DARS2 protects against neuroinflammation and apoptotic neuronal loss, but is dispensable for myelin producing cells

Abstract

Although mitochondria are ubiquitous, each mitochondrial disease has surprisingly distinctly different pattern of tissue and organ involvement. Congruently, mutations in genes encoding for different mitochondrial tRNA synthetases result in the development of a very flamboyant group of diseases. Mutations in some of these genes, including aspartyl-tRNA synthetase (DARS2), lead to the onset of a white matter disease—leukoencephalopathy with brainstem and spinal cord involvement, and lactate elevation (LBSL) characterized by progressive spastic ataxia and characteristic leukoencephalopathy signature with multiple long-tract involvements. Puzzled by the white matter disease phenotypes caused by DARS2 deficiency when numerous other mutations in the genes encoding proteins involved in mitochondrial translation have a detrimental effect predominantly on neurons, we generated transgenic mice in which DARS2 was specifically depleted in forebrain-hippocampal neurons or myelin-producing cells. Our results now provide the first evidence that loss of DARS2 in adult neurons leads to strong mitochondrial dysfunction and progressive loss of cells. In contrast, myelin-producing cells seem to be resistant to cell death induced by DARS2 depletion despite robust respiratory chain deficiency arguing that LBSL might originate from the primary neuronal and axonal defect. Remarkably, our results also suggest a role for early neuroinflammation in the disease progression, highlighting the possibility for therapeutic interventions of this process.

Introduction

Although mitochondrial diseases are often multisystemic, brain and muscle are the most commonly affected tissues. Central nervous system (CNS) involvement in mitochondrial disorders is clinically heterogeneous, manifesting as epilepsy, stroke-like episodes, migraine, ataxia, spasticity, extrapyramidal or hypophysial abnormalities, bulbar dysfunction, psychiatric abnormalities or neuropsychological deficits. Leucoencephalopathy has also been added to this long list of symptoms (1). Leucoencephalopathy with brain stem and spinal cord involvement and high brain lactate (LBSL) is a childhood or juvenile-onset disorder clinically characterized by slowly progressive cerebellar ataxia and spasticity with dorsal column dysfunction (2). LBSL is defined by a characteristic magnetic resonance imaging showing signal abnormalities in the cerebral white matter, specific brain stem and spinal cord tracts. Additionally, there are spectroscopic findings of increased lactate in the abnormal white matter in almost all affected individuals (1–3).

LBSL is caused by recessive mutations in the DARS2 gene, which encodes the mitochondrial aspartyl-tRNA synthetase (1). DARS2 belongs to the group of mitochondrial tRNA synthases (mtARSs), which differ from their cytoplasmic counterparts (4). These enzymes are responsible for the first step in mitochondrial protein synthesis that involves covalently attaching an amino acid to its cognate tRNA in a process referred to as tRNA charging or loading (4). How mutations in this ubiquitously expressed gene cause a disorder in which tracts of the central and peripheral nervous systems (PNS) are selectively affected is unexplained. An increased number of patients with DARS2 mutations indicates that the phenotypic spectrum in LBSL is much wider than originally assumed. Adult-onset oligosymptomatic cases were described as well as patients with infantile onset, rapid neurological deterioration and early demise (5). Almost all patients are compound heterozygous for different DARS2 mutations of which one is almost invariably a splice site mutation in intron 2, upstream of exon 3 (1). As a consequence, exon 3 is not included in the messenger RNA, leading to a frameshift, premature stop and absence of a functional protein (6). Other mutations in DARS2 include deletions, non-sense, missense and splice site mutations, for which very little, if any, molecular data are available. More recently, another white matter disease characterized by hypomyelination (HBSL—hypomyelination with brain stem and spinal cord involvement) has been associated with the mutations in DARS2 gene (7). Intriguingly, a majority of other mitochondrial diseases caused by mutations that disrupt mitochondrial protein synthesis usually lead to syndromes with devastating consequences in neurons, while DARS2 mutations leading to LBSL or HBSL suggest a primary defect in white matter (1,7). To shed more light on this important question, we generated transgenic mice in which DARS2 was specifically depleted in forebrain-hippocampal neurons or myelin-producing cells.

