Skip to main content
NCBI home page
The FASEB Journal logoLink to The FASEB Journal
. 2010 Mar;24(3):844–852. doi: 10.1096/fj.09-143974

Neuronal selenoprotein expression is required for interneuron development and prevents seizures and neurodegeneration

Eva K Wirth *, Marcus Conrad §, Jochen Winterer †, Christian Wozny †, Bradley A Carlson ‖, Stephan Roth *, Dietmar Schmitz †, Georg W Bornkamm §, Vincenzo Coppola , Lino Tessarollo , Lutz Schomburg ‡, Josef Köhrle ‡, Dolph L Hatfield ‖, Ulrich Schweizer *,‡,1
PMCID: PMC2830140  PMID: 19890015

Abstract

Cerebral selenium (Se) deficiency is associated with neurological phenotypes including seizures and ataxia. We wanted to define whether neurons require selenoprotein expression and which selenoproteins are most important, and explore the possible pathomechanism. Therefore, we abrogated the expression of all selenoproteins in neurons by genetic inactivation of the tRNA[Ser]Sec gene. Cerebral expression of selenoproteins was significantly diminished in the mutants, and histological analysis revealed progressive neurodegeneration. Developing interneurons failed to specifically express parvalbumin (PV) in the mutants. Electrophysiological recordings, before overt cell death, showed normal excitatory transmission, but revealed spontaneous epileptiform activity consistent with seizures in the mutants. In developing cortical neuron cultures, the number of PV+ neurons was reduced on combined Se and vitamin E deprivation, while other markers, such as calretinin (CR) and GAD67, remained unaffected. Because of the synergism between Se and vitamin E, we analyzed mice lacking neuronal expression of the Se-dependent enzyme glutathione peroxidase 4 (GPx4). Although the number of CR+ interneurons remained normal in Gpx4-mutant mice, the number of PV+ interneurons was reduced. Since these mice similarly exhibit seizures and ataxia, we conclude that GPx4 is a selenoenzyme modulating interneuron function and PV expression. Cerebral SE deficiency may thus act via reduced GPx4 expression.—Wirth, E. K., Conrad, M., Winterer, J., Wozny, C., Carlson, B. A., Roth, S., Schmitz, D., Bornkamm, G. W., Coppola, V., Tessarollo, L., Schomburg, L., Köhrle, J., Hatfield, D. L., Schweizer, U. Neuronal selenoprotein expression is required for interneuron development and prevents seizures and neurodegeneration.

Keywords: selenium, parvalbumin, excitability, schizophrenia, γ oscillation

Selenocysteine (Sec) is the 21st proteinogenic amino acid in mammals. It had initially been overlooked when the genetic code was elucidated, because it is encoded by the stop codon UGA and is present in only a small number of proteins (1). This rare amino acid contains a selenium (Se) atom in place of the sulfur atom of cysteine. The human and mouse genomes contain 25 and 24 genes encoding selenoproteins, respectively (2). Since the discovery of Se as an essential trace element in rats (3), Se deficiency has been related to numerous disorders in humans (4). Glutathione peroxidases (GPxs), thioredoxin reductases (Txnrds), and iodothyronine deiodinases (Dios) were the first mammalian selenoenzymes to be identified; hence, virtually all biological actions of Se were initially explained by their actions on peroxide degradation and thyroid hormone metabolism, respectively. Recently, selenoproteins have been implicated in protein folding, degradation of misfolded membrane proteins, and control of cellular calcium homeostasis, all processes known to be deregulated in neurodegenerative diseases (5). Dietary Se restriction in animals does not lead to spontaneous neurological damage, because the brain retains its Se during dietary Se shortage (6, 7). However, targeted inactivation of the main plasma Se transport protein, selenoprotein P (SePP), demonstrated that SePP is epistatic to brain Se content and brain selenoenzyme expression (8). Depending on dietary Se supply, Sepp−/− mice die prematurely before weaning and display seizures and ataxia (9). Recently, the lipoprotein receptor ApoER2 was identified as a SePP receptor in brain, and ApoER2−/− mice suffer neurodegeneration similar to Sepp-deficient mice when fed an Se-deficient diet (10, 11). In aggregate, these studies demonstrate that SePP/ApoER2-mediated Se uptake into brain is essential for normal brain function in mammals.

It is still unclear which selenoproteins mediate Se function in the brain. Moreover, the brain is composed of different cell types, such as neurons, astrocytes, oligodendrocytes, microglia, and endothelium, and dysfunction of any of these cell types may lead to neurological impairment or neurodegeneration. To address the role of selenoproteins in brain function, we have abrogated neuronal selenoprotein expression by Cre recombinase-mediated deletion of the gene encoding tRNA[Ser]Sec (gene symbol Trsp). Because of elimination of tRNA[Ser]Sec, biosynthesis of all selenoproteins is simultaneously disrupted in targeted cells. We show that neuron-specific ablation of selenoprotein expression causes a neurodevelopmental and -degenerative phenotype affecting the cerebral cortex and hippocampus, specifically the parvalbumin (PV)-positive interneuron population. Given the selective loss of PV neurons in certain neuropsychiatric diseases, the class of selenoproteins should be evaluated for its possible role in human disease.

MATERIALS AND METHODS

Animals

Mice were maintained according to local regulations as described previously for the Sepp-deficient mice generated in our laboratory (12). All animal experiments were approved by the local authorities in Berlin and Munich. Conditional Trsp-knockout mice (Trspfl/fl) have been described previously (13). These mice were crossed with transgenic mice expressing Cre recombinase under control of the tubulinα1 promotor (14) or under control of the CamKIIα promotor (15), yielding mice deficient in neuronal selenoprotein biosynthesis, Tα1-Cre/Trspfl/fl, and CamK-Cre/Trspfl/fl, respectively. Conditional Gpx4 mice have been described previously (16). Mutant mice and littermate controls were analyzed between postnatal day (P)3 and P15. Electrophysiological studies were performed on P10.

Enzyme assays

Forebrains were freshly dissected from postnatal mice and immediately frozen on dry ice. Brain tissue was powdered under liquid nitrogen using a dismembrator (Braun Melsungen, Melsungen, Germany) and enzymatic activities assessed as described previously (12).

