Intracellular α-ketoglutarate maintains the pluripotency of embryonic stem cells

Abstract

The role of cellular metabolism in regulating cell proliferation and differentiation remains poorly understood. For example, most mammalian cells cannot proliferate without exogenous glutamine supplementation even though glutamine is a non-essential amino acid. Here we show that mouse embryonic stem (ES) cells grown under conditions that maintain naive pluripotency are capable of proliferation in the absence of exogenous glutamine. Despite this, ES cells consume high levels of exogenous glutamine when the metabolite is available. In comparison to more differentiated cells, naive ES cells utilize both glucose and glutamine catabolism to maintain a high level of intracellular α-ketoglutarate (αKG). Consequently, naive ES cells exhibit an elevated αKG to succinate ratio that promotes histone/DNA demethylation and maintains pluripotency. Direct manipulation of the intracellular αKG/succinate ratio is sufficient to regulate multiple chromatin modifications, including H3K27me3 and ten-eleven translocation (Tet)-dependent DNA demethylation, which contribute to the regulation of pluripotency-associated gene expression. In vitro, supplementation with cell-permeable αKG directly supports ES-cell self-renewal while cell-permeable succinate promotes differentiation. This work reveals that intracellular αKG/succinate levels can contribute to the maintenance of cellular identity and have a mechanistic role in the transcriptional and epigenetic state of stem cells.

Extended Data Figure 1. Pluripotent stem cells can proliferate in the absence of glutamine when cultured in 2i/LIF medium

a, Doubling time of ESC-V19 cells cultured in serum/LIF (S/L) or 2i/LIF (2i/L). b, Growth curve of ESC-1 cells cultured in S/L or S/L/2i medium devoid of glucose. c, Samples of S/L (left) and 2i/L (right) media with and without glutamine were analyzed by gas chromatography-mass spectrometry. Representative chromatograms of the total ion count reveal a clear glutamine (Q) peak in +Q media (grey) and no detectable glutamine in –Q media (red). m, minutes. d, Teratoma formation from ESCs grown in 2i/L medium without glutamine for three days. Representative images of haematoxylin and eosin staining reveal neural tissue (ectoderm), hepatocytes and pancreatic acinar cells (endoderm) and smooth muscle (mesoderm). Scale bar, 200 μm. e, Growth curve of ESC-1 cells grown in glutamine-free 2i/L or 2i medium. f, Gene expression analysis confirms that EpiSCs, which represent post-implantation pluripotency, were generated from ESC-1 cells by culture with Fgf and Activin A. Transcript levels were assessed by qRT-PCR, normalized to Gapdh and expressed as a ratio of values of mESCs cultured in 2i/L medium. g, Growth curve of epiblast stem cells (EpiSCs) cultured in serum-free epiblast medium (serum-free medium containing FGF and Activin A, Fgf/ActA) with or without glutamine. h, Growth curve of an induced pluripotent (iPS) cell line derived from fibroblasts using Oct3/4 (O), Klf4 (K), and Sox2 (S) cultured in glutamine-free S/L or 2i/L medium. i, Doubling time of ESC-1 cells cultured in 2i/L medium in the presence and absence of glutamine. j, Growth curve ESC-V19 cells cultured in glutamine-free 2i/L medium in the presence or absence of 1 μM methyl-sulfoxide (MSO). k, ESC-V19 cells grown glutamine free S/L (left) or 2i/L (right) medium with or without 4 mM dimethyl-α-ketoglutarate (DM-αKG). For growth curve experiments, cells were seeded on day 0 in complete medium and then were changed to experimental medium on day 1 (indicated by red arrow). Data are presented as the mean ± s.d of triplicate wells from a representative experiment.

