Cyclin D1 extensively reprograms metabolism to support biosynthetic pathways in hepatocytes

Abstract

Cell proliferation requires metabolic reprogramming to accommodate biosynthesis of new cell components, and similar alterations occur in cancer cells. However, the mechanisms linking the cell cycle machinery to metabolism are not well defined. Cyclin D1, along with its main partner cyclin-dependent kinase 4 (Cdk4), is a pivotal cell cycle regulator and driver oncogene that is overexpressed in many cancers. Here, we examine hepatocyte proliferation to define novel effects of cyclin D1 on biosynthetic metabolism. Metabolomic studies reveal that cyclin D1 broadly promotes biosynthetic pathways including glycolysis, the pentose phosphate pathway, and the purine and pyrimidine nucleotide synthesis in hepatocytes. Proteomic analyses demonstrate that overexpressed cyclin D1 binds to numerous metabolic enzymes including those involved in glycolysis and pyrimidine synthesis. In the glycolysis pathway, cyclin D1 activates aldolase and GAPDH, and these proteins are phosphorylated by cyclin D1/Cdk4 in vitro. De novo pyrimidine synthesis is particularly dependent on cyclin D1. Cyclin D1/Cdk4 phosphorylates the initial enzyme of this pathway, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase (CAD), and metabolomic analysis indicates that cyclin D1 depletion markedly reduces the activity of this enzyme. Pharmacologic inhibition of Cdk4 along with the downstream pyrimidine synthesis enzyme dihydroorotate dehydrogenase synergistically inhibits proliferation and survival of hepatocellular carcinoma cells. These studies demonstrate that cyclin D1 promotes a broad network of biosynthetic pathways in hepatocytes, and this model may provide insights into potential metabolic vulnerabilities in cancer cells.

Keywords: BAY 2402234; aldolase; anaerobic glycolysis; cell cycle; cyclin D1; glyceraldehyde-3-phosphate dehydrogenase (GAPDH); liver regeneration; palbociclib; pentose phosphate pathway (PPP); purine; pyrimidine.

Figure 1

Modulation of glycolysis and mitochondrial respiration in live hepatocytes. Mouse AML12 hepatocytes were cultured in the presence or absence of 10% fetal bovine serum (FBS). Cells were treated with cyclin D1 (or control) siRNA or transduced with an adenovirus overexpressing cyclin D1 (or control) as indicated, and harvested after 48 h. A, Western blot of cyclin D1 expression. B, DNA synthesis are measured by BrdU uptake. C, basal and compensatory glycolysis as measured by ECAR. D, glucose uptake and media lactate content (normalized to cellular protein content). E, basal and maximal respiration and spare respiratory capacity as measured by OCR (∗p < 0.05; ∗∗p < 0.01; and ∗∗∗p < 0.001). ECAR, extracellular acidification rate; OCR, oxygen consumption rate.

Figure 2

Cyclin D1 binds metabolic proteins in the liver. Livers were transduced with cyclin D1-HA-FLAG (or a control vector encoding GFP), and lysates were subjected to tandem affinity purification followed by mass spectrometry. A, selected cell cycle proteins and metabolic enzymes binding to cyclin D1 by mass spectroscopy. The mouse gene name is shown for each. B, lysates of livers transduced with either full-length cyclin D1-HA-FLAG or a truncation mutant of the midportion of cyclin D1 (RD-FLAG (14)) were subjected to immunoprecipitation and elution from anti-FLAG beads as described in the Experimental procedures. The eluted proteins and the corresponding lysates were subjected to Western blot analysis of the indicated proteins. HA, hemagglutinin.

Figure 3

Analysis of the glycolytic pathway. AML12 hepatocytes were cultured as in Figure 1 in the presence of serum and [¹³C₆]-glucose for 24 h, followed by MS analysis of cell extracts and NMR of media (n = 3 per condition). A, abundance of key ¹³C isotopologues of glycolytic metabolites by IC-MS and media lactate by NMR. Values are normalized to cellular protein content for each replicate. B, Western blot of aldolase B and GAPDH. C, aldolase and GAPDH activity assays in cells with cyclin D1 siRNA (top), or in serum-deprived cells transduced with cyclin D1 as in Figure 1 (ADV-D1, bottom). D, nonradioactive in vitro kinase assay using recombinant cyclin D1/Cdk4, aldolase B, and GAPDH along with ATPγS. Western blot was performed using an antibody to thiophosphate ester (top, representing the phosphorylated protein) or the substrate proteins (bottom). ADV-D1, adenovirus expressing cyclin D1; Cdk4, cyclin-dependent kinase 4; IC-MS, ion chromatography with mass spectrometry; MS, mass spectrometry.

