Ganglioside expression in tissues of mice lacking β₂-microglobulin

Abstract

This study presents a comparative analysis of gangliosides from lymphoid (spleen and thymus) and other (brain, liver, lungs and muscle) tissues of C57BL/6 mice lacking the gene for β₂-microglobulin (β₂M), a constitutive component of the MHC class I molecule. Ganglioside fractions in the tissues of mice homozygous (β₂M−/−) and heterozygous (β₂M−/+) for the gene deletion were determined by high performance thin-layer chromatography (HPTLC), followed by immunostaining with specific polyclonal antibodies. Ubiquitous gangliosides G_m3(Neu5Ac) and G_m3(Neu5Gc) were the dominant gangliosides in the lungs of the control β₂M−/+ mice, whereas the homozygous knockout mice had substantially decreased expression of these structures. The lungs of the β₂M−/− mice also had reduced expression of T-lymphocyte-specific G_m1b-type gangliosides (G_m1b and GalNAc-G_m1b). β₂M-deficient mice also had more G_m1a and G_d1a gangliosides in the liver, and several neolacto-series gangliosides were increased in the brain and lungs. This study provides in vivo evidence that the β₂M molecule can influence the acquisition of a distinct ganglioside assembly in different mouse organs, implicating its non-immunological functions.

INTRODUCTION

β₂-microglobulin (β₂M) is a polypeptide that associates with polymorphic MHC class I proteins encoded by H-2K, H-2D and Qa/Tla loci in the mouse, or HLA-A,-B or-C loci in the human [1]. β₂M is necessary for the function of MHC class I proteins in the presentation of foreign antigens to the immune system [2]. Although class I MHC is primarily known for its function in cell-mediated immune recognition, recent findings suggest that class I MHC molecules may play a role in the re-modelling and plasticity of connections in the developing and mature mammalian central nervous system [3]. Mice with the β₂M gene disrupted by targeted mutation do not express β₂M protein and therefore, lack class I MHC antigen on the surface of most cells [4,5]. Mouse strains with this mutation have no obvious defects and breed normally [4,5]. They possess almost no CD8⁺ cytotoxic lymphocytes in the thymus, but their peripheral lymphoid organs can clear certain virus infections [6].

Gangliosides are acidic glycosphingolipids (GSLs) characterized by the presence of one or more sialic acid units in the oligosaccharide chain. The parent compounds are Neu5Ac and Neu5Gc, which play crucial roles in various biological functions [7]. GSLs are ubiquitous, highly conserved membrane components with an important biological role in cell surface recognition [8] and modulation of the function of a variety of membrane-associated proteins [9]. GSLs are assembled as ‘rafts’[10] or ‘glycosignalling domains’[11] in the outer leaflet of the plasma membrane, and these clustered GSLs, but not non-clustered GSLs, exert biological activities. Clusters of liquid-ordered GSL and cholesterol molecules in the exoplasmic leaflet of the T-cell plasma membrane constitute microdomains, which play an essential role in T-cell receptor signalling [12]. Surface cross-linking of many glycosylphosphatidylinositol-anchored proteins and G_m1 ganglioside leads to cell stimulation following the coalescence of microdomains and activation of the associated Src family protein tyrosine kinases [12]. There are a wealth of data on the specific roles of gangliosides in the regulation of the immune response. The examples of such effects are the inhibition of lymphoproliferation [13] and modulation of CD4 by helper T lymphocytes [14], suppression of cytotoxic activity of natural killer cells [14] and inhibition of monocyte accessory function [15].

GSLs may also interact with MHC gene products. MHC class I and II gene expression in astrocytes can be specifically suppressed by a mixture of brain gangliosides [16]. In addition, brain ganglioside G_T1b has been shown to modulate MHC class I gene expression in astrocytes [17]. As a first step in elucidating the interactions between MHC I and GSLs, we analysed ganglioside expression patterns in different organs from mice deficient in the β₂M gene compared with control mice heterozygous for the gene knockout. Specific polyclonal antibodies were used to detect three groups of gangliosides on high-performance thin-layer chromatograms [18]: G_M3, a ubiquitous ganglioside; lymphocyte-specific G_M1b-type gangliosides (G_M1b, GalNAc-G_M1b), G_M1a and G_D1a, and neolacto-series gangliosides, a newly-characterized ganglioside group.