Extended Data Figure 2. mESCs cultured with 2i demonstrate altered glucose and glutamine utilization

a, 2i enables glutamate synthesis from glucose-derived carbons. ESC-1 cells cultured in S/L, S/L/2i or 2i/L medium were incubated with medium containing [U-¹³C]glucose for four hours and the fraction of glutamate containing glucose-derived carbons is shown. b, ESC-1 cells were cultured for four hours in glutamine-free S/L or 2i/L medium containing [U-¹³C]glucose and the total amount of glutamate labeled by glucose-derived carbons is shown. c, Incorporation of ¹⁴C derived from [U-¹⁴C]glucose (¹⁴C-glc) (left) or derived from [U-¹⁴C]glutamine (¹⁴C-gln) (right) into total cellular protein after 48 hour incubations. p < 0.05 for ¹⁴C-glc, p = 0.1 for ¹⁴C-gln, calculated by unpaired two-tailed Student’s t-test. Data are presented as the mean ± s.d of triplicate wells (a,b) or ± s.e.m of quadruplicate wells (c) from a representative experiment.

Extended Data Figure 3. Regulation of histone methylation in 2i/LIF cells

a, Western blot analysis of ESC-1 cells grown in glutamine-free 2i/L medium for 24 hours with supplementation as indicated (DM-αKG, dimethyl-α-ketoglutarate). b,c, H3K27me3 ChIP-PCR of ESC-1 cells cultured in S/L (b) or 2i/L (c) medium with or without 30 μM UTX/JMJD3 inhibitor GSK-J4 for five hours. Data are presented as the mean ± s.e.m. of triplicate samples. *, p < 0.05 by unpaired Student’s two-tailed t-test.

Extended Data Figure 4. Generation of JMJD3 mutant cells

a, Schematic of targeting strategy for gRNAs to mouse Jmjd3 exon 17. gRNA sequences are highlighted in blue. b. Representative sequences from two clones used in this study. Sanger sequencing revealed indels as shown in schematic. Red dashes, deleted bases; red bases, insertions. gRNA is highlighted in blue and protospacer adjacent motif (PAM) sequences identified in green. Predicted cut site indicated by red triangle. Location of in-frame downstream stop is indicated on the right. c, An example chromatogram for clone JMJD3Δ/Δ-2 showing single base-pair insertions at predicted Cas9 cleavage site.

Extended Data Figure 5. αKG increases Tet activity in mESCs

a, Simplified schematic of the reaction mechanism of TET1/2 enzymes. b, Relative percent 5-methylcytosine (% 5-mC) in ESC-1 cells cultured in S/L medium with or without DM-αKG for 24 hours. Each data point represents a sample from triplicate wells of a representative experiment. c, Gene expression in ESC-1 cells cultured with DM-αKG or DM-succinate for three days. d, Gene expression in ESC-1 cells cultured in S/L medium with or without DM-αKG for four passages. e, Gene expression in wild-type or Tet1/Tet2 double knock out (KO) mESCs cultured with DM-αKG or DM-succinate for 72 hours. qRT-PCR data (c–e) was normalized to Actin or Gapdh and samples were normalized to the control group. Oct3/4 is not expected to change and is included as a control. Data are presented as the mean ± s.e.m. of triplicate wells.

Extended Data Fig. 6. αKG increases Nanog expression

a, Representative histogram of GFP intensity of Nanog-GFP cells treated with or without DM-αKG for three days. Grey represents background staining. b, ESC-1 cells were cultured in S/L medium with DM-αKG for four passages and then switched to medium containing the indicated amounts of DM-αKG (0.5 – 4 mM) or vehicle control (S/L) for three days. GFP expression (mean fluorescence intensity, M.F.I.) was determined by FACS. Data are presented as the mean ± s.d. of triplicate wells from a representative experiment.

Figure 1. 2i is necessary and sufficient to confer glutamine independence

a–f, Growth curves and representative images of ESCs grown in the absence of glutamine. Growth curves of ESC-V19 cells (a) and V6.5 ESC lines (ESC-1-4) (b) cultured in glutamine-free S/L or 2i/L medium. Phase images showing ESC-1 cells cultured in glutamine-free 2i/L (c) or S/L/2i (e) medium for 3 days. Top, brightfield (BF); bottom, alkaline phosphatase (AP) staining. Bar, 500 μm. d, Growth curve of ESC-V19 cells in glutamine-free S/L or S/L/2i medium. f, Growth curve of ESC-V19 cells cultured without glutamine in two serum-free media formulations containing N2 and B27 supplements, 2i/L and BMP4/L. g, Intracellular glutamate levels 8 hours after addition of medium with or without glutamine (Q). Data are presented as the mean ± s.d of triplicate wells from a representative experiment.

