ABSTRACT
l-2-Hydroxyglutarate aciduria (L-2-HGA) is an autosomal recessive neurometabolic disorder caused by a mutation in the l-2-hydroxyglutarate dehydrogenase (L2HGDH) gene. In this study, we generated L2hgdh knockout (KO) mice and observed a robust increase of l-2-hydroxyglutarate (L-2-HG) levels in multiple tissues. The highest levels of L-2-HG were observed in the brain and testis, with a corresponding increase in histone methylation in these tissues. L2hgdh KO mice exhibit white matter abnormalities, extensive gliosis, microglia-mediated neuroinflammation, and an expansion of oligodendrocyte progenitor cells (OPCs). Moreover, L2hgdh deficiency leads to impaired adult hippocampal neurogenesis and late-onset neurodegeneration in mouse brains. Our data provide in vivo evidence that L2hgdh mutation leads to L-2-HG accumulation, leukoencephalopathy, and neurodegeneration in mice, thereby offering new insights into the pathophysiology of L-2-HGA in humans.
INTRODUCTION
The rare, autosomal recessive neurometabolic disorders d-2-hydroxyglutaric aciduria (D-2-HGA) and l-2-hydroxyglutaric aciduria (L-2-HGA) are characterized by the accumulation of d-2-hydroxyglutarate (D-2-HG) and l-2-hydroxyglutarate (L-2-HG), respectively, in body fluids. Genetic characterization has shown that 50% of the D-2-HGA population and the majority of L-2-HGA patients harbor pathogenic mutations in D2HGDH and L2HGDH genes, respectively (1–3). The other half of D-2-HGA patients have a gain-of-function mutation in isocitrate dehydrogenase 2 (IDH2) at the residue of R140 (R140Q), which leads to abnormally high accumulation of D-2-HG (4). Based on phenotypic severity in patients, D-2-HGA is classified as mild type I (D2HGDH mutation) and severe type II (IDH2 mutation) (4, 5). Of note, 2-HG concentrations are 2- to 8-fold higher for type II D-2-HGA than type I D-2-HGA patients (4, 6, 7), suggesting that endogenous D2HGDH is insufficient to metabolize the excessive D-2-HG produced by mutated IDH2. Compared to D-2-HGA, L-2-HGA is more prevalent and severe and mainly affects the central nervous system (CNS) starting at childhood, leading to progressive hypotonia, tremor, epilepsy, leukoencephalopathy, mental retardation, psychomotor regression, and low-frequency brain tumors (8–10). The neurological symptoms in L-2-HGA patients may be due to the toxicity from L-2-HG accumulation, although the underlying mechanisms are largely unclear.
Emerging evidence has indicated that mitochondrial dysfunction and impairment of oxidative phosphorylation (OXPHOS) are involved in the pathology of various organic acidemias. Recently, 2-HG has been reported to impair the activities of two enzymes in OXPHOS, including cytochrome c oxidase and ATP synthase (11, 12), but whether the pathogenesis of 2-HGA is associated with impairment of OXPHOS remains unknown. Furthermore, 2-HG is now considered an oncometabolite (13–16), and numerous studies indicate that altered epigenetic regulation is a major mechanism underlying its oncogenic activity. The structural similarity of 2-HG to α-ketoglutarate (α-KG) enables it to act as an α-KG analog and inhibit the activity of multiple α-KG-dependent dioxygenases, including the JmjC domain-containing histone demethylases and the TET (ten-eleven translocation) family of cytidine hydroxylases (13, 17–19). Supporting this notion, a mouse Idh1-R132H knock-in study reveals increases of both DNA and histone methylations (3). In vitro studies demonstrate that L-2-HG is more potent than D-2-HG in suppressing the activity of α-KG-dependent dioxygenases (13, 17).
Although it has been over 30 years since the first report of L-2-HGA, the progression of the disease is poorly understood and little is known about its long-term impact on neural development and neural activity (20). Recently, Rzem et al. constructed an L2hgdh deletion mouse model and found extensive vacuolations in the central nervous system (CNS) (21). To study the biochemical and pathophysiological consequences of L-2-HG accumulation, we created L2hgdh knockout (KO) mice by piggyBac transposon-mediated insertion mutation of the L2hgdh gene. Our data demonstrate an age-dependent accumulation of L-2-HG in the cerebrum and alterations in a subset of histone methylations in the CNS of L2hgdh KO mice. Importantly, L2hgdh KO mice exhibit subcortical white matter abnormalities, recapitulating the typical clinical features of L-2-HGA patients. Moreover, L2hgdh KO mice exhibit dys/demyelination, extensive gliosis, expanded cell numbers of oligodendrocyte progenitor cells (OPCs), and microglia-mediated neuroinflammation. Finally, L2hgdh KO mice also show impaired adult hippocampal neurogenesis as well as age-dependent neurodegeneration.
RESULTS
L2hgdh deficiency causes 2-HG accumulation, reduced body weight, and premature death in mice.To understand the pathophysiology of L-2-HGA, we generated an L-2-HGA mouse model in which the L2hgdh gene was disrupted by insertion of a piggyback (PB) transposon between exon 1 and exon 2 (see Fig. S1 in the supplemental material). Western blot analysis demonstrated expression of L2ghdh protein across diverse mouse tissues (Fig. S2). Heterozygous insertion of the PB transposon led to reduced L2hgdh protein expression, whereas homozygous insertion resulted in a complete loss of L2hgdh protein in a variety of mouse tissues tested (Fig. 1A), confirming the deletion and functional inactivation of the L2hgdh gene.
Deletion of L2hgdh leads to 2-HG accumulation, reduced body weight, and premature death in mice. (A) L2hgdh protein is diminished in L2hgdh−/− mice. Western blotting of L2hgdh in tissues from cerebrum, testis, liver, kidney, and spleen of L2hgdh wild-type and piggyBac insertion heterozygous and homozygous mice. Beta-actin was used as a loading control (Ctr). (B and C) 2-HG is highly accumulated in urine (B) and plasma (C). 2-HG content in urine was measured by GC-MS and normalized to creatinine. (D) 2-HG accumulation in different tissues in L2hgdh KO mice. For each bar, tissues were collected from 3 or 4 mice at 2 to 3 months of age. (E) Temporal accumulation of 2-HG in postnatal mouse brains. Cerebral tissue at postnatal days 1, 3, 8, 14, 18, and 180 was collected for GC-MS analysis. (F) Age-dependent body weight loss in L2hgdh KO mice. For each time point and each genotype, n = 15. (G) Reduction of fat content in L2hgdh KO mice. Whole-body fat content was measured by a minispec (n = 8 to 10 per bar). (H) L2hgdh KO causes reduced locomotion in old mice. The open field test was applied to monitor mouse motor activity. Two age groups, 16 to 24 weeks and 45 to 55 weeks, were selected for testing. Each dot represents the movement of an individual mouse. (I) Premature death in L2hgdh KO mice. For wild-type and L2hgdh KO mice, n = 16 and 19, respectively. Quantifications are presented as means ± SEM and were analyzed by two-tailed t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant.
