Loss of pyruvate kinase M2 limits growth and triggers innate immune signaling in endothelial cells

Abstract

Despite their inherent proximity to circulating oxygen and nutrients, endothelial cells (ECs) oxidize only a minor fraction of glucose in mitochondria, a metabolic specialization that is poorly understood. Here we show that the glycolytic enzyme pyruvate kinase M2 (PKM2) limits glucose oxidation, and maintains the growth and epigenetic state of ECs. We find that loss of PKM2 alters mitochondrial substrate utilization and impairs EC proliferation and migration in vivo. Mechanistically, we show that the NF-κB transcription factor RELB is responsive to PKM2 loss, limiting EC growth through the regulation of P53. Furthermore, S-adenosylmethionine synthesis is impaired in the absence of PKM2, resulting in DNA hypomethylation, de-repression of endogenous retroviral elements (ERVs) and activation of antiviral innate immune signalling. This work reveals the metabolic and functional consequences of glucose oxidation in the endothelium, highlights the importance of PKM2 for endothelial growth and links metabolic dysfunction with autoimmune activation in ECs.

Fig. 1

Loss of endothelial PKM2 leads to mitochondrial dysfunction. a RT-qPCR analysis of PKM mRNA expression showing that PKM2 is more abundantly expressed in HUVECs (n = 3). b Western blot analysis of PKM1 and PKM2 expression in HUVECs, HEK293 and mouse skeletal muscle. c Relative energy charge ([ATP] + 1/2[ADP]/[ATP] + [ADP] + [AMP]) in control, PKM2^KD and PKM^KD ECs (n = 3). Incorporation of [U-13C₆]-glucose into m+3 lactate (d) and m+2 citrate (e) in control, PKM2^KD and PKM^KD ECs (n = 3). f Oxygen consumption rate (OCR) in control, PKM2^KD and PKM^KD ECs under basal conditions and in response to oligomycin, fluoro-carbonyl cyanide phenyl-hydrazone (FCCP) and antimycin A/rotenone (n = 6, data represent mean ± s.d.). g Incorporation of [U-13C₆]-glucose into m+2 α-ketoglutarate in control, PKM2^KD and PKM^KD ECs (n = 3). h Incorporation of [U-13C₆]-glucose into m+2 aspartate in control, PKM2^KD and PKM^KD ECs (n = 3). i Ratio of keto-leucine/isoleucine in control, PKM2^KD and PKM^KD ECs. j Incorporation of [U-13C₆]-glucose into m+3 alanine in control, PKM2^KD and PKM^KD ECs (n = 3). k Relative levels of different acylcarnitine species in control, PKM2^KD and PKM^KD ECs (n = 3). l Normalized MitoSOX fluorescence intensity in control, PKM2^KD and PKM^KD ECs. a, c–l Data represent means ± s.d., m+n represents the mass isotopomers for individual metabolites (***P < 0.001, **P < 0.01, *P < 0.05 by one-way analysis of variance (ANOVA) followed by Tukey’s HSD test)

Fig. 2

Loss of PKM2 leads to angiogenic sprouting defects. a EdU-positive cell numbers in control, PKM2^KD and PKM^KD ECs (n = 11). b Relative cell numbers at the indicated time points in control, PKM2^KD and PKM^KD ECs (n = 6). c Scratch wound assay in control and PKM2^KD ECs (n = 8). d Representative confocal projections of 48 hpf Tg(kdrl:EGFP) pkma2^+/+ pkmb^+/+ and pkma2^MZ pkmb^−/− zebrafish embryos (scale bar = 50 μm). e Percentage of intersegmental vessels (ISVs) that have failed to connect with neighbouring ISVs in 48 hpf zebrafish embryos (n = 8). Data in a–c and e represent means ± s.e.m. (***P < 0.001 by one-way analysis of variance (ANOVA) followed by Tukey’s HSD test (a, e) or two-tailed Student's t test (c))

