- Email: [email protected]
Gangliosides of human cerebrospinal fluid in various neurologic diseases
192
Journal of the Neurological Sciences, 105 (1991) 192-199 © 1991 Elsevier Science Publishers B.V. All rights reserved 0022-510X/91/$03.50
JNS 03613
Gangliosides of human cerebrospinal fluid in various neurologic diseases Milica Trbojevic-Cepe 1, Ivica Kracun 3, Anica Jusic 2 and Ivan Pavlicek 4 I Institute of Clinical Laboratory Diagnostics, : Clinic of Neurology and Department of Neuropathology. Clinical Hospital Center Rebro, Zagreb, 3 Department of Chemistry and Biochemistry, Zagreb University School of Medicine, 4 Department of Neurology, Medical Center Varazdin, Zagreb (Yugoslat~ia)
(Received 13 December 1990) (Revised, received 1 May, 1991) (Accepted 24 May, 1991) Cerebrospinal fluid; Gangliosides; Neurologic diseases; Quantification
Summary Simultaneous profile determination and quantification of human cerebrospinal fluid (CSF) gangliosides in various neurologic diseases (n = 71) was examined. Gangliosides were extracted with methanol/chloroform from clinically available amounts of CSF (4-5 ml), then separated and quantified by high-performance thin-layer chromatography (HPTLC) and direct densitometry. Based on chromatographic comparison with standards, the percentage of lipid-bound NeuAc positive fractions in 'normal' CSF samples were: GM1 (1I 3 NeuAc-GgOse4Cer) (3%); GD3 (II 3 NeuAc2-Lac-Cer) (4%); GDla (IV 3 NeuAc, 1I 3 NeuAc-GgOse4 Cer) (15%); X1 (3%); GDlb (II3(NeuAc)2-GgOse4 Cer) (16%); X2 (4%); GTlb (IV 3 NeuAc, II3(NeuAc)2-GgOse4-Cer) (40%); and GQlb (IV3(NeuAc)2, II3(NeuAc)2 - GgOse4-Cer (15%). Similarity between CSF and human cerebellar cortex, particularly in proportion of "b" series gangliosides (GQlb, GTlb, GDlb), could be observed. A higher proportion of GDla ganglioside, with decreased GQlb was found in infancy. The total ganglioside content (mean + 2 SD) varied between 645-894 p~g/l. Significant alterations of the CSF ganglioside profile, with an increase in less polar gangliosides, GM3 and GD3, correlated with the blood-brain barrier dysfunction (CSF hemorrhages, compressive syndrom), or some malignant processes (metastatic brain melanoma). A statistically significant increase in the content of total CSF gangliosides was found in the following groups of patients as compared to controls: (1) ischemic cerebrovascular accident (CVI) with good outcome (P < 0.02); (2) peripheral neuropathy and polyneuropathy (P < 0.001) and (3) intravertebral discopathy (P < 0.05). A significant decrease in the content of total CSF gangliosides was found in CVI group with lethal outcome (P < 0.05).
Introduction Very early, evidence was presented that lipid alterations in cerebrospinal fluid (CSF) correlated with certain neurologic diseases, such as lipidosis, or white matter destruction (Pilz, 1970). Although the overall lipid composition of CSF reflects that of plasma (Tichy, 1972) and their increased content mostly correlates with b l o o d - b r a i n barrier impairment (Pedersen, 1973a,b), certain similarities to brain lipids have been observed. H u m a n CSF has been shown to contain galactosyl-ceramide (Nagai and Kanfer, 1971) and ganglio-series gangliosides which represent brain glyco-
Correspondence to: Dr. Milica Trbojevic-Cepe, Institute of Clinical Laboratory Diagnostics, Clinical Hospital Center Rebro, Kispaticeva 12, Zagreb University School of Medicine, YU-41 000 Zagreb (Yugoslavia). Tel.: 0038 41 23 32 33 (354).
sphingolipids (Ledeen and Yu, 1972; Nagai et al., 1973; Davidsson et al., 1989). Gangliosides * are sialic acid-containing glycosphingolipids that are prominent constituents of nervous tissue and highly concentrated in the outer half of the plasma membrane (Wiegandt, 1967; Ledeen, 1978). They are long lived lipids and their profile depends on the developmental state of the cells and their growth condition (review Hakomori, 1981). Also, their involvement in certain hereditary diseases, e.g. gangliosidosis, has been documented (Sandhoff and Conzelmann,
* Designation according to IUPAC-IUB Recommendations (1977) Eur. J. Biochem., 79: 11-21. Ganglioside abbreviations follow the nomenclature system of L. Svennerholm (1963): GM3; II 3 NeuAcLac-Ger; GM1; II 3 NeuAc-GgOse4-Cer;GDla; IV 3 NeuAc-, II 3 NeuAc-GgOse4-Cer, GDlb; II 3 (NeuAc)2-GgOse4-Cer;GTlb; IV ~ NeuAc-, II 3 (NeuAc)2-GgOse4-Cer; GQlb; IV3 (NeuAc)2-II3 (NeuAc)2-GgOsea-Cer.
193 1984), but few reports have described their accumulation in CSF (Pullarkat et al,, 1981, Kotagal et al., 1986). It has not been clarified to date to what extent the other diseases of the central nervous system, i.e. degenerative diseases, turnouts, inflammatory or autoimmune processes, etc., are reflected in CSF. We developed a method for simultaneous profile determination and quantification of CSF gangliosides from clinically available amounts of CSF (Trbojevic and Kracun, 1990). The aim of this study was to follow up CSF ganglioside alterations in various neurologic diseases and determine their possible diagnostic value.
Materials and methods
Solvents, chloroform and methanol, were redestilled before use. A commercially available apparatus and collodium bags SM 132 00 from Sartorius GmbH (G6ttingen, F.R.G.) were used for ultrafiltration. Standard dialyzing membranes from Technicon Chemicals Co. (Orcq, Belgium) were used for dialysis. High performance thin-layer chromatographic plates (HPTLC) on silicagel 60 (10 x 10 cm) were purchased from Merck (Darmstadt, F.R.G.). Cronassial, a ganglioside mixture from Fidia (Abano Terme, Italy), diluted 1:10 with water (0.5 mg/ml of gangliosides), was used as standard for CSF ganglioside HPTLC determination. Lumbar and ventricular CSFs were obtained from patients with various neurologic diseases: ischemic cerebrovascular accident (n = 29), intracerebral hemorrhages penetrating into CSF (n = 3), neoplastic diseases (n = 7), motor neuron diseases (n = 5), peripheral neuropathy and polyneuropathy (n = 4) intravertebral discopathy (n = 9), multiple sclerosis (n = 7) and miscellaneous diseases (n = 7). Lumbar CSFs from adults: vertigo (n = 2), cephalea (n = 3), neurosis (n = 1) and TIA (transient ischemic attack) (n = 4), median age 39 years (range 23-55 years), were used as a "control" group.
Routine CSF studies After cytologic examination, CSF samples were centrifuged at 2000 x g for 10 rain. Total protein concentration was determined by the colorimetric method (Rieder, 1966). Hematogenic pigments in CSFs were determined by spectrophotometry (Kjellin and Soerstrom, 1974). The functional state of blood-brain barrier and intrathecally synthesized IgG were determined by evaluation graph according to Reiber (Reiber, 1980). Oligoclonal IgG bands were detected from native CSF samples by isoelectric focusing in ultrathin-layer PAGE, immunofixation and silver staining (Trbojevic et al., 1989).
CSF ganglioside extraction and purification by microdialysis In our procedure, the volume of native CSF samples (5 ml) was first reduced to 0.5 ml by ultrafiltration against water, semipermeable membrane rinsed with 0.250 ml of 0.077 mol/l saline and CSF concentrate volume adjusted to 0.750 ml. Gangliosides were extracted by chloroform/methanol (1:2, v/v). The solvent ratio was chloroform/ methanol/water, 1:2:0.75 (v/v). Instead of water phase, a CSF concentrate (0.75 ml) was used. After extraction procedure using a cyclomixer, the gangliosides were separated from other lipids by phase partition, adding 0.65 ml of water. Phases were separated by centrifugation (15 min, 3000 x g) and the upper phase was set aside. For second phase partition, on the lower phase 0.38 ml of methanol and 0.25 ml of water were added. Two upper phases were combined and evaporated to dryness. Crude CSF ganglioside extract was purified by microdialysis. The upper phase residue was dissolved in several portions of chloroform/methanol/water (60: 30:4.5, v/v) and quantitatively transferred to a microcup for dialysis (volume 100 izl). After solvent evaporation to dryness, the ganglioside extract was dissolved in 100 ~1 of water, the microcup covered with a dialyzing membrane, and dialyzed against water for about 4 hours. The dialyzed CSF gangliosides were evaporated to dryness.
