| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,1
,1,2,*,2
Departments of
Biochemistry,
* Biological Sciences and
Anatomy, The Yong Loo Lin School of Medicine, National University of Singapore, Singapore
2Correspondence: M.R.W., Yong Loo Lin School of Medicine, National University of Singapore, 8 Medical Dr., Block MD7, Singapore. E-mail: bchmrw{at}nus.edu.sg; W-Y.O., Yong Loo Lin School of Medicine, National University of Singapore, 4 Medical Dr., Block MD10, Singapore. E-mail: antongwy{at}nus.edu.sg
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: lipidomics • neurotoxicity • mass spectrometry • ceramide • N-acylated phosphatidylethanolamine
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Excitatory neurotransmitters are intimately linked to lipid metabolism. A major initial event after receptor stimulation is calcium influx into the nerve terminal, activation of phospholipases (A2, C, and D) resulting in degradation of membrane glycerophospholipids, and subsequent generation of second messengers such as arachidonic acid (AA), eicosanoids, platelet-activating factor, lysophosphatidic acid (LPA), diacylglycerol (DAG), and inositol 1,4,5 triphosphate, which are implicated in neuronal functions. However, it is this same signaling pathway that, when overstimulated, leads to excitotoxic brain damage (2
, 3)
. Deleterious changes in lipid homeostasis and signaling may be a key factor in the onset and progression of pathologies of the nervous system; indeed, intervention of these changes may confer neuroprotection. For instance, inhibition of phospholipase A2 and cyclooxygenase-2 after excitotoxic insults induced by kainate, a glutamate analog (4)
, was shown to prevent neural injury (5
6
7
8)
.
Elucidation of the molecular mechanisms of neurotoxicity thus is of great molecular and potential clinical interest. While many glutamate receptor antagonists have been developed, their use remains limited by narrow therapeutic windows that often are not accessible in clinical cases such as stroke patients. In addition, results obtained from the study of excitotoxicity may also lead the way, at least in part, to a better understanding of lipid metabolism during normal neuronal function as well as other neurological disorders that have been implicated with altered lipid metabolism for instance, Alzheimer’s or Niemann Pick disease (9
, 10)
.
Here we describe a mass spectrometry-based method for nontargeted profiling of lipids from tissue and cell extracts. The approach can in principle be used whenever two lipid extracts of similar composition are to be compared. Indeed, this is often the case in biological applications (11
, 12)
. We used this approach to determine the changes of hippocampal lipids during kainate induced neuronal injury. This is the first study that addresses lipid metabolism of neuronal damage in such a nontargeted fashion. Among the most affected lipids were phospholipids with polyunsaturated fatty acids as well as ceramides. In addition, for the first time in complex mixtures, we directly detect ions with masses that correspond to ethanolamine phospholipids with acylated headgroups. These lipids are precursors for endocannabinoids, potent lipid ligands of endocannabinoid receptors implicated in pain control.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
200 g were anesthetized by intraperitoneal (i.p.) injection of 1.2 mL of 7% chloral hydrate, and the cranial vault was exposed. Kainate (1 µl of a 1 mg/mL solution) was injected into the right lateral ventricle (coordinates: 1.0 mm caudal to bregma, 1.5 mm lateral to the midline, 4.5 mm from the surface of the cortex) using a microliter syringe. The needle was withdrawn 10 min later and the scalp was sutured. Experimental control rats were injected with 1 µl of saline instead of kainate (13)
. Five independent experiments were performed; for each experiment three animals were randomly assigned to each treatment group. All procedures involving animals were done in accordance with guidelines of the local Animal Care and Use Committee.
Internal standards
Synthetic lipids with fatty acyl compositions that are naturally low abundant were used as internal standards. Phosphatidic acid with C17 fatty acyl chains (diheptadecanoyl PA, 34:0-PA), C14-phosphatidylserine (dimyristoyl PS, 28:0-PS), C14-phosphatidyglycerol (dimyristoyl PG, 28:0-PG), C19-ceramide (C19-CER), and C12 sphingomyelin (C12-SPM) were obtained from Avanti Polar Lipids (Alabaster, AL, USA). Phosphatidylinositol with C8 fatty acyl chains (dioctanoyl PI, 16:0-PI) was obtained from Echelon Biosciences, Inc. (Salt Lake City, UT, USA). The internal standards were solubilized in chloroform at a stock concentration of 10 µg/µl.
