Correspondence to: Dr. Junling Wang, Department of Neurology, Xiangya Hospital, Central South University, Changsha, Hunan province, 410008, China. Fax: +86-(731)−84–327-332. E-mail:firstname.lastname@example.org or email@example.com
The spinocerebellar ataxias (SCAs) are a large clinically and genetically heterogeneous group of inherited neurodegenerative disorders that are characterized by dysfunction of the cerebellum, as shown by progressive loss of balance and motor coordination of gait and limbs (Duenas et al., 2006). Atrophies of the cerebellum and brainstem are the prominent features, but different combinations of degeneration in the cerebellum, spinal tracts, peripheral nerves, cerebral cortex, basal ganglia, optic nerve, and others are also observed in SCA (Durr, 2010). Several cellular and molecular mechanisms have been proposed to be critical for the development of SCAs, such as toxic accumulation of aggregates and intranuclear inclusions, transcription interference, altered calcium homoeostasis, altered glutamate transmission, mitochondrial dysfunction, RNA alterations, chromatin structure abnormalities, toxic accumulation of aggregates and RNA foci, cytoskeletal abnormalities, axonal transport deficits, and increased tau phosphorylation (Durr, 2010). In 2010, our group identified two missense mutations in the TGM6 gene in two Chinese SCA families, c.1550T>G transition (p.L517W) and c.980A>G transition (p.D327G), suggesting a novel causative gene for SCA, named as SCA 35 (Wang et al., 2010). In 2012, Li et al. (Li et al., 2013) reported a novel mutation in the TGM6 gene in another Chinese SCA family, c. 1528G>C transition (p.D510H), which was supportive of the TGM6 gene being causative gene for SCA 35. In addition, overexpression of the two missense mutations of TGM6 (D327G, L517W) dramatically increases the sensitivity of NIH3T3 cells to staurosporine-induced apoptosis via increasing the activity of caspases (Guan et al., 2013).
Transglutaminases (TGMs) form a family of structurally- and functionally related enzymes that post-translationally modify proteins by catalyzing a Ca2+-dependent transferase reaction between the c-carboxamide group of a peptide-bound glutamine residue and various primary amines (Yee et al., 1994). Nine distinctly expressed TGM genes are present in mammals (Fesus and Piacentini, 2002). Among them, TG 1–3 and 6 were discovered in human brain (Hadjivassiliou et al., 2008). In addition to SCA, TGs are also hypothesized to be involved in the pathogenesis of several other neurodegenerative diseases, including polyglutamine expansion diseases, Alzheimer's, Parkinson's and supranuclear palsy (Jeitner et al., 2009). However, despite extensive investigation over the last two decades, the physiologic or pathologic roles of TGs in the brain remain unclear.
Although the functions of the TG6 protein implicated in SCA are presently unknown, analysis of its expression pattern in the mammalian brain could facilitate our understanding of the mechanisms that underlie the neuropathology of the disease. Therefore, in this study we investigated the distribution of TG6 in adult mouse brain by immunohistochemistry and found that a large number of TG6-positive neurons were located in brain regions regulating locomotor activity.
MATERIALS AND METHODS
Animals and Tissue Processing
Twelve 2-month-old C57BL/6J mice (six males and six females) were deeply anesthetized with sodium pentobarbital (70 mg kg−1) by intraperitoneal injection, and perfused transcardially with 0.01M phosphate-buffer saline (PBS; pH 7.4), followed by 4% paraformaldehyde (PFA; Sigma–Aldrich, St. Louis, MO) in 0.1 M phosphate buffer (PB; pH 7.4). The brain and the spinal cord were removed, postfixed in 4% PFA for 16 h, then cryoprotected with 30% sucrose in PBS overnight at 4°C. Coronal sections (30-µm thick) were prepared with a cryostat (CM1950; Leica, Wetzlar, Germany). Glutamate acid decarboxylase 67 (GAD67)-GFP knockin mice (Tamamaki et al., 2003) were used to investigate whether TG6 was expressed in gamma-aminobutyric acid (GABA)-expressing neurons. In this mouse, GABAergic neurons are tagged with GFP, and thus GFP-positive cells reflect the GABAergic neurons. Animal care procedures and experimental protocols were approved by the Animal Studies Committee at the Tongji University School of Medicine in Shanghai, China.