MATERIALS AND METHODS

Animals

Mice lacking the gene for β₂M were originally generated by Zijlstra et al[4]. They were bred onto a C57BL/6 background and kept under standard housing conditions (laboratory rodent chow and water ad libitum, and a 12h light-dark cycle) at the Animal Facility of the Rijeka University School of Medicine, Rijeka, Croatia. Mice used in the study were either homozygous (β₂M–/–) or heterozygous (β₂M–/+) for the β₂M gene knockout. They came from the same litters, and the β₂M–/– phenotype was checked by the absence of CD8 antigen using flow cytometry. Lymphocytes from heterozygous β₂M–/+ mice had similar expression of MHC class I molecules and immunological characteristics as the lymphocytes from the wild-type (C57BL/6) mice [5].

For ganglioside analysis, females aged 10 weeks were killed by cervical dislocation under CO₂ anaesthesia; the tissues were dissected out and stored at – 20°C until GSL extraction.

Isolation of GSLs from tissues

Brain, liver, lungs, thigh muscle, spleen and thymus were dissected from 10 animals of each group. Animals were perfused with 20 ml cold 0·1 m phosphate-buffered saline, pH 7·4, immediately after CO₂ sacrification. Identical tissues were pooled, minced with a scalpel, suspended in distilled water in a 1/2 ratio (w/v), homogenized for 10 min with a dispersing tool (Polytron PT1200C, Kinematica AG, Littau/Luzern, Switzerland) and isolated according to standard procedures [19]. Briefly, GSLs were extracted with chloroform/methanol (2/1, v/v), chloroform/methanol (1/1, v/v) and chloroform/methanol (1/2, v/v) (10-fold volumes of the tissue wet weight) for 30 min with sonication. The combined extracts were evaporated and resuspended in chloroform/methanol/water (30/60/8, by volume); gangliosides were isolated by anion exchange chromatography on DEAE-Sepharose CL-6B (Pharmacia, Freiburg, Germany) [20]. The ganglioside fraction was incubated for 1 h at 37°C in aqueous 1 N NaOH to saponify phospholipids, followed by neutralization with acetic acid and dialysis. Gangliosides were further purified by adsorption chromatography on Iatrobeads 6RS-8060 (Macherey-Nagel, Düren, Germany) [21]. After column chromatography purification, the final ganglioside fractions were adjusted to defined volumes of chloroform/methanol (2/1) corresponding to 4 mg wet tissue weight/μl for the liver and 2 mg wet tissue weight/µl for all other tissues.

Thin-layer chromatography and reference gangliosides

The gangliosides were separated on silica gel 60 pre-coated high-performance thin-layer chromatography plates (HPTLC-plates, size 10 × 10 cm, thickness 0·2 mm, Merck Darmstadt, Germany) using chloroform/methanol/water (120/85/20, each by volume) with 2 mm CaCl₂, and visualized with resorcinol [22]. A ganglioside mixture composed of G_M3(Neu5Ac), IV³Neu5Ac-nLcOse₄Cer (IV³nLc4), IV⁶Neu5Ac-nLcOse₄Cer (IV⁶nLc4) and VI³Neu5Ac-nLcOse₆Cer (VI³nLc6) was isolated from human granulocytes as previously described [23]. G_M3(Neu5Gc) was purified from a mouse–mouse hybridoma [24]. Gangliosides from a murine T lymphoma YAC-1 and a lymphoreticular tumour cell line MDAY-D2 were isolated according to established procedures [25,26] (see Table 1). Human brain gangliosides were purchased from Supelco Inc. (Bellefonte, PA, USA).

Table 1.

Relative quantities of resorcinol-stained ganglioside fractions in the lungs of mice homozygous (−/−) or heterozygous (+/−) for the β₂MG gene knockout

	Percent total density*
Ganglioside fraction	−/− Mice	+/− Mice	Homo-hetero density ratio †
GM3(1)	12	20	0·4
GM3(2)	18	25	0·5
GM1	5	8	0·4
GD1a	25	9	1·7
GT1b	20	22	0·6
Not identified	20	16	0·8

^*Density of individual bands was expressed as the percentage of the total density for all bands.

^†Ratio of densitometric peaks for individual gangliosides fraction from homozygous (−/−) and heterozygous (+/−) mice.

Antibodies and cholera toxin B subunit (choleragenoid)

Polyclonal chicken anti-G_M3(Neu5Ac) and anti-G_M3(Neu5Gc) antibodies have been characterized in our previous reports [24,27]. Rabbit anti-GgOse₄Cer antiserum used for the detection of G_M1b and G_D1α has also been described in earlier publications [28,29]. Highly specific chicken anti-GalNAc-G_M1b antibody has been described in detail elsewhere [25,26], as well as chicken anti-nLcOse₄Cer antibody used for the analysis of neolacto-series gangliosides [27,30,31]. Cholera toxin B subunit (= choleragenoid) specific for ganglioside G_M1 was from Sigma (Deisenhofen, Germany) and goat anti-choleragenoid antiserum from Calbiochem (Frankfurt, Germany).