Our analysis revealed that loss of DARS2 in adult neurons leads to strong mitochondrial dysfunction accompanied by an early inflammation response and progressive loss of cells. Remarkably, DARS2 deficiency in adult myelin-producing cells, despite clear signs of high level of mitochondrial dysfunction, had no effect on survival, inflammatory response or myelination.

Results and Discussion

We recently demonstrated that DARS2 is an essential protein needed for early mammalian development, as depletion of DARS2 in the whole body leads to lethality around the time of organogenesis (8). To analyse the role of DARS2 in diverse neural cell types, we generated two different models where conditional Dars2 mice (Dars2^fl/fl) were bred with mice that postnatally express the Cre recombinase specifically in neurons or in myelin producing cells. Neuron-specific DARS2-deficient mice were obtained by mating Dars2^fl/fl to mice expressing Cre recombinase under the calcium/calmodulin-dependent kinase II alpha promoter (CaMKIIα-Cre) resulting in Dars2^fl/fl; CaMKIIα-Cre mice (referred to as Dars2^NEKO mice). In this model, Cre recombinase is active in forebrain, hippocampus and striatum neurons from postnatal Day 14, with maximal recombination at postnatal Day 29 (9). To deplete DARS2 in myelin-producing cells, we used mice that express Cre recombinase under the control of the tamoxifen-inducible proteolipid 1 promoter (Plp1-CreERT), which allows temporal control of the recombination (10). The resulting Dars2^fl/fl; Plp-CreERT animals (referred to as Dars2^MYKO mice) were injected intraperitoneally with tamoxifen for five consecutive days at 4 weeks of age. Recently, we demonstrated high sensitivity and specificity of the Cre expression in oligodendrocytes using this protocol, after crossing Plp1-CreERT mice with a reporter transgenic line expressing a floxed, mitochondrially targeted YFP (ROSA26^+/SmY) (11). Using this specific timing for the activation of Cre-mediated recombination in PLP positive cells, we ensured that DARS2 depletion occurs at about the same time in both models.

Progressive degeneration of cortex and hippocampus owing to massive cell loss observed in DARS2^NEKO mice

Initially, we followed the development of phenotypes by monitoring the growth curves in both transgenic lines. We did not observe any change in the growth curves of Dars2^MYKO mice, while Dars2^NEKO animals had a lower weight gain already at the time of weaning, and this became more prominent around 15 weeks of age (Supplementary Material, Fig. S1A). The gross morphology and weight of Dars2^NEKO brains did not change until around 24-26 weeks of age when obvious loss of brain mass occurred, with a strong atrophy of the forebrain (Fig. 1A). Early on, most mice showed no obvious behavioral abnormalities, but around 27-28 weeks of age some mice started scratching their necks and faces, resulting in self-inflicted wounds, at which point mice had to be sacrificed (Supplementary Material, Fig. S1B). In the rest of Dars2^NEKO cohort, progressive nature of the phenotype led to a shorter lifespan of up to 32-34 weeks of age. In contrast, neither changes in gross appearance, body weight nor behavior in Dars2^MYKO mice were observed (up until 80 weeks of age when the mice were terminated).

Figure 1

Analyses of cortex and hippocampus degeneration in control (Ctrl) and Dars2^NEKO (NEKO) mice. (A) Brain weights of Ctrl and NEKO mice at the indicated age; representative photographs of isolated brains from both knockout and control mice at the indicated age. Bars present mean levels±SD. Statistical difference was calculated by Student’s t-test; **P<0.005, ***P<0.001 (n=3 − 5). (B) Representative Nissl staining (Scale bars: 500 μm) of hippocampal regions of Ctrl and NEKO mice at indicated age. Close-ups are for the indicated areas. (C) Quantification of brain cortical thickness [mm] at the indicated age in Ctrl and NEKO mice; bars represent mean levels ±SD. Statistical difference was calculated by Student’s t-test; **P<0.005, ***P<0.001 (n=3–5). (D) Representative transmission electron micrographs (Scale bar: 0.5 μm) of Ctrl and NEKO 29-week-old mice.