Measurement of 2-HG levels revealed that the urinary excretion of L-2-HG was approximately 6 to 8 mmol/mmol creatinine in L2hgdh KO mice, and that the plasma levels of L-2-HG were about 50 μM (Fig. 1B and C). The L-2-HG concentrations in body fluids of L2hgdh KO mice were much higher than those of wild-type mice and were even comparable with those in L-2-HGA patients (2). Moreover, examination of L-2-HG in tissues of L2hgdh KO mice demonstrated that this metabolite was most abundant in the brain and testis, with L-2-HG concentrations of 4 mmol/kg tissue weight (Fig. 1D). In other tissues of L2hgdh KO mice, 2-HG was accumulated to a lesser extent, ranging from 40 μmol/kg tissue weight in fat to 400 to 800 μmol/kg tissue weight (Fig. 1D). Notably, we found an age-dependent increase of L-2-HG level in the cerebrum of postnatal L2hgdh KO mice (Fig. 1E). The L-2-HG level in the cerebrum was about 1 mmol/kg tissue weight at neonatal day 1, gradually increased to about 4 mmol/kg tissue weight at day 14, and thereafter was maintained at a high steady-state level (Fig. 1E). Neonatal L2hgdh KO mice exhibited no overt structural or cellular defects in the brain (data not shown).
Strikingly, L2hgdh KO but not wild-type mice older than 6 months exhibited a significant loss in body weight, decreasing by 30% in 1-year-old L2hgdh KO mice compared to wild-type controls (Fig. 1F). Time domain-nuclear magnetic resonance (TD-NMR) analysis revealed that loss of body weight was largely attributed to reduced fat content in older L2hgdh KO mice (Fig. 1G). Furthermore, the motor activity of L2hgdh KO mice and their wild-type littermates was evaluated with the open field test (OFT), and no difference in the moving distance was found between young L2hgdh KO mice and their wild-type littermates aged 16 to 24 weeks (Fig. 1H). In contrast, the distance moved by older L2hgdh KO mice was significantly decreased compared to that of their wild-type littermates aged 45 to 55 weeks (Fig. 1H). We also found that the majority of L2hgdh KO mice (n = 18 of 19) died at the age of 48 to 72 weeks (12 to 18 months), while most of the wild-type littermates were still alive at the same age (Fig. 1I). Collectively, these data indicate that deletion of L2hgdh leads to reduced motor activity, reduced body weight, and premature death in mice.
L2hgdh deficiency results in testicular atrophy and selective changes in histone methylation.Our observation that L-2-HG was most abundant in the testis and brain of L2hgdh KO mice (Fig. 1D) prompted us to examine the pathophysiological consequences of L-2-HG accumulation in these two tissue types. We found that L2hgdh KO mice exhibited testicular atrophy and that this phenotype was exacerbated at an advanced age (Fig. 2A and B). Histological analysis showed that the seminiferous tubules had normal anatomy and spermatogenesis in L2hgdh KO mice compared to the wild-type controls (Fig. 2C). L2hgdh KO male mice were fertile but less potent than their wild-type littermates (data not shown). Together, these data indicate that L2hgdh deficiency and high L-2-HG accumulation affect the organ size but not the development of the testis.
Testis atrophy and increased histone lysine methylation in L2hgdh KO mice. (A) Age-dependent testis atrophy in wild-type (Ctr) and L2hgdh KO mice. (B) Representative testes from L2hgdh KO mice and wild-type littermates at postnatal month 14. (C) HE staining of representative testis morphology at postnatal month 14. Scale bars at the lower and upper right are 200 μm and 50 μm, respectively. (D and E) Analysis of histone methylation of the testis (D) and cerebrum (E). Histone methylation on K4, K9, K27, K36, and K79 was determined and is presented. Quantification of histone modification was calculated from five independent samples and is presented as means ± SEM; data were analyzed by a two-tailed t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant.
It is known that 2-HG acts as an α-KG antagonist to inhibit the activity of α-KG-dependent dioxygenases, including the JmjC domain-containing histone demethylases and the TET family of DNA cytidine hydroxylases (17, 22). We therefore performed multiple reaction monitoring (MRM) based on liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis with isotopically labeled standards to quantify relative histone levels (23, 24). Our data revealed that H3K9me2 and H3K9me3 levels were significantly increased while H3K9me0 levels were reduced in the testis of L2hgdh KO mice (Fig. 2D). In addition, H3K27me2 levels significantly increased while H3K27me0 levels decreased moderately in the testis of L2hgdh KO mice (Fig. 2D). In contrast, other histone methylation markers, including H3K4me1/2/3 and H3K79me1/2/3, were not substantially changed in the testis of L2hgdh KO mice (Fig. 2D).
The effects of L2hgdh deficiency and L-2-HG accumulation on histone methylation markers was also observed in the brain, where L-2-HG accumulation is the highest (Fig. 1D). We found that H3K9me2 was significantly increased while the H3K9me0 level was reduced in the cerebrum of L2hgdh KO mice (Fig. 2E). As a result, the H3K9me2/H3K9me0 ratio increased by 37% in the cerebrum of L2hgdh KO mice compared to wild-type littermates. Moreover, H3K27me3 was significantly increased, and the H3K27me2 level accordingly was reduced in the cerebrum of L2hgdh KO mice (Fig. 2E). Furthermore, H3K36me2 increased significantly in the cerebrum of L2hgdh KO mice (Fig. 2E). In contrast, other histone methylation markers were not substantially altered in the cerebrum of L2hgdh KO mice, including H3K4me1/2/3 and H3K79me1/2/3 (Fig. 2E). L2hgdh deficiency had no effect on histone methylation in other tissues that showed only moderate accumulation of L-2-HG, such as the liver, kidney, and heart (Fig. S3A to C). Collectively, these findings suggest that L2hgdh deficiency leads to changes in a subset of histone methylation markers in mouse testis and brain, where L-2-HG accumulation is the highest.