TLC immunostaining

The overlay technique was carried out according to Magnani et al[32], with some modifications [18,33]. All GSL-specific antibodies were used at 1:1000 dilutions. Secondary rabbit anti-chicken IgG, goat anti-rabbit IgG and rabbit anti-goat IgG antisera, all affinity chromatography-purified and labelled with alkaline phosphatase (0·6 mg/ml), were purchased from Dianova (Hamburg, Germany) and used at 1:2000 dilution. Briefly, the silica gel was fixed with polyisobutylmethacrylate, and unspecific protein binding was blocked by a 15 min incubation of the plate with solution A (phosphate-buffered saline, PBS, supplemented with 1% bovine serum albumin and 0·02% NaN₃). The plates were then overlaid for 1 h with anti-GSL antibodies. Unbound antibodies were removed by washing each plate five times with solution B (0·05% Tween 21, 0·02% NaN₃ in PBS). After 1 h of incubation with secondary antibodies diluted 1:2000 in solution A, the plates were washed again, followed by two rinses with glycine buffer (0·1 m glycine, 1 mm ZnCl₂, 1 mm MgCl₂, pH 10·4), to remove phosphate. Bound antibodies were visualized with 0·05% (w/v) 5-bromo-4-chloro-3-indolylphosphate (Biomol, Hamburg, Germany) in glycine buffer. Each antibody analysis was performed three times, with identical results.

TLC immunostaining of neolacto-series gangliosides

Neuraminidase treatment of neolacto-series gangliosides with (α2–3)-substituted sialic acid is necessary prior to immunostaining with anti-nLcOse₄Cer antibody, whereas (α2–6)-sialylated neolacto-type gangliosides can be detected without enzyme treatment because sialylation at position 6 of the terminal galactose does not hinder recognition [30]. Fixed silica gel plates were incubated with 2·5 mU/ml Vibrio cholerae neuraminidase (Behring Werke AG, Marburg, Germany) for 2 h at 37°C in the buffer described above, followed by immunostaining with anti-nLcOse₄Cer antibody [30]. Neolacto-series gangliosides from human granulocytes, composed of G_M3(Neu5Ac), IV³nLc4, IV⁶nLc4 and VI³nLc6, served as references [23].

TLC immunostaining of G_M1b-type gangliosides

Terminally-sialylated gangliosides with a GgOse₄Cer backbone (G_M1b, G_D1α) were detected as described previously [26,28]. After chromatography of gangliosides and silica gel fixation, the plate was incubated for 2 h at 37°C with 2·5 mU/ml V. cholerae neuraminidase (Behring Werke AG) in 0·05 m sodium acetate, 9 mm CaCl₂, pH 5·5. Desialylated gangliosides were immunostained with polyclonal anti-GgOse₄Cer antibody.

TLC detection of G_M1a-type gangliosides

The TLC binding assay using choleragenoid for specific detection of G_M1 was developed by Magnani et al[34] and was used with modifications [29]. A fixed silica gel plate was overlaid with 250 ng/ml choleragenoid (Sigma) diluted in solution A for 2 h at room temperature, followed by goat anti-choleragenoid (Calbiochem) and secondary rabbit anti-goat IgG antibody incubation (both diluted 1:2000). Bound antibodies were visualized as described above. To reveal the presence of G_D1a, G_D1b and G_T1b, the plates were incubated with 5 mU/ml V. cholerae neuraminidase (2 h, 37°C) prior to combined choleragenoid immunostaining [35].

Semi-quantitative estimation of ganglioside fractions

For semi-quantitative estimation, resorcinol and immunostained chromatograms were scanned with a CD60 scanner (Desaga, Heidelberg, Germany) equipped with an IBM compatible personal computer and densitometric software. Ganglioside amounts corresponding to identical wet weight quantities of each tissue were applied, and bands were measured in reflectance mode at 580 nm (resorcinol) and 630 nm (indolylphosphate) with a light beam slit of 0·1 × 2 mm. These scanning parameters give a linear response in the range 0·15–3·0 µg sialic acid per band [36] and have proved to be convenient for a reproducible quantification of picomolar amounts of individual ganglioside species [37,38].