In agreement with previous results, no clear change in structure, morphology and size of Dars2^NEKO cortex and hippocampus was observed at 20 weeks of age. However, a significant decrease of the cortical thickness was detected at 28 weeks of age, accompanied by an apparent decrease in mean hippocampal area (Fig. 1B and C). Curiously, the number of cells within cortical and hippocampal area seemed increased (Fig. 1B), suggesting that a progressive atrophy is accompanied by cellular infiltration, most likely caused by an immune response. Conversely, a number of neurons within the stratum granulosum of the dentate gyrus (DG) or parts of the stratum pyramidale of the cornu ammonis (primarily CA1 and CA2) was prominently reduced in 28-week-old Dars2^NEKO mice (Supplementary Material, Fig. S1C and D). An increasing level of pyknotic cells with clear signs of chromatin condensation suggested an irreversible state of heading into cell death (Supplementary Material, Fig. S1C and E). Vacuolar lesions in the same brain regions of the 28-week-old Dars2^NEKO mice further implied cell loss (Supplementary Material, Fig. S1C–E). Transmission electron microscopy of the hippocampus area in the 28-week-old Dars2^NEKObrain revealed less dense mitochondria that presented an apparent loss of lamellar-shaped cristae as a clear sign of mitochondrial dysfunction in these cells (Fig. 1D).

Next, we used the TUNEL assay to detect nuclear DNA fragmentation, a hallmark of cells undergoing apoptosis. A massive increase in apoptosis occurred at 20 weeks of age (Fig. 2A and B;Supplementary Material, Fig. S2), although at this time no apparent morphological changes were detected in the same brain regions (Fig. 1B). Remarkably, apoptotic cells in distinct brain regions seem to appear at different dynamics; neurons in the second and third cortical layer of retrosplenial area (RS) undergo considerable apoptosis only around 20 weeks of age, while in the rest of the cortex massive apoptotic cell death is observed until 26 weeks of age (Fig. 2A and B;Supplementary Material, Fig. S2). At 28 weeks of age, almost no TUNEL-positive nuclei could be observed, suggesting that the majority of them had already been lost (Fig. 2A and B;Supplementary Material, Fig. S2).

Figure 2

Apoptotic cell death in cortical and hippocampal regions in control (Ctrl) and Dars2^NEKO (NEKO) mice. (A) Representative TUNEL staining of brain hemispheres (cortex and hippocampus regions) in Ctrl and NEKO mice at the indicated age. (B, C) Quantification of TUNEL positive cells per hemisphere in the region of cortex (B) and hippocampus (C) at the indicated age in Ctrl and NEKO mice. Bars present mean levels ±SD. Statistical difference was calculated by Student’s t-test; *P<0.05, ***P<0.001 (n=3).

In general, cortical neurons seem to be more vulnerable to the loss of DARS2 and mitochondrial translation than hippocampal neurons (Fig. 2A and 2C). The hippocampal CA1 region displayed a high occurrence of progressive apoptotic cells with increased age, while in the DG region, despite having a large number of COX deficient cells, only a few neurons seemed to undergo apoptosis. The variances in the cell death dynamics might be a result of differences in neuronal types and physiology, and therefore their reliance on mitochondrial respiratory function. Indeed, selective neuronal vulnerability to oxidative stress (12), Parkinson’s disease (13) or Alzheimer’s disease (14) was previously described suggesting distinctive metabolic profiles in different brain regions.

Strong mitochondrial deficiency and neuroinflammation precede loss of DARS2-deficient neurons

Severe respiratory chain deficiency was observed already at 15 weeks of age, and reached its peak in 20 weeks old Dars2^NEKO brains (Fig. 3A). The number of COX deficient cells markedly decreased from 24 weeks on, likely owing to the massive apoptosis of DARS2-deficient cells (Fig. 2A). As a result, 28-week-old brains had almost no COX(−)SDH(+) (Complex IV deficient/Complex II positive cells), but a severely reduced cortex. It also became apparent that, instead of cell bodies, Dars2^NEKO mice had holes throughout the hippocampus and cortex area (Fig. 3A). Analysis of mitochondrial respiratory chain (MRC) enzyme activities in neocortical samples demonstrated decreased activities of complexes containing critical mitochondrial DNA (mtDNA)-encoded subunits, in 28-week-old Dars2^NEKO mice (Fig. 3B). The observed alterations in the MRC levels and enzyme activities matched the reduction in the amount of fully assembled complex I (C I) and IV (C IV—COX) in Dars2^NEKO mice (Fig. 3C). The F1 subcomplex of complex V (C V) was also detected in Dars2^NEKO mitochondria, implying a defect in mitochondrial protein synthesis, as previously reported (8,15,16). Remarkably, at 28 weeks of age, mitochondrial de novo protein synthesis in Dars2^NEKO cortex appeared better than at 22 weeks of age (Fig. 3D). This is likely owing to vast loss of DARS2-deficient cells at 28 weeks, hence the majority of mitochondria in these samples stems from other cell types, unaffected by Cre-mediated deletion of Dars2.