Western blot

Equal amounts of protein from tissue homogenates (50 μg) were applied to SDS-PAGE and electroblotted on PVDF membrane by standard techniques. Even transfer of proteins was verified by Ponceau staining of the blotted membranes and by immunostaining for Txnrd2 (ATLAS; rabbit polyclonal 1:5000). Antibodies against selenoprotein M (SePM, mouse monoclonal 1:5000) and GPx1 and GPx4 (rabbit polyclonal 1:1000) were from Abcam (Cambridge, MA, USA) or were made in our laboratories as described previously for selenoproteins (R, H, and T). Immunoreactive bands were visualized on X-ray film by chemiluminescence.

Primary neuron culture and neuronal survival

For in vitro cultures of cortical neurons, brains were dissected on embryonic day 15, and isolated neurons were cultured in neurobasal medium on poly-l-lysine and collagen-coated 12-well plates. Because the commercial supplement (B27; Invitrogen, Carlsbad, CA, USA) contains an unspecified amount of vitamin E and selenite, a self-made supplement based on earlier formulations (17) was prepared. Accordingly, the Se+ medium contained 83 nM sodium selenite, a concentration sufficient to allow maximal expression of cellular GPx activity (unpublished results). Based on the same report, we added to Vit E+ medium 1 μg/ml each of α-tocopherol and α-tocopherol-acetate, similar to the content of 10% calf serum conditions. Synthetic media with Se and vitamin E not added are expected to contain at best trace impurities of these agents. In vitro differentiation of cortical interneurons in culture proceeded for 17 d and was assessed by double immunofluorescent labeling for GAD67 and PV or calretinin (CR). Photomicrographs were evaluated for the fraction of PV+ cells among the GAD67+ cells according to previous work (18); in addition, the density of CR+ and GAD67+ cells per mm2 was quantified. All experiments were replicated with at least 3 independent neuron cultures and means calculated. Cell death was assessed by lactic acid dehydrogenase activity in the medium and compared with a commerical calibrator solution (Greiner). MTT oxidation by metabolically active cells was assessed as an alternative measure for cellular survival. Several animals of each genotype were always cultured in parallel as indicated in the figure legends.

Electrophysiology

Hippocampal slices were prepared from P10 wild-type and knockout mice as described previously (19). Field potential recordings were performed with low-resistance patch pipettes filled with external solution placed in stratum radiatum of hippocampal area CA1. Schaffer collaterals were extracellularly stimulated with low-resistance patch-pipettes filled with external solution placed in stratum radiatum of area CA1. Data were digitized at 5 kHz, recorded, and analyzed with custom-made software in IGOR Pro (WaveMetrics Inc., Lake Oswego, OR, USA). Local field potential recordings were performed at 32 ± 1°C in an interface-type recording chamber. Glass microelectrodes with tip diameter of ∼5 μm were filled with ACSF before use. Extracellular signal was amplified ×1000 and filtered at 1 Hz to 2 or 5 kHz.

Immunohistochemistry

Brains from mouse pups were immediately fixed after dissection in 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4, as described previously (20). Free-floating sections were stained with the indicated antibodies at dilutions of 1:1000–1:5000 at 4°C overnight. Polyclonal rabbit α-PV and rabbit α-CR antibodies were from Swant (Bellinzona, Switzerland), mouse monoclonal α-GFAP antibodies were from Sigma (St. Louis, MO, USA), and α-NPY, α-SOM, and α-NeuN antibodies were from Millipore (Billerica, MA, USA). Horseradish peroxidase and diaminobezidine substrate were used in conjunction with the Vectastain ABC kit (Vector, Burlingame, CA, USA). Photomicrographs were taken at a Zeiss Axioskop II equipped with AxioCam MRc and operated with Axiovision software (Carl Zeiss, Oberkochen, Germany). The area comprising the somatosensory cortex was determined using Axiovision software (2–4 mm2), and neuronal profiles within were counted in several sections and both hemispheres (up to n=106/area for CR+ and up to n=474/area for PV+).

Statistical analysis

Student’s t test was used to compare differences between 2 groups. Multiple comparisons were assessed using ANOVA followed by Dunnett’s post hoc test. Calculations were done using Excel (Microsoft, Redmond, WA, USA) or GraphPad Prism software (GraphPad, San Diego, CA, USA). Data are expressed as means ± se. Statistical significance was defined as P < 0.05, P < 0.01, or P < 0.001, as indicated.

RESULTS

Loss of selenoprotein expression in neuron-specific Trsp-knockout brain leads to neurodegeneration

Because selenoprotein biosynthesis is strictly dependent on the availability of tRNA[Ser]Sec, targeted deletion of Trsp abrogates expression of all selenoproteins (Fig. 1A). Neuron-specific Trsp-deficient mice (Tα1-Cre/Trspfl/fl) were born at Mendelian frequency and did not display any obvious phenotype at birth. However, starting from P6, growth and weight gain of mutant mice was delayed compared to littermates, although they were fostered by their mothers, and fur growth was normal (not shown). After P8, a neurological phenotype became apparent and mutant animals did not gain postural control (Fig. 1B). Tα1-Cre/Trspfl/fl mice rarely survived beyond P12. Selenoprotein expression was significantly reduced in the brains of Tα1-Cre/Trspfl/fl mice as compared to controls. For instance, cytosolic GPx activity, a marker for selenoprotein expression, was diminished to 68% of controls in the forebrains of Tα1-Cre/Trspfl/fl mice at P9 (Fig. 1C). Similarly, cytosolic Txnrd activity was significantly reduced in P9 Tα1-Cre/Trspfl/fl brains to 66% of controls (from 3.8±0.4 to 2.5±0.1 nmol·mg−1·min−1; mean ± se; n=5–7; P<0.01, Student’s t test). Expression of other selenoproteins appeared more restricted to neurons, as Western blot analysis revealed almost a complete loss of selenoproteins R and H from the brains of Tα1-Cre/Trspfl/fl mice (Fig. 1D). Cre-mediated recombination and subsequent neurodegeneration occurred also in brainstem and cerebellum of Tα1-Cre/Trspfl/fl mice (not shown).

