PFKFB3-mediated Pro-glycolytic Shift in Hepatocellular Carcinoma Proliferation

Abstract

Background & aims: Metabolic reprogramming, in particular, glycolytic regulation, supports abnormal survival and growth of hepatocellular carcinoma (HCC) and could serve as a therapeutic target. In this study, we sought to identify glycolytic regulators in HCC that could be inhibited to prevent tumor progression and could also be monitored in vivo, with the goal of providing a theragnostic alternative to existing therapies.

Methods: An orthotopic HCC rat model was used. Tumors were stimulated into a high-proliferation state by use of off-target liver ablation and were compared with lower-proliferating controls. We measured in vivo metabolic alteration in tumors before and after stimulation, and between stimulated tumors and control tumors using hyperpolarized ¹³C magnetic resonance imaging (MRI) (h¹³C MRI). We compared metabolic alterations detected by h¹³C MRI to metabolite levels from ex vivo mass spectrometry, mRNA levels of key glycolytic regulators, and histopathology.

Results: Glycolytic lactate flux increased within HCC tumors 3 days after tumor stimulation, correlating positively with tumor proliferation as measured with Ki67. This was associated with a shift towards aerobic glycolysis and downregulation of the pentose phosphate pathway detected by mass spectrometry. MRI-measured lactate flux was most closely coupled with PFKFB3 expression and was suppressed with direct inhibition using PFK15.

Conclusions: Inhibition of PFKFB3 prevents glycolytic-mediated HCC proliferation, trackable by in vivo h¹³C MRI.

Keywords: Aerobic Glycolysis; Hyperpolarized (13)C MRI.

Figure 1

RFA of the liver promotes off-target HCC tumor growth. The volume change between baseline and 7 days post-stimulation/control (ΔVolume = Volume_after – Volume_before) in both stimulated (STI) and control/sham (Ctrl) group. The comparison was performed with a Mann-Whitney test.

Figure 2

h¹³C MRI of N1S1 orthotopic tumors within the left hepatic lobe of female Fischer rats.A, T2-weighted proton MRI (grayscale), with h¹³C pyruvate (middle panel) and h¹³C lactate (right panel) imaging overlaid in green, showing high lactate production within the tumor (white arrows) and no lactate production in kidney (red arrows). B, Example h¹³C pyruvate and h¹³C lactate signal over time in both control and stimulated tumors from baseline to 3 days post-stimulation/control. C, The baseline lactate flux in tumor (T) and normal liver (NL), expressed as LPR. D and E, LPR in tumor (D) and normal liver (E) from baseline to 3 days post-stimulation/control. F–G, Lactate K_PL (F) and lactate R1^eff (G) in tumor between baseline and 3 days post-stimulation/control. The P value was evaluated by paired t test.

Figure 3

¹³C alanine and¹³C bicarbonate flux in N1S1 tumor.A–B, Bicarbonate flux (BPR) (A) and alanine flux (APR) (B). T, tumor, NL, normal liver. C– D, APR in tumor (C) and normal liver (D) from baseline to 3 days post-stimulation/control. E–F, BPR in tumor (E) and normal liver (F) from baseline to 3 days post-stimulation/control. The P value was evaluated by paired t test. G–H, Correlation between Ki67 expression and post-stimulation APR (G) or post-stimulation ΔAPR (post-stimulation APR – baseline APR) (H). I–J, Correlation between tumor volume change ΔVolume and post-stimulation APR (I) or post-stimulation ΔAPR (J).

Figure 4

Stimulated normal liver by stimulation promotes HCC tumor proliferation.A, Ki67 expression in orthotopic tumors was examined by immunohistochemistry in both stimulated (STI) and control/sham (Ctrl) group. Representative images were provided as indicated. B, The volume change between baseline and 3 days post-stimulation/control (ΔVolume = Volume_after – Volume_before) in both stimulated and control/sham group. Analyses were performed using the Mann-Whitney test. C, Correlation of Ki67expression and volume change.