RESULTS

Expression of G_M3

Resorcinol-stained thin-layer chromatograms of gangliosides from the brain and lungs of β₂M–/– and control β₂M+/– mice showed twofold lower expression of G_M3 (1) and (2) fractions in the lungs of control β₂M+/– mice compared with those in β₂M–/– mice (Fig. 1a and Table 1). Mice heterozygous (+/–) for β₂M gene knockout were chosen as controls because they came from the same litters as homozygous mice and had similar immunological characteristics [5] they also had similar expression of gangliosides on resorcinol-stained thin-layer chromatograms (data not shown) as wild-type (C57BL/6) mice.

Fig. 1.

Resorcinol stain (a) and anti-G_M3(Neu5Gc) TLC immunostain (b) of ganglioside fractions from the brain and lungs of β₂M−/− and control β₂M+/− mice. (a) Shows chromatograms of ganglioside amounts corresponding to 10 mg brain (Br) and 60 mg lungs wet weight (Lu), whereas for TLC immunostaining (b), one third of ganglioside quantities were applied in the same order. In both panels, 3 µg G_M3(Neu5Gc) (lane S₁) and 10 µg human granulocyte gangliosides (lane S₂) were used as references.

The separation of individual murine gangliosides on HPTLC plates as double or triple bands is a common feature, due to the variation in the ceramide portion (C₁₆-or C₂₄-fatty acids) and oligosaccharide moiety (Neu5Ac and Neu5Gc) [24]. The upper band in our samples was composed of G_M3(Neu5Ac) with a C₂₄-fatty acid, and the lower band was G_M3(Neu5Ac) with a C₁₆-fatty acid and G_M3(Neu5Gc) with a C₂₄-fatty acid. In immuno-overlay assays, which can detect as little as 0·5 pmol (0·64–1·0 ng) of G_M3[33], the expression of G_M3(Neu5Gc) was similar in organs from homozygous and control heterozygous mice (data not shown), except for the lungs (Fig. 1b). Ganglioside extracts from the lungs of β₂M–/– mice did not bind anti-G_M3(Neu5Gc) antibody, whereas control β₂M+/– lung extracts had a band of C₂₄-fatty acid-substituted G_M3(Neu5Gc). Human granulocyte gangliosides (Fig. 1b, lane S₂) did not bind this antibody, indicating the high specificity of the anti-G_M3(Neu5Gc) antibody, which does not react with G_M3(Neu5Ac) or any ganglio-or neolacto-series gangliosides [24]. Staining with anti-G_M3(Neu5Ac) antibody showed a similar expression pattern, i.e. G_M3(Neu5Ac) was abundant in the lungs of β₂M+/– mice compared with the trace quantities found in the lungs of homozygotes (data not shown).

Expression of neolacto-series gangliosides

The anti-nLc4 antibody detects terminally (α2–3)-sialylated neolacto-series gangliosides (IV³nLc4 and VI³nLc6) after V. cholerae neuraminidase treatment, whereas visualization of IV⁶nLc4 is feasible without enzyme treatment [30]. The most prominent difference in the expression of these gangliosides between the two groups of mice was found in the brains and lungs. Brains from β₂M–/– mice showed an almost eightfold greater expression of the IV⁶nLc4(C₂₄) fraction compared with control β₂M+/– mice (Fig. 2 and Table 2).

Fig. 2.

Detection of neolacto-series gangliosides in the brain, liver, lungs, muscle and spleen of β₂M−/− and control β₂M+/− mice. Ganglioside amounts corresponding to 3·3 mg brain (Br), 13·3 mg liver (Li), 20 mg lungs and muscle (Lu and Mu, respectively) and 13·3 mg spleen wet weight (Sp) were chromatographed together with 2·5 µg human granulocyte gangliosides (lane S). The TLC immunostain was performed with anti-nLcOse₄Cer antibody after V. cholerae neuraminidase treatment.

Table 2.

Relative quantities of immunostained nLc-type ganglioside TLC fractions in the brain and lungs of mice homozygous (−/−) or heterozygous (+/−) for the β₂MG gene knockout

	Brain			Lungs
	Percent total density*		Percent total density*
Ganglioside fraction	−/− Mice	+/− Mice	Homo-hetero density ratio	−/− Mice	+/− Mice	Homo-hetero density ratio†
GM3(1)	10	–	–	14	27	0·5
GM3(2)	15	19	1·2	20	30	0·7
GM3(3)	4	–	–	4	7	0·5
IV3Neu5Ac-nLc4(4)	–	8	–	13	6	2·3
IV3Neu5Ac-nLc4(5)	11	10	1·1	11	8	1·3
IV6Neu5Ac-nLc4(6)	25	7	7·7	20	16	1·3
IV6Neu5Ac-nLc4(7)	–	5	–	17	6	3·0
VI3Neu5Ac-nLc6(8)	10	20	1·1	1	–	–
VI3Neu5Ac-nLc6(9)	6	–	1·1	–	–	–
Not identified	19	31	–	–	–	–

^*Density of individual bands was expressed as the percentage of the total density for all bands.