Besides histone demethylases, the TET family of DNA hydroxylases is another major pathologically relevant target of 2-HG (13, 25). TET enzymes catalyze three sequential oxidative reactions, converting 5-methylcytosine (5mC) first to 5-hydroxymethylcytosine (5hmC), then to 5-formylcytosine (5fC), and finally to 5-carboxylcytosine (5caC) (26–28). We next performed LC-MS analysis to measure and quantify 5mC and its oxidative derivatives in the cerebrum of L2hgdh KO mice and wild-type littermates. Our data demonstrated that L2hgdh deficiency and high accumulation of L-2-HG had a minor effect on the genomic levels of 5mC/5hmC/5fC in the cerebrum of L2hgdh KO mice (Fig. S4A to C). Previously, it was reported in an Idh1 mutant knock-in mouse model that accumulated D-2-HG in the brain was associated with reduced hydroxylation of hypoxia-inducible transcription factor 1 alpha (Hif-1α) as well as collagen type IV, leading to endoplasmic reticulum (ER) stress response, blood-brain barrier aberration, and hemorrhage (29). However, no sign of hemorrhage was found in the cerebrum of L2hgdh KO mice (data not shown). Furthermore, the levels of Hif-1α protein or eukaryotic initiation factor 2α (eIF2α) phosphorylation were not altered in the cerebrum of L2hgdh KO mice compared to their wild-type littermates (Fig. S4D). Moreover, the blood vessel distribution of collagen type IV was intense and continuous in the cerebrum of L2hgdh KO mice (Fig. S4E and data not shown), implying that hydroxylation modification and maturation of collagens are not disturbed by L2hgdh deficiency and L-2-HG accumulation.
L2hgdh deficiency leads to early-onset white matter abnormality.MRI (magnetic resonance imaging) analysis from previous reports suggest that the pathological changes in L-2-HGA patients are abnormalities of the subcortical cerebral white matter dentate nucleus, globus pallidus, putamen, and caudate nucleus (30). Using T2-weighted MRI analysis, we found that the signal intensity of the subcortical corpus callosum (cc) and striatum was consistently increased in L2hgdh KO mice compared to wild-type littermates (Fig. 3A). L2hgdh KO mice therefore display white matter abnormalities similar to those observed in L-2-HGA patients.
Deletion of L2hgdh leads to early-onset white matter abnormality and demyelination. (A) T2-weighted MRI. Hyperintense signals appeared in the external capsule and striatum in a coronal brain section of L2hgdh KO mice. The scale bar is 4 mm. (B) SRS microscopy showed a decreased lipid signal in the myelin fibers in the striatum of L2hgdh KO mice. Lipid and protein signals are represented as green and red, respectively. Scale bars at the lower left and right are 50 μm and 10 μm, respectively. (C) Luxol fast blue imaging of a whole coronal section showed reduced signal intensity in L2hgdh KO mice. The scale bar is 2 mm. (D) L2hgdh KO mice showed decreased FluoroMyelin staining in the myelin fibers of the striatum. Scale bars at the lower and upper right are 50 μm and 25 μm, respectively. (E and F) Western blotting of cerebrum samples for MBP (E) and quantification of results (F). ***, P < 0.0001. pb/pb, homozygous piggyback transposon insertion into the L2hgdh gene, equal to KO. (G) Transmission electron micrographs (TEM) of the myelin fibers in the striatum showed reduced myelination, naked axons, and vacuolation in L2hgdh KO mice. Mice aged 2 to 6 months were selected, and a representative image at 2 months is shown. Scale bars at the lower and upper right are 4 μm and 2 μm. (H) SRS microscopy for the myelin fibers of the striatum at postnatal ages p9, p12, p14, p21, p28, p60, p180, p360, and p480. The scale bar is 50 μm.
Changes of MRI signal intensity normally reflect alterations in chemical composition. Therefore, we used stimulated Raman scattering (SRS) microscopy to analyze the chemical composition of our biological samples (31, 32). While protein signal intensity did not change in the striatum of L2hgdh KO mice, lipid signal intensity was remarkably weakened in the white matter fiber of the striatum as well as corpus callosum in L2hgdh KO mice compared to wild-type littermates (Fig. 3B and data not shown). Since the white matter fiber of the striatum is mainly composed of axons and the surrounding myelin, we hypothesized that L2hgdh deficiency leads to widespread dys/demyelination in the brain. Supporting this notion, Luxol fast blue (LFB) staining of mouse brain cryosections revealed that the content of myelin was dramatically reduced in the cerebrum of L2hgdh KO mice, including the cortex, corpus callosum, and striatum (Fig. 3C). Likewise, FluoroMyelin staining showed significantly reduced lipid signal in the white matter fiber of the striatum as well as the corpus callosum in L2hgdh KO mice (Fig. 3D and data not shown). Consistent with this finding, the total protein level of MBP (myelin basic protein) was reduced by 60% in the cerebrum of L2hgdh KO mice compared to wild-type littermates (Fig. 3E and F). Furthermore, transmission electron microscopy (TEM) analysis showed reduced myelination intensity and naked axons in the white matter fibers of the striatum in L2hgdh KO mice (Fig. 3G). Together, these results suggest that L2hgdh deficiency leads to widespread dys/demyelination in mouse brain.
To determine the time point when dys/demyelination initially occurred, we utilized SRS microscopy to examine cryosections of mouse brains at different ages. We found that the lipid signal intensity was comparable between L2hgdh KO mice and wild-type littermates at postnatal day 9 (p9) and that the lipid signal intensity thereafter was gradually increased in wild-type mice but much more stagnant in L2hgdh KO mice (Fig. 3H). As a result, an obvious reduction in lipid signal intensity was detected in L2hgdh KO mice as early as postnatal day 12 (Fig. 3H).
To determine whether L-2-HG accumulation in L2hgdh KO mice impairs OPC myelination and early differentiation to oligodendrocytes, we isolated and cultured OPCs from the cerebral tissue of KO and control mice at postnatal day 7. The isolated OPCs were of high purity, as greater than 90% stained positive for the OPC marker NG2 (Fig. S5A). Early differentiation of L2hgdh KO OPCs induced by thyroid hormone T3 was unimpaired in vitro (Fig. S5B), indicating that L-2-HG affects later stages of OPC differentiation and myelination.
Besides dys/demyelination, an extensive spongiotic appearance was observed in multiple brain regions in L2hgdh KO mice, such as the brainstem, the basal ganglia (including the striatum), corpus callosum, hippocampus, inner layer of cortex, and cerebellar nucleus (Fig. S6). The appearance of vacuoles in the brainstem of L2hgdh KO mice was initially observed at postnatal day 8 and reached peak levels at the age of 2.5 to 3 weeks before gradually declining (Fig. S7A). This dynamic change in vacuolation was found in most of the examined brain regions with a close initiation time between postnatal days 8 and 11, with the exception being the dentate gyrus (DG) of the hippocampus, where a large vacuole was first observed in 1-year-old L2hgdh KO mice and became extremely severe thereafter (Fig. S7B).