Figure 3

Characterization of mitochondrial function in control (Ctrl) and Dars2^NEKO (NEKO) mice. (A) COX-SDH staining of brain hemispheres (cortex and hippocampus region) in Ctrl and NEKO mice at the indicated age. (B) Relative MRC complex activities in isolated cortex mitochondria of Ctrl and NEKO mice at 28 weeks of age. Bars present mean levels ±SD. Statistical difference was calculated by Student’s t-test; *P<0.05 (n=3). (C) Western blots for OXPHOS complexes (with indicated position of specific complexes) and representative Coomassie Brilliant Blue (CBB) stained BN-PAGE gel in cortex mitochondria from 28-week-old Ctrl and NEKO mice. (D) Representative gel of in organello translation of 22- and 28-week-old Ctrl and NEKO mice cortex mitochondria. CBB stained gel used as loading control.

The essential role of the OXPHOS function in neurons has been previously demonstrated in other mouse models. Tissue-specific depletion of Tfam in forebrain and hippocampus neurons (Tfam^L/L; CamKIIα-Cre) guided the reduction of mtDNA levels and strong respiratory chain deficiency from four months of age, resulting in a severe, progressive neurodegeneration and massive apoptosis (17). A complete loss of the complex I (C I) NDUFS4 subunit (Ndufs4^−/−) resulted primarily in a CNS defect and rapid development of a Leigh-like phenotype characterized by ataxia, blindness, retarded growth rate, lethargy and increased serum lactate, leading to premature death at about 7 weeks of age (18). Similarly, depletion of Cox10 in neurons (Cox10^L/L; CaMKIIα-Cre) rendered them respiratory-deficient already at 2 months of age in both the cortex and hippocampus areas, while the neurodegenerative phenotype progressed slowly until the ages of 8 and 12 months when the mice die (19,20). Typically, there is a time lag between the occurrence of OXPHOS deficiency and onset of neurodegeneration (17,20), during which neurons seems to compensate for the disruption of oxidative phosphorylation. Increased glycolysis and support from other cell types have been proposed to account for this (21). However, here we show that DARS2 depletion in neurons leads to a respiratory chain deficiency around 15 weeks of age that quickly culminates in the occurrence of massive cell death at 20 weeks, resulting in a severe brain atrophy (30% loss) within the next four weeks. Therefore, if existing, compensatory ATP supplementation by other cell types seems to be insufficient to maintain neuronal function for prolonged time periods in Dars2^NEKO mice.

In agreement with a strong OXPHOS defect and disturbed mitochondrial protein synthesis, increased levels of mitochondrial HSP70 chaperone and LONP1 protease were detected, suggesting higher activation of mitochondrial stress responses, possibly mitochondrial unfolded protein response (UPRmt), in 20-week-old Dars2^NEKO animals (Supplementary Material, Fig. S3A). Increased mitochondrial biogenesis might also contribute to the observed change, as suggested by higher SDH levels (Fig. 3C). Remarkably, the increase was statistically significant only in the hippocampus, possibly owing to a higher number of cells other than neurons in the isolated cortex samples (Supplementary Material, Fig. S3B and C).