^†Ratio of the densitometric peak for individual ganglioside fraction from homozygous (−/−) and heterozygous (+/−) mice.

The lungs of β₂M–/– mice strongly expressed double bands of (α2–3)-sialylated IV³nLc4 and (α2–6)-sialylated (IV⁶nLc4) neolacto-core gangliosides, compared with the amounts of these gangliosides in the lungs of the control mice (Fig. 2 and Table 2).

The anti-nLc4Cer antibody recognizes a neolactodiglycosyl residue (Galβ1–4GlcNAc) but can cross-react with the LacCer (Galβ1–4Glc residue). Thus, the anti-nLc4 overlay assay after V. cholerae neuraminidase pre-treatment, which cleaves neuraminic acid and leaves lactosylceramide accessible for antibody binding, also detects G_M3 ganglioside. This staining confirmed the reduced expression of G_M3 in the lungs of β₂M–/– mice compared with control β₂M+/– mice (Fig. 2).

Expression of G_M1b, G_D1α and GalNAc-G_M1b

Terminally-sialylated ganglio-series gangliosides G_M1b and G_D1α have been reported to be typical gangliosides of mouse leucocytes [20,39]. GalNAc-G_M1b, the elongation product of the G_M1b ganglioside, has been reported to be a marker of mature and/or stimulated T lymphocytes [40]. G_M1b and G_D1α are detectable on HPTLC plates with an anti-GgOse₄Cer antibody after V. cholerae neuraminidase treatment [25]. This combined treatment is highly specific and allows sensitive detection down to 10 ng of GgOse₄Cer core gangliosides in complex GSL mixtures [33]. Ganglioside fractions from murine lymphoma YAC-1 and lymphoreticular MDAY-D2 cell lines, known to express high levels of G_M1b-type gangliosides [25,26], were used as standards and their major gangliosides are listed in Table 3.

Table 3.

Major gangliosides from murine lymphoma YAC-1 and lymphoreticular MDAY-D2 cells

Ganglioside fraction	Fatty acid	Symbol*	Sialic acid	YAC-1	MDAY-D2
-II	24:0,24:1	GM2	Neu5Ac	−	+
-I	16:0	GM2	Neu5Ac	−	+
0	24:0,24:1	GM1a	Neu5Ac	−	+
I	16:0	GM1a	Neu5Ac	−	+
	24:0,24:1	*GM1b	Neu5Ac	+	+
II	16:0	*GM1b	Neu5Ac	+	+
	24:0,24:1	GalNAc-GM1b	Neu5Ac	+	+
	24:0,24:1	*GM1b	Neu5Gc	+	−
III	16:0	*GM1b	Neu5Gc	+	−
	16:0	GalNAc-GM1b	Neu5Ac	+	+
	24:0,24:1	GalNAc-GM1b	Neu5Gc	+	−
IV	16:0	GalNAc-GM1b	Neu5Gc	+	−
V	24:0,24:1	GD1a	Neu5Ac	−	+
VI	16:0	GD1a	Neu5Ac	−	+
VII	24:0,24:1	*GD1α	Neu5Ac	−	+
VIII	16:0	*GD1α	Neu5Ac	−	+

^*G_M1b-and G_D1α-gangliosides detectable by immunostaining with anti-GgOse₄Cer antibody after V. cholerae neuraminidase treatment are marked with asterisks; structural data drawn from Müthinget al[24,25].

MDAY-D2 cells express Neu5Ac-substituted-and lack Neu5Gc-containing gangliosides [26], whereas YAC-1 cells are characterized by high expression of Neu5Gc-substituted G_M1b-type gangliosides [25]. HPTLC immuno-overlay analysis of G_M1b and G_D1α showed a sevenfold lower expression of G_M1b in the lungs of β₂M–/– mice (Fig. 3,Table 4) separating at positions I (= G_M1b(Neu5Ac) with C₂₄-fatty acid) and II (=G_M1b(Neu5Ac) with C₁₆-fatty acid) of YAC-1 and MDAY-D2 reference gangliosides (see Table 3), indicating down-regulation of terminally-sialylated G_M1b ganglioside in the lungs of β₂M−/– mice. Both groups of mice showed similar expression of G_M1b in the spleen and thymus (Fig. 3). The same was true for the brain (data not shown), where G_M1b is known to be a minor constituent [29].