L2hgdh deficiency leads to extensive gliosis.Approximately half of L-2-HGA patients show macrocephaly (2). Consistent with this report, we observed that the net brain weight was significantly increased in L2hgdh KO mice compared to wild-type littermates above 1 year of age (Fig. 4A and B). Measurements of cerebral cortical surface area revealed a 20 to 25% increase in L2hgdh KO mice compared to wild-type littermates (Fig. 4D), with a 15 to 20% reduction in cerebellum surface area in L2hgdh KO mice (Fig. 4E). In contrast, the surface area of olfactory bulbs (OB) was unchanged in L2hgdh KO mice compared to the wild-type controls (Fig. 4C). Notably, 4′,6-diamidino-2-phenylindole (DAPI) staining for DNA revealed that the number of nuclei increased in the cerebrum of L2hgdh KO mice in an age-dependent manner (Fig. 4F). To confirm this observation, we quantified the number of nuclei in layer IV of the primary somatosensory cortex and beneath the external corpus in the brains of L2hgdh KO and wild-type littermates at different ages. There were no differences in the number of nuclei in the cortex or external corpus regions between L2hgdh KO mice and wild-type littermates at postnatal day 20 or 60 (Fig. 4G and H). Strikingly, the number of nuclei in the cortex and external corpus increased significantly (by 35 to 45% and 3.5- to 4.5-fold, respectively) in L2hgdh KO mice compared to wild-type littermates at postnatal day 240. Actually, a 2- to 2.5-fold increase in nuclear number was observed even earlier in the external corpus of L2hgdh KO mice at postnatal day 120 (Fig. 4H). Together, these findings suggest that L2hgdh deficiency leads to an age-dependent increase in cell number in the cerebrum, which may contribute at least in part to macrocephaly.
Deletion of L2hgdh increases glial cells and brain mass. (A) L2hgdh KO mice show increased brain weight at 14 months but not 4 months compared to their wild-type controls. Note that L2hgdh KO mice express RFP and have a red color due to the inserted transposon construct. For each group, n = 3. (B to E) Whole-mount images (B) and size quantification for OB (C), cerebrum (D), and cerebellum (E) of 14-month-old L2hgdh KO mice compared to their wild-type controls. For each group, n = 3. The scale bars are 1 cm. (F to H) Age-dependent increase in nuclear number in L2hgdh KO mice and quantification in the VI layer of primary somatosensory cortex (G) and the external capsule (H). DAPI was used for nuclear staining. ec, external capsule; CPu, striatum; ctx, cortex. The scale bar is 100 μm. (I to N) Increase in nuclear number was accompanied by reactive astrocytes and an increase in microglia but not neurons. Reactive astrocytes are represented by upregulation of Gfap in L2hgdh KO mice (J), and the quantification is shown (M). NeuN (I and L) and Iba1 (K and N) are representative markers for microglia and neurons, respectively. The scale bar for panels I to K is 200 μm. (O to Q) Immunostaining with oligodendrocyte lineage cell marker Olig2 (O), OPC cell marker NG2 (P), and mature oligodendrocyte marker cc-1 (Q), which is quantified in panels R to T. The scale bar for panels I to K is 200 μm. Quantifications are presented as means ± SEM and were analyzed by two-tailed t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant.
The CNS is composed of diverse cell types, including neurons, astrocytes, oligodendrocytes and their progenitor cells (OPCs), and microglia. To identify the cell types(s) responsible for increased nuclear number in the cerebrum of L2hgdh KO mice, coronal sections of mouse brains were immunostained with different cell-type-specific markers. Neurons are the major cell type in the CNS and account for about two-thirds of the total cell population in the adult mouse brain. We found that NeuN+ cells were barely detectable in the external corpus of L2hgdh KO mice or wild-type littermates (Fig. 4I). No differences in NeuN+ cell number were observed in the cortex between L2hgdh KO mice and wild-type littermates (Fig. 4L), suggesting that neurogenesis in the cerebrum, which mostly occurs in the embryonic stage, is not disturbed. In contrast, Gfap (an astrocyte activation marker) dramatically increased in several regions of the cerebrum in L2hgdh KO mice older than 4 weeks, including the inner layer cortex, white matter, and striatum (Fig. 4J, M, and data not shown). Consistent with this, quantitative reverse transcription-PCR (qRT-PCR) analysis showed that Gfap mRNA levels increased 2.5- to 3-fold in the cerebrum of L2hgdh KO mice compared to wild-type littermates, while that of Aldh1l1, another astrocyte marker, was unchanged (Fig. S8), suggesting that L2hgdh deficiency and L-2-HG accumulation lead to activation but not increased proliferation of astrocytes in mouse brain. Astrocyte activation is often concomitant with the activation of microglia, which are considered resident macrophage cells and act as the first line of innate immune defense in the CNS. Indeed, we found that Iba1+ cells increased by 3-fold in the external corpus of L2hgdh KO mice at postnatal day 120 (Fig. 4K and N). In addition, the population of Olig2-positive cells (an oligodendrocyte lineage marker) increased by 2-fold and accounted for ∼78% of all cells in L2hgdh KO mice at p120 (Fig. 4O and R). These data indicate that the expansion of oligodendrocyte lineage cells is a major contributor to increased nuclear number in L2hgdh KO mice. The oligodendrocyte lineage is comprised of OPCs. We found that the levels of Ng2-positive oligodendrocytes and CC1-positive OPCs were increased by 2- to 2.5-fold and 1.5- to 2-fold, respectively, in the external corpus of L2hgdh KO mice (Fig. 4P, Q, S, and T).
OPCs have been proposed to be the origin of gliomas (33, 34). The observed OPC expansion in the cerebrum of adult L2hgdh KO mice prompted us to investigate the oncogenic effect of L2hgdh deficiency and L-2-HG accumulation in the CNS. IDH1 mutant gliomas, which exhibit high levels of D-2-HG, often contain mutations in p53 (35, 36). We therefore crossed L2hgdh KO mice with p53 KO mice and found that L2hgdh and p53 double KO mice developed tumor types similar to those of p53 single KO mice, most of which were lymphoma and peripheral tumors but not brain tumors (data not shown). Notably, deletion of both L2hgdh and p53 genes did not significantly change tumor-free survival in mice compared to the single p53 deletion controls (Fig. S9). A likely explanation is that the p53 deletion produces a very strong tumor phenotype, which may mask the effect of L2hgdh deficiency on tumorigenesis. Together, our data suggest that L2hgdh deficiency and L-2-HG accumulation lead to an increased glial population and a neoplastic state via the expansion of OPC; however, deletion of L2hgdh does not induce an overtly malignant phenotype in mouse brains either alone or in combination with whole-body p53 deletion.