Lipid extraction
Rats were sacrificed 1 and 3 days postinjection. The rats were anesthetized with an i.p. injection of 7% chloral hydrate, followed by decapitation. Lesioned right hippocampi were rapidly removed and homogenized in 5 volumes of PBS using a small tissue blender. Lipids were then extracted with two different methods. First, we used a widely used modified protocol of Bligh and Dyer (14)
. Briefly, 400 µl of chloroform-methanol, 1:2 (v/v) was added to 20 mg of tissue homogenate and 5 µg of internal standards; 34:0-PA and 16:0-PI were added. After 10 min incubation on ice, 300 µl of chloroform and 200 µl of 1M hydrochloric acid (HCl) were added to the mixture and the lipids were isolated from the organic phase after centrifugation. The sample was vacuum dried (Thermo Savant SpeedVac), resuspended in 2 mL of chloroform-methanol, 1:1 (v/v), and used for analysis.
Sphingolipids were extracted as described by the method of Sullards et al. (15)
. Briefly, 20 mg of tissue homogenate was sonicated in 500 µl of methanol and 250 µl of chloroform spiked with 2 µg of C19-CER and 2 µg of C12-SPM. The mixture was sonicated at room temperature and incubated overnight at 48°C. After cooling to room temperature, 75 µl of 1M methanolic KOH was added and the mixture was sonicated for 30 s, followed by incubation at 37°C for 2 h. The solution was neutralized with acetic acid, followed by the addition of 1 mL chloroform and 2 mL water. The extract was centrifuged at 4000 g for 10 min and the organic phase was recovered. The sample was vacuum dried and used for sphingolipid analysis.
Mass spectrometry
Electrospray ionization mass spectrometry (ESI-MS) was performed on a Waters Micromass Q-Tof micro (Waters Corp., Milford, MA, USA) mass spectrometer. Typically, 2 µl of sample was injected for analysis. The capillary voltage and sample cone voltage were maintained at 3.0 kV and 50 V, respectively. The source temperature was 80°C and the nano-flow gas pressure was 0.7 bar. The mass spectrum was acquired from m/z 400 to 1200 in the negative ion mode with an acquisition time of 10 min; the scan duration was 1 s. The inlet system consisted of a Waters CapLC autosampler, and a Waters CapLC pump and chloroform-methanol 1:1 (v/v) at a flow rate of 3 µl/min was used as the mobile phase. Individual molecular species were identified using tandem mass spectrometry, and the collision energy used ranged from 25 to 80 eV.
Quantification of individual molecular species was performed using multiple reaction monitoring (MRM) with an Applied Biosystems 4000 Q-Trap mass spectrometer (Applied Biosystems, Foster City, CA, USA). Samples were directly infused using a Harvard syringe pump at a flow rate of 5 µl/min. In these experiments, the first quadrupole, Q1, was set to pass the precursor ion of interest to the collision cell, Q2, where it underwent collision-induced dissociation. The third quadruple, Q3, was set to pass the structure specific product ion characteristic of the precursor lipid of interest. Each individual ion dissociation pathway was optimized with regard to collision energy to minimize variations in relative ion abundance due to differences in rates of dissociation (16)
. Lipid concentrations were calculated relative to the relevant internal standards.
Data processing
Single-stage Q-Tof spectra were acquired using MassLynx 4.0 (Waters Corp., Milford, MA, USA). Plain text files of these scans were then transferred into Matlab (The MathWorks Inc., Natick, MA, USA) for processing (refer to supplementary Fig. S1, which displays the algorithm in Matlab format). For phospholipid profiling, the intensity value of each m/z was normalized to the tissue mass based on the 34:0-PA internal standard (normalized intensity). Correlation optimized warping (COW) was used as a preprocessing method in order to obtain precise alignment of the normalized MS data (17)
. For averaging of spectra from replicate independent samples, normalized data of each replicate were warped against a reference set. After alignment of the spectra, the intensity values of individual m/z were averaged to obtain one mean spectra for each condition (i.e., kainate and control treatment, cf. also
Fig. 2
). The averaged spectra for the two conditions can then be aligned and relative differences in the lipid compositions of two conditions can be computed by simple arithmetic division and representation as the ratios on a logarithmic scale (log10 ratios).