Endogenous peroxidase activity was blocked by immersing the slides in 3% hydrogen peroxide diluted with methanol for 15 min at room temperature. Sections were then pretreated with 0.01 M citrate buffer (pH 6.0) for 5 min at 95°C. After washing in PBS, they were incubated overnight at 4°C with goat anti-TG6 (G-16) antibody (1:600; sc-85959, Santa Cruz Biotechnology, Santa Cruz, CA) diluted in PBS containing 0.3% Triton X-100 and 1% bovine serum albumin, followed by incubation with biotinylated horse anti-goat IgG (1:500; Vector Laboratories, Burlingame, CA) for 3 hr and incubation with VECTASTAIN ABC rabbit-IgG KIT (1:200; Vector Laboratories) for 1 hr at room temperature. The bound antibodies were visualized using 3,3′-diaminobenzidine tetrahydrochloride as the chromogen.
To test the specificity of immunostaining, the TG6 antibody was pre-absorbed with the blocking peptide (1:2; sc-85959 P, Santa Cruz) overnight at 4°C in the antibody dilution buffer, and then the neutralized antibody was used for immunostaining as mentioned above. In addition, two another TG6 antibody, rabbit anti-TG6 (1:200; AV49194, Sigma-Aldrich, St. Louis, MO) and rabbit anti-TG6 (1:200; ab105725, Abcam, Suite B2304 Cambridge, MA) were also used. Immunoreactive signals were visualized by incubating sections with biotinylated goat anti-rabbit IgG (1:500; Vector Laboratories) antibodies for 3 hr, followed by incubation with Cy3-conjugated streptavidin (1:1,000; Jackson Immunoresearch) for 1 hr at room temperature. In addition, negative controls were also performed by omitting the TG6 antibody or replacing it with normal serum.
For double immunostaining, sections were incubated with a mixture of the goat anti-TG6 antibody (1:400; Santa Cruz Biotechnology) and one of the following antibodies: mouse anti-NeuN (1:1000; Millipore, Billerica, MA), rabbit anti- GFAP (1:500; Dako, Glostrup, Denmark), rabbit anti-GFP (1:2000; Invitrogen, Carlsbad, CA; used only in GAD67-GFP knockin mouse, see below), and mouse antityrosine hydroxylase (TH; 1:40,000; Sigma–Aldrich). Immunoreactive signals were visualized by incubating sections with a mixture of Alexa Fluor 488 donkey anti-mouse or donkey anti-rabbit IgG (1:500; Invitrogen) and biotinylated horse anti-goat IgG (1:500; Vector Laboratories) antibodies for 3 hr, followed by incubation with Cy3-conjugated streptavidin (1:1,000; Jackson Immunoresearch) for 1 hr at room temperature.
Microscopy and Imaging
Images were acquired on an epifluorescent Nikon microscope equipped with a Nikon Coolpix digital camera (DS-Ri1; Tokyo, Japan). All images were made into figures using Adobe Photoshop (Adobe Systems Incorporated, San Jose, CA) and only minor adjustments to the contrast and brightness settings were applied if necessary.
Three GAD67-GFP knockin mice were used for the quantitative analysis of TG6/GABA-double labeled neurons in the reticular thalamic nucleus and reticular part of the substantia nigra. The percentage of colabeled cells in a total population of labeled neurons was calculated by counting the number of double-labeled and single-labeled neurons. Six sets of consecutive 30-µm-thick coronal sections were collected from each brain. One set of sections was processed for the double immunostaining of TG6 and GFP and then used for cell counts. All cell count results are presented as mean ± SD.