L2hgdh deficiency results in reactive gliosis and neuroinflammation.As shown earlier in this study, a subset of histones showed increased methylation in the CNS of L2hgdh KO mice (Fig. 2E). Histone methylation not only affects chromatin status and transcription factor recruitment to regulate gene expression (37) but also has been linked to a number of neurological and psychiatric disorders (38). This prompted us to perform RNA-sequencing experiments in hippocampal tissue of mice at the age of postnatal day 60, when glial cell numbers have not yet increased. Surprisingly, there were only selective groups of genes whose transcriptional expression was different in the hippocampus of L2hgdh KO mice compared to wild-type littermates (Fig. 5A and B). For instance, the gene ontology (GO) item “oxidation reduction” is the top dysregulated pathway in the L2hgdh KO mice compared to wild-type littermates (Fig. 5C). In accordance with this, we found that the level of glutathione (GSH) was significantly reduced by 30% in the cerebrum of L2hgdh KO mice (Fig. 5D). In agreement with our histological findings (Fig. 3 and 4), we detected significant dysregulation of GO items involved in myelination, axon ensheathment, and cell proliferation in L2hgdh KO mice (Fig. 5C). In addition, genes involved in “immune response” and “chemotaxis” were also dysregulated (Fig. 5C). To verify the RNA-sequencing data, we performed qRT-PCR analysis and found that multiple genes involved in chemotaxis and immune response were indeed upregulated in the cerebrum of L2hgdh KO mice, including Clec7a, Itgax, Cybb, Tnf, and Ccl family genes Ccl3/4/6 but not Ccl2/5 (Fig. 5E). When mouse brain tissues were immunostained with Cd3 and Mpo (lymphocyte and neutrophil markers, respectively), we found that neither Cd3+ nor Mpo+ cells could be detected in the cerebrum of L2hgdh KO mice or wild-type littermates (Fig. S10). This indicated that the observed immune response triggered by chemotaxis in the cerebrum of L2hgdh KO mice is not caused by peripheral lymphocyte or neutrophil infiltration into the CNS. It should be noted that genes involved in the pathways of chemotaxis and immune response are predominantly expressed in microglia in the CNS (Fig. S11). Immunofluorescence staining demonstrated that the signal intensity of Cd68 was significantly upregulated in the cerebrum of L2hgdh KO mice at postnatal day 60 and strongly overlapped with microglia-specific marker Iba1 (Fig. 5F), indicating that the elevated immune response is mediated by microglia. Our data therefore suggest that L2hgdh deficiency leads to microglia activation prior to cell number proliferation in the adult mouse brain.
Deletion of L2hgdh results in disturbed redox status and neuroinflammation. (A) Gene expression profiling and clustering analysis of the hippocampus of L2hgdh KO mice and their wild-type controls. (B) MA plot of differential gene expression in the brains of L2hgdh KO and wild-type mice. The horizontal axis represents gene expression levels (FPKM), while the vertical axis represents relative fold changes in gene expression between L2hgdh KO and wild-type mice. Each dot represents one gene, and red dots indicate genes with differential expression using a cutoff (q value) of <0.1 (Benjamini-Hochberg method or false discovery rate). (C) Gene ontology (GO) analysis of differentially expressed genes. (D) L2hgdh KO reduces GSH levels in the cerebrum. Whole-cerebrum tissues of adult L2hgdh KO and wild-type mice were used for LC-MS/MS to measure GSH and GSSG levels. Three mice were used for each group. (E) qRT-PCR verification of differentially expressed genes categorized as the GO items “chemotaxis” and “immune response.” (F) Cd68 signal is upregulated in 2-month-old L2hgdh KO mice. Coimmunofluorescence of Cd68 and microglial marker Iba1 shows highly overlapped staining. Representative images were taken in the striatum. The scale bar is 30 μm. Quantifications are presented as means ± SEM and were analyzed by two-tailed t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant.
L2hgdh deficiency results in impaired adult hippocampal neurogenesis and late-onset neurodegeneration.We next investigated the impact of L2hgdh deletion and 2-HG accumulation on neurons, the major cell type in the CNS. Continuous neurogenesis throughout life occurs primarily in the subventricular zone of the lateral ventricles (SVZ) as well as the subgranular zone (SGZ) of the DG in the hippocampus. Notably, we found that the number of cells which were positive for nestin, a specific marker for neural progenitor cells (NPCs), decreased by 80% in the SGZ of 8-week-old L2hgdh KO mice compared to that in wild-type controls (Fig. 6A and B), suggesting that L2hgdh is important for maintaining the NPC pool in the hippocampal SGZ of the adult mouse brain. In addition, we also used specific cell lineage markers to test the effect of L2hgdh KO on the population of nonradial progenitor cells and newborn neurons. Immunofluorescence staining illustrated that the number of Tbr2-positive nonradial progenitor cells decreased by 70% in the SGZ of 8-week-old L2hgdh KO mice (Fig. 6C and D). Moreover, Dcx-positive newborn neurons were reduced by 60 to 70% in the same brain region in L2hgdh KO mice (Fig. 6E and F and Fig. S12). Together, these results provide in vivo evidence that adult hippocampal neurogenesis is impaired in L2hgdh KO mice.
Deletion of L2hgdh results in impaired neurogenesis, neurodegeneration, and cell death. (A) Nestin staining in neural progenitor cells in the adult SGZ. Shown are coronal images of the SGZ in 2-month-old L2hgdh KO and wild-type mice. Their density is quantified in panel B. (C and D) Reduction of Tbr2+ intermediate progenitors in 2-month-old L2hgdh KO mouse brains compared to controls. (E and F) Reduction of newborn neurons represented by Dcx+ cells in 2-month-old L2hgdh KO compared to control mouse brains. The scale bars for panels A, C, and E at the lower right are 100 μm, and the upper right insets are enlarged images with a scale bar of 50 μm. (G) L2hgdh KO reduces the self-renewal of neural progenitor cells in vitro. Neural progenitor cells were isolated from the hippocampus and cultured for four generations, with an initial seeding density of 50,000 cells/well. (H to J) The tripotent differentiation capacity of L2hgdh KO neural progenitor cells. Neural progenitor cells isolated from the hippocampus of L2hgdh KO and wild-type mice were induced to differentiate into neurons, astrocytes, and oligodendrocytes in vitro. Scale bars for panels H to J are 100 μm. (K to Q) Neurodegeneration caused by L2hgdh KO in the CNS. Serial coronal sections from 14-month-old L2hgdh KO and wild-type mice were used for Fluoro Jade-C staining (K), cleaved caspase 3 (L), and TUNEL staining (M). Representative images of the globus pallidus are shown. The scale bar is 100 μm. (O to Q) Respective quantifications of data from panels K to M. Quantifications are presented as means ± SEM and were analyzed by a two-tailed t test. ***, P < 0.001; **, P < 0.01; *, P < 0.05; n.s., not significant.
Furthermore, we analyzed NPCs isolated from neonatal mouse hippocampus and found that L2hgdh-deficient NPCs showed lower proliferation capacity than the wild-type controls (Fig. 6G). To determine the self-renewal capacity of these NPCs, primary neurospheres were individually dissociated into single cells and replated at clonal density for three additional passages. This replating assay revealed that L2hgdh-deficient NPCs also exhibited lower capacity for self-renewal (Fig. 6G). However, we found that the differentiation potential of NPCs was not affected by L2hgdh deletion (Fig. 6H to J). Upon differentiation in vitro, the neurospheres from L2hgdh KO mice were tripotent, with comparable abilities to generate neurons, astrocytes, and oligodendrocytes as evidenced by immunostaining of Tuj1, Gfap, and O4 markers, respectively (Fig. 6H to J). Thus, these data provide evidence that L2hgdh deficiency reduces the self-renewal capability of NPCs, thereby contributing to the impaired adult hippocampal neurogenesis in L2hgdh KO mice.