High-performance thin-layer chromatography (HPTLC) and CSF ganglioside quantification by direct densitometry Gangliosides were separated using the ascending "sandwich" HPTLC technique (1-mm thick plastic strips were used as gaskets) on pre-washed and activated (II0°C, 60 min) plates. The CSF ganglioside extract was dissolved in a few microliters of chloroform-methanol-water (60 : 30: 4.5, v/v) and quantitatively transferred onto a HPTLC plate. When densitometric quantification was required, three standards of 5, 10 and 15/zl of diluted Cronassial (corresponding to 2.5, 5 and 7.5 /~g of gangliosides, respectively) were spotted on the same HPTLC plate. The samples were applied as a 6-mm line. Chromatography was performed at 17-18 ° C. Double development in the same but freshly prepared solvent system, chloroform/ methanol/12 mmol/l MgCI 2 in H20/13.3 mol/1 NH 3 (60 : 40 : 9 : 0.5, v/v), was used. Development was stopped when the solvent front reached 1-1.5 cm from the upper edge of the plate. Gangliosides were visualized on the plate (as lipid-bound NeuAc) by spraying with resorcinol-HC1 reagent diluted with an equal volume of water (Svennerholm, 1957) and the plate was heated upside-down, as suggested by Ando et al. (1978). Chromatograms were scanned using an LKB 2202 U1-
194
GM3
GM2 GM1 GD3 GDla
GDlb GT 1 b
Xl x2
CIOlb
1 2 " C ro n a s s i a l "
a
b
c
d
e
f
g
h
i
J
k
1 2 3 "Cr o nassial"
Fig. 1. a-k: high-performance thin-layer chromatographic (HPTLC) separation of gangiiosides extracted from cerebrospinal fluid (CSF) of patients with cerebrovascular insult (CVI) (b-e), neurinoma (f), spinal cord tumor (j) in comparison to adult (a, h, k) and infant (i) controls. 1-3, samples correspond to 2.5 (1), 5 (2) and 7.5/zg of Cronassial-gangliosides. XI (GTla?); X2 (GD2?).
trascan Laser Densitometer with 2220 Recording Integrator. Total amounts of CSF gangliosides were calculated from a calibration curve constructed from the densitometer detector response (total area) plotted as a function of the amount of gangliosides in standards (Cronassial). Individual gangliosides were calculated from the calibration curve according to the corresponding ganglioside detector response (area), or from the relative percentage multiplied by the total ganglioside content. Results were statistically analysed by Student's t-test.
Results
Fig. 1 shows HPTLC chromatograms of CSF gangliosides from various neurologic patients. Individual fractions containing about 60 ng of gangliosides could be detected, if applied as a 6-mm line. In human CSF samples, eight to nine lipid-bound NeuAc positive fractions were present. In comparison to a standard ganglioside mixture, control (as well as pathologic) CSF contained mainly ganglio-series gangliosides (GQlb, GTlb, GDlb, GDla, GM1, approx. 90%). Only traces of other ganglioside species (GD3, approx. 4%) as well as unknown ganglioside fractions (Fig. 1, X1, X2) were detected. Fig. 2 (a-k) shows densitograms of the HPTLC ganglioside patterns in normal CSF and in various neurologic diseases. In addition, the ganglioside profile of the adult human cerebellar cortex and blood serum were demonstrated (Fig. 2b, c). Similarity between CSF and human cerebellar cortex, particularly in proportion of "b"-series gangliosides (GQlb, GTlb, GDlb), could be observed.
Blood serum contained mainly GD3 and GM3 ganglioside species (approx. 75%). However, a reduced function of blood-brain barrier led to a leakage of plasma gangliosides (GM3, GD3) into CSF Fig. 2e, f, g). Fig. 3 shows the frequency of distribution of total CSF ganglioside contents in control group and in various groups of neurologic diseases. A statistically significant increase in the content of total CSF gangliosides was found in the following groups of patients as compared to controls: (1) CVI (cerebrovascular accident) with good outcome (p < 0.02); (2) peripheral neuropathy and polyneuropathy (P < 0.001); and (3) intravertebral discopathy (P < 0.05). A significant decrease in the content of total CSF gangliosides was found in CVI group with lethal outcome (P < 0.05). Fig. 4 shows amounts of individual CSF gangliosides (~g/1, mean + SD) in control group and in various neurologic diseases. The following amounts of individual CSF gangliosides in control adults were found: GM1 (26 +_5), GD3 (30 + 5), GDla (113 + 12), GDlb (122_+ 10), GTlb (31l _+ 14) and GQlb (117 +_ 11) *. A statistically significant increase (P < 0.05) in the amounts of CSF predominant gangliosides generally correlated with the contents of total CSF gangliosides in the groups of patients examined. We observed an elevated concentration of CSF "b"-series gangliosides (GTlb, GDlb), particularly GQlb ganglioside in different neuropathies. In other
* Corresponding to: GM1 (16+_3 nmol/l), GD3 (16+_3 nmol/D, G D l a (60+- 6 nmol/l), GDIb (65-5 nmol/I), GTIb (141 +-6 nmol/1), and G Q l b (47 +-4 nmol/I), as calculated by using appropriate molecular mass to each individual ganglioside.
195
D
m
i
c) Normal plasma
b) Cerebellar cortex
a) Normal CSF
m
m
8
8
J d)
CSF
-
r~.urinOm (ventrieular)
e)
C'~F -
x~
z~
tumor reed. eplrmlls
f) C S F
- tumor
rood.
spinalls
g) C.SF - cerebral hemorrhage
m
m
/ h) I n f a n t ' s
CSF - purpura (11 months)
1)
C.SF - m u l t i p l e
selerosls
J) CSF - ALS (inherited)
k ) CSF - sy. amyotrophicum
Fig. 2. a-k: HPTLC of CSF gangliosides from different neurologic diseases (d-k), and control (a) as revealed by densitometry. The gangliosides from adult human cerebellum (b) and blood plasma (c) are presented for comparison.
hand, a decreased amount of X2 ganglioside fraction in CSF of all patients with CVI was detected. Percentage distribution of CSF ganglioside (%, mean + SD), contents of total CSF gangliosides (/zg/l, mean + SD) and proteins (g/l, mean + SD) in control group and in some individual neurologic diseases are shown quantitatively by Table 1. Significant alteration of the CSF ganglioside profile, with an increase in less polar gangliosides, GM3 and GD3 (but varying respective interrelations) and a decrease in polysialogangliosides, especially GQlb, were demonstrated in cerebral hemorrhages and some neoplastic brain diseases. It is important to point a high proportion of GD3 ganglioside (28%) in cerebral metastasis of melanoma. A higher proportion of GDla
ganglioside, with a decreased GQlb, was found in infancy (see also Figs. li, j and Figs. 2e, f, g, h).
Discussion Use of the HPTLC method and chromatography condition described, resulted in uniform ganglioside bands geometry, a prerequisite for ganglioside quantification within a mixture by direct densitometry. Employing direct densitometry, simultaneous quantitative results and ganglioside pattern (8-9 fractions) in individual CSF samples (4-5 ml) could be obtained (Trbojevic and Kracun, 1990). The CSF ganglioside composition in normal samples
196
:liL NO
Controls (adults)
Controls (adults)
(n = ~0)
(n = 10) 20 0
774 + 60 /ug/L
774 3 60 / u g / L
3~.
:f
CEREBROVASCULAR ACCIDENT - outcome good ,
ta
(n = 9 ) /o
: _ . ~ ! n a ! ! a !"--.
887 S 154 /~g/L**
MOTOR NEURON DISEASES (n : 5)
~ /
.
