The role of thioredoxin reductases in brain development

Abstract

The thioredoxin-dependent system is an essential regulator of cellular redox balance. Since oxidative stress has been linked with neurodegenerative disease, we studied the roles of thioredoxin reductases in brain using mice with nervous system (NS)-specific deletion of cytosolic (Txnrd1) and mitochondrial (Txnrd2) thioredoxin reductase. While NS-specific Txnrd2 null mice develop normally, mice lacking Txnrd1 in the NS were significantly smaller and displayed ataxia and tremor. A striking patterned cerebellar hypoplasia was observed. Proliferation of the external granular layer (EGL) was strongly reduced and fissure formation and laminar organisation of the cerebellar cortex was impaired in the rostral portion of the cerebellum. Purkinje cells were ectopically located and their dendrites stunted. The Bergmann glial network was disorganized and showed a pronounced reduction in fiber strength. Cerebellar hypoplasia did not result from increased apoptosis, but from decreased proliferation of granule cell precursors within the EGL. Of note, neuron-specific inactivation of Txnrd1 did not result in cerebellar hypoplasia, suggesting a vital role for Txnrd1 in Bergmann glia or neuronal precursor cells.

Figure 1. Txnrd1 expression is strongly reduced in Txnrd1-NS null mouse brain.

(A) Western blot analysis of cytoplasmic brain protein extracts, revealing strongly reduced Txnrd1 protein levels in brain tissue of Txnrd1-NS null mice as compared to control mice. 30 µg protein were separated per lane. α-tubulin served as loading control. (B) Thioredoxin reductase activities were significantly reduced in the cytosol of Txnrd1-NS null brain but not of control brain nor of liver tissue regardless of genotype.

Figure 2. NS-specific Txnrd1 knockout mice display restricted growth and striking movement disorders.

(A) Shown are a Txnrd1-NS null mouse (right) and a control littermate (left) at the age of 1.5 years. (B) Determination of body weight at various time points revealed impaired growth of Txnrd1-NS null mice compared to control littermates. Values are expressed as mean±SD. (C) Motor coordination performance was tested using the pole test. Shown is the time required to climb down the vertical pole on 5 consecutive trials. If a mouse fell down, slipped or was unable to climb down, a default value of 120 s was taken into account. (P = postnatal day).

Figure 3. Striking cerebellar hypoplasia in Txnrd1-NS null mice.

H&E staining of midline sagittal cerebellar sections from control (left column) and Txnrd1-NS null mice (right column) at embryonic stages E15.5, E18.5 and at postnatal days P1, P7 and P21 showed progressive differences in the foliation and formation of the molecular-, Purkinje cell- and granular layers during development of the cerebellum. In each photograph, the anterior part of the cerebellum is located to the left and the dorsal part to the top. At E15.5, no difference in cerebellar development is visible. At E18.5, the cerebellum is smaller and the formation of the primary fissure is slightly retarded in Txnrd1 null brains. The external granular layer (EGL), which is the source of cells for the granular layer, is hypoplastic particularly in the anterior cerebellar area in Txnrd1-NS null mice at P1. Higher magnification demonstrated clear differences in the EGL thickness at P1 and P7 as well as the ectopic localisation of Purkinje cells and the absence of a well-organized, trilaminar-stuctured cerebellum. Already from E18.5 onwards, the Txnrd1 null cerebella are smaller when compared to controls, and the laminar organisation of the mutant cerebellum becomes progressively more distorted in the anterior part. At P21, the posterior lobules X to VII of the knockout mice appear relatively well developed, whereas within lobules VI–V there is an abrupt transition from an apparently normally structured cortex to a disrupted cortex. Finally, in the abnormally structured anterior region (Lobules V–I) the IGL is missing and the PCs are ectopic. In addition, the IGL is missing and the PC are ectopically localised (abbreviations: EGL = External granular layer, PC = Purkinje cells, IGL = internal granular layer, Pf = primary fissure). Lobules are indicated by roman numerals I–X. Scale bar: E15,5: 100 µm; E18,5, P1: 200 µm; P7, P21: 0,5 mm.

Figure 4. Reduced proliferation in Txnrd1 null cerebellum.

(A) To visualize mitotic cells, immunohistochemial staining on paraffin-embedded sagittal sections of P7 cerebelli were performed with an anti-phosphorylated histone 3 (PH3) antibody. Sections were counterstained with Mayer's haemalaun. Scale bar: 200 µm. (B) Illustrated is the number of PH3-positive cells per visual field from E11.5 until completion of postnatal cerebellar development at P21. (C) Txnrd1-NS null mice show a reduced cerebellar size from P1 onwards. Dashed line indicates the measured area. (D) Adult null mice show a strongly reduced relative cell density of the granular layer in the anterior cerebellum and only a slight reduction in the posterior part. The number of Purkinje cells remained unaffected.

Figure 5. Glial disorientation, lobular fusion and ectopic localisation with reduced arborisation of Purkinje cells in anterior cerebellum of Txnrd1-NS null mice.

Immunohistochemical staining of paraffin-embedded sagittal cerebellar slides using anti-Nestin (P1), anti-GFAP, and anti-Calbindin (both P14) counterstained with Mayer's haemalaun. The anterior and posterior cerebellar areas, represented in the high power micrographs shown in the lower part of the figure, are indicated by a filled box or an open box, respectively. Bergman Glia, shown as anti-Nestin or anti-GFAP immunoreactive cells, is essential for neuronal migration during pre- and postnatal cerebellar development. Clew-like alignment and shortening of Bergman glia is found in the anterior cerebellum of the mutant in 1 day old mutants (upper centre) or 14 day old mutants (middle centre). Anti-GFAP immunoreactive Bergmann glial fibers are shorter, disoriented, and reduced in density in the affected anterior Txnrd1-NS null lobules compared to control animals which show a radial alignment of these cells towards the pial surface (middle left). Glial cells in the posterior area of the Txnrd1-NS null cerebellum appear normal (upper and middle right). Purkinje cells, which are immunoreactive with anti-Calbindin, are located in a monolayer and project their dendrites towards the cells of the molecular layer in control mice (lower left). In the dysmorphic anterior Txnrd1-NS null cerebellum, ectopic Purkinje cell bodies in numerous layers fill the merged lobules and show impaired dendritic arborisation (lower centre). Purkinje cells of the mutant posterior cerebellar area show the same features, although less pronounced (lower right). Scale bar: 25 µm, anti-Calbindin staining: 200 µm.

Figure 6. Mice with neuron-specific disruption of Txnrd1 are phenotypically indistinguishable from wild-type litter mates.

(A) Histological analysis of cerebellum revealed normal size and foliation pattern of the cerebellum of Tα1-Cre;Txnrd1 ^fl/fl mice (with neuron-specific disruption of Txnrd1, right panel) in comparison to control mice (left panel) at the age of 5 weeks. Purkinje cells are stained with an antibody against Calbindin. (B) Higher power magnifications from (A). Sections were counterstained with Nissl. (B) Reduced cytosolic thioredoxin reductase activity in neuron-specific Txnrd1 knockout mice. Residual activity most likely resides in glial cells. Scale bars: 250 µm