Measuring brain lipids

Abstract

The rapid development of analytical technology has made lipidomics an exciting new area and this review will focus more on modern approaches to lipidomics than on earlier technology. Although not fully comprehensive for all possible brain lipids, the intent is to at least provide a reference for the analysis of classes of lipids found in brain and nervous tissue. We will discuss problems posed by the brain because of its structural and functional heterogeneity, the development changes it undergoes (myelination, aging, pathology etc.) and its cellular heterogeneity (neurons, glia etc.). Section 2 will discuss the various ways in which brain tissue can be extracted to yield lipids for analysis and section 3 will cover a wide range of techniques used to analyze brain lipids such as chromatography and mass-spectrometry. In Section 4 we will discuss ways of analyzing some of the specific biologically active brain lipids found in very small amounts except in pathological conditions and section 5 looks to the future of experimental lipidomic modification in the brain. This article is part of a Special Issue entitled Brain Lipids.

Keywords: Brain lipids; Mass-spectrometry; Phospholipids; Sphingolipids.

Fig.1

A crude way of measuring lipids by magnetic resonance images of human brain . These show even lipid distribution in controls (A) and the abnormal lipid composition and shrinkage of the brain associated with neurodegeneration (B,C).

Fig 2. The phospholipid content of control human brain with age

Phosphatidylcholines (PC) are 50% at birth, declining to 25% at age 8 whereas sphingomyelin (SM) (Ceramidephosphorylcholine) increases from 2% to 16% over the same time period. In contrast, phosphatidylserine (PS) and phosphatidylethanolamines (PE) show modest increases over the same time-period. Extracted brain lipids were resolved by 2D HPTLC and qualified by lipid phosphorus analysis according to Rouser [1]. Such analyses do not take into account any regional differences in lipids, fatty acid chain length heterogeneity degree of unsaturation, plasmalogen content and other metabolic modifications.

Fig 3

Panel A: HPLC analysis of a brain with a genetic deficiency of N-acetyl-b-D-galactosaminidase A (commonly called Tay-sachs disease. Lane 1 (NORM) is control brain gangliosides from the same amount of brain tissue as Lane 2. Lane 2 (AFF) shows GM2 ganglioside to be >95% of total gangliosides compared to control For a detailed discussion of ganglioside structures see [2,3]. For a detailed discussion of how GM2 is quantified see [6-8]. Panel B: Luxol stained gray matter from a brain with a lipid storage disease showing morphological evidence of massive lipid accumulation in neurons, which become grossly swollen and dysfunctional around 6 months of age. . Panel C: shows unstained gray matter from a brain with a genetic deficiency of tripeptidylpeptidase 1 (CLN2), late infantile Batten disease showing massive neuronal storage of autofluorescent pigment which is brain-specific and clinically similar to Tay-Sachs disease (initial normal behavior followed by a rapid decline in function after 6 months) but the lipid storage material remains uncharacterized

Fig 4. Relative amounts of [³H]Man (5-9 residues)GlcNAc₂ oligosaccharides liberated from stored Dolichol oligosaccharides in gray matter from patients with 3 forms of Batten disease

CLN1 (infantile) patient with a complete deficiency of palmitoyl:protein thioesterase 1, and CLN3 (Juvenile) patient. CLN1 patients show all 5 major forms whereas Man7 and Man8 forms are less in CLN2 and CLN3 forms.

Fig 5

Quantification of the major sphingomyelin (SM C18:1/18:0) (A) and major galactosylceramide (C18:1/24:1) in mouse brain after Fractionation to isolate microdomains (lipid Rafts) (Fraction 4). The major form of BMP(22:6/22:6) is essentially absent from lipid microdomains.

Fig 6

Quantification of 3 forms of phosphatidylcholine (PC) showing A: C16:0/C16:0 mainly in microdomains (Fraction 4), B: PtdChol C16:0/18:1 equally in both and C: PtdChol C16:0/18:2 predominantly in non microdomains.

