Some anti-ganglioside antibodies used in this study are licensed for commercial distribution by the Johns Hopkins University with Dr. Ronald L. Schnaar entitled to a share of royalty received. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: KV MH. Performed the experiments: KV MH BV IVD. Analyzed the data: KV MH RLS. Contributed reagents/materials/analysis tools: IVD RLS. Wrote the paper: KV RLS MH.
Gangliosides - sialic acid-bearing glycolipids - are major cell surface determinants on neurons and axons. The same four closely related structures, GM1, GD1a, GD1b and GT1b, comprise the majority of total brain gangliosides in mammals and birds. Gangliosides regulate the activities of proteins in the membranes in which they reside, and also act as cell-cell recognition receptors. Understanding the functions of major brain gangliosides requires knowledge of their tissue distribution, which has been accomplished in the past using biochemical and immunohistochemical methods. Armed with new knowledge about the stability and accessibility of gangliosides in tissues and new IgG-class specific monoclonal antibodies, we investigated the detailed tissue distribution of gangliosides in the adult mouse brain. Gangliosides GD1b and GT1b are widely expressed in gray and white matter. In contrast, GM1 is predominately found in white matter and GD1a is specifically expressed in certain brain nuclei/tracts. These findings are considered in relationship to the hypothesis that gangliosides GD1a and GT1b act as receptors for an important axon-myelin recognition protein, myelin-associated glycoprotein (MAG). Mediating axon-myelin interactions is but one potential function of the major brain gangliosides, and more detailed knowledge of their distribution may help direct future functional studies.
Gangliosides, sialic acid-containing glycosphingolipids, are expressed widely in vertebrate tissues but at particularly high abundance in the brain, where they are major cell surface determinants on nerve cells. Four ganglioside structures, GM1, GD1a, GD1b and GT1b (
Cer, ceramide; LacCer, lactosylceramide. GM1, GD1a, GD1b and GT1b comprise up to 97% of all gangliosides in the human central nervous system (boxed area).
Knowledge of the distribution of major gangliosides in the brain informs theories about their functions. The distributions of major brain gangliosides in rodent and human CNS have been studied using chemical and immunohistochemical methods
To reassess ganglioside distribution in the adult mouse CNS, we used highly specific IgG-class monoclonal antibodies (mAb) raised against each of the major brain gangliosides. Since mice fail to raise a robust IgG response to self-gangliosides, we successfully raised these mAb’s in mice genetically engineered to lack complex gangliosides (
Fourteen brains and spinal cords were obtained from 3-month-old wild type (C57Bl/6 strain) female mice. The protocol was approved by Ethical Committee of University of Osijek School of Medicine and Johns Hopkins University Animal Care and Use Committee. Johns Hopkins is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC) International. The protocol also complied with Directive 2010/63/EU of the European Parliament and of the Council on the protection of animals used for scientific purposes.
Mice were deeply anesthetized with isoflurane and consequently transcardially perfused with phosphate buffered saline (PBS) and 4% paraformaldehyde (PFA) in PBS. The brains and spinal columns were dissected, additionally fixed in 4% PFA for 24 hours and cryoprotected in gradients of sucrose (10–30% w/v). After cryoprotection, spinal cords were dissected from the spinal column and the specimens (brains and spinal cords) snap frozen by immersion in cold 2-methylbutane.Tissues were cut on a cryostat and immunohistochemistry was performed using 20–35 µm thick free-floating sections.
The primary antibodies used in this study are listed in
Antibodies to myelin-associated glycoprotein (anti-MAG, Chemicon, Temecula, CA; MAB1567, clone 513) and myelin basic protein (anti-MBP, QED Bioscience, San Diego, CA) were used as markers of myelinated fibers. Anti - MAG antibody detects MAG in immunohistochemistry on frozen sections, immunocytochemistry and Western blot and is routinely evaluated by the manufacturer by staining of myelinated fibers of rat hippocampus (manufacturer’s technical sheet). In our hands, anti - MAG antibody stained all white matter tracts in mouse CNS at the concentration of 1.2 µg/ml. No staining was present when anti-MAG antibody was used to stain
Anti - MBP antibody reacts with residues 130–136 in human myelin basic protein (21.5 kD and 18.5 kD molecular forms) and it also recognizes primate, rabbit, sheep, goat, rat and mouse myelin basic protein (manufacturer’s technical sheet). In our hands, anti-MBP antibody was shown to stain all white matter tracts in mouse CNS at the concentration of 1.5 µg/ml.