Fig. 3

Acute deletion of Pkm2 in mouse endothelial cells leads to angiogenic sprouting defects. a Representative confocal projections of PECAM, ERG and EdU immunostained retinas from control and Pkm2^iEC-KO mice (A, Artery; V, Vein). b Radial outgrowth in control and Pkm2^iEC-KO mice (n = 6). c Representative confocal projections of angiogenic front vessels in PECAM, ERG and EdU-stained retinas from control and Pkm2^iEC-KO mice. d ERG⁺ ECs, EdU⁺ ECs, EC area per field and branch point density at the angiogenic front of control and Pkm2^iEC-KO mice (n = 6). All analyses in the mouse retina were performed on postnatal day 7. Data in b and d represent means ± s.e.m. (**P < 0.01, *P < 0.05 by two-tailed Student's t test). Scale bars in a and c = 200 μm

Fig. 4

Loss of PKM2 leads to coordinated metabolic and protein expression changes. a Analysis of steady-state metabolite levels in control vs PKM2^KD ECs identifies the TCA cycle, pyrimidine metabolism and cysteine and methionine metabolism as significantly regulated pathways. b Metabolomics profile of control vs PKM2^KD ECs (log₂(fold change) vs −log₂(p value), two-tailed Student's t test, n = 3). c Heat map of protein expression for key enzymes in the pyrimidine synthesis and serine, glycine, one-carbon metabolic networks in control vs PKM2^KD ECs

Fig. 5

RELB-P53 suppresses endothelial growth in the absence of PKM2. a EdU and phospho-Histone H3 positive cell numbers in control, RELB^KD, PKM2^KD and PKM2^KD/RELB^KD ECs (n = 6). b Western blot analysis of RELB, P53, P21 and PKM2 expression in control, RELB^KD, PKM2^KD and PKM2^KD/RELB^KD ECs. c EdU and phospho-Histone H3 positive cell numbers in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs (n = 12). d Western blot analysis of RELB, P53, P21 and PKM2 expression in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs. e mRNA expression levels of key enzymes for pyrimidine synthesis in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs (RNA-seq analysis, n = 3). f Relative N-carbamoyl aspartate levels, UTP levels, incorporation of [U-13C₆]-glucose into m+3 and m+5 UTP, and relative dTTP levels in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs (n = 3). a, c Data represent means ± s.e.m., e, f data represent means ± s.d., m+n represents the mass isotopomers for individual metabolites (***P < 0.001, **P < 0.01, *P < 0.05 by one-way analysis of variance (ANOVA) followed by Tukey’s HSD test)

Fig. 6

Loss of PKM2 impairs methylation capacity, reduces DNA methylation and leads to the expression of endogenous retroviral elements. a Western blot analysis of MAT2A, P53, P21 and PKM2 in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs. b Relative levels of S-adenosylmethionine and S-adenosylhomocysteine and SAM/SAH ratio in control, P53^KD, PKM2^KD and PKM2^KD/P53^KD ECs (n = 3). c Percentage 5-methylcytosine (5mC) levels in control and PKM2^KD ECs (n = 4). d Heat map of protein expression for components of the cellular response to viral infection in control and PKM2^KD ECs. e Representative confocal projections of total double-stranded RNA (dsRNA) staining in control and PKM2^KD ECs (scale bar = 20 μm). f Normalized fluorescence intensity for dsRNA staining in control and PKM2^KD ECs (n = 7). g Relative mRNA expression of the indicated endogenous retroviruses in control and PKM2^KD ECs (n = 3). h Restriction digestion of bisulfite-treated DNA amplified from the MLT1B and MER4D genomic loci from control and PKM2^KD ECs (U, undigested/unmethylated DNA; D, digested/methylated DNA). i Relative mRNA expression of the indicated endogenous retroviruses and interferon-stimulated genes in control, RELB^KD, PKM2^KD and PKM2^KD/RELB^KD ECs. b, c, f, g, i Data represent means ± s.d. (***P < 0.001, **P < 0.01, *P < 0.05 by one-way analysis of variance (ANOVA) followed by Tukey’s HSD test)