Figure 1.

Figure 1.

Open in a new tab

Abrogation of neuronal selenoprotein biosynthesis leads to neurodegeneration. A) Schematic representation of selenoprotein biosynthesis. Sec is inserted in response to UGA codons in selenoprotein mRNAs, which are characterized by a selenocysteine insertion sequence (SECIS) element in the 3′-UTR. Genetic deletion (KO) of Trsp prevents translation of functional selenoproteins such as GPx and Txnrd, as indicated by the slash over tRNA[Ser]Sec. B) Lack of postural control and growth defect in P12 Tα1-Cre/Trspfl/fl mice. C) Cerebral activitiy of the prototype selenoenzyme cellular GPx in P9 Tα1-Cre/Trspfl/fl is significantly decreased (n=5–7). **P < 0.01; Student’s t test. D) Western blot analysis of selenoprotein expression in cerebral cortex. Selenoprotein expression is significantly reduced in homogenates from P12 Tα1-Cre/Trspfl/fl mice, suggesting that neurons are the predominant site of selenoprotein expression in the brain. Equal amounts of protein (50 μg) were applied to each lane (n=3–4 animals).

To restrict Cre-mediated recombination to cortical and hippocampal neurons and possibly prolong survival, we employed a second Cre-transgenic mouse strain. The neurological phenotype of CamK-Cre/Trspfl/fl mice was less severe than that of Tα1-Cre/Trspfl/fl mice. These mutants were able to walk, albeit they often tumbled and fell. However, their life span was only moderately extended, and they usually did not survive beyond P13–P15. At P14, cytosolic GPx activity in forebrain homogenates of CamK-Cre/Trspfl/fl mice was reduced to 70% of control and Txnrd activity to 58% of control (n=5–7; P<0.05, Student’s t test) paralleling the reductions in activity seen in Tα1-Cre/Trspfl/fl mice. When touched by a littermate or by the experimenter, CamK-Cre/Trspfl/fl mice went through a bout of uncontrolled, seizure-like movements as did Tα1-Cre/Trspfl/fl mice.

The overt behavioral phenotype of mice lacking neuronal selenoprotein expression was suggestive of major neuronal damage. To address this, we performed a histological analysis of CamK-Cre/Trspfl/fl mouse brains. At P14, Nissl staining revealed massive loss of neuronal cells in the cerebral cortex (Supplemental Fig. 1). Cortical neurons degenerated progressively from deeper to more superficial layers. Expression of the neuronal marker NeuN was lost in hippocampal pyramidal and interneurons as well as in cortical neurons (Supplemental Fig. 1). Astrogliosis, a stereotypic astrocytic reaction to neurodegeneration, followed the pattern of neuronal loss as judged by the analysis of GFAP immunoreactivity (Supplemental Fig. 1).

Impaired cortical interneuron development

A histological study using a number of neuronal markers revealed the absence of PV+ interneurons in cortex and hippocampus of Trsp mutant mice (Fig. 2 and not shown). We thus asked whether PV+ interneurons had already degenerated or whether the development of these interneurons was specifically impaired. Therefore, the development of major classes of interneurons was studied in the cerebral cortex from P8 to P15. In control mice, the first cortical PV+ cells were observed at P8, and their density increased from P11 to P15. In CamK-Cre/Trspfl/fl mice, expression of PV was not detectable at any time point investigated (Fig. 2A, E). This indicates that cortical PV+ interneuron development and/or maturation depends on selenoprotein expression. Interneurons expressing somatostatin-14 (SOM) or neuropeptide Y (NPY) are developmentally related to PV+ neurons and also arise in the medial ganglionic eminence. Their numbers, however, were similar in control and Trsp mutant cortex (Fig. 2B, C). Another class of GABAergic interneurons, CR+ cells, originates in the caudal ganglionic eminence and follows a different specification program. These cells were present in normal number and distribution in the cortex of Trsp mutants at P8, but they progressively degenerated thereafter (Fig. 2D, F). Similar results were found in the hippocampus, and PV+ neurons in the globus pallidus maintained strong PV expression at all time points investigated (not shown). We thus concluded that differentiation of PV+ interneurons is specifically impaired in the absence of selenoprotein expression.

Figure 2.

Figure 2.

Open in a new tab

Development of cortical interneurons is disrupted in Trsp-deficient brain. Expression of interneuron markers in somatosensory cortex in CamK-Cre/Trspfl/fl and control mice at P8, P11, and P15. A–D) Parvalbumin expression is not detected at any time point in Trsp-deficient mice (A), whereas other interneuron markers, such as somatostatin 14 (B), neuropeptide Y (C), and calretinin (D) appear normal. E, F) Quantitification of PV+ (E) and CR+ (F) neuronal profiles in somatosensory cortex, respectively. Solid bars, control; shaded bars, CamK-Cre/Trspfl/fl. **P < 0.01, ***P < 0.001; 2-sided Student’s t test. Scale bars = 100 μm (A–C); 200 μm (D).

Spontaneous epileptiform activity in mice lacking neuronal selenoproteins

Lack of PV+ interneurons or their dysfunction might be responsible for neurological findings such as uncoordinated seizure-like movements in CamK-Cre/Trspfl/fl mice. To investigate changes in neuronal excitability or network function, we performed electrophysiological recordings on hippocampal slices from P10 animals. At this age, CamK-Cre/Trspfl/fl mice already displayed a clearly visible neurological phenotype, although they did not exhibit tissue damage as determined by Nissl staining. To assess the strength of synaptic transmission, we compared the size of the presynaptic fiber volley (input) with the slope of the EPSP (output) in stratum radiatum of hippocampal area CA1. We found that synaptic transmission was not altered in CamK-Cre/Trspfl/fl mice (n=6) in comparison to littermate controls (n=9; Fig. 3A), confirming that tissue integrity in the mutant was still preserved at this age.

Figure 3.

Figure 3.