Neuronal degeneration frequently leads to a local change in physiological conditions that in turn promotes an inflammatory response and activation of microglia, which are the resident immune cells of the nervous system. Indeed, we detected high levels of activated microglia in the cortical and hippocampal regions of Dars2^NEKO mice, starting from 15 weeks of age (Fig. 4A). The number of infiltrated microglia in the hippocampal CA1 and DG regions dramatically increased with age, with an almost complete loss of recognizable structure at 32 weeks (Fig. 4A and Supplementary Material, Fig. S4A). On the contrary, the level of activated microglia in the cortex (SC and RS) peaked at 20 weeks (Fig. 4A and Supplementary Material, Fig. S4A), coinciding with the highest levels of OXPHOS deficiency observed in these Dars2^NEKObrains. The decline of reactive microgliosis over time coincides with a more robust, apoptotic neuronal loss in the cortex layers than observed in the hippocampus (Fig. 4A and 2A). Reactive astrocytes were increasingly detected from 15 weeks of age in both, cortex and hippocampal region of Dars2^NEKO mice, further emphasizing the presence of a neuroinflammatory response (Fig. 4B and Supplementary Material, Fig. S4B). The highest level of GFAP positive cells was observed around 20 weeks, and was persistently present until the end of the lifespan at around 32 weeks (Fig. 4B and Supplementary Material, Fig. S4B).

Figure 4

Neuroinflammatory processes in cortical and hippocampal areas of control (Ctrl) and Dars2^NEKO (NEKO) mice. (A and B) Representative immunofluorescence merged Z-stack images of IBA1 positive microglia (A) and GFAP positive astrocytes (B) in somatosensory cortex (SC) and the hippocampal CA1 region in Ctrl and NEKO mice of the indicated age.

Certainly, neuroinflammation is increasingly recognized as a contribution to processes underlying neurodegeneration. Primarily activated in response to diseases, toxins or mechanical injury, neuroinflammation is strongly associated with demyelinating diseases (22). It is now clear that the microgliosis and astrogliosis primarily intended to clear up debris and support neurons in distress, when chronically activated, can convert from a beneficial program into a cytotoxic insult promoting neuronal death (22). This might be one of the mechanisms to explain the abrupt worsening of the Dars2^NEKO phenotype after 20 weeks of age when we could observe massive micro- and astrogliosis.

Remarkably, activated microglia were detected in both the cortex and hippocampus already at 15 weeks of age before any cell death could be observed, suggesting that respiratory-deficient cells emit specific signals leading to activation of neuroinflammatory processes. We previously showed that DARS2 depletion in heart and skeletal muscle leads to accumulation of unfolded/misfolded proteins that serve as the first signal causing activation of stress signals and systemic changes in metabolism (8). It is tempting to speculate that the buildup of unfolded and/or unassembled respiratory chain subunits owing to the DARS2 depletion, and therefore imbalanced mitochondrial de novo protein synthesis acts as an early signal for the cell autonomous activation of stress responses, which might also trigger neuroinflammatory response. This would be in agreement with previous studies where the activation of microglia coincided with the accumulation of misfolded proteins in Alzheimer’s disease, amyloid lateral sclerosis or Huntington’s disease (23).

DARS2 depletion in myelin-producing cells does not affect the number of oligodendrocytes nor myelin production, despite strong OXPHOS deficiency

In agreement with our initial observation, DARS2 depletion in adult myelin producing cells (Dars2^MYKO mice) did not affect cell survival and therefore cell numbers at either 18 or 28 weeks (Supplementary Material, Fig. S5A). Moreover, the level of myelin basic protein, which is one of the most abundant proteins in myelin sheaths, was unchanged in Dars2^MYKO mice (Supplementary Material, Fig. S5B). We also analyzed myelination in the PNS and found no difference in the thickness of myelin sheaths or appearance of Schwann cells (myelin producing cells in PNS) on both sciatic nerve and spinal cord (Supplementary Material, Fig. S5C).

Puzzled by the initial results, we wondered if DARS2 depletion causes respiratory chain deficiency in myelin producing cells as observed in neurons. Therefore, we analyzed OXPHOS function, again turning to COX-SDH staining. A large number of COX-deficient cells were detected in the striatum of Dars2^MYKO mice, indicating the presence of strong OXPHOS deficiency (Fig. 5A). In the corpus callosum, the largest white matter structure in the brain, we detected COX-deficient oligodendrocytes in Dars2^MYKO mice, and in agreement with results on PNS, no changes in the myelination (Fig. 5C). Furthermore, a very prominent stream of OXPHOS deficient axons could be detected in 15-week-old Dars2^NEKO mice (Fig. 5B), which, remarkably, also did not affect the myelination of the corpus callosum (Fig. 5C and D).