~
8 5 6 - + 188 / u g / L
41"
4~
.... ,ifl
- outcome Door (n : 12)
511111111
~ ,
LIM'
PERIPHERAL NEUROPATHY AND POLYNEUROPATHY (n : 4)
1087 + 102 /ug/L***
725 3 282 /ug/L
~
;" ~2 INTRAVERTEBRAL
, o _-, 2
871 -+ 153 /ug/L t
- outcome l e t h a l
~'
LIM:
(n : 8)
, :/
. . . . . . .~ -
6 4 5 - + 222 / u g / L * ~o
z
41"
MULTIPLE SCLEROSIS 2
* p < 0.05 ** p < 0.02 *** p < 0.001
41-
DISCOPATHY
~
(n = 7)
781 + 125 /ug/L
Mean
Fig. 3. T o t a l c o n t e n t o f C S F g a n g l i o s i d e s ( / z g / l , m e a n + S D ) a n d f r e q u e n c y distribution o f g a n g l i o s i d e s ( / x g / I ) in v a r i o u s n e u r o l o g i c d i s e a s e s a n d c o n t r o l g r o u p . * P < 0.05; * * P < 0.02; * * * P < 0.001.
resembles more the brain ganglioside profile, particularly to cerebellar cortex with the predominance of GTlb ganglioside. This finding could be related to evidence obtained in mutant mice that cerebellar Purkinje cells are enriched in ganglioside GTlb (Seyfried et al., 1984). Our results also demonstrated GDla ganglioside to be increased (and GQlb to be decreased) in infancy, which may reflect high proportion of GDla gangliosides in the infant human brains (Kracun et al., 1986). Elevated proportions of GM3 and less GD3 gangliosides in CSF pattern correlate with a more pronounced dysfunction of bloodbrain barrier, such as in cerebral hemorrhages or some compressive syndroms. The penetration of blood serum into CSF attributed to elevated GM3 and GD3 gangliosides since serum contains mainly GM3 (54%) and GD3 (30%) gangliosides
(Kundu et al., 1985). Ledeen and Yu (1972) reported on the profile of ventricular CSF gangliosides differing from that of lumbar CSF, exhibiting marked NeuAc positive fractions in the GD3 and GM3 area. We did not find ventricular CSF to contain marked concentration of GM3 and GD3 in patients with neurinoma. So, we believe that elevated GD3 and GM3 in ventricular CSF described by Ledeen and Yu (1972) correlate to astrocytoma since the sample was taken from a child with hydrocephalus due to astrocytoma in the region of cerebellum. On the other hand, it has been demonstrated that astrocytoma contains accumulated GD3 and GM3 gangliosides (Traylor and Hogan, 1980; Markovic et al., 1988). In addition, we found high content of GD3 ganglioside (and less GM3) in lumbar CSF from a patient with metastatic brain melanoma free of sign of blood-brain barrier damage correlating
197
200
"
GOlb 50I
500
::::::::::::::::::::::::::::::::::::::::::::::::::::::
-
GTlb
100
I
90-
........
, _.__L_t
_'; .........
.....
.........~--}--?--4 .................. z_.
w
X2
10I
N 200
-
GDIb 50-
O
!
30-
o
X1
10-
I 200
-
.,~ T
; .......................... • . . . . . . . ~o_
*
..........t_~_}
~.__~_ . . . . . . .
~......
GDla
......... ~...............: 3 ............~ .......
I
so-
,ol I
GD3
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
50-
lOt
o~,,~ °"°'jJ
I I
~.¢ ~o go .¢teO,t 0¢. .,~oo'ROh/ Ot~z. G
n, •
,i, 0
q%~" G
Fig. 4. The concentration of individual gangliosides 0 x g / L ) in CSF of patients with various neurologic diseases and control group. * P < 0.05; X1, X2 = see Fig. 1.
well with accumulation of GD3 ganglioside in human melanoma (Pukel et al., 1982). Reported values on the total ganglioside content in CSF differ markedly (Ledeen and Yu, 1972; Nagai et al., 1974; Hirabayashi et al., 1986; Davidsson et al., 1989), probably due to different analytical procedures.
Our results on the content of gangliotetraose series of gangliosides in control adults (291-353 nmol/l) is comparable with the most recent results of Davidsson et al. (1989) (179-491 nmol/l) using very sensitive H P T L C / immunoblot techniques. The increased total CSF content of gangliosides in
198 TABLE I P E R C E N T A G E D I S T R I B U T I O N ( % M E A N +_ SD) A N D A M O U N T S INDIVIDUAL NEUROLOGIC PATIENTS
Controls (adults)
HPTLC (%)
(n = 101
GQIb
Controls (adults) (n = 101
Intracerebral hemorrhages penetrating in CSf
Artificial hemorrhages in C S F Neoplastic diseases: Tumor med. spinalis
Tumor cerebri a (neurinoma) Angioma (14 yrs old) Melanoma meta cerebri
GTlb
O F T O T A L C S F G A N G L I O S 1 D E S ( p , g / I , M E A N _+ SD) IN S O M E
Total C S F X2
GDlb
X1
GDla
15.2 +_ 1.4
40.2
4.5
15.7
+ 1.8
+ 1.3
+ 1.3
2.6 + 0.9
14.6 _+ 1.5
12.8
GD3
3.9 + 0.7
GM1
3.3 + 1
GM3
--
Total CSF
gangliosides
proteins
(~g/I)
(g/I)
774 _+ 60
0.32 +_ (I.04
940 b
2.0
-
36.2 b
3.1
12.9 b
2.0
14.2
6.0 b
3.0
8.2 b
35.0 b
3.8
13.8
2.2
13.8
6.9 h
2.8
13.5 b
810
2.8
7.1 b
30.2 b
2.7
13.1
2.9
13.6
8.2 b
2.2
20.0 h
840
4.1
12.3 b 6.5 b
35.4 b 26.7 b
4.0 2.1
14.3 12.0 b
2.0 3.8
13.1 13.4
5.9 b 7.2 h
2.8 2.1
10.2 b 26.2 b
880 920 b
1.8 4.4
10.2 b 9,6 b 5.7 b
42.9 37.2 33.1 b
3.1 7.0 4.1
15.8 14.9 14.6
1.0 2.2
14.9 14.1 11.5 b
2.1 3.8 1.8 b
-1.9 b 14.2 h
710 580 620 b
0.9 1.7 9.3
10.3 b
42.0
--
17.8 b
_
18.0 b
6.1 b
1.7 b
230 b
0.27
10.4 b 3.0 b
36.7 29.7 b
5.7 2.2
15.8 11.0 b
3.1 2.0
18.8 b 12.2
4.2 28.5 b
5.3 2.4
-9.0 b
930 b 1 030 b
0.37 0.50
12.4 15.0 19.8 b 15.7
41.0 44.6 b 40.2 42.2
7.0 6.1 5.2 6.8
18.6 b 14.2 16.0 17.9 b
1.3 2.1 2.8 1.9
12.8 10.3 b 10.2 b 9.8 b
3.9 4.0 3.0 3.1
3.0 3.7 2.8 2.6
---
930 b 860 1 130 b 680
0.36 0.52 0.42 0.34
14.8
43.6
--
15.2
3.6
15.0
4.3
3.5
-
680
0.30
1 200 1 060 960 1 130
10.0 b 9.3 b 15.0 b 4.1
9.8
Motor neuron diseases: ALS ALS
(inherited) Spinal muscular atr. Sy. amyotrophicum ALS
Peripheral neuropathy and polyneuropathy: Polyneuropathy 18.2 b 37.1 Charcot-Marie-Tooth Brachial neuropathy
16.3 19.2 b 20.1 b
38.0 40.2 42.0
6.1 7.5 b 6.8 2.2
15.2 16.3 13.0 15.8
2.1 1.2 1.1 2.1
13.9 14.1 13.0 12.1
3.4 4.3 4.1 2.7
4.0 2.3 2.6 3.0
-----
Miscellaneous diseases: Epilepsy grand real Contusio cerebri (5 yrs) Purpura (13 months old) Pediatric pooled CSFs
8.5 11.4 9.0 11.5
47.8 b 42.6 36.8 37.5
2.2 5.0 4.2 5.2
16.6 12.0 b 14.3 13.9
3.8 _ 3.7 2.9
14.4 19.2 b 24.0 b 20.0 b
4.1 3.8 3.0 4.0
2.6 6.0 b 5.0 b 5.0 b
_ _ _
b b b b
b b b b
0.38 0.34 0.36 0.45
720 240 b 960 b 660
0.41 0.21 0.26 0.22
" V e n t r i c u l a r C S F (hydrocephalus). b Out fo control range.