Construction of TGM6 Expression Vectors
Full-length human TGM6 complementary DNA (cDNA) cloned into the mammalian expression vector pCAGGS-myc, and was used as the TGM6 expression vectors. The plasmid was sequenced to confirm its identity.
Cell Culture and DNA Transfection
The TGM6 expression vector was transfected into the PC12 cells with Lipofectamine 2000 (Invitrogen). PC12 cells were grown at 37°C under a 5% CO2 atmosphere in Dulbecco's modified Eagle's medium (Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT) and 100 U mL−1 of penicillin/streptomycin. Cells were harvested 48 hr after the transfection, and protein extracts were processed for Western blot.
Western Blot Assays
Proteins were separated by 10% SDS–PAGE and then transferred onto a polyvinylidene difluoride membrane (MilliporeBillerica, MA). The following primary antibodies were used: goat anti-TG6 (G-16) antibody (sc-85959, Santa Cruz Biotechnology), TG6 antibody preabsorbed with the blocking peptide (1:2; sc-85959 P, Santa Cruz), and mouse anti-GAPDH antibody (Santa Cruz Biotechnology). Goat anti-mouse and donkey anti-goat IgG-HRP antibodies (Kangchen) were used as secondary antibodies, and detection of binding signals was performed using the ECL plus kit (Millipore).
Identification of the Specificity of TG6 Antibody
Three anti-TG6 antibodies (Santa Cruz Biotechnology; Sigma; Abcam) were used in the initiation of the study. Although the distribution patterns of TG6 immunoreactivity shown by the three antibodies were similar in mouse brain (Fig. 1A,C,D), the antibody from Santa Cruz was the best one in term of the signal intensity and background staining. When the antibody (Santa Cruz) was pre-absorbed with the blocking peptide, almost no immunoreactivity was detected (Fig. 1B). In addition, Western Blot data showed that the anti-TG6 antibody (Santa Cruz) recognized a 79 kDa protein band, which is consistent with estimated TG6 molecular weight (79 kDa; Fig. 1E) and this band was largely reduced when it was pre-absorbed with blocking peptide, (Lane 3, Fig. 1E). These results suggest the specificity of the TG6 antibody, and it was used in our investigation of TG6 immunoreactivity in mouse brain.
Expression Patterns of TG6 in Adult Mouse Brain
Although TG6 immunoreactivity was widely distributed throughout the adult mouse brain, it showed a relatively restricted expression pattern. Immunoreactivity was more abundant in the subcortical regions and brain stem, with minimal neocortex staining. Double immunostaining with NeuN showed that all TG6-positive cells were immunoreactive for NeuN and no TG6 immunoreactivity colocalized in cells immunostained with the antibody against GFAP, which is an astrocytic marker. Thus, TG6 demonstrated a neuron-specific expression pattern in mouse brain. In addition, TG6 immunoreactivity was localized in the cytoplasm of neuronal cell bodies, and no TG6 immunoreactivity was found to be localized in blood vessels in the brain.
TG6 Expression in the Olfactory System, Cerebral Cortex, and Basal Forebrain
In the olfactory system, abundant TG6-positive neurons were detected in the mitral cell layer of the olfactory bulb and the external part of the anterior olfactory nucleus, while a small number of stained cells were scattered throughout the other layers of the olfactory bulb (Fig. 2A–C). The accessory olfactory bulb and the other parts of the anterior olfactory nucleus did not demonstrate TG6 immunoreactivity.
In the cerebral cortex, TG6-positive neurons were only detected in the cingulate and piriform cortices (Fig. 3A,B). In the hippocampal formation, TG6 immunoreactivity was detected in the CA3 region, while CA1, CA2 and the dentate gyrus demonstrated no immunoreactivity (Fig. 3C). On the other hand, many immunopositive neurons were distributed in the subiculum, a transitional region between the hippocampus and cerebral cortex (Fig. 3D). In the ventral forebrain, a large number of TG6-positive neurons were distributed in the medial septal nucleus and nuclei of both vertical and horizontal limbs of the diagonal band (Fig. 2D–G), while a moderate number of TG6-positive neurons were present in the ventral pallidum (Fig. 2F,H), bed nucleus of stria terminalis, and substantia innominata (Fig. 4A,C). In the basal ganglia, many neurons with intense TG6 immunoreactivity were distributed in the lateral and medial globus pallidus, whereas the caudate and putamen, located lateral to the lateral globus pallidus, contained no TG6 immunoreactivity (Figs. 3E,F and 4D,F).