Open in a new tab

Spontaneous epileptiform activity in acute hippocampal slices in vitro. A) Unaltered input/output function of the CA1-fEPSP in Trsp mutant mice. Top panel: representative fEPSP traces recorded from 2 individual control and CamK-Cre/Trspfl/fl slices showing similarly increased responses to higher stimulation intensity applied in the stratum radiatum. Bottom panel; summary graph showing fEPSP slope vs. FV amplitude. B) Interictal discharges in hippocampal slices of CamK-Cre/Trspfl/fl mice. Representative traces of local field potentials in slices of control and Trsp mutant animals, recorded in stratum pyramidale of hippocampal CA1 in 10 μM carbachol (CCh). Bottom panel: magnification of representative traces. C) Summary plot of untreated and CCh treated slices from wild-type (n=10/group) and Trsp mutant animals (untreated, n=10; CCh treated, n=6).

We next addressed whether disruption of Trsp leads to alterations of hippocampal network oscillations. In untreated acute hippocampal slices of P10 animals, profound differences in network synchronization between CamK-Cre/Trspfl/fl and control mice were observed (Fig. 3B). In slices from mutant animals (n=10; n=3), we detected spontaneous interictal discharges (epileptiform activity) in 40% of the slices tested. In contrast, we did not observe any spontaneous epileptiform activity in control slices (n=10; n=3). Furthermore, we investigated the effects of cholinergic activation on the network rhythms in the hippocampus, which receives a major cholinergic input from the medial septum/diagonal band (21). Bath application of carbachol (5–20 μM) induced epileptiform activity in 66% of the slices of the CamK-Cre/Trspfl/fl mice, whereas in the control animals just one single slice showed interictal discharges on application of 20 μM carbachol (Fig. 3C). These electrophysiological findings suggest that deficiency of neuronal selenoproteins results in pathological network synchronization, compatible with compromised PV+ interneuron development.

Maturation of PV interneurons in vitro depends on Se and vitamin E

Given the influence of selenoproteins on PV expression in vivo, we wanted to investigate whether interneuron development in vitro is also modulated by selenoproteins. Because Trsp-deficient neurons do not survive for a sufficiently long period of time in culture (see Supplemental Fig. 2), we cultured wild-type neurons in the presence or absence of Se and vitamin E. After 17 d in vitro (DIV), cultures were stained for PV, GAD67, and CR (Fig. 4A). Withdrawal of Se from the medium alone did not significantly reduce the fraction of PV+ interneurons in the presence of high vitamin E concentrations; however, deprivation of Se and vitamin E significantly reduced the number of PV-expressing GAD67+ cells (Fig. 4B). Because the absolute density of GAD67+ cells was not influenced by Se and vitamin E, cell death cannot account for the reduced PV+ expression among interneurons (Fig. 4C). As a further control, we quantified the density of CR+ cells, some of which are GAD67+ interneurons, but most comprise cholinergic interneurons. Again, the density of CR+ cells was not changed by Se or vitamin E in the culture medium (Fig. 4D). We thus concluded that the loss of PV+ interneurons may not represent cell death, but reflect a failure to properly differentiate into the mature PV+ phenotype. The interaction of Se and vitamin E pointed to a possible role of the selenoenzyme phospholipid hydroperoxide GPx4 in the maturation of PV+ cells.

Figure 4.

Figure 4.

Open in a new tab

In vitro interneuron differentiation depends on Se and vitamin E (Vit E) in the culture medium. A) Representative images of PV+, GAD67+, and CR+ interneurons in cortical neuron culture at 17 d in culture. B) Fraction of PV+ cells among GAD67+ interneurons is reduced in cultures deprived of both Se and Vit E. C, D) Density of GAD67+ interneurons (C) and CR+ neurons (D) does not depend on Se and Vit E. Data are means of 3 independent cultures performed in triplicate. +, 83 nM selenite or 1 μg/ml each of α-tocopherol and α-tocopherol-acetate; −, exclusion of selenite or Vit E. **P < 0.01; ANOVA with Dunnett’s post hoc test.

Neuronal Gpx4 mutants partially resemble Trsp-mutant mice

GPx4 is a critical regulator of cellular survival and interferes with a lipoxygenase-dependent apoptotic pathway (16). Gpx4-deficient mice do not survive until birth (16, 22). To assess the role of Gpx4 in brain development in relation to a total loss of neuronal selenoproteins, we generated CamK-Cre/Gpx4fl/fl mice and compared them with CamK-Cre/Trspfl/fl mice. Gpx4-deficient mice displayed the same growth defect as Trsp-mutant mice (not shown). Their neurological phenotype was milder, but CamK-Cre/Gpx4fl/fl mice were still hyperexcitable and displayed an awkward gait. Nissl staining at P13 revealed that cell death was mostly concentrated in the CA3 region of the hippocampus as opposed to a complete degeneration of hippocampal neurons as observed in Trsp-deficient mice (Fig. 5A, B). Nevertheless, the number of PV+ interneurons was significantly reduced in the somatosensory cortex of CamK-Cre/Gpx4fl/fl mice (Fig. 5C, D, F), higher than in Trsp-mutant mice (Fig. 2). As in Trsp-deficient animals, this developmental defect of prospective PV+ interneurons was subtype specific, because the density of CR+ cortical interneurons was not reduced in Gpx4-mutant mice (Fig. 5E, F). The normal number of CR+ cells in Gpx4-deficient mice as opposed to Trsp-deficient mice further demonstrated the greater tissue damage in Trsp-mutant animals. This finding suggested that, in addition to GPx4, at least one more selenoprotein is essential for neurons. Since we have shown that Gpx4-deficient neurons and fibroblasts can be rescued from cell death by increased vitamin E in the culture medium (16), we tested whether vitamin E is also able to rescue Trsp-mutant neurons in culture. Survival of Trsp mutant neurons in vitro was severely compromised after 9 DIV, but could be rescued by addition of vitamin E (Supplemental Fig. 2A, B). In contrast, vitamin E was not able to sustain Trsp-deficient neurons for 14 DIV (Supplemental Fig. 2C, D). This finding is compatible with the notion that GPx4 is not the only essential selenoprotein in neurons.

Figure 5.

Figure 5.

