As part of a study of glycolipids, we have analyzed

Transcription

1 J. Lipid Research, October, 9F Voume 2, Number 4 Fatty acids of human brain cerebrosides* NORMAN S. RADIN~ and YUKIO AKAHORIt Biochemistry Department, Northwestern University Medica Schoo, Chicago 11, Iinois, and Menta Heath Research Institute, University of Michigan, Ann Arbor, Michigan [Received for pubication February 28, SUMMARY Four regions of two human brains were anayzed for the individua cerebroside acids: cerebra cortex white and gray matter, cerebeum (mainy gray matter), and corpus caosum. The tota ipids were extracted from each section and passed through a coumn of Forisi to remove choestero and phosphatides. The crude cerebrosides were next passed through a coumn of mixed ion exchange resins to remove cerebroside sufate and then pursed by eution from a siicic-acid coumn. The acids were ceaved from the cerebrosides and separated into four casses: norma saturated, hydroxy saturated, norma unsaturated, and hydroxy unsaturated. The individua acids were then anayzed by gas-iquid chromatography. The cerebroside content of each brain region differed somewhat, as did the reative contents of hydroxy acids. However, the distribution of the acids within each cass was rather independent of brain ocation. The norma saturated acids contained stearic and ignoceric as the major acids, but fairy arge amounts of the CZ2, CZ3, and Cz6 acids were aso present. The hydroxy saturated acids were simiar, but contained itte hydroxystearic acid. The unsaturated acids of both casses contained the Cz, acid as the major constituent, together with considerabe amounts of the Czs and Cza acids. As part of a study of gycoipids, we have anayzed different regions of human brain. Athough it has been known for a ong time that there is a higher cerebroside concentration in white matter than in gray matter, nothing is known of the content of the individua fatty acids in the cerebrosides of different portions of the brain. A procedure for such anaysis had been deveoped for rat brain (1) and this has been modified for use with human brain. The main modification invoved repacement of the saponification step by the use of siicic-acid chromatography. EXPERIMENTAL Sampes of frozen, unfixed human brain weighing 7 to g were extracted and the ipid extracts purified with Forisi and ion exchange resins as previousy described (1). The moderatey pure cerebrosides were then prepared for appication to a coumn of siicic acid' by making an intimate mixture with a sma por- * Supported in part by Grant B-179 from the Nationa Institute of Neuroogica Diseases and Bindness, United States Pubic Heath Servie. t Present address: Menta Heath Research Institute. University of Michigan, Ann Arbor, Mich. 1 Bio-Rad Laboratories, Richmond, Caif. Prepared according to the method of Hirsch and Ahrens (2) and used without further treatment. 335 tion of the adsorbent. This was done by dissoving the ipids in choroform-methano 2 : 1 (v/v), adding the siicic acid (1.2 g/100 mg ipids), and removing most of the sovent with a rotary vacuum evaporator. The remaining sovent was removed by drying in air overnight and then for severa hours at high vacuum (3). A surry of additiona siicic acid ( g/100 mg cerebrosides) was prepared in Skeysove B (mainy hexane). The surry was swired minutes to remove trapped air and poured in portions, with tapping, into a ong gass coumn (1.8 or 2.2 mm inside diameter). The coumn was fitted at the bottom with gass woo, sand, and a very fine neede vave. The preparation of the packing was speeded by appying about 5 psi air pressure to the top of the coumn. The ength of the packed region ranged between 70 and 95 cm. The siicic acid was conditioned in the coumn after the method of Weiss (4) by passing through choroformmethano 2: (v/v), 5.5 m/gram siicic acid, then choroform, 4 m/gram siicic acid. The sovents were pumped through by means of a precision punger-type pump, after the method of Spackman et a. (5). The means of connecting are shown in Figure 1. The ipid sampe, which had been dried in siicic acid, was then added as a surry in choroform and pumping was begun with choroform-methano 92:8 (v/v) at a fow rate of about 25 m1 per hour. Fractions of 15 m1