TG6 Expression in the Diencephalon
The hypothalamus was one of the brain regions that contained a high density of TG6-positive neurons, particularly in the lateral hypothalamic area (Fig. 4A,B,D,G,J), arcuate nucleus (Fig. 4G,H), ventromedial hypothalamic nucleus (Fig. 4D,E,G,H) and dorsomedial hypothalamic nucleus (Fig. 4G). On the other hand, the suprachiasmatic nucleus, supraoptic nucleus and paraventricular nucleus did not contain TG6-positive neurons (Fig. 4A,B). In addition, the subthalamic nucleus and zona incerta, which are two brain regions located between the thalamus and hypothalamus, contained numerous TG6-positive neurons (Fig. 4G,I,J,L).
In the thalamus, the anterodorsal thalamic nucleus (Fig. 5A,B), mediodoral thalamic nucleus (Fig. 5D), ventral posteromedial thalamic nucleus, ventral posterolateral thalamic nucleus (Fig. 5D–F), and paraventricular thalamic nucleus (Fig. 5D) contained a large number of TG6-positive neurons. In contrast, only a few immunopositive neurons were observed in the reticular thalamic nucleus (Fig. 5A,C–E), lateral habenular nucleus (Fig. 5D), anteromedial thalamic nucleus (Fig. 5A,C) and dorsal lateral geniculate nucleus (Fig. 5D,F). A few TG6-positive neurons were also observed in other thalamic nuclei, such as the anteroventral thalamic nucleus (Fig. 5A,C) and medial geniculate body (Fig. 6A).
TG6 Expression in the Brainstem, Cerebellum, and Spinal Cord
Widespread TG6 expression was observed in the brainstem. In the dorsal midbrain (tectum), many neurons with weak TG6 immunoreactivity were found in both the superior and inferior colliculi (Fig. 6D,E). A similar expression pattern was also observed in the deep mesencephalic nucleus (Fig. 6A,B), red nucleus (Fig. 6A), pariaqueductal grey (Fig. 6D), dorsal raphe nucleus, and nucleus of Darkschewitsch located in the ventral pariaqueductal grey at the level of rostral superior colliculus (Fig. 6A,B). In the ventral midbrain, two regions contained numerous neurons with intense TG6 immunoreactivity, the substantia nigra (Fig. 6A,C) and ventral nucleus of the lateral lemniscus (Fig. 6D,F). Other regions in the midbrain, such as the interpeduncular nucleus did not have TG6-immunopositive staining (Fig. 6A).
In the tegmentus of the pons, TG6 immunoreactivity was observed in the dorsal tegmental nucleus, locus coeruleus, and parabrachial region surrounding the superior cerebellar peduncle (Fig. 7A,B). At the caudal level of the pontine tegmentum, intense to moderate immunostaining was present in the prepositus nucleus and all subnuclei of the ventibular nucleus (Fig. 7D). Ventrally, moderate to weak TG6 immunoreactivitiy was observed in the principal sensory trigeminal nucleus, motor trigeminal nucleus (Fig. 7A,C), superior olivary nucleus, nucleus of the trapezoid body (Fig. 7A), facial nucleus (Fig. 7D,G), and pontine nucleus (Fig. 6G). Many neurons in the cochlear nucleus were also positively stained with the TG6 antibody (Fig. 7D,E). In addition, the reticular formation, including the raphe nuclei, contained many neurons with moderate TG6 immunoreactivity in the pons (Fig. 7A,D).