Open in a new tab

Neuron-specific Gpx4 inactivation resembles Trsp inactivation. A) Hippocampal cell death at P13 is limited to CA3 in CamK-Cre/Gpx4fl/fl, but widespread inCamK-Cre/Trspfl/fl (arrows). B) Somatosensory cortex. C) Selective reduction of the density of parvalbumin-positive interneurons in CamK-Cre/Gpx4fl/fl mice at P13. D) magnified view of C. E) Density and survival of CR+ interneurons is not reduced on genetic inactivation of Gpx4. F) Quantification of PV+and CR+interneurons in somatosensory cortex at P13. Solid bars, control; open bars, CamK-Cre/Gpx4fl/fl. ***P < 0.001; 2-sided Student’s t test. Scale bars = 200 μm (A–C, E); 50 μm (D).

DISCUSSION

The aim of this study was to define whether neurons require selenoproteins, try to characterize the pathomechanism of the neurological phenotype, and possibly identify which selenoprotein is crucial for the observed phenotypes. Here we demonstrate that selenoproteins are essential for neuronal function in vitro and in vivo. Beyond a simple role for neuronal survival, we provide evidence that selenoprotein function is particularly important for the function and differentiation of cortical inhibitory PV+ interneurons, and we show that isolated deficiency in Gpx4 alone is sufficient to provoke the interneuron phenotype in vivo.

Because mice deficient in the Se transport protein, SePP, or its receptor ApoER2 develop signs of neurodegeneration, including ataxia and seizures (9, 11, 12, 23,24,25), it is possible that reduced neuronal GPx4 activity may be involved in their neurological phenotypes. SePP is not only a plasma protein, but expressed in the human brain and represents a constituent of cerebrospinal fluid (26). Moreover, early studies on primary cortical neuron culture identified SePP as a neurotrophic activity from fetal calf serum (27, 28). In our hypothetical model, SePP/ApoER2 provides Se for neuronal selenoprotein biosynthesis. If this Se uptake mechanism is impaired, GPx4 biosynthesis may fall under a pathological threshold.

Recently we have shown at the transcript level that cerebral cortex, hippocampus, cerebellum, and olfactory bulb express the highest numbers of selenoproteins (29). In the current study, we confirm on the protein level that, within the brain, several selenoproteins are primarily expressed in neurons, and on neuronal Trsp deletion, GPx4, SePH, and SePR fall below the limits of detection. Accordingly, GPx and Txnrd activities are reduced to similar extents in Tα1-Cre/Trspfl/fl and CamK-Cre/Trspfl/fl mice. The incomplete loss of ubiquitous GPx and Txnrd activities in mutant brains suggests that non-neuronal cells, which are not affected directly by neuron-specific Trsp ablation, contribute significantly to whole-brain GPx and Txnrd activity. It remains an important question which selenoproteins are essential or important for brain development and function and in which molecular pathways they function. Because it has been proposed that altered redox metabolism is a causative factor in the etiology of various neurodegenerative disorders, knockout mice for Gpx1, the classical and most abundant glutathione peroxidase in mammals, were generated. Although Gpx1 was shown to protect against ischemic and neurotoxic insults, Gpx1−/− mice did not develop spontaneous neurological deficits (reviewed in refs. 6, 30). Functional redundancy with non-Se-dependent enzymes may be one reason for this observation. Other selenoproteins, such as deiodinases, play a surprisingly small role in brain function as shown by targeted disruption of their genes, Dio1–3. We have therefore genetically inactivated the essential selenoproteins, Txnrd1 and Txnrd2, in neurons and in neural progenitors. Only deletion of Txnrd1 in neural progenitors resulted in a detectable phenotype, cerebellar hypoplasia, and a lamination defect (31). Neuron-specific Gpx4 mutants resemble Trsp-mutant mice in many distinctive aspects. However, it is evident that Trsp deficiency causes more severe phenotypes than Gpx4 deficiency alone. For example, neurodegeneration in cortex and hippocampus is more advanced and more widespread in Trsp- than in Gpx4-mutant mice. Similarly, the survival of Gpx4-deficient, but not Trsp-deficient, neurons in culture can be rescued by addition of vitamin E. We speculate, therefore, that GPx4 is not the only essential selenoprotein in neurons. Interestingly, several cell types can apparently survive and function in the absence of selenoprotein expression: breast cells (13), hepatocytes (32), and podocytes (33). Therefore, our findings imply that selenoproteins are not mere housekeeping genes, but exert essential and specific functions for some cell types, including neurons.

A most intriguing finding of this study is the specific delay or block of PV+ cortical interneuron development. These interneurons share their developmental origin with NPY+ and SOM+ interneurons (34), whose numbers and timing of appearance are apparently not altered in CamK-Cre/Trspfl/fl mice. A developmental block or selective degeneration before maturation of the PV+ class of GABAergic interneurons is consistent with the functional impairment of the inhibitory system and spontaneous epileptiform activity in vitro. The unchanged slope of EPSP over the amplitude of the fiber volley suggested that the glutamatergic cells responded normally and again pointed to a dysfunctional GABAergic system. The electrophysiological findings were supported by the apparent hyperexcitable and jerky phenotype of the mutant mice. Recently conditional inactivation of the transcription factor Nkx2.1, which is expressed in cortical interneuron precursors, revealed a selective loss of PV+ cortical neurons in association with seizures in juvenile mice (35). Hyperexcitability and seizures are also features of Sepp-deficient mice and patients with childhood epilepsy associated with Se deficiency (36, 37). Mice carrying a hypomorphic allele of Trsp are also hyperexcitable and have reduced numbers of PV+ neurons in the cerebral cortex (38).

PV+ neurons, in particular, are sensitive to oxidative stress, and induction of oxidative stress with the NMDA receptor antagonist, ketamine, leads to loss of PV expression and impaired γ oscillations (18, 39). In schizophrenia, γ oscillations are impaired, and the number of PV+ interneurons is reduced (40). Interestingly, the levels of glutathione, the major cofactor of GPx4 and other glutathione-dependent enzymes, are reduced in the brains of schizophrenic patients, and mutations in enzymes involved in glutathione biosynthesis are frequent in schizophrenic patients (41). Our finding of a selective developmental deficit of PV+ interneurons in Gpx4-mutant mice is thus consistent with a specific role of GPx4 for PV expression in interneurons. Our model, however, cannot rule out a paracrine mechanism in which (GPx4-deficient) cortical neurons express less BDNF, a known factor promoting PV interneuron maturation (42). To test this hypothesis, one would need to inactivate Gpx4 in an interneuron-specific manner.