Figure 5

Correlation between tumor growth and¹³C-lactate flux post-stimulation.A–B, There is positive correlation between Ki67 expression and post-stimulation LPR and ΔLPR (post-stimulation LPR – baseline LPR). C–D, There is positive correlation between Ki67 expression and post-stimulation K_PL and ΔK_PL (post-stimulation K_PL – baseline K_PL). E–F, There is positive correlation between tumor volume change ΔVolume and post-stimulation LPR and ΔLPR. G–H, There is positive correlation between tumor volume change ΔVolume and post-stimulation K_PL and ΔK_PL.

Figure 6

Metabolic profiles of N1S1 tumors.A, Multivariate analysis of metabolomic data using 3-D PLS-DA model. B, Total heat map demonstrating contrasting metabolic profiles of N1S1 tumors of control vs post-stimulated groups.

Figure 7

Quantitative metabolome profiling of post-stimulated N1S1 tumors.A, A volcano plot shows significantly different metabolites in stimulated tumors compared with controls. Sig.down, Metabolites significantly lower in stimulated tumor; Sig.up, metabolites significantly higher in stimulated tumor; Unsig, no difference between 2 groups. B–C, Lactate (B) and alanine levels (C) in stimulated (STI) tumors vs controls (Ctrl). Comparisons were made using the Mann-Whitney test. D–E, Focused heat maps demonstrating upregulation of glycolysis/ gluconeogenesis and downregulation of the PPP in stimulated tumors.

Figure 8

mRNA expression of glycolysis-related genes in stimulated/non-stimulated N1S1 tumors and liver tissue.A, Representative gene expression patterns in tumor (T), liver tissue adjacent to RFA/sham site (periablational) (P), and normal liver (N). B–F, Relative HIF-1 (B), GLUT1 (C), PKM (D), HK2 (E), and GPT1 (F) gene expression levels within tumor, periablational, and normal liver, in stimulated (STI) and control (Ctrl) groups. Densitometry quantification of band intensity is presented as a percentage of relative densitometry normalized to the CYPA gene. Analyses were performed using 2-way ANOVA.

Figure 9

PFKFB3 expression in N1S1 tumors.A, Representative gene expression patterns in tumor (T), stimulated or nearby sham liver tissue (S), and normal liver (N). PFKFB3 gene expression levels in stimulated (STI) and control (Ctrl) groups. Densitometry quantification of band intensity is presented as a percentage of relative densitometry normalized to the cyclophilin A (CYPA) gene. Analyses were performed using 2-way ANOVA. B–D, PFKFB3 immunohistochemistry staining demonstrating different levels of expression within tumor, stimulated/sham area, and normal liver. Quantitative immunohistochemistry analysis in tumors (C), and stimulated area (STI area), and normal liver (NL) (D). Analyses were performed using the Mann-Whitney test. E–I, Correlation PFKFB3 expression with Ki67 (E), post-stimulated LPR (F), ΔLPR (G), post-stimulated lactate K_PL (H) or ΔK_PL (I).

Figure 10

PFK15 inhibits N1S1 cell viability.A, Flow cytometric dotplot analysis of Annexin V/PI staining in heat-treated cells 2 days following heat treatment. B, Cellular proliferation was measured by the MTT assay in heat-treated cells at 24 hours, 48 hours, and 72 hours. Values are expressed as mean and standard deviation. C, Cell viability was measured by the MTT assay in heat-treated cells treated with the indicated concentrations of PFK15 for 72 hours. Values are expressed as mean and standard deviation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001; compared with the 37 °C group. Analyses were performed using 2-way ANOVA.

Figure 11

Inhibition of PFKFB3 in N1S1 with PFK15.A–B, The lactate flux (A) and lactate K_PL (B) in control (Ctrl), stimulated (STI), and stimulation with PFK15 (STI+PFK15) group. The P value was evaluated by paired t test. C, PFKFB3 gene expression levels within tumor, stimulated area/sham liver tissue (STI/Sham area), and normal liver (NL). Densitometry quantification of band intensity is presented as a percentage of relative densitometry normalized to the cyclophilin A (CYPA) gene. Analyses were performed using 2-way ANOVA. D, Immunohistochemistry shows PFKFB3 expression in tumor and normal liver. Analyses were performed using the Mann-Whitney test.