Figure 5

Characterization of mitochondrial function in control (Ctrl) and Dars2 ^MYKO (MYKO) mice. (A) Representative COX-SDH staining of brain striatum in Ctrl and MYKO mice at 18 weeks of age (Scale bars, 100 μm). (B) Representative COX-SDH staining of corpus callosum with the close-ups in NEKO and MYKO mice at 18 weeks of age. (C) Representative images of Gallyas’ myelin staining of the brain in 28-week-old Ctrl, MYKO and NEKO mice. (D) Representative immunofluorescence merged Z-stack images of MBP positive cells in corpus callosum of Ctrl and NEKO mice at the indicated age. (E) Simple composite phenotype scoring system of NEKO and MYKO mice compared with Ctrl at the indicated age. Bars represent mean±SD.

DARS2 deficiency in forebrain and hippocampal neurons induced strong OXPHOS deficiency and neuroinflammatory response (Fig. 4 and Supplementary Material, Fig. S4). Despite clear evidence of strong OXPHOS deficiency in oligodendrocytes, no signs of neuroinflammatory response were detected in Dars2^MYKO mice (Supplementary Material, Fig. S6).

Finally, using simple composite phenotype scoring system we tested to what extent DARS2 depletion affects motor skills in our two models (24). In accordance with previous results, Dars2^MYKO mice did not show any signs of motor dysfunction (Fig. 5E). In contrast, DARS2 deficiency in neurons led to progressive loss of motor skills characterized by tremor, ataxia and age-dependent kyphosis (Fig. 5E).

Our results demonstrate that DARS2 deficiency in forebrain and hippocampal neurons leads to devastating cell death and severe neurodegeneration preceded by an early neuroinflammatory response (Supplementary Material, Fig. S7). The loss of DARS2 in oligodendrocytes, despite leading to strong respiratory deficiency, does not trigger any of these responses on either a cellular or physiological level (Supplementary Material, Fig. S7). Remarkably, a recent study showed that the most common DARS2 patient mutation, located in a splice-site of intron 2, has the strongest effect causing exon 3 exclusion in neuronal cell lines (6). Therefore, it is tempting to conclude that LBSL in DARS2 patients is a consequence of a primary neuronal deficiency, as neurons seems to be more affected with the splice-site mutation and here we show that they are also much more sensitive to a Dars2 depletion than myelin producing cells. However, one has to keep in mind that our models induce the post-natal ablation of DARS2, whereas LBSL is typically owing to a hypomorphic allele which partially impairs the function of DARS2 throughout the embryonic and brain development. Therefore, a detrimental role of the mutation during oligodendrocyte-mediated myelinization of the developing brain cannot be excluded in the pathogenesis of the disease.

This is in agreement with a study showing that the most common DARS2 patient mutation located in a splice-site of intron 2, has the strongest effect causing exon 3 exclusion in neuronal cell lines (6).

Why would these two cell types react so differently to the loss of respiratory chain function? A different metabolic profile of adult oligodendrocytes and Schwann cells, enabling them to survive primarily using glycolysis for energy production when mitochondrial respiration is impaired, might be the main reason. In support of this hypothesis, Schwann-cell-specific depletion of Tfam (Tfam^L/L, P₀-Cre), the mitochondrial transcription factor A gene, which is essential for mtDNA transcription and maintenance, did not affect cell proliferation or survival, despite severe mtDNA depletion and respiratory chain abnormalities (25). Oligodendrocyte-specific deletion of Cox10 (Cox10^L/L; Cnp1-Cre), an essential assembly factor for complex IV, also did not lead to axonal degeneration, demyelination or cell death (26). In fact, recent data suggest that oligodendrocytes’ mitochondria may be essential for specialized functions relevant for myelin maintenance, such as lipid synthesis or fatty acid oxidation, rather than for ATP production (27). In agreement with this idea, a complete ablation of the m-AAA protease using the same promoter and Cre-induction paradigm as in this study (Afg3l1^−/−; Afg3l2^L/L; Plp1-Cre) resulted in progressive motor dysfunction and demyelination, owing to rapid oligodendrocyte cell death (11). In light of our results, this suggests that the imbalance of m-AAA protease substrates, other than the ones related to respiratory chain function, is detrimental for the myelin-producing cells of the central and PNS (11).