|
|
Immunoperoxidase labeling
Four kainate-injected rats were killed 3 days postinjection. The rats were deeply anesthetized by i.p. injection of 1.5 µl of 7% chloral hydrate and perfused through the left cardiac ventricle with a solution of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed and a block consisting of the posterior two-thirds of the forebrain, including the hippocampi, was dissected out. The blocks were sectioned coronally at 100 µm using a vibrating microtome. The sections were divided into sets for cresyl fast violet (Nissl) staining and immunohistochemical staining using a mouse monoclonal antibody (mAb) to ceramide (Sigma, St. Louis, MO, USA.). Sections intended for immunohistochemical analyses were washed for 3 h in PBS to remove any traces of fixative and immersed for 1 h in a blocking solution of 5% normal horse serum in PBS. They were then incubated overnight with a mouse mAb to ceramide (diluted 1:100 in PBS). The sections were washed three times with PBS and incubated for 1 h at room temperature in a 1:200 dilution of biotinylated horse antimouse IgG (Vector, Burlingame, CA, USA). Sections were then reacted for 1 h at room temperature with an avidin-biotinylated horseradish peroxidase complex, and the reaction were visualized by treatment for 5 min in 0.05% 3, 3-diaminobenzidine tetrahydrochloride (3,3'-diaminobenzidine) solution in Tris buffer containing 0.05% hydrogen peroxide. The color reaction was stopped with several washes of Tris buffer, followed by PBS. Sections were mounted on glass slides and lightly counterstained with methyl green. Control sections were incubated with preimmune mouse serum or PBS instead of primary antibody (Ab), and showed an absence of immunostaining.
Statistical analysis
Comparison of the means of control and treatment group was performed. Data represent the mean of at least 3 independent samples ± SE from individual animals. The difference between control and treatment groups was determined statistically using Student’s t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Figure 1
A shows a typical mass spectrum of a total lipid extract derived from rat brain tissue. In this case, the lipid mixture was infused into a quadrupole-Time of flight (Q-Tof) mass spectrometer by electrospray ionization in negative mode. Ions of many prominent phospholipids can readily be detected and tentatively assigned based on theoretical calculation of lipid masses. Some of the most prominent ions detected include phosphatidylethanolamines, PE (18:1 PE, 20:4 PE, 22:6 PE and 38.4 PE), phosphatidylserines, PS (18:0 PS, 36:1 PS, 40:6 PS), and phosphatidylinositol PI (38:4 PI) at m/z 478, 500, 524 (PE), 766 (PE), 524, 788, 834 (PS), and 885 (PI), respectively. Figure 1B
shows replicate analysis of the same extract that has been spiked with a small amount (0.2 µg) of synthetic phosphatidylglycerol (28:0-PG). Hence, the overall appearance of the mass spectrum is very comparable, yet the exogenously added PG leads to a small increase in the ion intensity at m/z 665 (Fig. 1B
, inset), the expected mass of 28:0-PG. Note that Tof analysis yields sharp signals with readily resolved isotopic peaks (e.g., Fig. 1A, B
, inset).
To identify differences in such high-resolution mass spectra in a nontargeted fashion, we developed a chemometric approach that allows comparison of differences in ion intensities of all detectable ions between two experimental conditions (Fig. 2
A). Slight differences in m/z values of ions between experiments are common due to small drifts in experimental conditions such as variations during Tof measurements. Note, for example, the slight differences in m/z values of some of the ions in Fig. 1A, B
(e.g., 478.30 vs. 478.28, 524.28 vs. 524.27). To correct for this, we used a previously developed chemometric method based on correlation optimized warping (COW) (17)
. COW builds on piecewise stretching and compression of spectra along the m/z axis to correct for drifts (Fig. 2B
). Such warping does not affect the peak intensities (17)
.
Using such chemometric warping, it is possible to almost perfectly align spectra from multiple replicates of experimental condition A (A1... An) and condition B (B1... Bn). Once spectra are aligned they can be further processed arithmetically. A differential profile, DP, which is the ratio of averaged replicate spectra from condition A <A> and averaged replicate spectra from condition B, <B>, can be computed as a function of m/z. In the case of the spiking experiment mentioned above, this yields a differential profile that displays a single sharp peak at m/z 665, the expected mass of 28:0-PG (Fig. 1C
).