Fig. 3.

Resorcinol stain (a) and TLC immunostaining (b) of G_M1b-type gangliosides from the lungs, spleen and thymus of β₂M−/− and β₂M+/− mice. In (a), ganglioside amounts corresponding to 60 mg (lungs, Lu) and 40 mg wet weight (spleen and thymus, Sp and Th, respectively) were chromatographed together with 10 µg YAC-1 (lane S₁) and 10 µg MDAY-D2 gangliosides (lane S₂). In (b), one third of ganglioside quantities were applied in the same order. YAC-1 and MDAY-D2 gangliosides are marked with roman numerals from II to VIII and their structures are given in Table 3. Crosses indicate impurities that did not stain with resorcinol.

Table 4.

Relative quantities of immunostained G_M1b-type ganglioside TLC fractions in the lungs of mice homozygous (−/−) or heterozygous (+/−) for the β₂MG gene knockout

	Percent total density*
Ganglioside fraction	−/− Mice	+/− Mice	Hetero-homo density ratio†
GM1b (I)	46	39	7·1
GM1b (II)	54	44	6·8
Not identified	–	17	–

^*Density of individual bands was expressed as the percentage of the total density for all bands.

^†Ratio of the densitometric peak for individual ganglioside fraction from heterozygous (+/−) and homozygous (−/−) mice.

Anti-GalNAc-G_M1b antibody does not bind to other ganglio-or neolacto-series gangliosides [26], and it does not discriminate between Neu5Ac-and Neu5Gc-substituted GalNAc-G_M1b. In the HPTLC immuno-overlay assay of ganglioside extracts from the lungs, spleen and thymus of β₂M–/– mice, this antibody stained only a single band of GalNAc-G_M1b in the spleen fraction (Fig. 4) at position III (=GalNAc-G_M1b(Neu5Ac) with C₁₆-fatty acid and/or GalNAc-G_M1b(Neu5Gc) with C₂₄-fatty acid) of reference YAC-1 and MDAY-D2 gangliosides (see Table 3). Two positive GalNAc-G_M1b bands were detected in the lungs of the control β₂M+/– mice at positions II and III (Fig. 4), most likely GalNAc-G_M1b(Neu5Ac) with C₂₄-or C₁₆-fatty acid, respectively. These bands were absent in the lunggangliosides of β₂M–/– mice (Fig. 4). This indicated that GalNAc-G_M1b expression, like that of its precursor G_M1b, was down-regulated in the lungs of β₂M–/– mice. The brain, liver and muscle of both types of mice failed to express GalNAc-G_M1b (data not shown).

Fig. 4.

Anti-GalNAc-G_M1b TLC immunostaining of ganglioside fractions from the lungs, spleen and thymus of β₂M−/− and control β₂M+/− mice. Ganglioside amounts corresponding to 20 mg lungs (Lu) and 13·3 mg spleen and thymus wet weight (Sp and Th, respectively) were chromatographed together with 3 µg YAC-1 (lane S₁) and 10 µg MDAY-D2 gangliosides (lane S₂). YAC-1 and MDAY-D2 gangliosides are marked with roman numerals according to Table 3.

Expression of G_M1a-type gangliosides

The expression patterns of G_M1a-type gangliosides in the tissues of mice homo-or heterozygous for the β₂M gene knockout were very similar (data not shown), except for the liver tissue (Fig. 5). Resorcinol staining of liver gangliosides from β₂M+/– mice, which have been bred onto the C57BL/6 background, confirmed previous reports that G_M1 and G_D1a gangliosides cannot be detected by resorcinol staining of ganglioside extracts from the liver of C57BL/6 mice [41]. However, traces of G_M1 and G_D1a gangliosides were clearly visible after resorcinol staining of ganglioside extracts from the liver of β₂M–/– mice (Fig. 5a). Immunostaining with cholera toxin, which detects down to 0·01 ng (6·5 fmol) G_M1[33], showed that G_M1 was expressed in the liver of both β₂M–/– and β₂M+/– mice, but that the expression in β₂M–/– mice was fourfold greater. Immunostaining with cholera toxin after V. cholerae neuraminidase, which also detects G_D1a, G_D1b and G_T1b gangliosides, showed that the expression of G_D1a was also four times stronger in the liver of β₂M–/– than β₂M+/– mice (Fig. 5c and Table 5).