For detection of brainstem catecholaminergic neurons we used an antibody against tyrosine hydroxylase (TH) (anti- tyrosine hydroxylase, AB152, Millipore, Billerica, MA) at a dilution of 1: 1000. Each batch of anti-TH antibody is routinely evaluated by the manufacturer by Western Blot on PC12 lysates where it labels a single band at approximately 62 kDa (reduced) corresponding to tyrosine hydroxylase (manufacturer’s technical sheet). In our hands, anti–TH antibody stained catecholaminergic neurons of medial and lateral parabrachial nuclei, locus coeruleus, caudal ventrolateral medulla and solitary nucleus.
All steps of ganglioside immunohistochemistry were performed at 4°C and without the use of any detergent. The tissues were blocked in 1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO), 5% goat serum (Invitrogen, Carlsbad, CA) in PBS. Immunohistochemistry of myelin-associated glycoprotein (MAG) and myelin basic protein (MBP) was performed using 1% Triton X-100 in the blocking solution.
After washing the free-floating tissues, primary antibody binding was detected with biotin-SP-AffiniPure goat anti-mouse IgG (H+L) (Jackson Immunoresearch Labs., West Grove, PA) at 1 µg/ml final concentration followed by Vector Elite peroxidase kit (Vector Laboratories, Burlingame, CA) and developed with SIGMA
Co-localization experiments were performed using affinity-purified rabbit anti- tyrosine hydroxylase (TH) antibody and monoclonal anti-ganglioside antibodies as previously described. Primary TH antibody binding was detected using biotinylated goat anti-rabbit IgG (H+L) (Jackson), and streptavidin Alexa Fluor® 488 conjugate (Invitrogen, Carlsbad, CA). Anti-ganglioside antibodies were detected using Alexa Fluor® 546 goat anti-mouse IgG (H+L) (Invitrogen).
To determine ganglioside expression in white matter tracts, five mice were subjected to anterograde labeling. Craniotomy was performed on anesthetized mice and sensorimotor cortex was injected with one injection of 0.2 µl 10% (w/v) 10,000 MW biotinylated dextran amine (BDA, Invitrogen). After 2 weeks mice were perfusion-fixed, brains and spinal cords were recovered as described above, cryosections prepared, and gangliosides detected using immunohistochemistry. BDA-labeled axons were detected with streptavidin Alexa Fluor® 488 conjugate, whereas anti-ganglioside antibodies were detected using Alexa Fluor® 546 goat anti-mouse IgG.
Fluorescent immunohistochemical images were obtained using a Zeiss LSM 510 inverted confocal microscope (The Johns Hopkins School of Medicine Microscope Facility), assembled in CorelDraw 12 software and assembled images adjusted for contrast, intensity and brightness.
Qualitative analysis of immunohistochemical reactivity to gangliosides GD1a and GT1b was performed on images taken with the same exposure times by two independent observers. The relative expression (+++, strong signal; ++, moderate signal; +, weak signal, −, no signal) of gangliosides GD1a and GT1b was compared to negative control on each mouse brain region listed in
The expression patterns of gangliosides GM1, GD1a, GD1b and GT1b were studied using immunohistochemistry on adult C57Bl/6 mouse brains and spinal cords cut in three planes (coronal, sagittal and horizontal). GD1b expression is abundant in both gray and white matter throughout the mouse brain (
The negative control (A) was performed by omitting the primary antibody. Scale bar: 5000 µm.