Reactive gliosis and neuroinflammation could be strong driving forces to facilitate neurodegeneration (39, 40). To examine neurodegeneration in L2hgdh KO mice, we performed Fluoro Jade-C (FJC) staining, which is widely used to label degenerating neurons (41). The number of FJC-positive cells increased remarkably in the cerebrum, especially the basal ganglia regions, including the globus pallidus and striatum, of 14-month-old L2hgdh KO mice compared to the wild-type controls (Fig. 6K and O and data not shown). In accordance with this, terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL)-positive and cleaved caspase 3-positive signals were much stronger in the cerebrum of 14-month-old L2hgdh KO mice than in wild-type littermates (Fig. 6L, M, P, and Q), indicating that apoptosis is highly elevated in L2hgdh-deficient mouse brains.
DISCUSSION
In this study, we provide in vivo evidence that mutation of the detoxification enzyme L2hgdh, likely due to L-2-HG accumulation, causes abnormal myelination, disturbed glia/adult neural stem cell homeostasis, and increased cell death in the central nervous system. The major findings of this study, as summarized in Fig. 7, are as follows. Deletion of the L2hgdh gene leads to a time-dependent accumulation of the enzyme's substrate, L-2-HG, from ∼30 μM in control mouse tissues up to 4,000 μM in brain and testis, whereas most other tissues accumulate L-2-HG to ∼500 μM. The earliest lesions in the central nervous system, which begin at postnatal week 2, include dysmyelination and vacuolation and are mainly but not exclusively myelin-rich regions of the corpus callosum (cc) and striatum fibers. In the dentate gyrus (DG), vacuoles are initially formed at 1 year of age but are exacerbated with age. The temporal heterogeneity of vacuolation suggests that there is more than one mechanism contributing to vacuolation in different brain regions. Gliosis represented by neuroinflammation of microglia and astrocyte activation starts as early as week 3 to 4. The neuroinflammation of microglia is strongly associated with the vacuolation (except in the DG), as both phenotypes decrease with time. L2hgdh deletion also increases the adult glial cell population, including oligodendrocytes, OPCs, and microglia. The increased number of oligodendrocytes, possibly from increased proliferation and differentiation of progenitor OPCs, might be a compensatory effect in response to reduced myelination. L-2-HG accumulation may compromise the maintenance of the neural progenitor pool, leading to reduced neurogenesis without affecting the tripotent differentiation potential of NPCs in vitro. The long-term burden of high L-2-HG levels could further induce apoptosis and neurodegeneration, eventually leading to decreased motor activity, reduced body weight, and death.
Schematic illustration of disease onset in L2ghdh KO mice. The red arrows indicate activation or increase, while blue arrows indicate inactivation or reduction.
Our current study is in line with previous work conducted by Rzem et al., demonstrating that L-2-HG accumulates most significantly in the brain and testis of L2hgdh KO mice (21). We further show that high L-2-HG burden led to epigenetic disruption and progressive atrophy of the testis. In previous studies, L2hgdh KO mice exhibit increased distance of movement in OFT; however, this was not observed in our current study. This may be due to age differences of mice used for comparison, as we also observed increased motor activity in younger mice (data not shown). We speculate that the long-term burden of high L-2-HG could severely compromise motor activity, as most L2hgdh KO mice older than 8 months completely lost movement under routine feeding conditions. Moreover, this is in line with clinical reports that L-2-HGA patients exhibit delayed motor development, frequently accompanied by loss of milestones like unassisted walking (2).
Both dys/demyelination and vacuolation occur during the second week after birth in the L2hgdh KO mice. This time point closely correlates with the peak accumulation of L-2-HG. Considering that most peripheral tissues contain less than 1 mM L2-HG and are quite normal in L2hgdh KO mice, intracellular L-2-HG levels may need to reach a higher level (i.e., >1 mM) to cause pathological changes. Postnatal week 2 is crucial for CNS pathology in L2hgdh KO mice, as it is a critical time point for OPC differentiation, oligodendrocyte maturation, and axon myelination in the mouse brain (42, 43). Myelination is relatively normal at postnatal day 9 but the deficiency occurs around postnatal day 12, suggesting that early development, differentiation, and maturation of oligodendrocytes are normal. It is possible that L-2-HG affects mostly the later stages of oligodendrocyte maturation and myelination.
Vacuolation of the CNS is a common feature in mice with high levels of either L-2-HG or D-2-HG, as this phenotype is also observed in mutant IDH2-R140Q or IDH2-R172K knock-in mice, which accumulate D-2-HG (44). We speculate that neuroinflammation and microglia activation are a host defense reaction against vacuolation for the following reasons: (i) the vacuolation appears earlier and is followed by neuroinflammation and microglia activation; (ii) the activated microglia are recruited and highly proliferative around the vacuolation region in corpus callosum (cc) and myelin fiber of striatum (CPu); (iii) when mice are older than 1 year, the vacuolation phenotype becomes less severe. Concomitantly, the neuroinflammation and microglia activation are diminished.
The alpha-KG-dependent dioxygenase family consists of more than 60 members, and they catalyze hydroxylation reactions on diverse substrates, including histone and DNA demethylation (18). Both D-2-HG and L-2-HG are structurally similar to alpha-KG and have been shown to inhibit dioxygenases (13, 17, 45). Although L-2-HG is more potent than D-2-HG in inhibiting these enzymes, not every alpha-KG-dependent dioxygenase exhibits the same sensitivity to D-2-HG or L-2-HG. In fact, L-2-HG inhibits different dioxygenases with a wide range of potencies in vitro (17). The enzymes with the lowest 50% inhibitory concentrations (IC50s) were histone demethylases KDM4A/JMJD2A (IC50 of 26 ± 3 μM) for H3K9 and H3K36, KDM4C/JMJD2C (IC50 of 97 ± 24 μM) for H3K9 and H3K36, and KDM2A/FBXL11 (IC50 of 48 ± 15 μM) for H3K36. In contrast, the enzyme PHD2, which is responsible for Hif-1α hydroxylation and ubiquitination-mediated degradation, has a relatively high IC50 of 419 ± 150 μM (17). As a competitive inhibitor, the IC50 of L-2-HG depends on the concentration of alpha-KG and other cosubstrates used in the enzymatic assays. Therefore, we take caution in comparing the in vitro IC50 and in vivo effectiveness of L-2-HG in dioxygenase inhibition. In the L2hgdh KO mice, neither ER stress nor a global increase of Hif-1α was detected. These phenotypes are different from those of the IDH1-R132H knock-in mice (29). Interestingly, a significant increase in global methylation of H3K9, H3K27, and H3K36, but not H3K4 or H3K79, was found in the brain and testis, two regions with the highest L-2-HG accumulation in L2hgdh KO mice. We speculate that the L-2-HG accumulated in the brain and testis of L2hgdh KO mice selectively inhibits some members of the JmjC domain-containing histone demethylases.
