Transcriptional and Translational Downregulation of Thioredoxin Interacting Protein Is Required for Metabolic Reprogramming during G(1)

Abstract

Growth factor signaling drives increased glucose uptake and glycolysis-the Warburg effect-that supports macromolecular synthesis necessary for cell growth and proliferation. Thioredoxin interacting protein (TXNIP), a direct and glucose-induced transcriptional target of MondoA, is a potent negative regulator of glucose uptake and utilization. Thus, TXNIP may inhibit cell growth by restricting substrate availability for macromolecular synthesis. To determine TXNIP's contribution to metabolic reprogramming, we examined MondoA and TXNIP as cells exit quiescence and enter G(1). Serum stimulation of quiescent immortal diploid fibroblasts resulted in an acute upregulation of glucose uptake and glycolysis coinciding with downregulation of TXNIP expression. Ectopic expression of either MondoA or TXNIP restricted cell growth by blocking glucose uptake. Mechanistically, Ras-MAPK and PI3K/Akt signaling inhibit TXNIP translation and MondoA-dependent TXNIP transcription, respectively. We propose that the coordinated downregulation of MondoA transcriptional activity at the TXNIP promoter and inhibition of TXNIP translation are key components of metabolic reprogramming required for cells to exit quiescence.

Keywords: MondoA; TXNIP; metabolism; quiescence; transcription.

Figure 1.

TXNIP downregulation correlates with metabolic reprogramming and cell growth. (A) Increased glucose uptake in primary murine T-lymphocytes after 4- and 24-hour stimulation with α-CD3/α-CD28 (mean ± SD of duplicate biological samples) coincides with (B) TXNIP downregulation determined by Western blotting. (C) BrdU incorporation was determined in immortalized wild-type and TXNIP-null MEFs growing asynchronously in complete media or serum-starved for 48 hours. Data are presented as percentage of total BrdU incorporation relative to BrdU incorporation in asynchronous cells (mean ± SD of triplicate experiments). TXNIP expression and metabolic parameters were characterized during the G₀/G₁ transition in BJ hTERTs (D-F). (D) Expression of the indicated proteins was determined by Western blotting. Serum withdrawal for 72 hours (“Q”) was sufficient to synchronize cells in G₀ (p27^Kip1 stabilization), and cyclin A stabilization at 20 hours postserum addition preceded S-phase at 22 hours (determined by BrdU labeling, not shown) (asynchronous cells, “A”). (E) Expression of TXNIP mRNA determined by RT-qPCR. Data are presented as fold change relative to expression in asynchronous cells (“A”) normalized to RPL19 mRNA (mean ± SD of triplicate samples). (F) Glucose uptake was determined in quiescent cells (G₀) and at the time points indicated following serum addition (G₁) (mean ± SD of triplicate biological samples). (G) Glycolytic flux was determined in G₀ and at the time points indicated (G₁) (mean ± SD of triplicate biological samples).

Figure 2.

Ectopic TXNIP expression restricts glucose uptake and glycolysis and blocks G₁ progression. (A) Expression of the indicated proteins in BJ hTERTs under indicated growth conditions determined by Western blotting. (B) Glucose uptake was determined in control and TXNIP-mCherry–infected BJ hTERTs in G₀ and following 8 hours’ serum treatment. Data are presented as fold change relative to G₀ (mean ± SD; *P < 0.05, paired Student t test, n = 3). (C) Glycolytic flux was measured for conditions as in B. Each data point represents triplicate assay points from duplicate biological samples from 3 independent experiments (*P < 0.05, paired Student t test, n = 3). (D) BrdU labeling of control or TXNIP-mCherry–infected BJ hTERTs in G₀ and following 24-hour serum treatment (G₁/S). Data are presented as percentage BrdU labeling relative to total cells counted (mean ± SD; **P < 0.02, paired Student t test, n = 3).

Figure 3.