The ion response of such exogenously added lipids is linear as a function of the amount of spiked material for a variety of different chemistries (e.g., PG, PS, and ceramides; Fig. 1C
inset and data not shown). Thus, for samples of comparable similarity, which typically are at the source in many biological applications, this approach allows for detection of differences in lipid compositions. While this analysis is not as quantitative as, for example, multiple reaction monitoring (see below) it certainly has its advantages. Most important, the above method allows for nontargeted screening of multiple lipid species in complex lipid extracts.
We next used this method to analyze differences in lipid extracts derived from hippocampi of kainate-treated and control animals. Rats were injected with 1 µg of kainate and hippocampi were removed 1 and 3 days postinjection. Typically, the right hippocampus was injected with kainate, which leads to pronounced neuronal damage characterized by apoptotic and necrotic degeneration (13
, 18)
. The left part of the hippocampus is generally much less affected judged by Nissl staining (data not shown). Representative lipid profiles of control and kainate-treated rat hippocampus are shown in Fig. 3
A, B (3 days treatment). The differential profile was generated as described above (Fig. 3C
). Downward/upward reflections of ions indicate that levels of these lipids drop/rise during exposure to kainate. While the extent of animal response is variable, the overall differences in lipid profiles are quite consistent. Among the most pronounced differences are 1) an increase in levels of ions with m/z 536, 564, and 592, 2) a drop in signal of ions with m/z of 742, 766, 834, and 883, and 3) increased signals in m/z 984, 1004, 1012, and 1028 (Fig. 3C
). Some of these m/z correspond to readily visible peaks in the single mass spectrum (e.g., 564, 834) whereas others, such as 1004, 1028, 1056, are minor signals (Fig. 3A, B
, inset). To investigate the time dependence of the response we analyzed tissue extracts that were harvested after 1 day of kainate exposure. The lipid profiles at days 1 and 3 postinjection were qualitatively similar, although the extent of the changes at day 1 was less pronounced compared to day 3 (data not shown). Thus, within the limit of detection, these results suggest that excitotoxicity induced by kainate follows a generally progressive pattern.
|
Product ion analysis using tandem mass spectrometry and collision-induced dissociation (MSMS) is a powerful method to characterize fragment ions that allows for a more rigorous analysis of a parent ion of interest. Ions that showed at least a 1.5-fold difference in the differential profile between the five independent experiments were selected for MSMS in order to identify the underlying lipid molecular species. The differential profile in Fig. 3C
shows changes in a region 700 <m/z < 900 where most phospholipids ionize. Indeed, product ion analysis leads to the identification of prominent phospholipids, primarily PE, PS, and PI, with mostly polyunsaturated fatty acyl chains (Fig. 3C
and Fig. 4
). In addition, although the major PI, 38:4 PI (m/z 885) seemed to be less affected, other, less abundant species of PI were found to decrease with kainate treatment. Figure 4
shows a summary of the results compiled from 5 independent experiments with a total of 15 animals. Only results that have been consistently observed in all animals were included.
|
At m/z values > 1000, the most striking differences between kainate-treated and control hippocampi are increased signals of lipids with m/z 984, 1004, 1012, 1028, and 1056. Based on theoretical predictions, these ions cannot easily be assigned based on common membrane phospholipid structures. The product ion spectra produce fragments commonly found in phospholipids, including fatty acyls (m/z 255, 283, and 327) and a dehydrated phosphoglycerol moiety (m/z 153). In addition, fragments at m/z 700 and 718 were observed for MSMS of m/z 1028 (Fig. 3D
). We isolated the m/z 718 ion using the linear trap function available on our mass spectrometer and analyzed the fragments (MS3). In these spectra m/z 255, 283 were present in addition to a signal at m/z 196 (data not shown), consistent with an ethanolamine lipid parent compound containing palmitic (16:0) and stearic (18:0) acid. This indicated that the m/z 1028 precursor of m/z 718 indeed could be a PE that carries an additional fatty acyl moiety.