Both L-2-HG and D-2-HG are normal metabolic by-products (46). Their intracellular levels in most normal cells and tissues are maintained at low levels through the actions of L2HGDH and D2HGDH (5, 47, 48). Currently, few physiological roles have been associated with either enantiomer of 2-HG. High levels of 2-HG have been implicated in pathogenesis of acidurias, as found in the patients with mutations in L2HGDH or D2HGDH. Moreover, D-2-HG acts as an oncometabolite in various human cancers harboring mutations in IDH1 or IDH2. Approximately 80% of secondary glioblastomas, 20% of AML, 50% of chondrosarcomas, and 10 to 20% of cholangiocarcinomas accumulate high levels of D-2-HG due to IDH1 or IDH2 mutations (49–54). Furthermore, L-2-HG is elevated in clear cell renal cell carcinoma (ccRCC) due to reduced expression of the L2HGDH gene. In line with the oncogenic function of 2HG, we found a massive expansion of OPCs in L2hgdh KO mice, suggesting a neoplastic effect of L-2-HG. However, no brain malignancy was found in the L2hgdh KO mice, indicating that accumulation of L-2-HG alone is insufficient to induce glioma in mice.
L2hgdh KO mice display deficiencies in adult neurogenesis; a similar phenotype has also been observed in Tet1 deletion mice (55). Tet1 is highly expressed in the CNS and plays an important role in DNA demethylation (55, 56). In Tet1-deficient mice, hypermethylation and reduced expression are observed in genes involved in neuronal progenitor proliferation (55). Similar to the Tet1 KO mice, neural progenitor cells (NPCs) isolated from the hippocampus of L2hgdh KO mice showed reduced capacity for self-renewal; however, this reduction was modest and may not explain the dramatic reduction in neurogenesis in vivo. Additional causes, such as increased apoptosis, were observed in the L2hgdh KO brain and may contribute to the neurogenesis defects in the L2hgdhd mice. The L2hgdh KO mice presented in this report provide a valuable model for studying the role of L-2HG in the pathophysiology of human diseases.
MATERIALS AND METHODS
Mice and ethics statement.FVB background L2hgdh knockout (KO) mice (L2hgdh−/−) were generated by the Institute of Developmental Biology and Molecular Medicine (IDM) at Fudan University using the piggyBac transposon insertion method. Briefly, a modified piggyBac transposon harboring the red fluorescent protein (RFP) gene was inserted between exon 1 and exon 2 of the L2hgdh allele. The genomic insertion site is chr12:70798148, and the corresponding insertion DNA sequence is 5′-AATTGTGATTCAAAAATAACATTTCCAGGAATGGAAGCTATACTATGTTATAAAATGCAAAAGCAAGATAGCAAGATACCTCTGTAACTTGGTGCCGTTGGCGTTTGTATTGGATAGTTGTTTGCTGTGAGGCC-3′. The underlined sequence, AATT, is the insertion site. Knockout efficiency was confirmed by genomic PCR, protein Western blotting, and accumulation of the L2HGDH substrate l-2-hydroxyglutarate. Whole-body p53 KO mice (B6.129-Trp53tm1Srcmo) were purchased from Shanghai Biomodel Organism and verified by genomic PCR. Trp53+/tm1Srcmo mice were crossed with L2hgdh+/− mice to produce double heterozygous mice (F1). The F1 mice were then intercrossed to produce double-gene-knockout mice and the corresponding control mice with Trp53 knockout and wild-type L2hgdh.
For all of the following experiments, littermate mice were used unless stated otherwise. Mice were bred under standard husbandry conditions at the Fudan University mouse facility on a 12-h light/dark cycle, and all experiments were performed in accordance with the Animal Care and Use Committee at Fudan University.
Neural stem cell assay.Hippocampus cells were dissected from neonatal mouse brain, dissociated using a cell dissociation reagent (A1110501; Thermo), and filtered through a 40-μm cell filter (Falcon). Cells were quantified by hemocytometer and plated on 6-well ultralow attachment plates (CLS3471-24EA; Corning) at a density of 50,000 cells/well to form neurospheres. The cell growth medium consisted of Dulbecco's modified Eagle's medium nutrient mixture F-12 (10565-018; Thermo) supplemented with N-2 supplement (17502-048; Thermo), B27 supplement minus vitamin A (12587-010; Thermo), epidermal growth factor (EGF; PHG0311; Thermo), fibroblast growth factor (FGF)-basic (PHG0026; Thermo), and penicillin-streptomycin (15070-063; Thermo). Cultured neurospheres were cultured for 7 days before dissociation for cell counting. The tissue/neurosphere dissociation was continued for 4 passages and labeled as primary passage (1st), secondary passage (2nd), 3rd passage, and 4th passage.
The in vitro differentiation of neural stem cells was performed with NeuroCult differentiation medium (05704; Stem Cell). Briefly, neurospheres from secondary passage at day 3 to ∼5 were dissociated to single cells and plated onto 8-well glass chambers (154941; Nunc) in differentiation medium. After a 7-day culture, cells were fixed and used for immunostaining.
TEM analysis.For TEM, mice were perfused sequentially with phosphate-buffered saline (PBS) and prewarmed 2% glutaraldehyde–2% paraformaldehyde (PFA) in PBS. After perfusion, tissues were fixed overnight and coronally sectioned at 100-μm thickness by a vibratome (VT1200S; Leica). Tissues were then fixed in OsO4 for 1 h and embedded in epoxy resin. Ultrathin sections were cut with a diamond knife to 75 nm (EM UC6; Leica) and visualized using TEM (JEM-1230; JEOL).
OFT.The OFT comprehensively assessed behavioral and locomotor activity levels of rodents, which can be correlated with locomotive function. Briefly, each mouse was dropped in the same place of a square arena (60-cm length) and video tracked for 10 min. The total distance of movement was measured and quantified automatically.
Antibodies.Antibodies against L2hgdh (15707-1-AP; Proteintech), beta-actin (A00702; GenScript), Hif-1α (610958; BD), eIF2α (9722; CST), and phospho-eIF2α (Ser51) (9721; CST) were used for Western blotting. Gfap (MAB360; Millipore), Iba1 (019-441; Wako), Olig2 (AB9610; Millipore), Ng2 (AB53320; Millipore), CC-1 (OP80-100UGCN; Millipore), collagen type IV (600-401-106; Rockland), Cd68 (14-0681-80; eBioscience), nestin (MAB2736; R&D), Tbr2 (ab23345 Abcam), doublecortin (Dcx) (sc-241380; Santa Cruz), and cleaved caspase-3 (Asp175) (5A1E) (CST 9664) were used for immunofluorescence or immunohistochemistry.