Dominant-active MondoA blocks G₁ progression. (A) Determination of MondoA, TXNIP, and β-tubulin expression in control and MondoA knockdown BJ hTERTs by Western blotting. (B) Expression of TXNIP, MondoA, ΔN237NLSMondoA, Hif1α, and β-tubulin in BJ hTERTs under indicated growth conditions determined by Western blotting. (C) Glucose uptake was determined in control and ΔN237NLSMondoA-expressing BJ hTERTs in G₀ and following 8 hours’ serum treatment. Data are presented as fold change relative to G₀ (mean ± SD; *P < 0.05, paired Student t test, n = 4). (D) Glycolytic flux was measured for conditions as in C. Each data point represents triplicate assay points from duplicate biological samples from 3 independent experiments (mean ± SD). (E) BrdU labeling of control or ΔN237NLSMondoA-expressing BJ hTERTs in G₀ and following 24-hour serum treatment (G₁/S). Data are presented as percentage BrdU labeling relative to total cells counted (mean ± SD; **P < 0.02, paired Student t test, n = 3).

Figure 4.

MondoA-dependent TXNIP transcription is acutely inhibited by serum. (A) Determination of TXNIP mRNA at the indicated times in G₀ in the presence of 5 µg/mL actinomycin D (●) or in the presence of serum (G₁) without actinomycin D (■) at the indicated time points by RT-qPCR. Data are presented as TXNIP message abundance relative to mRNA in G₀ (mean ± SD of triplicates from duplicate biological samples). Curve fitting in GraphPad Prism (plateau followed by one phase decay) revealed slopes for the exponential portion of each curve (y = 2.1139e^−0.006x, ● and y = 1.8702e^−0.013x, ■), which suggest comparable rates of decay. (B) TXNIP mRNA levels were determined by RT-qPCR after TSA treatment (100 ng/mL, 5 hours) (mean ± SD of triplicate assay points of biological duplicates). (C) ChIP was used to determine MondoA occupancy at the TXNIP promoter in BJ hTERTs under the indicated growth conditions. The data are presented as mean ± SD of 3 biological replicates (*P < 0.05, Student t test, n = 3).

Figure 5.

TXNIP expression is refractory to upregulation in early G₁. TXNIP, MondoA, and β-tubulin expression was determined by Western blotting (A) and RT-qPCR (B) following 2DOG treatment. There was 20 mM 2DOG added to cells 3 hours prior to the time points indicated. Data in A are that of a representative experiment. Data in B are presented as fold change relative to expression in untreated quiescent cells normalized to RPL19 mRNA (mean ± SD of triplicate samples).

Figure 6.

TXNIP translation is acutely inhibited by serum in early G₁. TXNIP, MondoA, and β-tubulin levels in BJ hTERTs were determined following treatment of quiescent cells with 10% serum (A) or 50 µg/mL cycloheximide (B) for the time points indicated by Western blotting.

Figure 7.

Acute growth factor signaling downregulates TXNIP expression. (A-C) TXNIP, MondoA, and β-tubulin levels were determined in BJ hTERTs by Western blotting under the conditions indicated. (C) There were 10 µM U0126 or 100 nM wortmannin added where indicated, and phospho-MAPK and phospho-Akt levels were also determined. (D) TXNIP mRNA levels were determined by RT-qPCR following 4 hours’ treatment under conditions indicated. Data are presented as fold change relative to expression in G₀ normalized to RPL19 mRNA (mean ± SD of triplicate samples of biological duplicates).

Figure 8.

Acute Ras-MAPK activation downregulates TXNIP translation. TXNIP, MondoA, phospho-MAPK, and β-tubulin levels were determined in BJ hTERTs infected with pBABE-puro-ER (vector) and pBABE-puro-ER:Ras^G12V (Ras^G12V) by Western blotting. Quiescent cells (G₀) were treated with 10% serum or increasing concentrations (64.5 nM, 645 nM, and 6.45 µM) of 4-hydroxytamoxifen (4OHT) as indicated for 4 hours.