Figure 4

Regulation of the PPP by cyclin D1. Hepatocytes were cultured as in Figure 3. A, the abundance of unlabeled and fully ¹³C labeled (M6 or M5) PPP metabolites as determined by IC-UHR FTMS. B, Western blot of G6PD. C, G6PD enzyme activity assay using a commercial kit. G6PD, glucose-6-phosphate dehydrogenase; IC-UHR FTMS, ion chromatography coupled with ultra high-resolution Fourier transform mass spectrometry; PPP, pentose phosphate pathway.

Figure 5

Cyclin D1 promotes purine synthesis. Hepatocytes were cultured as in Figure 3. A, abundance of ¹³C labeled adenine in AXP (AMP, ADP, and ATP combined) by NMR. B, ¹³C-labeled isotopologues of ATP by IC-UHR FTMS. C, abundance of ¹³C-enriched purine rings of ATP, GFP, and IMP, calculated by the sum of the ¹³C M6–M8 isotopologues. D, changes in the level of mRNA transcripts of genes involved in purine synthesis induced by cyclin D1 siRNA (relative to control siRNA) from our prior RNA-seq study in AML12 hepatocytes cultured under these conditions (ref. (14)). E, Western blot of ATF4 and Psat1. Relative expression is shown in the graphs below. F, expression of Mthfd2, Ppat, and Psat1 proteins by reverse phase protein array analysis. IC-UHR FTMS, ion chromatography coupled with ultra high-resolution Fourier transform mass spectrometry.

Figure 6

Regulation of pyrimidine synthesis and CAD by cyclin D1. Hepatocytes were cultured as in Figure 3. A, amount of ¹³C-UXP (UMP, UDP, and UTP combined) at 5 and 6 positions by NMR. B, diagram of the pyrimidine synthesis pathway. C, abundance of ¹³C isotopologues of UTP by MS. D, concentration of N- carbamoyl-L-aspartate isotopologues by MS. E, abundance of ¹³C-aspartate by NMR (left) and MS (right). F, expression of total CAD, phospho-Ser1859 CAD, and ph-S6K1 by Western blot in AML12 hepatocytes. G, nonradioactive in vitro kinase assay using recombinant cyclin D1/Cdk4 along with CAD-HA-FLAG isolated from fasting mouse liver and ATPγS. Western blot was performed using an antibody to thiophosphate ester. H, Western blot analysis of resting (0 h) and regenerating liver 42 h after PH in control (GFP) or hepatocyte-specific KO (Cre) mice. The relative expression of total and phospho-Ser1859 CAD are shown. I, Phos-Tag Western blot of total and phospho-Ser1859 CAD from the PH model. Cyclin D1 KO (Cre) was associated with a less phosphorylated form of total and phospho-Ser1859 CAD compared to control mice (GFP). CAD, carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase; Cdk4, cyclin-dependent kinase 4; HA, hemagglutinin; MS, mass spectrometry; PH, partial hepatectomy.

Figure 7

Effect of Cdk4 and DHODH inhibition in HCC cells. HuH7 cells were cultured in the presence of vehicle, palbociclib, and BAY2402234 as indicated. A, DNA synthesis as determined by BrdU uptake in cells treated with palbociclib (50 nM) and/or BAY2402234 (5 nM). B, Western blot of HuH7 cells cultured as in panel A. C, viability of cells treated with BAY2402234 at the indicated concentrations with and without palbociclib (5 μM, left), or with palbociclib at the indicated concentrations with and without BAY2402234 (5 nM, right). Cdk4, cyclin-dependent kinase 4; DHODH, dihydroorotate dehydrogenase; HCC, hepatocellular carcinoma.