cerebrovascular insult with good outcome, peripheral neuropathies and polyneuropathies, and in intravertebral discopathies cannot be explained. At this time, it would be warranted to relate elevated CSF gangliosides to regenerating factors (Gorio et al., 1980; Toffano et al., 1983; Cuello et al., 1989) or to the presence of gangliosides antibodies seen in many types of peripheral neuropathies. Further, we are inclined to believe that the finding of a decreased total content of gangliosides in CVI with lethal outcome results from a severe and irreversible brain damage and a decreased ganglioside metabolism. It should also be considered, however, that the development of edema in CVI and increase in
intracranial pressure may accelerate the bulk-flow mechanism in arachnoid villi leading to faster ganglioside elimination from CSF. Nevertheless, non-cerebral complications of the underlying disease are quite frequently present in CVI, which may account for great differences in the total CSF content of gangliosides in CVI with poor outcome. Analyzing 100 ml of pooled CSF, Nagai et al. (1973) have pointed to an increased content of gangliosides in multiple sclerosis. These report were not confirmed by our results obtained on individual CSF samples. However, in one patient with multiple sclerosis two lumbar punctures were performed, i.e. in the active stage of the disease and in the stage of remission (time interval
199
3 months). In the active stage of the disease, higher values of gangliosides in CSF were recorded (970/~g/1), and in the stage of remission it was 670/~g/l. Accordingly, the amount of gangliosides in CSF in multiple sclerosis may correlate with the dynamic of disease. Our results on CSF gangliosides in control individuals (n = 10) and different neurologic patients (n = 71) demonstrated that CSF gangliosides reflect some brain diseases. Acknowledgement This work was supported by Scientific Fund of Croatia (No. 1.08.03.05.39.). The authors wish to thank Mrs Radojka Barbaric for skilful technical assistance and Antonija Redovnikovic, B.A., for editorial work.
References Ando, S., N. Chang and R.K. Yu (1978) High peformance thin-layer chromatography and densitometric determination of brain ganglioside composition of several species. Anal. Biochem., 89: 437450. Cuello, A.C., L. Garafalo, R.L. Kenigsberg and D. Maysinger (1989) Ganglioside potentiate in vivo and in vitro effects of nerve growth factor on central cholinergic neurons. Proc. Natl. Acad. Sci., 86: 2056-2060. Davidsson, P., P. Fredman and L. Svennerholm (1989) Gangliosides and sulphatides in human cerebrospinal fluid: quantitation with immunoaffinity techniques. J. Chromatogr., 496: 279-289. Gorio, A., G. Carmignato, L. Facci and M. Finesso (1980) Motor nerve sprouting induced by gangliosides treatment. Possible implication for gangliosides on neuronal growth. Brain Res., 297: 236-241. Hakomori, S.-I. (1981) Glycosphingolipids in cellular interaction, differentiation and oncogenesis. Annu. Rev. Biochem., 50: 733764. Hirabayashi, Y., K. Koketsu, H. Higashi, Y. Suzuki, M. Matsumoto, M. Sugimoto and T. Ogawa (1986) Sensitive enzyme-immunostaining and densitometric determination of ganglioseries gangliosides on thin-layer plate: pmol detection of gangliosides in cerebrospinal fluid. Biochim. Biophys. Acta, 876: 178-182. Kjellin, K.G. and C.E. S6derstrfim (1974) Diagnostic significance of CSF spectrometry in cerebrovascular diseases. J. Neurol, Sci., 23: 359-369. Kotagal, S., D.A. Wenger, H. Alcala, C. Gomez and S. Horesstein (1986) AB variant GM2 gangliosidosis; cerebrospinal fluid and neuropathological characteristics. Neurology, 36: 438-440. Kracun, I., H. Rosner and C. Cosovic (1986) In: (Eds.), Tettamanti, G., Ledeen, R.W., Saudhoff, K., Nagai, Y., Toffano, G. Gangliosides and Neuronal Plasticity, Fidia Research Series, Liviana Press, Springer-Verlag, Padova-Berlin-Heidelberg-New YorkTokyo, pp. 67-76. Kundu, S.K., I. Diego, S. Osovitz and D.M. Marcus (1985) Gangliosides of human plasma. Arch. Biochem. Biophys., 238: 388-4130. Ledeen, R.W. (1978) Ganglioside structures and distribution: are they localised at the nerve ending? J. Supramol. Struct., 8: 1-17. Ledeen, R.W. and R.K. Yu (1972) Gangliosides of CSF and plasma:
Their relation to the nervous system. Adv. Exp. Med., and Biol., 19: 77-93. Markovic, I., M. Heifer, J. Rogan-Grgas, T. Vladovic-Relja, G. Kracun, C. Cosovic and I. Kracun, 1988) Gangliosides composition of different human brain gliomas. Period. Biol., 90: 97-104. Nagai, Y. and J.N. Kanfer (1971) Composition of human cerebrospinal fluid cerebrosides. J. Lipid Res, 12: 143-148. Nagai, Y., J.N. Kanfer and W.W. Tourtellotte (1973) Preliminary observations of gangliosides of normal and multiple sclerosis cerebrospinal fluid. Neurology, 23: 945-948. Pedersen, H.E. (1973a) Cerebrospinal fluid cholesterol. Quantitative analysis of unesterified cholesterol and normal values. Acta Neurol. Scand., 49: 626-638. Pedersen, H.E. (1973b) Cerebrospinal fluid phospholipids. Quantitative determination of fractionated phospholipids and normal values in the cerebrospinal fluid. Acta Neurol. Scand., 49: 639-648. Pilz, H. (1970) Die Lipide des normalen und pathologischen Liquor cerebrospinalis. Springer-Verlag, Berlin-Heidelberg-New York. Pukel, C.S., C.S. Lloyd, K.O. Lloyd, L.R. Travassos and W.G. Dippold (1982) GD3, a prominent ganglioside of human melanoma. J. Exp. Med., 155: 1133-1147. Pullakart, R.K., H. Reha and N.G. Beratis (1981) Accumulation of GM2 ganglioside in cerebrospinal fluid of a patient with the variant AB of infantile GM2 gangliosides. Pediatrics, 68: 106-108. Reiber, H. (1980) Discrimination between different blood - CSF barrier dysfunctions and inlammatory reactions of the CNS by a recent evaluation graph for the protein profile of cerebrospinal fluid. J. Neurol,, 224: 89-99. Rieder, H.P. (1966) Eine neue Modifikation zur Methode der Eiweissbestimmung Mittels Cu-(II) und phenolreagens nach FolinCiocalteu. Klin. Wschr., 44: 1036-1039. Sandhoff, K. and E. Conzelmann (1984) The biochemical basis of gangliosidosis. Neuropediatrics, 15: 85-92. Seyfried, T.N., D.J. Bernard and R.K. Yu (1984) Cellular distribution of gangliosides in the developing mouse cerebellum: analysis using staggerer mutant. J. Neurochem. 43: 1152-1162. Svennerholm, L. (1957) Quantitative estimation of sialic acid. Biochim. Biophys. Acta, 24: 604-611. Svennerholm, L. (1963) Chromatographic separation of human brain gangliosides. J. Neurochem., 10, 613-623. Tichy, J. (1972) Vergleichende Untersuchungen der Lipide im Blut and Liquor cerebrospinalis. Zschr. Inn. Med., 27: 21-25. Toffano, G., G. Savoini, F. Moroni, G. Lombardi, L. Calza and L.F. Agnati (1983) GM1 ganglioside stimulates the regeneration of dopaminergic neurons in the central nervous system. Brain Res., 261: 163-166. Traylor, T.D. and E.L. Hogan (1980) Gangliosides of human cerebral actrocytomas. J. Neurochem., 34: 126-131. Trbojevic-Cepe, M., Z. Poljakovic, N. Vrkic and I. Bielen (1989) Detection of oligoclonal bands in unconcentrated CSF by isoelectric focusing in ultrathin polyacrylamide gel, direct antiserum immunofixation and silver nitrate staining. J. Clin. Chem. Clin. Biochem., 27: 211-216. Trbojevic-Cepe, M. and I. Kracun (1990) Determination of gangliosides in human cerebrospinal fluid by high-performance thin-layer chromatography and direct densitometry. J. Clin. Chem. Clin. Biochem., 28: 863-872. Wiegandt, H. (1967) The subcellular localization of gangliosides in the brain. J. Neurochem., 14: 671-674.