Fig 7. HPTLC of lower phase lipids

Standards are PC (phosphatidylcholine), lPC (lyso-PC), SM (Sphingomyelin), Cer (Ceramides), GC (Galactosyl ceramides), and S (sulfatides). Lanes 6-8 CLN2 – Batten disease, CD (Canavans disease), C (Control) Total Lipids Lanes 9-11 sphingolipids after alkaline methanolysis to remove glycerophospholipids. Note: the loss of myelin lipids (GC& S) in Canavans Leukodystrophy but not from CLN2 brain. The individual lipids are quantified by densitometry by reference to known amounts of standards (SM etc.).

Fig 8. Three Ages of Mass-Spectrometry – Lipidomics

Panel A: University of Pittsburg 1967: LKB 9000 gas-chromotograph (GLC) mass spectrometer with Ryhagge separator and manual identification of m/e peaks. Panel B: Michigan State University 1969: LKB 9000 GLC-MS with computerized (punch card) data analysis (PDP8I occupying most of the room. Panel C: University of Chicago 2005: HPLC hybrid triple quadruple/linear ion trap MS with electrospray ionization and a much better data acquisition system!

Fig 9

Inter-related pathways of sphingolipids and dihydrosphingolipids which can all be quantified by various types of mass-spectrometry. Quantification depends largely on how fast the brain tissue can be extracted following removal of the brain.

Fig 10. Shotgun MS reveals novel sphingolipids such as hydroxyglucosylceramide unique to oligodendrocytes

Panel A Lipid analyses of PE, PC, lysoPC, PS, and PG in brain from a mouse with null ceramide :galactosyl transferase showing minimal lipid changes in five phospholipids. Panel B: Shotgun MS lipid analysis of SM (no change) but GlyCer – big reduction, Sulfo-GalCer (sufatide-absent) but ceramide no change. The reason that some GlyCer is detected in CGT-/- brain is that Glucosylceramide appears identical to GalCer by MS/MS but with mainly C22/24 alpha-hydroxy fatty acids, and is exclusively found in oligodendrocytes. GalCer and GlcCer can be resolved by HPTLC either after impregnating TLC plates with sodium tetraborate or using HPTLC to resolve GlcCer from Galcer (normally in excess by 500:1) as a faster moving band.

Fig 11. Variability of sphingolipid fatty acid composition between rat neurons (N) and oligodendrocytes (O)

Panel A: Ceramide (Cer) showing increased C₁₆ in oligodendrocytes (O) and increased C 24:0 and 24:1 fatty acids in neurons ( N). Panel B: sphingomyelin (SM) showing increased C₁₈ in Neurons and increased C24:1 in Oligodendrocytes. Panel C: Galactosyl ceramides (GalC) found only in Oligodendrocytes (O) and glucosyl ceramides (GLC Cer) found only in Oligodendrocytes.

Fig. 12

Shotgun Lipidomic analysis of phosphatidylethanolamine species from normal mouse brain and mice overexpressing Akt and having 10% increased myelin. Samples were generously supplied by Dr. Wendy Macklin, University of Colorado Medical Center, Denver, CO.

Fig 13

Different metabolic fates of exogenous C6-ceramide (diagonal bars) and C6- dihydroceramide (open bars) in neonatal rat hippocampal neuron cultures as described in [26]. Quantification was done by HPLC/MS/MS.

Fig. 14

Exogenously added lipids and lipopeptides can be measured in brain tissue after delivery attached to Coated Quantum dots. The amount of lipid delivered to the brain can be quantified by image J/Fuji software (NIH). Panel A: Low resolution image of part of the Ventricle of a 6 day old chick embryo, 2 days after injection of QDs with a CL4 coating and a lipopeptide cargo (His6-G2-Pro9WG[palmitoyl]DapVKIKK) (JB577) into Day 4 embryonic spinal column. The uniform distribution of QD-CL4s into neuroblasts can be readily seen and this persists until day 11 when the QDs are concentrates in the developing choroid plexus and start to disappear from the brain. Panel B: Addition of the same lipopeptide to a neonatal rat hippocampal slice preparation (62) showing preferential labeling of Pyramidal neurons and essentially no labeling of glial cells. Red is QDs, blue is DAPI nuclear staining and green is Nissl body staining for neurons. The amount of lipid introduced into the brain tissue can be measured by Fuji software in an unbiased manner.