In the medulla oblongata, intense TG6 immunoreactivity was observed in its dorsal part, including the dorsal column nucleus (gracile and cuneate nuclei) and external cuneate nucleus (Fig. 8A,D,G,H). Moderate to weak immunostaining was observed in the nucleus of the solitary tract (Fig. 8A,D,E,G,H), dorsal motor nucleus of the vagus nerve (Fig. 8E), the oral and interpolar subnuclei of the spinal trigeminal nucleus (Fig. 8A,B,D), and hypoglossal nucleus (Fig. 8D,E,G,H). The reticular formation, including the lateral reticular nucleus (Fig. 8D,F) and raphe nucleus, also showed moderate TG6 immunoreactivity (Fig. 8D). In addition, the rostal part of the inferior olivary nucleus was weakly immunostained (Fig. 8A,C), but its caudal part showed intense immunoreactivity (Fig. 8D). On the other hand, the caudal subnucleus of the spinal trigeminal nucleus was not immunoreactive with the TG6 antibody (Fig. 8G,I).
In the cerebellar cortex, Purkinje cells showed weak immunoreactivity, while the molecular and granule cell layers were negative for TG6 immunostaining (Fig. 7H). In the deep cerebellar nucleus, however, many neurons were intensely immunostained with TG6 antibody (Fig. 7D,F). In the spinal cord, TG6 immunoreactivity was only observed in the ventral horn, including the motor neurons (Fig. 8J,K).
TG6 was Expressed in Neurons but not Glia Cells
TG6 immunoreactivity was not detected in the white matter, suggesting that TG6 was expressed in a neuron selective manner. To further confirm this, we performed double immunostaining with NeuN, a panneuronal marker. This showed that all TG6-positive cells were costained with NeuN antibody in mouse brain (Fig. 9A–A″). However, double immunostaining with GFAP, which is specifically expressed in astrocytes, showed that none of TG6-positive neurons were positive for GFAP (Fig. 9B–B″). These results demonstrated that TG6 was exclusively expressed in neurons in adult mouse brain.
TG6-Expressing Cells were GABAergic in the Substantia Nigra
The substantia nigra is the key region in control of locomotion through its projections to the basal ganglia (Biggs and Starr, 1997). We found that many TG6-positive neurons were located in its reticular part, which is where many GABAergic inhibitory neurons are located (Brazhnik et al., 2008). To examine whether TG6-positive neurons were GABAergic neurons, we performed double staining for TG6 and GFP in GAD67-GFP knockin mice. The results indicated that the vast majority of TG6-expressing cells were GFP positive, showing that TG6 was expressed in GABAergic neurons in the reticular part of the substantia nigra (Fig. 10A–A″). Cell counts showed that the TG6/GABA-colabeled neurons in the reticular part constituted 97.35% ± 2.82% of the GABAergic neurons and 96.75% ± 2.89% of the TG6-positve neurons. In addition, double immunostaining of TG6 and TH showed that although TG6 was located in both the compact and reticular parts of the nigra, the TG6-positive neurons in the compact part were not positive for TH immunoreactivity (Fig. 10C–C″).
It is known that the vast majority of neurons in the reticular thalamic nucleus are GABAergic (Brazhnik et al., 2008). In this study, many neurons with intense TG6 immunoreactivity were also present in this nucleus. We therefore examined whether TG6 and GABA colocalized in neurons of this region. We found that all TG6-positive cells were co-stained with GFP antibody (Fig. 10B–B″), and TG6/GABA double-labeled neurons corresponded to 95.77% ± 2.62% of the GABAergic neurons in the reticular thalamic nucleus.
A previous study has investigated TG6 expression in mouse central nervous system at embryonic stages, and identified an abundant expression of TG6 adult brain (Thomas et al., 2013). Overall, our results were consistent with this report, and we provided comprehensive immunohistochemical mapping data throughout the whole brain regions. Robust TG6 expression in brain regions involved in locomotive control, such as the globus pallidus, subthalamic nucleus, sustantia nigra, ventricular nucleus, deep cerebellar nucleus, supports the hypothesis that TG6 is associated with the neurodegenerative disease and TG6 is the causative gene for SCA 35 (Wang et al., 2010).