Fig. 5.

Resorcinol stain (a), TLC immunostain with cholera toxin B subunit (B) and TLC immunostain with cholera toxin B subunit after V. cholerae neuraminidase treatment (c) of ganglioside fractions from the liver of β₂M−/− and control β₂M+/− mice. (a) Shows chromatograms of ganglioside amounts corresponding to 40 mg whereas for TLC immunostaining (b and c), ganglioside amounts corresponding to 1·3 mg liver wet weight were applied in the same order. As a reference, 10 µg human brain gangliosides (lane S) were applied at panel a and 250 ng in panels (b) and (c).

Table 5.

Relative quantities of ganglioside TLC fractions stained with cholera toxin after neuraminidase treatment in the liver of mice homozy-gous (−/−) or heterozygous (+/−) for the β₂MG gene knockout

	Percent total density*
Ganglioside fraction	−/− Mice	+/− Mice	Homo-hetero density ratio†
GM1	38	32	3·9
GD1a	37	29	4·2
GD1b	13	22	1·4
GT1b	12	17	1·9

^*Density of individual bands was expressed as the percentage of the total density for all bands.

^†Ratio of the densitometric peak for individual ganglioside fraction from homozygous (−/−) and heterozygous (+/−) mice.

DISCUSSION

Immunochemical analysis of ganglioside expression in the tissues of β₂M knockout mice showed that the β₂M gene may be involved in the acquisition of distinct ganglioside assemblies in different mouse organs. GSLs have been implicated in immune cell circulation and/or localization of cells in lymphoid organs [9]. Our data indicate that the interaction between GSLs and immune cells is reciprocal and that the expression of functional class I MHC molecules on the surface of the cells may regulate the expression of GSLs.

Only traces of G_M1b in the lungs and G_D1α in the spleen and thymus were detected in β₂M–/– mice compared with abundant expression in the control β₂M+/– mice. This indicated a strong down-regulation of terminally-sialylated gangliosides with a GgOse₄Cer backbone in the absence of β₂M. G_M1b is a typical ganglioside of murine T cells [20,42,43] and macrophages [39], and G_M1b expression in the lungs may be related to the presence of T cells and macrophages in this organ [44]. GalNAc-G_M1b, the elongation product of G_M1b, was found in considerable quantities in the spleen and lungs, and in trace quantities in the thymus of β₂M+/– mice. This correlates with previous findings by us and others that GalNAc-G_M1b is a marker of activated T cells and macrophages [40,43,45]. Similar to G_M1b, GalNAc-G_M1b expression was considerably reduced in the lungs and thymus of the β₂M–/– mice, indicating that β₂M is indeed involved in the acquisition of this ganglioside by lymphoid cells.

High amounts of terminally-sialylated nLcOse₄Cer were detected in both β₂M–/– and control β₂M+/– mice, particularly in the brain, lungs and spleen, and to a lesser extent in the muscle and liver. However, very little is known about the cell type-specific expression of neolacto-series gangliosides in murine organs. Increased expression of ganglioside IV⁶nLc4(C₂₄) in the brain of β₂M–/– compared with the control β₂M+/– mice may be related to the role of class I MHC molecules in the central nervous system [3]. It has been shown that class I MHC molecules are expressed by specific sets of neurones that undergo activity-dependent changes [46], and that mice lacking either β₂M or CD3ζ exhibit abnormalities in connections between these neurones [3]. Neolacto-series gangliosides are well characterized in acoustic neurinoma samples [47] and the peripheral nervous system [48] in humans, but additional biochemical and morphological studies are needed to address the possible role of neolacto-series gangliosides in the murine central nervous system.

G_M3 is a ubiquitous ganglioside and has been implicated in the regulation of tyrosine kinase activity and its signal transduction [49,50]. G_M3 also has immunomodulatory actions, such as the inhibition of natural killer cell activity [14] and the induction of CD4 internalization in human peripheral blood T lymphocytes [51]. The major difference in ganglioside expression among the five tissues analysed in our study was the almost complete absence of G_M3 in the lungs of β₂M–/– compared with the control β₂M+/– mice. G_M3 is abundant in adult lungs, and its appearance changes during pre-natal and post-natal development [52,53]. Its role in the lungs is unknown but is probably not related to the interaction with pulmonary surfactant apoproteins [53]. The lungs of adult β₂M–/– mice lack obvious morphological pathology (our unpublished data), but this does not exclude a possible role for the β₂M molecule in lung homeostasis.