The maturation of PV+ interneurons is also delayed in hypothyroid rodents (43) and in mice expressing a dominant negative thyroid hormone receptor (44). Because all deiodinases are selenoenzymes, one may argue that Trsp-deficient mice are hypothyroid, and therefore interneuron maturation is impaired. However, Gpx4 deficiency replicates the cortical interneuron phenotype of Trsp-deficient mice, thereby rendering deiodinase deficiency unlikely as the primary cause for the interneuron phenotype.

How can redox state or lipid peroxidation influence interneuron differentiation? Elucidating the precise mechanism will require further study, but it is known that the differentiation of neural precursor cells is modulated by cellular redox state (45). Recently it was shown that activity of the redox sensing histone deacetylase Sirt1 reduces proliferation and promotes astroglial differentiation of neural progenitor cells (46). It remains to be shown whether PV+ interneuron differentiation is blocked by Sirt1 action and whether Sirt1 can sense lipid peroxides. We have recently reported that GPx4 acts in a 12/15-lipoxygenase apoptotic pathway in fibroblasts, and we blocked cell death in the absence of GPx4 with a lipoxygenase inhibitor (16). However, primary neurons do not tolerate the inhibitor at effective concentrations, and thus we were not able to probe the role of this pathway in interneuron differentiation.

Given the essential roles of selenoproteins in neurons, the question arises whether metabolic disturbances of Se utilization or selenoprotein biosynthesis may contribute to neurodegenerative disease in humans. The epileptic phenotype of Sepp-deficient mice suggests that a moderate reduction of brain Se content suffices to impair brain function. We can only speculate whether a slight reduction in brain Se content over decades may precipitate neurodegeneration in humans.

Supplementary Material

Supplemental Data
fj.09-143974_index.html (698B, html)

Acknowledgments

The authors thank Dr. Claudia Iserhot, who performed initial physiological studies. SiJie Zhang performed some of the neuron cultures. CamK-Cre mice were generously provided by Günther Schütz (Deutsches Krebsforschungszentrum, Heidelberg, Germany). The authors gratefully acknowledge the technical assistance of Vartitér Seher, Anita Kinne, Silke Kappler, Antje Kretschmer, and Heidi Förster. Dorette Freyer (Experimental Neurology, Charité Universitätsmedizin Berlin), initially helped with primary neuron culture. Funding for this study was provided by Deutsche Forschungsgemeinschaft Scho849/2–2, SFB 665/A7, Ko922/11–1, and Charité Universitätsmedizin Berlin and in part by the Intramural Research Program of the U.S. National Institutes of Health, National Cancer Institute, Center for Cancer Research.