GD1a shows strong immunoreactivity in all layers of olfactory bulb and accessory olfactory bulb (
Horizontal sections of olfactory bulb (A, A’, A’’; mitral cell layer of olfactory bulb is pointed out with black arrows). Coronal sections of adult mouse brain at coordinates: interaural line 3.22 mm, bregma –0.58 mm (B, B’, B’’) and interaural line 2.10 mm, bregma –1.7 mm (C, C’, C’’). Hippocampal formation is boxed in C’, arrowhead in C’ shows zona incerta and asterix in C’’ indicates medial and lateral habenular nuclei. Higher magnifications of CA1 and CA3 fields of pyramidal cell layer of hippocampus (D1–D5 and E1–E5, respectively) and granule cell layer of dentate gyrus (F1–F5). Sagittal sections of zona incerta (ZI), subthalamic nucleus (STh) and substantia nigra (SN) (G-G’’). The negative controls were performed by omitting the primary antibody (A, B, C, D1, E1, F1, G). Amy, amygdala; CA1, CA1 field of pyramidal cell layer of hippocampus; CA3, CA3 field of pyramidal cell layer of hippocampus; CPu, caudate putamen (striatum); Ctx, cortex; df, dorsal fornix; DG, dentate gyrus; EPl, external plexiform layer of olfactory bulb; fi, fimbria of the hippocampus; Gl, glomerular layer of olfactory bulb; GlA, glomerular cell layer of accessory olfactory bulb; GP, globus pallidus; GrA, granule cell layer of accessory olfactory bulb; GrO, granule cell layer of olfactory bulb; ic, internal capsule; MiA, mitral cell layer of accessory olfactory bulb; opt, optic tract; Pa, paraventricular hypothalamic nucleus; Pir, piriform cortex; Rt, reticular nucleus (prethalamus); st, stria terminalis; vhc, ventral hippocampal commissure. Scale bars = 1000 µm in A-A’’, G-G’’; 4000 µm in B-C’’ and 50 µm in D1–F5.
Strong expression of gangliosides GD1a, GD1b and GT1b is found in all fields and layers of cerebral cortex (
Both striatum (caudoputamen) and globus pallidus show strong immunoreactivity to gangliosides GD1a (
All amygdaloid nuclei show strong immunoreactivity to gangliosides GD1a (
Major brain gangliosides are differentially distributed in the hippocampal formation. Anti-GM1 immunostaining is limited to alveus hippocampi and few other fibers found in oriens layer and lacunosum moleculare layer of hippocampus, as well as in polymorph layer of dentate gyrus. GM1 is absent from pyramidal cells of the CA1 and CA3 fields of Ammon’s horn (
Gangliosides GD1a and GT1b are expressed in some white matter tracts of the telencephalon, but not others. Consistent with their expression on myelinated axons, gangliosides GD1a and GT1b immunodetection was found to be satisfactory only if tracts were cut longitudinally instead of perpendicularly. For that reason, brains cut in three planes (coronal, sagittal and horizontal) were analyzed. A subset of fibers in the corpus callosum show high immunoreactivity for GD1a (
Commissural fibers and corticospinal tract are labeled with BDA (C-N; green). A subset of GD1a and GT1b expressing fibers are detected in the middle of corpus callosum (A, B; arrowheads). Cell nuclei are stained with DAPI (blue). The negative control is performed by omitting the primary antibody (E, J). Scale bars = 500 µm in A, B and 50 µm in C-N.