We thus hypothesized that these series of ions (i.e., m/z 1028, 1056, etc.) could be N-acetylated PEs (NAPEs). Indeed, our product ion analysis is consistent with published product ion spectra of synthetic NAPE (19)
that show, at least in part, a similar MSMS pattern (Fig. 3C
). Compared to synthetic and pure standards there are, however, additional ions present in these MSMS spectra (e.g., m/z 153), which indicates that the ions in our mixtures might be composites of isobaric compounds. Alternatively, or in addition, variations in experimental conditions (e.g., different mass spectrometers, collision energies, etc.) may account for some of these differences.
Product ion analysis of m/z 564 yields pronounced ions at m/z 237, 263, 282, 308, 324. Fragment ions were also observed at m/z 534 and 516 (Fig. 5
A). Based on theoretical calculations, m/z 564 can be attributed to a major ceramide species, d18:1/18:0 ceramide. Indeed, product ion analysis of synthetic d18:1/18:0 ceramide standard leads to a tandem mass spectrum that is almost identical to the one obtained from endogenous m/z 564 (Fig. 5B
). An ion at m/z 600 that was also picked up in our assay corresponds to the chloride adduct of d18:1/18:0 ceramide. Other ceramide species d18:1/16:0, d18:1/18:1, and d18:1/20:0 ceramides were found to be elevated as well (Figs. 4
and 5)
.
|
While aberrant phospholipid metabolism has been reported extensively in this model, the role of sphingolipids is much less understood. Ceramides and their metabolites are important signaling molecules that regulate multiple cellular functions, including cell growth, migration, and death (20)
. The major brain ceramide species could be readily detected using the above extraction method but other sphingolipids are suppressed, possibly by the presence of phospholipids, which make up the bulk of biomembranes. We thus subjected the lipid extracts to alkaline hydrolysis (which hydrolyzes ester linkages, common to most phospholipids, much more effectively than amide bonds found in sphingolipids) in order to examine the effect of kainate on the sphingolipids in rat hippocampus (Fig. 5C, D
). In general, ceramides ionize in a region between 500 <m/z < 650, sphingomyelins between 600 <m/z < 900 and glycosylated ceramides between 650 <m/z < 1000. Again, there was a demonstrable increase in ceramide levels based on visual inspection of the mass spectra (Fig. 5C, D
). Moreover, there were pronounced differences in levels of sphingomyelin and hexosyl-ceramides, both of which are metabolically closely linked to ceramides (see also Fig. 6
B). Levels of the most prominent sphingomyelins are reduced while those of hexosyl-ceramides are elevated during kainate treatment (Fig. 5C, D
).
|
We next used multiple reaction monitoring to fully quantify several of the sphingolipid molecular species in kainate-treated and control hippocampal lipid extracts. We observed a progressive increase in ceramide levels with time of kainate exposure, and there was no apparent selectivity for any ceramide molecular species (Fig. 5E
). We also found a moderate increase in hexosylceramides (Fig. 5F
) and decreases in abundant sphingomyelin molecular species (Supplementary Fig. S2), respectively. These results are consistent with the findings of the chemometric approach using Q-Tof profiling.
Since ceramide by itself may have a functional role during kainate-induced injury, we attempted to investigate its spatial localization using immunohistochemistry. Thin sections of rat brain were stained with primary Ab against ceramides (21)
(Fig. 6)
. The saline-injected control hippocampus showed very light or no immunolabeling for ceramide (Fig. 6A
). In contrast, kainate-injected animals showed loss of neurons in the cornu ammonis (CA) fields 1 and 3 of the hippocampus, as seen in Nissl sections (data not shown) and increased immunoreactivity to ceramide. Notably, increased labeling was observed primarily in degenerating neurons as opposed to other cell types in the hippocampus (Fig. 6B
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 23)
. It is indeed becoming increasingly evident that many disorders of the nervous system involve deregulation of lipid metabolism.
Altered phospholipid (24)
and sphingolipid metabolism (25
, 26)
, increased lipid peroxidation (13
, 27)
, and cholesterol synthesis (13)
have been described in models of ischemia related to the kainate model used here. Collectively, these results clearly indicate that multiple pathways involving lipid metabolism are affected by excessive stimulation of glutamate receptors. Using lipidomic profiling, it is becoming possible to determine the levels of many lipids, including membrane constituents (28
, 29)
as well as soluble lipid mediators (30
, 31)
, with great sensitivity and selectivity. However, no platform for simultaneous determination in a single experiment is currently available. Instead, these approaches use highly selective and optimized conditions for the different classes of lipids.