In summary, we provide strong in vivo evidence in support of the hypothesis that Dars2 dysfunction, causing LBSL through disturbances in the white matter of both central and PNS, arises from a primary neuronal or axonal deficiency and not from a defect of myelin-producing cells. Remarkably, our results also highlighted neuroinflammatory processes, which coincide with respiratory chain deficiency and precede neuronal cell loss as possible yet unexplored targets of therapeutic interventions in LBSL and potentially other mitochondrial diseases.

Experimental Procedures

Generation of Dars2 mice

Forebrain-specific DARS2-deficient mice (NEKO) were generated by mating Dars2^loxP/loxP animals with transgenic mice expressing Cre recombinase under the control of CaMKIIα promoter (9). MPC-specific DARS2-deficient mice (MYKO) were created also by using Dars2^loxP/loxP animals, mated with mice containing Plp1 driven CreER recombinase (10), whereas the activation of recombination was done by application of tamoxifen. Tamoxifen (T5648, Sigma) was dissolved in a corn oil/ethanol (9:1) mixture at a final concentration of 10 mg/ml. A total of 1 mg tamoxifen was administrated by intraperitoneal injection once a day for five consecutive days to 4 weeks old mice. All genotypes were acquired by PCR. Animal protocols were in accordance with guidelines for humane treatment of animals and were reviewed by the Animal Ethics Committee of the Nord-Rheine Westphalia, Germany.

Histological analysis

Mice were anesthetized with the combination of ketamine/xylazine, given intraperitoneally. Perfusion was performed intracardially with phosphate buffered saline (PBS) followed by 4% paraformaldehyde (PFA) in 0.1 M PBS (pH 7.4). Isolated brains were post-fixed in 4% PFA at 4 °C (for transmission electron microscopy analysis, post-fixed in 2% glutaraldehyde at 4 °C) overnight and then stored at 4 °C in 0.05% sodium azide-PBS (NaN₃-PBS) until further analysis. After perfusion, coronal sections were cut on Leica VT1200S with a thickness of 30–40 µm. Free floating sections were kept at 4 °C in 0.05% NaN₃-PBS.

For cryostat sections, freshly isolated brains were directly embedded in Tissue-Tek (Sakura) and frozen using dry ice. Coronal sections were cut on a Leica CM1850 cryostat with a thickness of 7 µm. Sections were directly mounted onto microscope slides and stored at −20 °C. Gallyas’ staining was performed as previously described (28).

Nissl staining was performed on vibratome sections, which were mounted on microscope slides using cromalin solution. The sections were incubated for 45 s in Nissl solution and washing step was followed by dehydration with different ethanol concentrations.

H&E staining was performed on cryosections using Mayer‘s Hematoxylin solution (Sigma), and sections were mounted with Entellan (Millipore).

TUNEL staining was performed on vibratome sections via the ApopTag Plus Peroxidase In situ Apoptosis Detection Kit (Millipore) according to the manufacturer’s protocol.

COX/SDH staining was performed on cryosections, with 40 min of incubation in Cytochrome C solution (0.8 ml 3, 3 diaminobenzidine tetrahydrochloride, 0.2 ml 500 µM cytochrome c, a few grains of catalase), and 40 min Succinate solution (0.8 ml 1.875 mM nitroblue tetrazolium, 0.1 ml 1.3 M sodium succinate, 0.1 ml 2 mM phenazine methosulphate, 0.01 ml 100 mM sodium azide) at 37 °C in a humid chamber. Sections were washed in PBS, dehydrated with increasing ethanol concentrations (75% for 2 min, 95% for 2 min, 100% for 10 min), air dried, and mounted in D.P.X. Light microscopy images were acquired using Leica SCN400 slidescanner. Cortical thickness and hippocampal areas were measured in coronal brain sections of the same rostro-caudal position in all animals. To quantify the number of neurons in the hippocampal DG and CA1 regions, neurons were counted in equally sized areas in the respective regions (n = 4 per mouse).