References

  1. Hatfield D L, Gladyshev V N. How selenium has altered our understanding of the genetic code. Mol Cell Biol. 2002;22:3565–3576. doi: 10.1128/MCB.22.11.3565-3576.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Kryukov G V, Castellano S, Novoselov S V, Lobanov A V, Zehtab O, Guigo R, Gladyshev V N. Characterization of mammalian selenoproteomes. Science. 2003;300:1439–1443. doi: 10.1126/science.1083516. [DOI] [PubMed] [Google Scholar]
  3. Schwarz K, Foltz C M. Selenium as an integral part of factor 3 against dietary necrotic liver degeneration. Nutrition. 1957;15:255–264. [PubMed] [Google Scholar]
  4. Rayman M P. The importance of selenium to human health. Lancet. 2000;356:233–241. doi: 10.1016/S0140-6736(00)02490-9. [DOI] [PubMed] [Google Scholar]
  5. Andersen J K. Oxidative stress in neurodegeneration: cause or consequence? Nat Med. 2004;10:S18–S25. doi: 10.1038/nrn1434. [DOI] [PubMed] [Google Scholar]
  6. Schweizer U, Schomburg L, Savaskan N E. The neurobiology of selenium: lessons from transgenic mice. J Nutr. 2004;134:707–710. doi: 10.1093/jn/134.4.707. [DOI] [PubMed] [Google Scholar]
  7. Behne D, Hilmert H, Scheid S, Gessner H, Elger W. Evidence for specific selenium target tissues and new biologically important selenoproteins. Biochim Biophys Acta. 1988;966:12–21. doi: 10.1016/0304-4165(88)90123-7. [DOI] [PubMed] [Google Scholar]
  8. Burk R F, Hill K E. Selenoprotein P: an extracellular protein with unique physical characteristics and a role in selenium homeostasis. Annu Rev Nutr. 2005;25:215–235. doi: 10.1146/annurev.nutr.24.012003.132120. [DOI] [PubMed] [Google Scholar]
  9. Schweizer U, Michaelis M, Köhrle J, Schomburg L. Efficient selenium transfer from mother to offspring in selenoprotein-P-deficient mice enables dose-dependent rescue of phenotypes associated with selenium deficiency. Biochem J. 2004;378:21–26. doi: 10.1042/BJ20031795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Olson G E, Winfrey V P, Nagdas S K, Hill K E, Burk R F. Apolipoprotein E receptor-2 (ApoER2) mediates selenium uptake from selenoprotein P by the mouse testis. J Biol Chem. 2007;282:12290–12297. doi: 10.1074/jbc.M611403200. [DOI] [PubMed] [Google Scholar]
  11. Burk R F, Hill K E, Olson G E, Weeber E J, Motley A K, Winfrey V P, Austin L M. Deletion of apolipoprotein E receptor-2 in mice lowers brain selenium and causes severe neurological dysfunction and death when a low-selenium diet is fed. J Neurosci. 2007;27:6207–6211. doi: 10.1523/JNEUROSCI.1153-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Schomburg L, Schweizer U, Holtmann B, Flohe L, Sendtner M, Köhrle J. Gene disruption discloses role of selenoprotein P in selenium delivery to target tissues. Biochem J. 2003;370:397–402. doi: 10.1042/BJ20021853. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kumaraswamy E, Carlson B A, Morgan F, Miyoshi K, Robinson G W, Su D, Wang S, Southon E, Tessarollo L, Lee B J, Gladyshev V N, Hennighausen L, Hatfield D L. Selective removal of the selenocysteine tRNA [Ser]Sec gene (Trsp) in mouse mammary epithelium. Mol Cell Biol. 2003;23:1477–1488. doi: 10.1128/MCB.23.5.1477-1488.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Coppola V, Barrick C A, Southon E A, Celeste A, Wang K, Chen B, Haddad E, Yin J, Nussenzweig A, Subramaniam A, Tessarollo L. Ablation of TrkA function in the immune system causes B cell abnormalities. Development. 2004;131:5185–5195. doi: 10.1242/dev.01383. [DOI] [PubMed] [Google Scholar]
  15. Casanova E, Fehsenfeld S, Mantamadiotis T, Lemberger T, Greiner E, Stewart A F, Schutz G. A CamKIIalpha iCre BAC allows brain-specific gene inactivation. Genesis. 2001;31:37–42. doi: 10.1002/gene.1078. [DOI] [PubMed] [Google Scholar]
  16. Seiler A, Schneider M, Förster H, Roth S, Wirth E K, Culmsee C, Plesnila N, Kremmer E, Radmark O, Wurst W, Bornkamm G W, Schweizer U, Conrad M. Glutathione peroxidase 4 senses and translates oxidative stress into 12/15-lipoxygenase dependent- and AIF-mediated cell death. Cell Metab. 2008;8:237–248. doi: 10.1016/j.cmet.2008.07.005. [DOI] [PubMed] [Google Scholar]
  17. Brewer G J, Cotman C W. Survival and growth of hippocampal neurons in defined medium at low density: advantages of a sandwich culture technique or low oxygen. Brain Res. 1989;494:65–74. doi: 10.1016/0006-8993(89)90144-3. [DOI] [PubMed] [Google Scholar]
  18. Kinney J W, Davis C N, Tabarean I, Conti B, Bartfai T, Behrens M M. A specific role for NR2A-containing NMDA receptors in the maintenance of parvalbumin and GAD67 immunoreactivity in cultured interneurons. J Neurosci. 2006;26:1604–1615. doi: 10.1523/JNEUROSCI.4722-05.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Breustedt J, Vogt K E, Miller R J, Nicoll R A, Schmitz D. Alpha1E-containing Ca2+ channels are involved in synaptic plasticity. Proc Natl Acad Sci U S A. 2003;100:12450–12455. doi: 10.1073/pnas.2035117100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Wirth E K, Roth S, Blechschmidt C, Holter S M, Becker L, Racz I, Zimmer A, Klopstock T, Gailus-Durner V, Fuchs H, Wurst W, Naumann T, Brauer A, de Angelis M H, Köhrle J, Gruters A, Schweizer U. Neuronal 3′, 3, 5-triiodothyronine (T3) uptake and behavioral phenotype of mice deficient in Mct8, the neuronal T3 transporter mutated in Allan-Herndon-Dudley syndrome. J Neurosci. 2009;29:9439–9449. doi: 10.1523/JNEUROSCI.6055-08.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fisahn A, Pike F G, Buhl E H, Paulsen O. Cholinergic induction of network oscillations at 40 Hz in the hippocampus in vitro. Nature. 1998;394:186–189. doi: 10.1038/28179. [DOI] [PubMed] [Google Scholar]
  22. Yant L J, Ran Q, Rao L, Van Remmen H, Shibatani T, Belter J G, Motta L, Richardson A, Prolla T A. The selenoprotein GPX4 is essential for mouse development and protects from radiation and oxidative damage insults. Free Radic Biol Med. 2003;34:496–502. doi: 10.1016/s0891-5849(02)01360-6. [DOI] [PubMed] [Google Scholar]
  23. Hill K E, Zhou J, McMahan W J, Motley A K, Atkins J F, Gesteland R F, Burk R F. Deletion of selenoprotein P alters distribution of selenium in the mouse. J Biol Chem. 2003;278:13640–13646. doi: 10.1074/jbc.M300755200. [DOI] [PubMed] [Google Scholar]
  24. Hill K E, Zhou J, McMahan W J, Motley A K, Burk R F. Neurological dysfunction occurs in mice with targeted deletion of the selenoprotein p gene. J Nutr. 2004;134:157–161. doi: 10.1093/jn/134.1.157. [DOI] [PubMed] [Google Scholar]