Here we present a unique approach for nontargeted profiling of lipids in complex mixtures. The method does not require any prior knowledge of the identity or chemical (and fragmentation) property of the lipids in the mixture. It is also particularly sensitive to changes in intensities that arise from ions with low intensity, which is often the case for signaling compounds. Conceptually, the approach follows comparative analysis of paired samples, which is reminiscent of other system scale methodologies. The results are plotted in "up-and-down" format, and for some lipid chemistries the signal intensities are linearly dependent on ion concentration (Fig. 1)
.
An illustrative example of the powerful nature of this approach is the detection of changes in ions that can be attributed to NAPE. Ions at m/z 984, 1012, and 1028 are very low abundant and fall below detection by visual inspection of mass spectra (Fig. 3A, B
, inset). In the differential profile they do show up as one of the most up-regulated signals however (Fig. 3C
). Their masses are consistent with that of N-acylated phosphatidylethanolamines with a (total fatty acyl carbon number:double bonds) of 52:0, 54:0 and 56:6. Indeed, tandem mass spectrometry reveals ions that correspond to PE and PE fragments, including corresponding fatty acyl moieties (Fig. 3D
).
These lipids and their breakdown product, the N-acylated ethanolamines (NAEs), are found at very low levels under normal physiological conditions but increase, for example, upon NMDA stimulation (32)
and during neuronal injury, including stroke (33)
, indicating potential neuroprotective roles. N-Arachidonoylethanolamine (20:4 NAE), anandamide, is an endogenous ligand of cannabinoid receptors (34)
. The mechanism by which KA receptor activation leads to increased NAPEs levels is not known but is likely to be related to changes in calcium levels and activation of enzymes that lead to the generation (and subsequent breakdown) of NAPE (33)
. It is tempting to speculate that accumulation of these lipids reflect neuroprotective mechanisms (NAPE) as well as pain control (anandamides, i.e., further metabolism of NAPE). In fact, many aspects of the biosynthesis and metabolism of endocannabinoids remain to be discovered. The results presented here may contribute to such efforts, for example, in attempts of therapeutic targeting of enzymes including the N-acylphosphatidylethanolamine-hydrolyzing phospholipase D (35
,36)
.
Production of compounds that act in the regulation of inflammation also occurs via breakdown of phospholipids. Generation of PUFAs is a source of inflammatory products such as the eicosanoids. The role of excessive phospholipid metabolism as a result of overactivation of cPLA2 has been discussed and reviewed extensively (37
,38)
. Effects of such PLA activation could include altered structural properties of synaptic membranes and regulation of enzymes and gene expression. This is the first time that the precise molecular species of phospholipids are measured directly in this model for neuronal damage. Indeed, upon kainate action, not all phospholipid species respond equally to phospholipase activation. Instead, molecular species with polyunsaturated fatty acids (PUFA), arachidonoyl (20:4), docosapentaenoic (22:5), and docosahexaenoic (22:6) phospholipids are affected the most (Figs. 3
, 4)
.
Phosphoinositides (phosphorylated derivatives of phosphatidylinositol) are precursors for second messengers (IP3 and DAG) and act as signaling molecules themselves via recruitment of cytosolic factors to cellular membranes. We used a described mass spectrometry method (39)
to measure the levels of these compounds during kainate treatment. PIP and PIP2 are rapidly eliminated after exposure to kainate (Supplementary Fig. S3). Since maintenance of phosphoinositides requires continuous phosphorylation, the loss of these lipids might reflect the apoptotic and necrotic environment of the hippocampus under kainate treatment.
We consider the Q-Tof-based metabolite profiling as performed here an excellent initial screening tool. It is particularly powerful when used in combination with alternative and complementary approaches, such as proteomics, genomics (see also below). Here we chose to use immunohistochemistry in order to obtain information on the spatial distribution of ceramide in the hippocampus. At least at early stages of neuronal damage (day 3 postinjection) accumulation of ceramide is largely restricted to neurons rather than glia cells (Fig. 6)
. The majority of ceramide staining is found in the cell bodies. Indeed, it has been proposed that ceramides can act in neuronal apoptosis via pathways that lead to mitochondrial dysfunction. The even distribution of ceramide staining is consistent with such spatially delocalized mode of action. Immunofluorescence experiments of neuronal cultures could address at higher resolution the precise site of ceramide accumulation.