Figure 2. 2i/L alters glucose and glutamine utilization

a, Analysis of glucose uptake (left), glutamine uptake (center) and lactate secretion (right). b, Intracellular metabolite levels. Bars, mean of n = 4 (a) or n = 3 (b) replicate wells ± s.d. from representative experiments. c, Schematic of the TCA cycle including entry points for glucose- and glutamine-derived carbons. Isotope tracing was performed for metabolites shown in red. d,e, Fraction of each metabolite labeled by ¹³C derived from [U-¹³C]glutamine (¹³C-gln) (d) or derived from [U-¹³C]glucose (¹³C-glc) (e) over time (0–12 hours, h). Averages ± s.e.m. of three independent experiments are shown. f,g, Glutamine (f) and glucose (g) flux through αKG and malate pools. Averages ± s.e.m. of flux calculated for three independent experiments (shown in Figs. 2d,e) are shown. *, p < 0.05; **, p < 0.005; ***, p < 0.0005. p values determined by unpaired two-tailed Student’s t-tests.

Figure 3. Histone demethylation is regulated by intracellular α-ketoglutarate in embryonic stem cells

a, GC-MS analysis of the αKG/succinate ratio in four ESC lines (ESC-1-4) grown in either S/L or 2i/L medium. b, Western blot of ESC-1 and ESC-2 cells grown in 2i/L medium with or without glutamine for three days. Molecular weight marker (in kDa) is shown. c, Simplified schematic of the reaction mechanism of αKG-dependent dioxygenases. d, ESC-1 cells grown in S/L in the presence of increasing amounts of GSK-J4 for 24 hours. e, H3K27me3 ChIP-PCR of ESC-1 cells cultured in S/L or 2i/L containing 30 μM of GSK-J4 for five hours. Values represent fold-change (GSK-J4/control) at individual bivalent domain genes (n=14). f, H3K27me3 ChIP-PCR of CRISPR/Cas9 edited cells JMJD3Δ/Δ-1 (left) and JMJD3Δ/Δ-2 (right) cultured in S/L or 2i/L. Values represent fold-change (JMJD3Δ/Δ cells relative to control cells) at individual bivalent domain genes (n=10). Bars represent mean values. p values determined by unpaired two-tailed Student’s t-test (e,f). g, The ratio of αKG/succinate in ESC-1 cells grown in S/L or 2i/L medium with 1 μM or 5 μM of GSK-J4 or GSK-J5 for three hours. **, p < 0.001; ***, p < 0.0001 as determined by 2-way ANOVA with Sidak’s multiple comparisons post-test (a,g). Data are presented as the mean ± s.d (a) or s.e.m. (g) of triplicate wells from a representative experiment.

Figure 4. α-ketoglutarate promotes the maintenance of pluripotency

a,b, Colony formation assay using ESC-1 cells. Cells were plated at clonal density and media changed to experimental media containing either DM-αKG or DM-succinate on day 2 and then analyzed 4 days later by alkaline phosphatase (AP) staining and scored for number of differentiated, mixed and undifferentiated colonies. a, Representative brightfield images of AP-stained colonies. b, Quantification of colonies. DM-αKG has more undifferentiated colonies than vehicle or DM-succinate treated wells, ***, p < 0.0001, calculated by 2-way ANOVA with Tukey’s multiple comparisons post-test. c, Mean GFP intensity of Nanog-GFP cells treated for three days with or without DM-αKG. Data are presented as the mean ± s.e.m. (b) or 95% confidence intervals (c) of triplicate wells from a representative experiment.