2 33G ItADIN AKD AKAHOIZI each were coected and evaporated to dryness. The fractions comprising the major peak were pooed and anayzed for gaactose content by the phosphoric acidanthrone method (6). The cerebrosides thus obtained (141 to 242 mg) were spit by refuxing under a very sow stream of nitrogen with 10 m of concentrated WC1 and m1 of methano. After 4 hours of heating, m1 of water was added and the mixture of fatty acids and esters was extracted with five 40-m1 portions of ether. The ether extract was washed with water, evaporated to dryness under vacuum with the aid of touene (to prevent spashing), and the residue esterified with dimethoxypropane, HCI, and methano (7). Foowing remova of the sovents under vacuum, the esters were separated into norma and hydroxy esters by chromatography with Forisi. Each fraction was then further separated into saturated and unsaturated groups with the aid of mercuric acetate (). The mercuric acetate reaction was modified by aowing the reaction mixture to sit overnight at room temperature instead of heating. The hydroxy esters were converted to methy ethers and a four fractions were anayzed by gas-iquid chromatography (1). DISCUSSION O F THE METHODS Isoation of the PuriJied Cerebrosides. Our previous work with rat brain cerebrosides had discosed the presence of an impurity characterized by a carbony group, as shown by infrared spectra. The impurity coud be removed by treatment with akai and then ion exchange resins. However, this method faied in the case of human cerebrosides and we turned to siicic acid, which had recenty become avaiabe in a much improved form. The siicic acid satisfactoriy removed the carbony impurity and gave cerebrosides of reasonaby good gaactose and fatty acid content (Tabe 1). The sphingosine vaues were high, part,icuary in the second group of anayses (see beow). The siicic-acid coumn yieded some minor fractions, as we as a sma amount of ipid traiing from the cerebroside peak. The traiing portion, which was discarded, contained ony about 1% of the cerebroside hexose. It was found that bubbes appeared in the siicic-acid coumn uness severa pounds of pressure were maintained throughout the packing. This pressure was maintained by constricting the outet with the neede vave at the bottom. In our earier work we used a piece of Neoprene tubing and a screw camp to adjust the resistance, but the sweing action of the sovent made it necessary to adjusthe camp frequenty. The neede vave aso needs one or two adjustments, but this is very easiy done. It shoud be pointed out that this type of coumn restrictor withstands pressure much better and gives fow adjustments much more easiy than the TefonQ stopcocks we have tried, and is in effect cheaper than these stopcocks because it can be removed readiy and paced on the outet of any other coumn fitted with quarter-inch gass capiary tubing. The other metering vave (Fig. 1) was incuded in the FIG. 1. Pumping system for ipid chromatography. A. Connection to compressed air, for priming the pump. B. Reservoir fask for sovent. Capiary tube, C, is offset to aow pacing magnetic stirrer under fask when gradient eution is desired. C. Capiary gass tubing, nominay 6 to 7 mm diameter, chosen to fit through /q" Tefon Swageok ferrues. D. Neoprene tubing, previousy extracted with choroform-ethano in a Soxhet extractor. E. Swageok ebow, stainess stee, #4W , with Tefon ferrues, #403-1 and (Crawford Fitting Co., Ceveand 10, Ohio). F. Chromatographic minipump, /8" punger, 29 rpm, maximum capacity 124 m/hour (Miton Roy Co., Phiadephia 18, Pa.). G. Swageok tee, #2W3TMT416. H. Swageok adapter, #201-A-2F-316. I. Sma pressure gage, /g" pipe connection. J. Short piece of I/g" outside diameter stainess stee tubing. K. Metering vave, Nupro SSZMA (Nucear Products Co., Ceveand 10, Ohio). L. Stainess stee hypodermic neede tubing, 20 gauge. One end is sodered into the /8" stee tubing; the other end is cemented into a piece of capiary gms tubing with epoxy cement. The capiary tubing is previousy seaed to an 18/7 socket. M. Neoprene sheet rubber gasket, thick, made with two cork borers. The 18/7 joint is hed together with two or three spring camps. N. Chromatographic coumn, made of standard or heavy wa tubing. 0. Gass woo and sand; base for packing. P. Metering vave, Nupro SS4M, with /,'' Tefon Swageok ferrues. Q. Gass woo fiter, to keep dirt out of pump. / I