  25. Valentine W M, Abel T W, Hill K E, Austin L M, Burk R F. Neurodegeneration in mice resulting from loss of functional selenoprotein P or its receptor apolipoprotein E receptor 2. J Neuropathol Exp Neurol. 2008;67:68–77. doi: 10.1097/NEN.0b013e318160f347. [DOI] [PubMed] [Google Scholar]
  26. Scharpf M, Schweizer U, Arzberger T, Roggendorf W, Schomburg L, Köhrle J. Neuronal and ependymal expression of selenoprotein P in the human brain. J Neural Transm. 2007;114:877–884. doi: 10.1007/s00702-006-0617-0. [DOI] [PubMed] [Google Scholar]
  27. Yan J, Barrett J N. Purification from bovine serum of a survival-promoting factor for cultured central neurons and its identification as selenoprotein-P. J Neurosci. 1998;18:8682–8691. doi: 10.1523/JNEUROSCI.18-21-08682.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kaufman L M, Barrett J N. Serum factor supporting long-term survival of rat central neurons in culture. Science. 1983;220:1394–1396. doi: 10.1126/science.6857258. [DOI] [PubMed] [Google Scholar]
  29. Zhang Y, Zhou Y, Schweizer U, Savaskan N E, Hua D, Kipnis J, Hatfield D L, Gladyshev V N. Comparative analysis of selenocysteine machinery and selenoproteome gene expression in mouse brain identifies neurons as key functional sites of selenium in mammals. J Biol Chem. 2008;283:2427–2438. doi: 10.1074/jbc.M707951200. [DOI] [PubMed] [Google Scholar]
  30. Schweizer U, Bräuer A U, Köhrle J, Nitsch R, Savaskan N E. Selenium and brain function: a poorly recognized liaison. Brain Res Brain Res Rev. 2004;45:164–178. doi: 10.1016/j.brainresrev.2004.03.004. [DOI] [PubMed] [Google Scholar]
  31. Soerensen J, Jakupoglu C, Beck H, Förster H, Schmidt J, Schmahl W, Schweizer U, Conrad M, Brielmeier M. The role of thioredoxin reductases in brain development. PLoS ONE. 2008;3:e1813. doi: 10.1371/journal.pone.0001813. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Schweizer U, Streckfuss F, Pelt P, Carlson B A, Hatfield D L, Köhrle J, Schomburg L. Hepatically derived selenoprotein P is a key factor for kidney but not for brain selenium supply. Biochem J. 2005;386:221–226. doi: 10.1042/BJ20041973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Blauwkamp M N, Yu J, Schin M A, Burke K A, Berry M J, Carlson B A, Brosius F C, III, Koenig R J. Podocyte specific knock out of selenoproteins does not enhance nephropathy in streptozotocin diabetic C57BL/6 mice. BMC Nephrol. 2008;9:7. doi: 10.1186/1471-2369-9-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Wonders C P, Anderson S A. The origin and specification of cortical interneurons. Nat Rev Neurosci. 2006;7:687–696. doi: 10.1038/nrn1954. [DOI] [PubMed] [Google Scholar]
  35. Butt S J, Sousa V H, Fuccillo M V, Hjerling-Leffler J, Miyoshi G, Kimura S, Fishell G. The requirement of Nkx2–1 in the temporal specification of cortical interneuron subtypes. Neuron. 2008;59:722–732. doi: 10.1016/j.neuron.2008.07.031. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Weber G F, Maertens P, Meng X Z, Pippenger C E. Glutathione peroxidase deficiency and childhood seizures. Lancet. 1991;337:1443–1444. doi: 10.1016/0140-6736(91)93130-2. [DOI] [PubMed] [Google Scholar]
  37. Ramaekers V T, Calomme M, Vanden Berghe D, Makropoulos W. Selenium deficiency triggering intractable seizures. Neuropediatrics. 1994;25:217–223. doi: 10.1055/s-2008-1073025. [DOI] [PubMed] [Google Scholar]
  38. Carlson B A, Schweizer U, Perella C, Shrimali R K, Feigenbaum L, Shen L, Speransky S, Floss T, Jeong S J, Watts J, Hoffmann V, Combs G F, Gladyshev V N, Hatfield D L. The selenocysteine tRNA STAF-binding region is essential for adequate selenocysteine tRNA status, selenoprotein expression and early age survival of mice. Biochem J. 2009;418:61–71. doi: 10.1042/BJ20081304. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Behrens M M, Ali S S, Dao D N, Lucero J, Shekhtman G, Quick K L, Dugan L L. Ketamine-induced loss of phenotype of fast-spiking interneurons is mediated by NADPH-oxidase. Science. 2007;318:1645–1647. doi: 10.1126/science.1148045. [DOI] [PubMed] [Google Scholar]
  40. Lewis D A, Hashimoto T, Volk D W. Cortical inhibitory neurons and schizophrenia. Nat Rev Neurosci. 2005;6:312–324. doi: 10.1038/nrn1648. [DOI] [PubMed] [Google Scholar]
  41. Gysin R, Kraftsik R, Sandell J, Bovet P, Chappuis C, Conus P, Deppen P, Preisig M, Ruiz V, Steullet P, Tosic M, Werge T, Cuenod M, Do K Q. Impaired glutathione synthesis in schizophrenia: convergent genetic and functional evidence. Proc Natl Acad Sci U S A. 2007;104:16621–16626. doi: 10.1073/pnas.0706778104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Huang Z J, Kirkwood A, Pizzorusso T, Porciatti V, Morales B, Bear M F, Maffei L, Tonegawa S. BDNF regulates the maturation of inhibition and the critical period of plasticity in mouse visual cortex. Cell. 1999;98:739–755. doi: 10.1016/s0092-8674(00)81509-3. [DOI] [PubMed] [Google Scholar]
  43. Gilbert M E, Sui L, Walker M J, Anderson W, Thomas S, Smoller S N, Schon J P, Phani S, Goodman J H. Thyroid hormone insufficiency during brain development reduces parvalbumin immunoreactivity and inhibitory function in the hippocampus. Endocrinology. 2007;148:92–102. doi: 10.1210/en.2006-0164. [DOI] [PubMed] [Google Scholar]
  44. Wallis K, Sjogren M, van Hogerlinden M, Silberberg G, Fisahn A, Nordstrom K, Larsson L, Westerblad H, Morreale D E, Shupliakov O, Vennstrom B. Locomotor deficiencies and aberrant development of subtype-specific GABAergic interneurons caused by an unliganded thyroid hormone receptor alpha1. J Neurosci. 2008;28:1904–1915. doi: 10.1523/JNEUROSCI.5163-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Smith J, Ladi E, Mayer-Proschel M, Noble M. Redox state is a central modulator of the balance between self-renewal and differentiation in a dividing glial precursor cell. Proc Natl Acad Sci U S A. 2000;97:10032–10037. doi: 10.1073/pnas.170209797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Prozorovski T, Schulze-Topphoff U, Glumm R, Baumgart J, Schroter F, Ninnemann O, Siegert E, Bendix I, Brustle O, Nitsch R, Zipp F, Aktas O. Sirt1 contributes critically to the redox-dependent fate of neural progenitors. Nat Cell Biol. 2008;10:385–394. doi: 10.1038/ncb1700. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Data
fj.09-143974_index.html (698B, html)
fj.09-143974_09-143974SuppData.zip (2.2MB, zip)

Articles from The FASEB Journal are provided here courtesy of The Federation of American Societies for Experimental Biology

RESOURCES