Journal of the Neurological Sciences, 105 (1991) 192-199 © 1991 Elsevier Science Publishers B.V. All rights reserved 0022-510X/91/$03.50
JNS 03613
Gangliosides of human cerebrospinal fluid in various neurologic diseases Milica Trbojevic-Cepe 1, Ivica Kracun 3, Anica Jusic 2 and Ivan Pavlicek 4 I Institute of Clinical Laboratory Diagnostics, : Clinic of Neurology and Department of Neuropathology. Clinical Hospital Center Rebro, Zagreb, 3 Department of Chemistry and Biochemistry, Zagreb University School of Medicine, 4 Department of Neurology, Medical Center Varazdin, Zagreb (Yugoslat~ia)
(Received 13 December 1990) (Revised, received 1 May, 1991) (Accepted 24 May, 1991) Cerebrospinal fluid; Gangliosides; Neurologic diseases; Quantification
Summary Simultaneous profile determination and quantification of human cerebrospinal fluid (CSF) gangliosides in various neurologic diseases (n = 71) was examined. Gangliosides were extracted with methanol/chloroform from clinically available amounts of CSF (4-5 ml), then separated and quantified by high-performance thin-layer chromatography (HPTLC) and direct densitometry. Based on chromatographic comparison with standards, the percentage of lipid-bound NeuAc positive fractions in 'normal' CSF samples were: GM1 (1I 3 NeuAc-GgOse4Cer) (3%); GD3 (II 3 NeuAc2-Lac-Cer) (4%); GDla (IV 3 NeuAc, 1I 3 NeuAc-GgOse4 Cer) (15%); X1 (3%); GDlb (II3(NeuAc)2-GgOse4 Cer) (16%); X2 (4%); GTlb (IV 3 NeuAc, II3(NeuAc)2-GgOse4-Cer) (40%); and GQlb (IV3(NeuAc)2, II3(NeuAc)2 - GgOse4-Cer (15%). Similarity between CSF and human cerebellar cortex, particularly in proportion of "b" series gangliosides (GQlb, GTlb, GDlb), could be observed. A higher proportion of GDla ganglioside, with decreased GQlb was found in infancy. The total ganglioside content (mean + 2 SD) varied between 645-894 p~g/l. Significant alterations of the CSF ganglioside profile, with an increase in less polar gangliosides, GM3 and GD3, correlated with the blood-brain barrier dysfunction (CSF hemorrhages, compressive syndrom), or some malignant processes (metastatic brain melanoma). A statistically significant increase in the content of total CSF gangliosides was found in the following groups of patients as compared to controls: (1) ischemic cerebrovascular accident (CVI) with good outcome (P < 0.02); (2) peripheral neuropathy and polyneuropathy (P < 0.001) and (3) intravertebral discopathy (P < 0.05). A significant decrease in the content of total CSF gangliosides was found in CVI group with lethal outcome (P < 0.05).
Introduction Very early, evidence was presented that lipid alterations in cerebrospinal fluid (CSF) correlated with certain neurologic diseases, such as lipidosis, or white matter destruction (Pilz, 1970). Although the overall lipid composition of CSF reflects that of plasma (Tichy, 1972) and their increased content mostly correlates with b l o o d - b r a i n barrier impairment (Pedersen, 1973a,b), certain similarities to brain lipids have been observed. H u m a n CSF has been shown to contain galactosyl-ceramide (Nagai and Kanfer, 1971) and ganglio-series gangliosides which represent brain glyco-
Correspondence to: Dr. Milica Trbojevic-Cepe, Institute of Clinical Laboratory Diagnostics, Clinical Hospital Center Rebro, Kispaticeva 12, Zagreb University School of Medicine, YU-41 000 Zagreb (Yugoslavia). Tel.: 0038 41 23 32 33 (354).
sphingolipids (Ledeen and Yu, 1972; Nagai et al., 1973; Davidsson et al., 1989). Gangliosides * are sialic acid-containing glycosphingolipids that are prominent constituents of nervous tissue and highly concentrated in the outer half of the plasma membrane (Wiegandt, 1967; Ledeen, 1978). They are long lived lipids and their profile depends on the developmental state of the cells and their growth condition (review Hakomori, 1981). Also, their involvement in certain hereditary diseases, e.g. gangliosidosis, has been documented (Sandhoff and Conzelmann,
* Designation according to IUPAC-IUB Recommendations (1977) Eur. J. Biochem., 79: 11-21. Ganglioside abbreviations follow the nomenclature system of L. Svennerholm (1963): GM3; II 3 NeuAcLac-Ger; GM1; II 3 NeuAc-GgOse4-Cer;GDla; IV 3 NeuAc-, II 3 NeuAc-GgOse4-Cer, GDlb; II 3 (NeuAc)2-GgOse4-Cer;GTlb; IV ~ NeuAc-, II 3 (NeuAc)2-GgOse4-Cer; GQlb; IV3 (NeuAc)2-II3 (NeuAc)2-GgOsea-Cer.
193 1984), but few reports have described their accumulation in CSF (Pullarkat et al,, 1981, Kotagal et al., 1986). It has not been clarified to date to what extent the other diseases of the central nervous system, i.e. degenerative diseases, turnouts, inflammatory or autoimmune processes, etc., are reflected in CSF. We developed a method for simultaneous profile determination and quantification of CSF gangliosides from clinically available amounts of CSF (Trbojevic and Kracun, 1990). The aim of this study was to follow up CSF ganglioside alterations in various neurologic diseases and determine their possible diagnostic value.
Materials and methods
Solvents, chloroform and methanol, were redestilled before use. A commercially available apparatus and collodium bags SM 132 00 from Sartorius GmbH (G6ttingen, F.R.G.) were used for ultrafiltration. Standard dialyzing membranes from Technicon Chemicals Co. (Orcq, Belgium) were used for dialysis. High performance thin-layer chromatographic plates (HPTLC) on silicagel 60 (10 x 10 cm) were purchased from Merck (Darmstadt, F.R.G.). Cronassial, a ganglioside mixture from Fidia (Abano Terme, Italy), diluted 1:10 with water (0.5 mg/ml of gangliosides), was used as standard for CSF ganglioside HPTLC determination. Lumbar and ventricular CSFs were obtained from patients with various neurologic diseases: ischemic cerebrovascular accident (n = 29), intracerebral hemorrhages penetrating into CSF (n = 3), neoplastic diseases (n = 7), motor neuron diseases (n = 5), peripheral neuropathy and polyneuropathy (n = 4) intravertebral discopathy (n = 9), multiple sclerosis (n = 7) and miscellaneous diseases (n = 7). Lumbar CSFs from adults: vertigo (n = 2), cephalea (n = 3), neurosis (n = 1) and TIA (transient ischemic attack) (n = 4), median age 39 years (range 23-55 years), were used as a "control" group.
Routine CSF studies After cytologic examination, CSF samples were centrifuged at 2000 x g for 10 rain. Total protein concentration was determined by the colorimetric method (Rieder, 1966). Hematogenic pigments in CSFs were determined by spectrophotometry (Kjellin and Soerstrom, 1974). The functional state of blood-brain barrier and intrathecally synthesized IgG were determined by evaluation graph according to Reiber (Reiber, 1980). Oligoclonal IgG bands were detected from native CSF samples by isoelectric focusing in ultrathin-layer PAGE, immunofixation and silver staining (Trbojevic et al., 1989).
CSF ganglioside extraction and purification by microdialysis In our procedure, the volume of native CSF samples (5 ml) was first reduced to 0.5 ml by ultrafiltration against water, semipermeable membrane rinsed with 0.250 ml of 0.077 mol/l saline and CSF concentrate volume adjusted to 0.750 ml. Gangliosides were extracted by chloroform/methanol (1:2, v/v). The solvent ratio was chloroform/ methanol/water, 1:2:0.75 (v/v). Instead of water phase, a CSF concentrate (0.75 ml) was used. After extraction procedure using a cyclomixer, the gangliosides were separated from other lipids by phase partition, adding 0.65 ml of water. Phases were separated by centrifugation (15 min, 3000 x g) and the upper phase was set aside. For second phase partition, on the lower phase 0.38 ml of methanol and 0.25 ml of water were added. Two upper phases were combined and evaporated to dryness. Crude CSF ganglioside extract was purified by microdialysis. The upper phase residue was dissolved in several portions of chloroform/methanol/water (60: 30:4.5, v/v) and quantitatively transferred to a microcup for dialysis (volume 100 izl). After solvent evaporation to dryness, the ganglioside extract was dissolved in 100 ~1 of water, the microcup covered with a dialyzing membrane, and dialyzed against water for about 4 hours. The dialyzed CSF gangliosides were evaporated to dryness.