3 CEREBROSIDES BRAIN 337 pumping system to aow raising the resistance further, thereby'increasing the accuracy of the pump (5). We adjusted this vave to keep the pump pressure at about 20 psi. There is an appreciabe voume of iquid in the pressure gauge; this tends to diffuse into the main stream of sovent rather sowy when a change is made to a sovent of higher density. For this reason we evacuated the pumping ine for 15 minutes before changing from choroform-methano to choroform. The sampe of cerebrosides was added to the siicicacid coumn as a dried mixture in a itte siicic acid because the ipid was too insoube to be disodged from the gassware by choroform. Addition as a soution in choroform-methano (4) and eution with choroform (of increasing methano content) gives two cerebroside peaks instead of one. It is evident that this technique causes part of the cerebrosides to move with the initia portion of choroform-methano, and the remainder of the cerebrosides remain adsorbed unti the methano content of the euting iquid is high enough to move TABLE 1. Cerebra cortex, white Cerebra cortex, gray Cerebeum Corpus caosum Theoretica vauest ANALYTICAL DATA FOR PURIFIED CEREBROSIDES Gaactose per cent 20.9t B 2 Fatty Acids* per cent Sphingosine per cent * As methy esters. t In paired figures, top figure = brain from 70-year-od man; bottom figure = brain from 44-year-od man. 1 Cacuated from typica fatty acid anayses, assuming cerebrosides are monohydrate; MW = 835. them as an equiibrating sampe. The finding by Weiss (4) that ignoceric acid coud be isoated from the first peak and cerebronic acid from the second suggests that there is a preferentia remova of the ess poar cerebrosides in the sovent used for addition. This expanation was confirmed2 by anayzing two such cerebroside fractions with the Forisi method. B0t.h 2 Y. Kishimoto and N. S. Radin. Unpubished work. fractions contained norma acids, the content being somewhat ower in the second fraction. spitting of the Cerebrosides. Ethano-aqueous HC1 has been used by Skipski et a. (8) for spitting off the acids of cerebrosides, and such a system is far more convenient than dry HC1 mixtures. However, the formation of ethy esters woud make gas-iquid chromatography more difficut, so we substituted methano for ethano and raised the refux period to 4 hours. Severa trias had indicated this period was sufficient, but the ow yieds of esters from our second brain (Tabe 1, coumn 2) suggest that this period is too short. Sweeey and Moscatei (9) have used 5 to 6 hours with methano-hc1. The crude esters obtained by ether extraction of the hydroyzates weigh more than expected from theory and evidenty contain some sphingosine, extracted possiby as the fatty acid sat. When the esters are separated with Forisi, the sphingosine is hed back on the coumn and can be euted with acetic acid-choroform-methano 3 :4:4 (v/v). In the case of our second brain, the sighty ow yieds of esters were accompanied by a high yied of this sphingosine fraction. It seems ikey that this fraction incuded ceramides. Athough it is possibe that incompete hydroysis may cause preferentia spitting of one cass of cerebrosides, the data for the fatty acid anayses do not indicate that this is an important effect. The vaues in Tabe 1 for sphingosine were obtained by combining the' weights of the two ong-chain base fractions obtained by (a) taking the residua water-hc1-methano soution after remova of esters, making it akaine (Tropaeoin 0 indicator), and extracting with ether; and (b) euting the Forisi coumn with acetic acid-choroform-methano and washing the choroform with water, akai, and water. The pooed sphingosine fractions were anayzed by Dr. Chares C. Sweeey, of the University of Pittsburgh, by gas-iquid chromatography (9). It shoud be noted that the isoation method described here coud be used equay we with much smaer amounts of brain, since no crystaization steps are invoved. The arge amounts were used mainy to hep in vaidating the individua steps. DISCUSSION OF THE DATA The data obtained from the steps prior to gas chromatography are shown in Tabe 2. Tota Lipids and Cerebrosides. Coumns 1 and 2, in