Our previous study showed that walking difficulty and cerebellar dysarthria were early clinical signs, while ataxia of the upper extremities was a relatively late symptom in SCA 35 families (Wang et al., 2010). The onset age ranged from 40 to 48 years and disease duration varied from 5 to 31 years (Wang et al., 2010). In the report by Li et al. (2013), patients presented with similar symptoms to those reported by our patients, including unsteady gait, spasticity, ataxia, hyperreflexia, and Babinski's sign, while the onset age of the patients ranged from childhood to 40s. These clinical symptoms are more likely related to extracerebellar pathology because imaging and neuropathological data show that major alterations are present in other parts of the motor system outside the cerebellum in SCA (van Gaalen et al., 2011).
Consistent with the clinical observations from SCA patients, TG6-positive neurons were abundantly distributed in both the cerebellum and other brain regions involved in motor control in this study. In the cerebellum, TG6-positive neurons were present in both the cerebellar cortex and cerebellar nuclei. In the cerebellar cortex, TG6 was only expressed in Purkinje cells, which are the only neurons responsible for sending output from the cortex. On the other hand, TG6 was expressed throughout the deep cerebellar nuclei which receive input from Purkinje cells and send output information to other brain regions. In addition, numerous TG6-positive neurons were contained in the pontine nucleus and vestibular nucleus, which send many mossy fiber axons to the cerebellum (Coesmans et al., 2004). The inferior olivary nucleus, which sends climbing fibers to the cerebellum (Devor, 2000), also contained TG6-positve neurons. In the view of direct efferent targets of the cerebellum, many TG6-positive neurons were located in the red nucleus and tectum. Taken together, TG6 was expressed in the cerebellum and those brain regions with direct neuronal connections with the cerebellum.
In addition to the globus pallidus and caudate putamen, the ventral pallidus, substantia innominata, substantia nigra, and subthalamic nucleus are also regarded as a part of the basal ganglia in terms of their functions in motor control. These regions, together with the cerebellum, are regarded as the key components of the extrapyramidal system (Bostan and Strick, 2010). Interestingly, in this study, except for the caudate putamen, TG6 was expressed in numerous neurons in these regions. Mutations of the TGM6 gene may lead to functional impairment and finally neuronal loss in these brain regions as SCA development progresses. An in vivo study has shown that intraventricular injection of anti-TG2/3/6 cross-reactive antibody provoked ataxia in mice (Boscolo et al., 2010). Taken together, the high distribution density of TG6-positive neurons in the extrapyramidal system supports the hypothesis that TGM6 is the causative gene in the development of SCA 35 (Wang et al., 2010).
Double immunostaining showed that TG6 was expressed in the reticular part of the sustantia nigra and reticular thalamic nucleus, which are the two brain regions containing a high density of GABAergic neurons (Brazhnik et al., 2008). GABA is the main inhibitory neurotransmitter in the brain. Our group has reported that the inhibitory amino acids, including GABA, in the cerebrospinal fluid of SCA patients is reduced (Yang et al., 1999), which may reflect an impairment or loss of GABAergic neurons in SCA patients.
Recently, several reports showed TG enzymes are also related to other neurodegenerative diseases. It has been reported that TG activity is significantly elevated in the affected cerebral regions in Alzheimer disease, Huntington disease and supranuclear palsy (Jeitner et al., 2009). In addition, increased TG2 protein is observed in the cerebrospinal fluid of Alzheimer disease (Bonelli et al., 2002) and Parkinson disease subjects (Vermes et al., 2004). It is likely that in addition to SCA TG6 may be also involved in the development of other neurodegenerative diseases.
In summary, we have described the detailed TG6 expression pattern in adult mouse CNS for the first time, and our data provide a morphological basis for studying the role of TG6 in the nervous system and neurodegenerative disorders, especially for exploring the pathogenesis of SCA35 in the future.