High-performance thin-layer chromatography (HPTLC) and CSF ganglioside quantification by direct densitometry Gangliosides were separated using the ascending "sandwich" HPTLC technique (1-mm thick plastic strips were used as gaskets) on pre-washed and activated (II0°C, 60 min) plates. The CSF ganglioside extract was dissolved in a few microliters of chloroform-methanol-water (60 : 30: 4.5, v/v) and quantitatively transferred onto a HPTLC plate. When densitometric quantification was required, three standards of 5, 10 and 15/zl of diluted Cronassial (corresponding to 2.5, 5 and 7.5 /~g of gangliosides, respectively) were spotted on the same HPTLC plate. The samples were applied as a 6-mm line. Chromatography was performed at 17-18 ° C. Double development in the same but freshly prepared solvent system, chloroform/ methanol/12 mmol/l MgCI 2 in H20/13.3 mol/1 NH 3 (60 : 40 : 9 : 0.5, v/v), was used. Development was stopped when the solvent front reached 1-1.5 cm from the upper edge of the plate. Gangliosides were visualized on the plate (as lipid-bound NeuAc) by spraying with resorcinol-HC1 reagent diluted with an equal volume of water (Svennerholm, 1957) and the plate was heated upside-down, as suggested by Ando et al. (1978). Chromatograms were scanned using an LKB 2202 U1-
194
GM3
GM2 GM1 GD3 GDla
GDlb GT 1 b
Xl x2
CIOlb
1 2 " C ro n a s s i a l "
a
b
c
d
e
f
g
h
i
J
k
1 2 3 "Cr o nassial"
Fig. 1. a-k: high-performance thin-layer chromatographic (HPTLC) separation of gangiiosides extracted from cerebrospinal fluid (CSF) of patients with cerebrovascular insult (CVI) (b-e), neurinoma (f), spinal cord tumor (j) in comparison to adult (a, h, k) and infant (i) controls. 1-3, samples correspond to 2.5 (1), 5 (2) and 7.5/zg of Cronassial-gangliosides. XI (GTla?); X2 (GD2?).
trascan Laser Densitometer with 2220 Recording Integrator. Total amounts of CSF gangliosides were calculated from a calibration curve constructed from the densitometer detector response (total area) plotted as a function of the amount of gangliosides in standards (Cronassial). Individual gangliosides were calculated from the calibration curve according to the corresponding ganglioside detector response (area), or from the relative percentage multiplied by the total ganglioside content. Results were statistically analysed by Student's t-test.
Results
Fig. 1 shows HPTLC chromatograms of CSF gangliosides from various neurologic patients. Individual fractions containing about 60 ng of gangliosides could be detected, if applied as a 6-mm line. In human CSF samples, eight to nine lipid-bound NeuAc positive fractions were present. In comparison to a standard ganglioside mixture, control (as well as pathologic) CSF contained mainly ganglio-series gangliosides (GQlb, GTlb, GDlb, GDla, GM1, approx. 90%). Only traces of other ganglioside species (GD3, approx. 4%) as well as unknown ganglioside fractions (Fig. 1, X1, X2) were detected. Fig. 2 (a-k) shows densitograms of the HPTLC ganglioside patterns in normal CSF and in various neurologic diseases. In addition, the ganglioside profile of the adult human cerebellar cortex and blood serum were demonstrated (Fig. 2b, c). Similarity between CSF and human cerebellar cortex, particularly in proportion of "b"-series gangliosides (GQlb, GTlb, GDlb), could be observed.
Blood serum contained mainly GD3 and GM3 ganglioside species (approx. 75%). However, a reduced function of blood-brain barrier led to a leakage of plasma gangliosides (GM3, GD3) into CSF Fig. 2e, f, g). Fig. 3 shows the frequency of distribution of total CSF ganglioside contents in control group and in various groups of neurologic diseases. A statistically significant increase in the content of total CSF gangliosides was found in the following groups of patients as compared to controls: (1) CVI (cerebrovascular accident) with good outcome (p < 0.02); (2) peripheral neuropathy and polyneuropathy (P < 0.001); and (3) intravertebral discopathy (P < 0.05). A significant decrease in the content of total CSF gangliosides was found in CVI group with lethal outcome (P < 0.05). Fig. 4 shows amounts of individual CSF gangliosides (~g/1, mean + SD) in control group and in various neurologic diseases. The following amounts of individual CSF gangliosides in control adults were found: GM1 (26 +_5), GD3 (30 + 5), GDla (113 + 12), GDlb (122_+ 10), GTlb (31l _+ 14) and GQlb (117 +_ 11) *. A statistically significant increase (P < 0.05) in the amounts of CSF predominant gangliosides generally correlated with the contents of total CSF gangliosides in the groups of patients examined. We observed an elevated concentration of CSF "b"-series gangliosides (GTlb, GDlb), particularly GQlb ganglioside in different neuropathies. In other
* Corresponding to: GM1 (16+_3 nmol/l), GD3 (16+_3 nmol/D, G D l a (60+- 6 nmol/l), GDIb (65-5 nmol/I), GTIb (141 +-6 nmol/1), and G Q l b (47 +-4 nmol/I), as calculated by using appropriate molecular mass to each individual ganglioside.
195
D
m
i
c) Normal plasma
b) Cerebellar cortex
a) Normal CSF
m
m
8
8
J d)
CSF
-
r~.urinOm (ventrieular)
e)
C'~F -
x~
z~
tumor reed. eplrmlls
f) C S F
- tumor
rood.
spinalls
g) C.SF - cerebral hemorrhage
m
m
/ h) I n f a n t ' s
CSF - purpura (11 months)
1)
C.SF - m u l t i p l e
selerosls
J) CSF - ALS (inherited)
k ) CSF - sy. amyotrophicum
Fig. 2. a-k: HPTLC of CSF gangliosides from different neurologic diseases (d-k), and control (a) as revealed by densitometry. The gangliosides from adult human cerebellum (b) and blood plasma (c) are presented for comparison.
hand, a decreased amount of X2 ganglioside fraction in CSF of all patients with CVI was detected. Percentage distribution of CSF ganglioside (%, mean + SD), contents of total CSF gangliosides (/zg/l, mean + SD) and proteins (g/l, mean + SD) in control group and in some individual neurologic diseases are shown quantitatively by Table 1. Significant alteration of the CSF ganglioside profile, with an increase in less polar gangliosides, GM3 and GD3 (but varying respective interrelations) and a decrease in polysialogangliosides, especially GQlb, were demonstrated in cerebral hemorrhages and some neoplastic brain diseases. It is important to point a high proportion of GD3 ganglioside (28%) in cerebral metastasis of melanoma. A higher proportion of GDla
ganglioside, with a decreased GQlb, was found in infancy (see also Figs. li, j and Figs. 2e, f, g, h).
Discussion Use of the HPTLC method and chromatography condition described, resulted in uniform ganglioside bands geometry, a prerequisite for ganglioside quantification within a mixture by direct densitometry. Employing direct densitometry, simultaneous quantitative results and ganglioside pattern (8-9 fractions) in individual CSF samples (4-5 ml) could be obtained (Trbojevic and Kracun, 1990). The CSF ganglioside composition in normal samples
196
:liL NO
Controls (adults)
Controls (adults)
(n = ~0)
(n = 10) 20 0
774 + 60 /ug/L
774 3 60 / u g / L
3~.
:f
CEREBROVASCULAR ACCIDENT - outcome good ,
ta
(n = 9 ) /o
: _ . ~ ! n a ! ! a !"--.
887 S 154 /~g/L**
MOTOR NEURON DISEASES (n : 5)
~ /
.