4 338 RADIX AND AKAHORI TABLE 2. ANALYTICAL DATA FOR LIPIDS AND FATTY ACID CLASSES* Tota Tota Norma Hydroxy Lipids Cerebrosides Acids Acids ~. " per cent of wet brain per cent of cerebrosides Cerebra cortex, white O Cerebra cortex, gray Cerebeum Corpus 27.8 caosum * See footnotes to Tabe 1 for expanatory notes. 5 Hydroxy/ Norma Saturated Norma/ Tota Norma ". saturated Hydroxy/ Tota Hydroxy agreement with the work of many others, show that white matter contains much more ipid and cerebrosides than gray matter. The difference is greater in the case of cerebrosides. Cerebeum resembes gray matter in this respect. The two brains differ in that the ipid and cerebroside concentrations are greater in the oder brain, especiay in the corpus caosum. Our figures for cerebroside concentrations are about one-third ower than those found by other workers (10, 11, 12), due mainy to the incusion of cerebroside sufate in tbe earier anayses. Hydroxy Versus Norma Acids. The vaues found for these two casses of acids are shown in coumns 3 and 4, and the ratios of the two casses in coumn 5. As in rat brain cerebrosides (13), the hydroxy acids predominate. Cerebeum again resembes gray cortex, I I[ 24 B 25 both showing somewhat greater predominance of hydroxy acids. Large individua differences are apparent in coumn 5, but we cannot state now whether these can be correated with age. Saturated Versus Unsaturated Acids. To conserve space, we present ony the ratios of the concentrations. In coumn 6 we see that t'he degree of unsaturation in the norma acids is quite simiar in both brains in a brain regions except gray cortex, where the saturated acids predominate. The degree of unsaturation is higher than that found in rat cerebrosides (13). In the case of the hydroxy acids (coumn 7), there is marked simiarity in a brain regions of any one brain. However, the younger brain consistenty shows ess unsaturation. Gas-Liquid Chromatographic Data. Figures 2 to 5 I I8 i A B FIG. 2. Distribution of each norma saturated acid in brain cerebrosides, cacuated m percentage of the tota norma saturated esters. Each bar consists of four sections, corresponding (eft to right) to cortex white, cortex gray, cerebeum, and corpus caosum. Figure 2A shows the resuts for the oder brain; Figure 2B, for the younger brain. The trace of C9 acid in Fig. 2B is in the cerebeum and the C21 acid is in the cortex gray

5 CEREBROSIDES BRAIN 339 c 9 % Lr % r "L A 24 D 25 IO 322 FIG. 3. Distribution of each hydroxy saturated acid. See Figure 2 for further description. show the resuts of gas-iquid chromatographic anaysis of each fraction. Excuded from the graphs are the onger and shorter acids that were found. Most of these occur in concentrations ess than 1%) with a few over 2%. A sight distortion in the percentages arises from the use of esters and ester ethers, which have moecuar weights differing from the acids originay present in the cerebrosides. The norma saturated acids contain stearic and ignoceric acids as their major components, but appreciabe amounts of the C22, (223, and C25 acids are aso present. As in rat cerebrosides (), the content of odd-numbered acids is unusuay high. This high content expains much of the variabiity in anaytica data reported for the CH acids isoated from brain by conventiona methods. The hydroxy saturated acids aso show a high content of CZ3, and C, acids. It has been found in rat brains that the concentrations of the odd-numbered acids increased with age (13)) and the finding that the C23 hydroxy acid is the second major acid is consistent with the rat data. In contrast to the norma saturated acids, there is ony a trace of the C18 acid, except in gray cortex. In this region there is a ower vaue for cere- 23 B bronic acid that is accompanied by reativey high vaues (1.1% to 2.3%) for the Cm, C,, Cis, and C14 hydroxy acids. This characteristic of gray cortex was found in both brains. The unsaturated acids of both casses are notabe for their high content of C25 and CB acids, especiay in comparison with the shorter acids. The C, acids, however, pay a much arger roe in the unsaturated acids, as noted in our work with rat acids. Kenk and Leupod (14) have isoated the norma unsaturated acid. Trace amounts of u.nknown