The elucidation of mechanisms/enzymes underlying these changes may provide greater insight into the cascade of events involved in neurodegeneration. Thus, characterization of enzyme expression and activity (e.g., of enzymes that break down sphingolipids, such as sphingomyelinase, or enzymes involved in de novo synthesis of ceramides, such as serine-palmitoyl-transferase) could be the next logical steps but are beyond the scope of this study. While our metabolite profiling data support the hypothesis that some of the accumulated ceramide stems from a breakdown of sphingomyelin, we cannot rule out other contributions such as increased elevated rates of de novo synthesis. Increased signals in glycosylated forms of ceramides (Fig. 5)
suggest a general shift in sphingolipid levels toward metabolic pathways that could compensate for elevated ceramide levels. This is supported by pharmacological experiments that show that the hexosyl-ceramide 
ceramide equilibrium is not a major contributing factor to ischemic neuronal cell death (40)
. Since most molecular species of ceramides are elevated (Fig. 5)
, it is unlikely that a more subtle mode of regulation based on species with distinct fatty acyl and sphingoid base compositions is involved as opposed to findings in ceramide signaling in human neck cancer (41)
.
Note that we have not attempted to measure levels of anandamides, which might have to be extracted with different methods, as such small molecule metabolites with high solubility in aqueous solutions have a tendency to escape during organic extraction. Nor did we attempt to measure other branches of the pathway such as phosphorylation or breakdown of ceramide to ceramides-phosphate and sphingosine/sphingosine-phosphate in our organic extracts. Instead, the method was originally developed as an overview screening protocol for membrane (phospho)lipids, the majority of which ionize well in ESI and negative mode. Surprisingly, we find it also works well for lipids with different chemical structures such as the ceramides and even nonpolar components such as acylated phospholipids. Hence, it will be interesting to extend such metabolite analysis to soluble lipid mediators (e.g., PAF, NAEs, etc.).
This method should be applicable to other modes of ionization (e.g., APCI, MALDI, photoionization) and mass spectrometric detection. Indeed, future applications in combination with ultra-high resolution mass analysis, e.g., FT-ICR mass spectrometry (42)
will lead to enhanced mass resolution. This together with the capabilities of such mass spectrometers to allow unambiguous identification of ions based on very accurate mass measurements make this method a most promising new tool for lipidomics research.
Collectively, the current work 1) describes a method for nontargeted profiling of lipids from tissue and cell extracts that is particularly useful for detection of "large changes on small signals" (signaling lipids), and 2) provides qualitative and quantitative direct profiling of a large and diverse set of lipids during kainate-induced neurotoxicity. In addition to phospholipids, and sphingolipids (ceramides, sphingomyelin and glycosylated ceramides) we find acylated phosphatidylethanolamines to be altered. 3) Accumulation of ceramides is restricted primarily to neurons at these early stages of neuronal degeneration.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 27, 2005. Accepted for publication January 27, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Fei, G. Shui, B. Gaeta, X. Du, L. Kuerschner, P. Li, A. J. Brown, M. R. Wenk, R. G. Parton, and H. Yang Fld1p, a functional homologue of human seipin, regulates the size of lipid droplets in yeast J. Cell Biol., February 6, 2008; 180(3): 473 - 482. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Astarita, F. Ahmed, and D. Piomelli Identification of biosynthetic precursors for the endocannabinoid anandamide in the rat brain J. Lipid Res., January 1, 2008; 49(1): 48 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Shui, A. K. Bendt, K. Pethe, T. Dick, and M. R. Wenk Sensitive profiling of chemically diverse bioactive lipids J. Lipid Res., September 1, 2007; 48(9): 1976 - 1984. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Aguado, E. Romero, K. Monory, J. Palazuelos, M. Sendtner, G. Marsicano, B. Lutz, M. Guzman, and I. Galve-Roperh The CB1 Cannabinoid Receptor Mediates Excitotoxicity-induced Neural Progenitor Proliferation and Neurogenesis J. Biol. Chem., August 17, 2007; 282(33): 23892 - 23898. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. Watson Thematic review series: Systems Biology Approaches to Metabolic and Cardiovascular Disorders. Lipidomics: a global approach to lipid analysis in biological systems J. Lipid Res., October 1, 2006; 47(10): 2101 - 2111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