The pattern of ganglioside expression in β₂M knockout mice partly overlapped the pattern we observed in TNFR1(p55) knockout mice [54]; in both models, there was decreased expression of G_M3 and G_M1b-type gangliosides in the lungs. This is not surprising because TNFR1-dependent responses include stimulation of class I MHC antigen expression on the cell surface [55].

Mice lacking the β₂M gene had greater expression of G_M1a and G_D1a in the liver than the control animals. The locus controlling liver G_M1(NeuGc) expression has been mapped 1 cm centromeric to H-2K [56] and named Ggm-1 [57]. In humans, β-N-acetyl-galactosamine β-1,3-galactosyltransferase (β3Gal-T4), which is involved in G_M1/G_D1 ganglioside synthesis, has been mapped to 6p21.3, next to the HLA-B associated transcript-3 [58]. These data indicate a close relationship between the enzymes involved in glycosphingolipid synthesis and the MHC. Thus, a possible explanation for the increased G_M1/G_D1 expression in β₂M–/– mice could be the increased activity of the G_M1/G_D1 ganglioside synthase gene as part of the compensatory change in MHC gene transcription. This is not likely, since the disruption of the β₂M gene prevents the synthesis of β₂M mRNA but does not interfere with the transcription or stability of MHC class I-heavy chain mRNA [4]. Increased β3Gal-T4 activity in β₂-microglobulin knockout mice may instead be related to disruption in the cellular trafficking of this enzyme in cells lacking β₂-microglobulin. β₂-microglobulin associates with molecules other than the classical MHC gene. These non-classical class I genes, termed MHC class IB genes, have functions unrelated to the immune system [59–61]. They encode β₂M-associated cell-surface molecules, most of which have yet to be assigned a function. β₂M knockout mice have an iron overload, which has been ascribed to the altered expression of non-classical MHC class I gene products that control iron absorption [59–61]. It is tempting to postulate that β3Gal-T4 is such a β₂M-associated molecule. The enzyme is normally located either in the Golgi apparatus or at the surface of the cell [62,63]. It is possible that β3Gal-T4 travels to the cell surface in a complex with β₂-microglobulin. When this transfer is blocked, the enzyme stays in the Golgi and, especially in metabolically-active hepatocytes, produces high quantities of G_M1a and G_D1a gangliosides from the substrates which are more accessible in the Golgi than in the extracellular space. Such an association between β3Gal-T4 and β₂-M is only hypothetical and awaits experimental proof. Combined biochemical, histological and cytological approaches [42] to ganglioside synthesis and expression in mouse tissues would be the best approach to elucidate the relationship between β₂M and non-classical MHC gene products.

APPENDIX

The designation of the glycosphingolipids follows the IUPAC-IUB recommendations [64] and the nomenclature of Svennerholm [65]. Lactosylceramide or LacCer, Galß1–4Glcß1–1Cer; gangliotriaosylceramide or GgOse₃Cer, GalNAcß1–4Galß1– 4Glcß1–1Cer; gangliotetraosylceramide or GgOse₄Cer, Galß1– 3GalNAcß1–4Galß1–4Glcß1–1Cer; lacto-N-neotetraosylceramide or nLcOse₄Cer or nLc4, Galß1–4GlcNAcß1–3Galß1–4Glcß1– 1Cer; lacto-N-norhexaosylceramide or nLcOse₆Cer or nLc6, Galß1–4GlcNAcß1–3Galß1–4GlcNAcß1–3Galß1–4Glcß1–1Cer; G_M3, II³Neu5Ac-LacCer; G_M2, II³Neu5Ac-GgOse₃Cer; G_M1 or G_M1a, II³Neu5Ac-GgOse₄Cer; G_M1b, IV³Neu5Ac-GgOse₄Cer; GalNAc-G_M1b, IV³Neu5Ac-GgOse₅Cer; G_D3, II³(Neu5Ac)₂-LacCer; G_D1α, IV³Neu5Ac,III⁶Neu5Ac-GgOse₄Cer; G_D1a, IV³Neu5Ac,II³Neu5Ac-GgOse₄Cer; G_D1b, II³(Neu5Ac)₂-GgOse₄Cer; G_T1b, IV³Neu5Ac,II³(Neu5Ac)₂-GgOse₄Cer. Only Neu5Ac-substituted gangliosides are presented in this list of abbreviations.