Most thalamic nuclei abundantly express gangliosides GM1 (
Medial and lateral habenular nuclei show strong expression of GD1b (data not shown) and GT1b (
All hypothalamic nuclei, including paraventricular and medial preoptic nuclei and anterior hypothalamic area, abundantly express gangliosides GM1 (
While GT1b is expressed both in subthalamic nucleus and zona incerta (
In the mouse cerebellum, GM1 is detected only in white matter (
Distribution of gangliosides GM1 (B, green, asterix denotes the white matter of cerebellum), GD1a (C, green), GD1b (D, green) and GT1b (E, green) in mouse cerebellum (low magnification). Double immunohistochemistry on GD1a (H, L, P, T; red) or GT1b (I, M, Q, U; red) and TH (G, H, I, K, L, M, O, P, Q, S, T, U; green) in cerebellum (F-I), red nucleus (J-M), raphe magnus nucleus (N-Q) and raphe obscurus nucleus (R-U). The negative controls were performed by omitting the anti-ganglioside antibody (A, G, K, O, S) or both anti-ganglioside and anti-TH antibody (F, J, N, R). Cell nuclei are stained blue using DAPI. gr, granular cell layer; mol, molecular layer; P, Purkinje cell layer; RMg, raphe magnus; RN, red nucleus; ROb, raphe obscurus. Scale bars = 200 µm in A-E, J-M; 50 µm in F-I and 100 µm in N-U.
Gangliosides GM1, GD1b and GT1b are present in most nuclei of the brainstem and in white matter tracts. In midbrain, moderate expression of GT1b is found both in reticular and compact part of substantia nigra (
The raphe magnus and pallidus nuclei show strong expression of GT1b (
Among autonomic nuclei of the brainstem, strong GD1a immunoreactivity is detected in periaqueductal gray matter, locus coeruleus (
Double immunohistochemistry with anti-tyrosine hydroxylase (TH) antibody (B-D, F-H, J-L, N-P, green) and anti-GD1a (C, G, K, O; red) or anti-GT1b (D, H, L, P; red). Locus coeruleus (A-D), lateral parabrachial nucleus (E-H), caudal ventrolateral medulla (I-L) and solitary nucleus (M-P). The negative controls (A, E, I, M) were performed by omitting primary antibodies. Cell nuclei are stained blue with DAPI. CVL, caudal ventrolateral medulla; LC, locus coeruleus; LPB, lateral parabrachial nucleus; Sol, solitary nucleus. Scale bar = 100 µm.
MAG is shown for comparison and is expressed in myelinated fibers (D). The negative control was performed by omitting the primary antibody (A). CI, caudal interstitial nucleus of the medial longitudinal fasciculus; ECu, external cuneate nucleus; Gi, gigantocellular reticular nucleus; icp, inferior cerebellar peduncule; IO, inferior olive; IRt, intermediate reticular nucleus; LPGi, lateral paragigantocellular nucleus; MVe, medial vestibular nucleus; PCRt, parvicellular reticular nucleus; Pr, prepositus nucleus; py, pyramidal tract; ROb, raphe obscurus nucleus; RPa, raphe pallidus nucleus; Sol, solitary nucleus; SpVe, spinal vestibular nucleus; sp5, spinal trigeminal tract; Sp5I – spinal trigeminal nucleus. Scale bar = 2000 µm.
Moderate expression of GD1a and GT1b is also found in mesencephalic, principal and spinal trigeminal nuclei, as well as in pontine, medullary, parvicellular and intermediate reticular nuclei (
GD1b is strongly expressed in inferior olivary, vestibular nuclei and solitary nucleus and to some extent in gigantocellular reticular nucleus, parvicellular reticular nucleus and intermediate reticular nucleus (
Whereas GD1b is expressed in both white and gray matter of the spinal cord (
The negative control is performed by omitting the primary antibody (A, A’, A’’). Black arrowheads denote the propriospinal white matter tracts. White arrows point to the corticospinal tract. Black arrow points to the Rexed laminae I and II. Scale bar = 500 µm.