~
8 5 6 - + 188 / u g / L
41"
4~
.... ,ifl
- outcome Door (n : 12)
511111111
~ ,
LIM'
PERIPHERAL NEUROPATHY AND POLYNEUROPATHY (n : 4)
1087 + 102 /ug/L***
725 3 282 /ug/L
~
;" ~2 INTRAVERTEBRAL
, o _-, 2
871 -+ 153 /ug/L t
- outcome l e t h a l
~'
LIM:
(n : 8)
, :/
. . . . . . .~ -
6 4 5 - + 222 / u g / L * ~o
z
41"
MULTIPLE SCLEROSIS 2
* p < 0.05 ** p < 0.02 *** p < 0.001
41-
DISCOPATHY
~
(n = 7)
781 + 125 /ug/L
Mean
Fig. 3. T o t a l c o n t e n t o f C S F g a n g l i o s i d e s ( / z g / l , m e a n + S D ) a n d f r e q u e n c y distribution o f g a n g l i o s i d e s ( / x g / I ) in v a r i o u s n e u r o l o g i c d i s e a s e s a n d c o n t r o l g r o u p . * P < 0.05; * * P < 0.02; * * * P < 0.001.
resembles more the brain ganglioside profile, particularly to cerebellar cortex with the predominance of GTlb ganglioside. This finding could be related to evidence obtained in mutant mice that cerebellar Purkinje cells are enriched in ganglioside GTlb (Seyfried et al., 1984). Our results also demonstrated GDla ganglioside to be increased (and GQlb to be decreased) in infancy, which may reflect high proportion of GDla gangliosides in the infant human brains (Kracun et al., 1986). Elevated proportions of GM3 and less GD3 gangliosides in CSF pattern correlate with a more pronounced dysfunction of bloodbrain barrier, such as in cerebral hemorrhages or some compressive syndroms. The penetration of blood serum into CSF attributed to elevated GM3 and GD3 gangliosides since serum contains mainly GM3 (54%) and GD3 (30%) gangliosides
(Kundu et al., 1985). Ledeen and Yu (1972) reported on the profile of ventricular CSF gangliosides differing from that of lumbar CSF, exhibiting marked NeuAc positive fractions in the GD3 and GM3 area. We did not find ventricular CSF to contain marked concentration of GM3 and GD3 in patients with neurinoma. So, we believe that elevated GD3 and GM3 in ventricular CSF described by Ledeen and Yu (1972) correlate to astrocytoma since the sample was taken from a child with hydrocephalus due to astrocytoma in the region of cerebellum. On the other hand, it has been demonstrated that astrocytoma contains accumulated GD3 and GM3 gangliosides (Traylor and Hogan, 1980; Markovic et al., 1988). In addition, we found high content of GD3 ganglioside (and less GM3) in lumbar CSF from a patient with metastatic brain melanoma free of sign of blood-brain barrier damage correlating
197
200
"
GOlb 50I
500
::::::::::::::::::::::::::::::::::::::::::::::::::::::
-
GTlb
100
I
90-
........
, _.__L_t
_'; .........
.....
.........~--}--?--4 .................. z_.
w
X2
10I
N 200
-
GDIb 50-
O
!
30-
o
X1
10-
I 200
-
.,~ T
; .......................... • . . . . . . . ~o_
*
..........t_~_}
~.__~_ . . . . . . .
~......
GDla
......... ~...............: 3 ............~ .......
I
so-
,ol I
GD3
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
50-
lOt
o~,,~ °"°'jJ
I I
~.¢ ~o go .¢teO,t 0¢. .,~oo'ROh/ Ot~z. G
n, •
,i, 0
q%~" G
Fig. 4. The concentration of individual gangliosides 0 x g / L ) in CSF of patients with various neurologic diseases and control group. * P < 0.05; X1, X2 = see Fig. 1.
well with accumulation of GD3 ganglioside in human melanoma (Pukel et al., 1982). Reported values on the total ganglioside content in CSF differ markedly (Ledeen and Yu, 1972; Nagai et al., 1974; Hirabayashi et al., 1986; Davidsson et al., 1989), probably due to different analytical procedures.
Our results on the content of gangliotetraose series of gangliosides in control adults (291-353 nmol/l) is comparable with the most recent results of Davidsson et al. (1989) (179-491 nmol/l) using very sensitive H P T L C / immunoblot techniques. The increased total CSF content of gangliosides in
198 TABLE I P E R C E N T A G E D I S T R I B U T I O N ( % M E A N +_ SD) A N D A M O U N T S INDIVIDUAL NEUROLOGIC PATIENTS
Controls (adults)
HPTLC (%)
(n = 101
GQIb
Controls (adults) (n = 101
Intracerebral hemorrhages penetrating in CSf
Artificial hemorrhages in C S F Neoplastic diseases: Tumor med. spinalis
Tumor cerebri a (neurinoma) Angioma (14 yrs old) Melanoma meta cerebri
GTlb
O F T O T A L C S F G A N G L I O S 1 D E S ( p , g / I , M E A N _+ SD) IN S O M E
Total C S F X2
GDlb
X1
GDla
15.2 +_ 1.4
40.2
4.5
15.7
+ 1.8
+ 1.3
+ 1.3
2.6 + 0.9
14.6 _+ 1.5
12.8
GD3
3.9 + 0.7
GM1
3.3 + 1
GM3
--
Total CSF
gangliosides
proteins
(~g/I)
(g/I)
774 _+ 60
0.32 +_ (I.04
940 b
2.0
-
36.2 b
3.1
12.9 b
2.0
14.2
6.0 b
3.0
8.2 b
35.0 b
3.8
13.8
2.2
13.8
6.9 h
2.8
13.5 b
810
2.8
7.1 b
30.2 b
2.7
13.1
2.9
13.6
8.2 b
2.2
20.0 h
840
4.1
12.3 b 6.5 b
35.4 b 26.7 b
4.0 2.1
14.3 12.0 b
2.0 3.8
13.1 13.4
5.9 b 7.2 h
2.8 2.1
10.2 b 26.2 b
880 920 b
1.8 4.4
10.2 b 9,6 b 5.7 b
42.9 37.2 33.1 b
3.1 7.0 4.1
15.8 14.9 14.6
1.0 2.2
14.9 14.1 11.5 b
2.1 3.8 1.8 b
-1.9 b 14.2 h
710 580 620 b
0.9 1.7 9.3
10.3 b
42.0
--
17.8 b
_
18.0 b
6.1 b
1.7 b
230 b
0.27
10.4 b 3.0 b
36.7 29.7 b
5.7 2.2
15.8 11.0 b
3.1 2.0
18.8 b 12.2
4.2 28.5 b
5.3 2.4
-9.0 b
930 b 1 030 b
0.37 0.50
12.4 15.0 19.8 b 15.7
41.0 44.6 b 40.2 42.2
7.0 6.1 5.2 6.8
18.6 b 14.2 16.0 17.9 b
1.3 2.1 2.8 1.9
12.8 10.3 b 10.2 b 9.8 b
3.9 4.0 3.0 3.1
3.0 3.7 2.8 2.6
---
930 b 860 1 130 b 680
0.36 0.52 0.42 0.34
14.8
43.6
--
15.2
3.6
15.0
4.3
3.5
-
680
0.30
1 200 1 060 960 1 130
10.0 b 9.3 b 15.0 b 4.1
9.8
Motor neuron diseases: ALS ALS
(inherited) Spinal muscular atr. Sy. amyotrophicum ALS
Peripheral neuropathy and polyneuropathy: Polyneuropathy 18.2 b 37.1 Charcot-Marie-Tooth Brachial neuropathy
16.3 19.2 b 20.1 b
38.0 40.2 42.0
6.1 7.5 b 6.8 2.2
15.2 16.3 13.0 15.8
2.1 1.2 1.1 2.1
13.9 14.1 13.0 12.1
3.4 4.3 4.1 2.7
4.0 2.3 2.6 3.0
-----
Miscellaneous diseases: Epilepsy grand real Contusio cerebri (5 yrs) Purpura (13 months old) Pediatric pooled CSFs
8.5 11.4 9.0 11.5
47.8 b 42.6 36.8 37.5
2.2 5.0 4.2 5.2
16.6 12.0 b 14.3 13.9
3.8 _ 3.7 2.9
14.4 19.2 b 24.0 b 20.0 b
4.1 3.8 3.0 4.0
2.6 6.0 b 5.0 b 5.0 b
_ _ _
b b b b
b b b b
0.38 0.34 0.36 0.45
720 240 b 960 b 660
0.41 0.21 0.26 0.22
" V e n t r i c u l a r C S F (hydrocephalus). b Out fo control range.