Western blotting.Mouse tissues were lysed in radioimmunoprecipitation assay (RIPA) buffer supplemented with a cocktail of protease inhibitors. Proteins were blotted by following standard protocols.
Tissue preparation and histology analysis.Frozen and paraffin-embedded sections were utilized for histological analysis. Wild-type and KO littermates were perfused with PBS and then 4% PFA in PBS at various time points. For frozen sections, brains were fixed in 4% PFA overnight at 4°C and transferred sequentially to 15% and then 30% sucrose after the brain sank to the bottom. Brains were embedded and frozen in OCT compound (Tissue-Tek), and 40-μm serial sections were coronally prepared by frozen microtome (CM1950; Leica). For paraffin-embedded sections, postfixed brains were sagittally or coronally prepared at 4-μm thickness. Slides from histologically comparable positions were picked for immunohistochemistry or immunofluorescence staining. To visualize myelinated fibers, frozen sections of brain tissue were processed using Luxol fast blue staining (S3382; Sigma) or FluoroMyelin fluorescent myelin staining (F34651; Thermo) according to standard protocols. For hematoxylin-eosin (HE) staining, paraffin-embedded tissue sections (4 μm) were used. To identify degenerating neurons, Fluoro-Jade C histofluorescent staining was performed according to the manufacturer's protocol (AG325-30MG; Millipore).
IHC, IF, and confocal imaging.Immunocytochemistry (IHC) and immunofluorescence (IF) were performed on paraffin sections and frozen sections, respectively. Visualization of primary antibodies was performed by the avidin-biotin-horseradish peroxidase system for IHC and Alexa Fluor 488/647-conjugated secondary antibodies (Invitrogen) for IF. The primary antibodies used in this study are listed in “Antibodies,” above. Sections were visualized under a fluorescence laser scanning confocal microscope (A1; Nikon).
Whole-body fat composition analysis.The fat composition of mice was measured nondestructively by Bruker's minispec whole-body composition analyzer by following the manufacturer's instructions. The method is based on time domain nuclear magnetic resonance (TD-NMR).
MRI analysis.We carried out T2-weighted MRI in vivo at 7.0 T using a Bruker BioSpin MRI GmbH system according to standard procedures.
Metabolite extraction and GC-MS analysis.The GC-MS analysis was performed as previously described (45). Standard curves of commercial L-2-HG and creatinine were used to quantify L-2-HG and creatinine in the samples.
Measurement of GSSG and GSH levels.Oxidized glutathione (GSSG) and reduced glutathione (GSH) levels were determined by using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Briefly, tissues were homogenized by using 10-fold 80% (vol/vol) chilled methanol and then centrifuged at 12,000 rpm for 10 min at 4°C. The supernatant then was subjected to LC-MS/MS analysis using a Shimadzu LC system coupled with a 4000-qtrap triple-quadrupole mass spectrometer (AB Sciex). A Phenomenex NH2 column (inner diameter, 50 mm by 2.0 mm; particle size, 5 μm) was used. The mass spectrometer was optimized and set up in selected reaction monitoring (SRM) scan mode for monitoring the [M-H] of GSSG (m/z 611.6→306.2) and GSH (m/z 306.2→142.8). Analyst software was used for analysis.
RNA isolation and qRT-PCR analysis.Total RNA was isolated from specific brain regions using TransZol Up (TransGen) by following the manufacturer's instructions. RNA was reverse transcribed with random primer (N9) (TransGen) and proceeded to qRT-PCR with gene-specific primers by SYBR Premix EX tag (TaKaRa). The relative expression abundance of specific genes was calculated by normalization to the beta-actin control. Primer sequences can be provided upon request.
Quantitative analysis of 5mC and its derivatives.Quantification of genomic cytosine and its derivatives was performed by LC-MS/MS as previously described (57).
Gene expression analysis.Next-generation sequencing was performed using standard methods. Briefly, total RNA was extracted using TransZol Up (TransGen). Samples were quantified with an Agilent 2100 Bioanalyzer. Library preparation and RNA sequencing were conducted using TruSeq and a HiSeq 2500 platform (Illumina Inc., San Diego, CA) at WuXi AppTec (Shanghai, China).
For sequencing result analysis, paired-end reads were aligned to the mouse genome (mm9) after quality filtering using Mapsplice, and we performed the quantitation with RSEM. Hierarchical clustering was performed with the open-source application Cluster 3.0 and visualized with Java TreeView. The differential expression analysis was carried out using DESeq2. We performed GO analysis with iPathwayGuide.
Histone methylation profiling.Histone methylation was quantified by MRM-based LC-MS/MS (LC-MRM-MS), as described previously (23, 24). In short, the core histones were acid extracted from tissue with 0.4 M H2SO4 and precipitated with trichloroacetic acid (TCA), followed by 2 washes with ice-cold acetone. The histones were treated with N-hydroxysuccinimide-propionate and digested with trypsin. Finally, the digested peptides were concentrated to dryness. Prior to LC-MRM-MS, a mixture of isotope-labeled histone peptides was added as internal standards to the samples, and the histone peptide concentration in the samples was calculated by the peak area ratio of histone peptides to internal standard peptides.
Statistical methods.Results are presented as means ± standard errors of the means (SEM) unless otherwise specified. Statistical analysis was performed using two-tailed unpaired Student's t test. P values less than 0.05 were considered significant.
Accession number(s).The full data set has been deposited in the GEO public database under accession number GSE89806.
ACKNOWLEDGMENTS
We thank members of the Fudan MCB laboratory for their valuable inputs and support throughout this study and Jian Hu, Yuan Zhu, and Vivian Fu for critical discussion and reading of the manuscript.
This work was supported by the CAS Strategic Priority Research Program (XDB02050400), NSFC grants (91432111) to Z.Q, by the 973 Program (no. 2012CB910303 to D.Y. and no. 2012CB910101 to K.-L.G.), the NSFC grant (no. 81372198 and no. 81522033 to D.Y.), the NSFC Program of International Cooperation and Exchanges (no. 81120108016 to Y.X.), and the Shanghai “Phosphor” Science Foundation, China (no. 14QA1400600 to D.Y.). This work was also supported by NIH grants (GM067113 and CA1638311 to Y.X.; CA196878 and GM51586 to K.-L.G.).
FOOTNOTES
- Received 5 September 2016.
- Returned for modification 4 November 2016.
- Accepted 24 January 2017.
- Accepted manuscript posted online 30 January 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/MCB.00492-16.
- Copyright © 2017 American Society for Microbiology.