In this study, we report a detailed expression analysis of major brain gangliosides (GM1, GD1a, GD1b and GT1b) in the adult mouse CNS. Previously, the distribution of major brain gangliosides was studied by using immunohistochemistry and biochemistry
Although gangliosides are major cell surface structures on all neurons, their immunohistochemical analysis is technically complicated by their lipid nature. Fixation of tissues with organic solvents
To help mitigate these issues we used highly specific IgG-class monoclonal antibodies produced in
A complementary method for analysis of ganglioside distribution in CNS tissue is imaging mass spectrometry (IMS). IMS gives detailed information on the ceramide core of the ganglioside and is able to assess multiple ganglioside species at a time in a single tissue slice. However, it has the well documented potential problem of partial loss of sialic acid during ionization and differentiation between ganglioside isomers (such as GD1a and GD1b) requires more detailed (MSn) analyses
Our results show that GD1b and GT1b are widely expressed throughout the mouse CNS, whereas the immunoreactivity to GM1 is predominately in white matter tracts and immunoreactivity to GD1a significantly decreases caudal to the superior colliculus. Particularly strong expression of GD1a, GD1b and GT1b is found in olfactory bulbs, neocortex, basal ganglia, amygdaloid nuclear complex, septal regions, thalamic nuclei, hypothalamic nuclei and hippocampal formation.
Strong expression of GD1b and GT1b is found in all three layers of cerebellum (molecular layer, Purkinje cell layer and granular cell layer) and cerebellar nuclei, whereas GD1a is moderately expressed in cerebellar cortex, but completely absent from cerebellar nuclei. Our results also show that GM1 is limited to white matter in the cerebellum.
Among nuclei of brainstem, strong immunoreactivity to GD1a is detected in periaqueductal gray matter, locus coeruleus, subcoeruleus nucleus, medial and lateral parabrachial nucleus, Kolliker-Fuse nucleus, rostral and caudal ventrolateral medulla and solitary nucleus. Interestingly, those are all autonomic nuclei that are also positive for tyrosine hydroxylase. It is important to note that these nuclei send their projections to the spinal cord and that they are involved in cardiovascular reflexes and respiratory control but also receive sensory, nociceptive and visceral input from periphery. In concordance with these results, previous biochemical studies have also shown the expression of GD1a and GT1b in locus coeruleus, as well as in dorsal raphe nucleus, laterodorsal tegmentum and pedunculopontine tegmentum
On the other hand, immunoreactivity to GD1a is weak or absent in the reticular nucleus of thalamus, habenular nuclei, zona incerta, subthalamic nucleus, red nucleus, superior and inferior colliculus, cochlear nuclei, vestibular nuclei, dorsal tegmental nucleus, motor nuclei of the trigeminal nerve, facial nucleus, hypoglossal nucleus, gracile nucleus, cuneate nucleus, external cuneate nucleus and ventral horn of spinal cord (for detailed comparison of GD1a and GT1b expression see
Another interesting finding is that certain white matter tracts such as corpus callosum, the main comissural pathway of telencephalon, and corticospinal tract express all four major gangliosides, while others, e.g. white matter of cerebellum, do not. It is not clear from our data if certain gangliosides are on axolemma or on myelin membrane. However, biochemical studies have shown that GM1 is part of myelin membrane, while CNS axolemma contains all four major brain gangliosides
A goal in documenting the distribution of major brain gangliosides is to inform theories about their functions. Based on studies of the
MAG is also known to inhibit axon outgrowth
Two recent
The studies of the distribution of major brain gangliosides in DRGs have shown that most neurons express GT1b, while GD1a is expressed mainly in small neurons (15–30 µm diameter), involved in nociception, and only in 10–20% of larger neurons
It is important to note that gangliosides GD1a and GT1b are strongly expressed in cerebral cortex, corticospinal tract and certain brainstem nuclei that comprise ascending and descending tracts of spinal cord. However GD1a and GT1b expression is weak in gracile and cuneate nuclei that comprise dorsal column pathway. This knowledge may be helpful in designing treatments that address MAG inhibition of axonal regeneration through manipulation of these particular gangliosides. Along the same line, pathway-specific treatments may be designed based on the specific MAG receptor expression profiles. This notion is supported by the findings that treatments that target a single molecule often fail to evoke robust axonal regeneration after spinal cord injury
Primary antibodies used in the study.
(DOCX)
Qualitative analysis of immunohistochemical reactivity to gangliosides GD1a and GT1b. +++, strong signal; ++, moderate signal; +, weak signal, −, no signal.
(DOCX)