cerebrovascular insult with good outcome, peripheral neuropathies and polyneuropathies, and in intravertebral discopathies cannot be explained. At this time, it would be warranted to relate elevated CSF gangliosides to regenerating factors (Gorio et al., 1980; Toffano et al., 1983; Cuello et al., 1989) or to the presence of gangliosides antibodies seen in many types of peripheral neuropathies. Further, we are inclined to believe that the finding of a decreased total content of gangliosides in CVI with lethal outcome results from a severe and irreversible brain damage and a decreased ganglioside metabolism. It should also be considered, however, that the development of edema in CVI and increase in
intracranial pressure may accelerate the bulk-flow mechanism in arachnoid villi leading to faster ganglioside elimination from CSF. Nevertheless, non-cerebral complications of the underlying disease are quite frequently present in CVI, which may account for great differences in the total CSF content of gangliosides in CVI with poor outcome. Analyzing 100 ml of pooled CSF, Nagai et al. (1973) have pointed to an increased content of gangliosides in multiple sclerosis. These report were not confirmed by our results obtained on individual CSF samples. However, in one patient with multiple sclerosis two lumbar punctures were performed, i.e. in the active stage of the disease and in the stage of remission (time interval
199
3 months). In the active stage of the disease, higher values of gangliosides in CSF were recorded (970/~g/1), and in the stage of remission it was 670/~g/l. Accordingly, the amount of gangliosides in CSF in multiple sclerosis may correlate with the dynamic of disease. Our results on CSF gangliosides in control individuals (n = 10) and different neurologic patients (n = 71) demonstrated that CSF gangliosides reflect some brain diseases. Acknowledgement This work was supported by Scientific Fund of Croatia (No. 1.08.03.05.39.). The authors wish to thank Mrs Radojka Barbaric for skilful technical assistance and Antonija Redovnikovic, B.A., for editorial work.
References Ando, S., N. Chang and R.K. Yu (1978) High peformance thin-layer chromatography and densitometric determination of brain ganglioside composition of several species. Anal. Biochem., 89: 437450. Cuello, A.C., L. Garafalo, R.L. Kenigsberg and D. Maysinger (1989) Ganglioside potentiate in vivo and in vitro effects of nerve growth factor on central cholinergic neurons. Proc. Natl. Acad. Sci., 86: 2056-2060. Davidsson, P., P. Fredman and L. Svennerholm (1989) Gangliosides and sulphatides in human cerebrospinal fluid: quantitation with immunoaffinity techniques. J. Chromatogr., 496: 279-289. Gorio, A., G. Carmignato, L. Facci and M. Finesso (1980) Motor nerve sprouting induced by gangliosides treatment. Possible implication for gangliosides on neuronal growth. Brain Res., 297: 236-241. Hakomori, S.-I. (1981) Glycosphingolipids in cellular interaction, differentiation and oncogenesis. Annu. Rev. Biochem., 50: 733764. Hirabayashi, Y., K. Koketsu, H. Higashi, Y. Suzuki, M. Matsumoto, M. Sugimoto and T. Ogawa (1986) Sensitive enzyme-immunostaining and densitometric determination of ganglioseries gangliosides on thin-layer plate: pmol detection of gangliosides in cerebrospinal fluid. Biochim. Biophys. Acta, 876: 178-182. Kjellin, K.G. and C.E. S6derstrfim (1974) Diagnostic significance of CSF spectrometry in cerebrovascular diseases. J. Neurol, Sci., 23: 359-369. Kotagal, S., D.A. Wenger, H. Alcala, C. Gomez and S. Horesstein (1986) AB variant GM2 gangliosidosis; cerebrospinal fluid and neuropathological characteristics. Neurology, 36: 438-440. Kracun, I., H. Rosner and C. Cosovic (1986) In: (Eds.), Tettamanti, G., Ledeen, R.W., Saudhoff, K., Nagai, Y., Toffano, G. Gangliosides and Neuronal Plasticity, Fidia Research Series, Liviana Press, Springer-Verlag, Padova-Berlin-Heidelberg-New YorkTokyo, pp. 67-76. Kundu, S.K., I. Diego, S. Osovitz and D.M. Marcus (1985) Gangliosides of human plasma. Arch. Biochem. Biophys., 238: 388-4130. Ledeen, R.W. (1978) Ganglioside structures and distribution: are they localised at the nerve ending? J. Supramol. Struct., 8: 1-17. Ledeen, R.W. and R.K. Yu (1972) Gangliosides of CSF and plasma:
Their relation to the nervous system. Adv. Exp. Med., and Biol., 19: 77-93. Markovic, I., M. Heifer, J. Rogan-Grgas, T. Vladovic-Relja, G. Kracun, C. Cosovic and I. Kracun, 1988) Gangliosides composition of different human brain gliomas. Period. Biol., 90: 97-104. Nagai, Y. and J.N. Kanfer (1971) Composition of human cerebrospinal fluid cerebrosides. J. Lipid Res, 12: 143-148. Nagai, Y., J.N. Kanfer and W.W. Tourtellotte (1973) Preliminary observations of gangliosides of normal and multiple sclerosis cerebrospinal fluid. Neurology, 23: 945-948. Pedersen, H.E. (1973a) Cerebrospinal fluid cholesterol. Quantitative analysis of unesterified cholesterol and normal values. Acta Neurol. Scand., 49: 626-638. Pedersen, H.E. (1973b) Cerebrospinal fluid phospholipids. Quantitative determination of fractionated phospholipids and normal values in the cerebrospinal fluid. Acta Neurol. Scand., 49: 639-648. Pilz, H. (1970) Die Lipide des normalen und pathologischen Liquor cerebrospinalis. Springer-Verlag, Berlin-Heidelberg-New York. Pukel, C.S., C.S. Lloyd, K.O. Lloyd, L.R. Travassos and W.G. Dippold (1982) GD3, a prominent ganglioside of human melanoma. J. Exp. Med., 155: 1133-1147. Pullakart, R.K., H. Reha and N.G. Beratis (1981) Accumulation of GM2 ganglioside in cerebrospinal fluid of a patient with the variant AB of infantile GM2 gangliosides. Pediatrics, 68: 106-108. Reiber, H. (1980) Discrimination between different blood - CSF barrier dysfunctions and inlammatory reactions of the CNS by a recent evaluation graph for the protein profile of cerebrospinal fluid. J. Neurol,, 224: 89-99. Rieder, H.P. (1966) Eine neue Modifikation zur Methode der Eiweissbestimmung Mittels Cu-(II) und phenolreagens nach FolinCiocalteu. Klin. Wschr., 44: 1036-1039. Sandhoff, K. and E. Conzelmann (1984) The biochemical basis of gangliosidosis. Neuropediatrics, 15: 85-92. Seyfried, T.N., D.J. Bernard and R.K. Yu (1984) Cellular distribution of gangliosides in the developing mouse cerebellum: analysis using staggerer mutant. J. Neurochem. 43: 1152-1162. Svennerholm, L. (1957) Quantitative estimation of sialic acid. Biochim. Biophys. Acta, 24: 604-611. Svennerholm, L. (1963) Chromatographic separation of human brain gangliosides. J. Neurochem., 10, 613-623. Tichy, J. (1972) Vergleichende Untersuchungen der Lipide im Blut and Liquor cerebrospinalis. Zschr. Inn. Med., 27: 21-25. Toffano, G., G. Savoini, F. Moroni, G. Lombardi, L. Calza and L.F. Agnati (1983) GM1 ganglioside stimulates the regeneration of dopaminergic neurons in the central nervous system. Brain Res., 261: 163-166. Traylor, T.D. and E.L. Hogan (1980) Gangliosides of human cerebral actrocytomas. J. Neurochem., 34: 126-131. Trbojevic-Cepe, M., Z. Poljakovic, N. Vrkic and I. Bielen (1989) Detection of oligoclonal bands in unconcentrated CSF by isoelectric focusing in ultrathin polyacrylamide gel, direct antiserum immunofixation and silver nitrate staining. J. Clin. Chem. Clin. Biochem., 27: 211-216. Trbojevic-Cepe, M. and I. Kracun (1990) Determination of gangliosides in human cerebrospinal fluid by high-performance thin-layer chromatography and direct densitometry. J. Clin. Chem. Clin. Biochem., 28: 863-872. Wiegandt, H. (1967) The subcellular localization of gangliosides in the brain. J. Neurochem., 14: 671-674.