Immunostaining

Free-floating sections were washed extensively and pretreated for 30 min in 2% Triton X-100/PBS. Sections were permeabilized and blocked in 0.4% Triton X-100, 10% goat serum in PBS for 1 h at room temperature. Primary antibodies (APC 1:400, OP80, Calbiochem), MBP (1:1000, SMI94, Covance), (GFAP 1:2000, NeoMarkers), IBA1 (1:3000, WAKO Life Sciences) were incubated overnight in 0.4% Triton X-100, 5% goat serum in PBS at 4 °C. After washing with PBS, secondary antibodies anti-mouse Alexa Fluor 488 (1:2000, Invitrogen), anti-rabbit Alexa Fluor 546 (1:2000, Invitrogen) were applied in 5% goat serum in PBS for 2 h. Finally, the sections were washed in PBS and mounted using FluorSave Reagent (Calbiochem).

Light microscopy images were acquired by Leica SCN400 slide scanner (Leica Microsystems). Evaluation and processing was done with SlidePath Gateway Client software (version 2.0). Fluorescent images were acquired by an Axio-Imager M2 microscope, equipped with Apotome 2 (Zeiss). Images were processed with AxioVision software (version 4.8.2).

Transmission electron microscopy

Coronal semi-thin sections (1 µm) were cut from hippocampus. After treatment with osmium tetroxide, the tissue was embedded in Epon (Fluka). For ultrastructural analyses, the tissue of interest was selected for electron microscopy after examination of semi-thin sections by light microscopy. Ultrathin sections (70 nm) were cut, collected on 200 mesh copper grids (Electron Microscopy Sciences), and stained with uranium acetate (Plano GMBH) and lead citrate (Electron Microscopy Sciences).

Western blot analysis

Protein lysates were obtained from either homogenized tissue or isolated mitochondria, and subsequently subjected to western blot analysis as described previously (29). MEFs were lysed with the RIPA buffer followed by sonication and lysate clearance. Western blot analysis was performed using the following primary antibodies at indicated concentrations: mtHSP70 and LONP1 (1:1000, Abcam), Actin (1:5000, Santa Cruz), NDUFA9, SDHA, UQCRC1, MTCO1 (1:1000, Invitrogen), ATP5A1 (1:1000, MitoSciences). All secondary antibodies were purchased from Sigma Aldrich and used in 1:2000 dilution. Western blot quantification was performed using the ImageJ software.

Blue native polyacrylamide gel electrophoresis (BN-PAGE)

BN-PAGE was carried out using the NativePAGE Novex Bis-Tris Mini Gel system (Invitrogen) according to the manufacturer’s specifications. Proteins were transferred onto a PVDF (polyvinylidene fluoride) membrane, and immunodetection of mitochondrial OXPHOS complexes was performed (CI: NDUFA9; CII: SDHA; CIII: UQCRC1; CIV: MTCO1; CV: ATP5A1).

Analyses of de novo translation in isolated mitochondria

In organello translation was performed as previously described in mitochondria from heart and liver (8). In organello proteins synthesis was performed for 1 h in presence of ³⁵S-met at 37 °C. Mitochondria were lysed by incubation in SDS-PAGE loading buffer. Translation products were separated by SDS-PAGE, the gel was stained with Coomassie Brilliant Blue R-250; incubated in Amplify solution (GE Healthcare) and newly synthesized polypeptides were detected by autoradiography.

Respiratory chain enzyme activities

The measurements of respiratory chain enzyme complex activities and citrate synthase activity were performed as previously described (30).

Phenotype scoring system

The protocol for simple composite scoring system was done according to previously described protocol (24).

Statistical analysis

Two-tailed unpaired Student’s t test was used to determine statistical significance. Error bars represent standard deviation (SD). Unless otherwise indicated, all experiments were performed on three biological replicates.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

The work was supported by grants of the European Research Council (ERC- StG-2012-310700) and German Research Council (DFG - TR1018/3-1). M.A. received scholarships from Marie Curie ITN - Marriage. The authors wish to thank Alexandra Kukat for critically reading the manuscript and Katherine Dodel for language proofing.

Conflict of Interest statement. None declared.

Funding

This work was supported by grants from Fritz Thyssen Foundation (Az. 10.11.1.221) and German Research Council (DFG - TR 1018/3-1). M.A. obtained an Early Stage Researcher Fellowship through Marie Curie ITN “Marie Curie Aging Network” (316964).