acids were observed in the two casses of hydroxy acids, mainy in the C19 region. Particuary interesting is a comparison of the different brain regions with respect to the distribution pattern of the individua fatty acids in each cass. This may be done by examining the profie of each bar in the bar graphs. Taking into consideration the variabiity intrinsic in the anaytica method, one can see a marked simiarity between the various brain regions. The simiarity is even more marked when one reviews the arge differences in cerebroside concentrations and ratios of hydroxy to norma acids (coumns 2 and 5, Tabe 2). Ony one serious exception to this simiarity appears,

6 340 RADIN AND AKAHORI % v greaty with brain ocation is we known. It woud appear, then, that ony one ce type synthesizes the cerebroside acids (and presumaby the cerebrosides too). The concept of a singe ce type as the source of myein has been proposed by Korey (16). This hypothesis aso serves to expain the simiarity in degree of saturation in the norma and hydroxy acids of the four brain areas (gray cortex again being the exception). However, one must then assume that the reative production of hydroxy acids is infuenced by the ocation of the ces. Oh r d IC FIG. 4. Distribution of each norma unsaturated acid (younger brain ony). See Figure 2 for further description. in Figure 2B, where the stearic and ignoceric acid proportions differ somewhat in the gray cortex and cerebeum. There appears to be a sma but consistent difference in a the cerebroside acids of gray cortex, where there is a itte ess of the C24 acids and more of the shorter acids. The effect is most noticeabe, as mentioned before, in the hydroxy saturated acid group. Baker (15) has recenty reported finding that the shorter acids of the nervous system (ester-inked) become more and more predominant with increasing distance from the center of the brain, a finding consistent with the above. Simiary consistent is the finding3 that hydroxystearic acid is an important component of the spina cord cerebrosides (8). Athough much more work is needed to carify the significance of our data, we woud ike to offer a tentative hypothesis to expain the simiarity in acid distribution between brain sections. If two or more types of ces synthesize the cerebroside acids, one woud expect that each type woud yied a different distribution pattern, that the two types of ces woud occur in ratios differing with their ocation in brain, and that the distribution pattern woud therefore differ from region to region. That the distribution of ce types does vary 3 A. K. Hajra and N. S. Radin. Unpubished work. IO t n FIG. 5. Distribution of each hydroxy unsaturated acid (younger brain ony). See Figure 2 for further description. Anaysis of the four sphingosine sampes from one of the brains, kindy carried out by Dr. Sweeey, showed that 95% or more of the bases was sphingosine, about 2% was dihydrosphingosine, and a trace consisted of an unknown ong-chain base. Here too the four brain areas were very simiar and our postuate of a singe ce type may be extended to cover the synthesis of the ong-chain bases. The authors wish to t.hank Dr. Harod Koenig, Veterans Administration Research Hospita, Chicago, for his kind aid in obtaining the brain sections. REFERENCES 1. Kishimoto, Y., and N. S. Radin. J. Iiipzd Research 1 : 72, Hirsch, J., and E. H. ihrens,,jr. J. Bio. Chem. 233 : 311, 1958.

7 BRAIN CEREBROSIDES Radin, N. S. Ana. Chem. 28: 542, Weiss, B. J. Bw. Chem. 223: 523, Spackman, D. H., W. H. Stein and S. Moore. Ana. Chem. : 1190, G. Radin, N. S., F. B. Lavin and J. R. Brown. J. Bio. Chem. 217: 789, Radin, N. S., A. K. Hajra and Y. Akahori. J. Lipid Research 1 : 2, Skipski, V. P., S. M. Arfin and M. M. Rapport. Brch. Bwchem. Biophys. 82 : 487, Sweeey, C. C., and E. A. Moscatei. J. Lipid Research 1 : 40, Cumings, J. N., H. Goodwin, E. M. Woodward, and G. Curzon. J. Neurochem. 2: 289, Brante, G. Acta Physio. Scand. 18: Suppementum 63, Tabe IX, Johnson, A. C., A. R. McNabb and R. J. Rossiter. Biochem. J. 43 : 573, Kishimoto, Y., and N. S. Radin. J. Lipid Research 1: 79, Kenk, E., and F. Leupod. 2. physio.chem., Hoppe- Seyer s 281 : 208, Baker, R. W. R. Biochem. J. 78 : 29P, Korey, S. R. A.M.A. Arch. Neurot. 2: 140,1960.