acids can directly cause inflammation and/or insulin resistance in neurons as they do in other cell types. In the current study, we show that two hypothalamic neuronal cell lines, N43/5 and GT1–7, do not show evidence of inflammation in response to direct exposure to saturated fatty acids, although prolonged exposure to fatty acids, even when complexed to albumin, did cause ER stress, ERK activation, and apoptosis. In addition, in contrast to a recent report using a different cell line (17), saturated fatty acid exposure did not cause insulin resistance in either hypothalamic neuronal cell line employed in the current studies. These results suggest that neurons may not be direct targets for the effect of saturated fatty acids to cause hypothalamic inflammation and insulin resistance, although whether specific neuronal subtypes or neuronal circuits are sensitive to this effect cannot be excluded by this study.

Recent studies have highlighted the effect of palmitate exposure to cause inflammation and insulin resistance in cultured endothelial cells, hepatocytes, and skeletal muscle cells (3, 13, 14). Multiple pathways have been invoked, but the primary focus has been the ability of saturated fatty acids to bind TLR4 and subsequently activate NF-κB through a MyD88-dependent mechanism. Moreover, in vivo studies have confirmed the importance of this signaling pathway to the effect of HF diet to cause vascular inflammation in a mouse model of DIO (13). However, in our current study, even prolonged palmitate exposure (24 h) did not activate inflammatory signaling in cultured hypothalamic neurons despite the fact that the N43/5 cells demonstrated robust inflammatory activation by LPS, evidence of an intact TLR4/MyD88 signaling pathway. Thus prolonged palmitate exposure does not cause inflammation even in cultured hypothalamic neurons that demonstrate intact TLR4 signaling, suggesting that these neurons are less sensitive to fatty acid-induced inflammation than are endothelial, muscle, or liver cells.

The only evidence for induction of inflammatory signaling by fatty acids in our study occurred in conjunction with apoptosis and cell death from prolonged exposure to high concentration of long-chain saturated fatty acids despite the fact that the fats were conjugated to albumin to negate their known action as solvents at the cell membrane. Besides activation of NF-κB signaling, palmitate is known to cause ER stress, ERK activation, and apoptosis in adipocytes (9) and hepatocytes (16). Accordingly, in our study, cultured hypothalamic neurons demonstrated induction of ER stress, ERK activation, and caspase-3 induction in response to palmitate exposure, confirming a recent study that prolonged exposure to long-chain fatty acids can induce cellular stress and apoptosis in cultured neurons (17). This finding is important to consider when interpreting the in vivo effect of icv administration of a fatty acid mixture to cause hypothalamic inflammation (18), and future studies of icv fatty acid administration will be needed to determine whether regional hypothalamic inflammation results from direct toxic effects of the fatty acids on neurons or other hypothalamic cell types.

In this study, we demonstrate that cultured neuronal cells, like peripheral cell types, become insulin resistant when exposed to an acute inflammatory mediator (TNFα). In nonneuronal cell types, saturated fatty acids may trigger insulin resistance via many mechanisms, including TLR4 activation, mitochondrial dysfunction, reactive oxygen species formation, and ER stress and the unfolded protein response (5, 11, 13). Our study shows that palmitate exposure did not result in insulin resistance at time points and doses that were clearly associated with ER stress. Interestingly, a very recent report using a different hypothalamic cell line (mHypoE-44) also documented palmitate-induced ER stress but only found insulin resistance after prolonged palmitate exposure, which, surprisingly, was not reversed by preventing ER stress (17). In light of these observations, an explanation for why palmitate exposure did not induce insulin resistance in N43/5 and GT1–7 cells might be that the mHypoE-44 cell line appears to be derived from a “nutrient-sensing” neuron, a small population of neurons (25) that express markers found in pancreatic β-cells such as glucokinase, Kir6.2, and sulfonylurea receptors. In contrast to a previous report (2), we did not find any evidence of nutrient-sensing capability in N43/5 or GT1–7 cells (by qPCR and by functional sulfonylurea receptor assay; data not shown), and this difference in neuronal glucose sensing may account for the lack of palmitate-induced insulin resistance in our studies. Importantly, although all three cell lines (N43/5, GT1–7, and HypoE-44) are thought to be derived from hypothalamic neurons that express either POMC or neuropeptide Y/AgRP, only a small fraction of these key neuronal regulators of energy balance are capable of nutrient sensing (25). Thus, it is conceivable that a very limited number of hypothalamic neurons (4, 25) are endowed with highly specialized qualities that confer sensitivity to the effect of palmitate to cause insulin resistance (and potentially inflammation) despite the lack of any evidence generated by our efforts to support this. From this perspective, additional studies are justified to determine whether such neurons truly exist and if so to identify mechanisms that confer this sensitivity.

Although palmitate-induced inflammation is implicated in the pathogenesis of neuronal leptin resistance as well as insulin resistance, we were unable to investigate this possibility in our studies since neither the N43/5 nor GT1–7 neuronal cell lines expressed the signaling form of the leptin receptor (LepRb) or demonstrated leptin-mediated signal transduction (including phosphorylation of signal transducer and activator of transcription 3 protein and induction of suppressor of cytokine signaling 3 mRNA) (data not shown). Therefore, further work is needed to determine the effect of saturated fatty acid exposure on neuronal leptin resistance.

Neurons are terminally differentiated cells that have no or very limited capacity to divide. A line of immortalized cells, which can divide without limit, may thus respond to nutrient excess differently than do nonimmortalized cells. To address this issue, we treated the neuronal cell lines when plates were near confluence, and division was slowed by contact inhibition. In addition, we tested the effect of palmitate exposure on primary neuronal cell cultures and found no effect of palmitate exposure on neuronal inflammation, supporting our conclusion that direct fatty acid exposure does not cause inflammation in hypothalamic neurons.

We caution that this study does not exclude the possibility that, in vivo, neuronal inflammation occurs via either direct or indirect mechanisms in the setting of DIO. Previous studies suggest that nonneuronal cells may play a key role in this process. For example, glia preferentially take up fatty acids in the central nervous system and release degradation products such as ketones and acetoacetyl-CoA to be used as fuel by the neurons (10). Thus, in the setting of DIO, nonneuronal cells exposed to high levels of fatty acids may be subject to nutrient excess-induced pathology and may release paracrine mediators that secondarily cause inflammation in neurons. Alternately, hypothalamic inflammation may occur via an endocrine mechanism triggered by circulating DIO-related peripheral inflammatory signals.

In conclusion, we report that prolonged saturated fatty acid exposure does not cause inflammation in cultured hypothalamic neurons. These findings suggest that DIO-related hypothalamic neuronal inflammation is not due to direct exposure of neurons to increased hypothalamic concentration of saturated fatty acids or may be limited to specific neuronal subtypes. In addition, although ER stress was induced by palmitate, this effect did not cause insulin resistance, indicating that neither ER stress nor saturated fatty acid exposure is sufficient to reliably cause insulin resistance in cultured neurons. Future studies are warranted to determine whether DIO causes neuronal inflammation via a mechanism independent of saturated fat or indirectly via the action of saturated fat on other central nervous system cell types. In addition, investigation is needed to determine whether sensitivity to saturated fat-induced inflammation/insulin resistance is related to specific neuronal characteristics (e.g., nutrient sensing).

GRANTS

This work was supported by grants from the National Institute of Diabetes and Digestive and Kidney Diseases to B. E. Wisse (DK-61384 and DK-71784), M. W. Schwartz (DK-068384, DK-052989, and DK-083042), and F. Kim (DK-073878).

DISCLOSURES

The authors have nothing to disclose.