Mutant Kras- and p16-regulated NOX4 activation overcomes metabolic checkpoints in development of pancreatic ductal adenocarcinoma

Abstract

Kras activation and p16 inactivation are required to develop pancreatic ductal adenocarcinoma (PDAC). However, the biochemical mechanisms underlying these double alterations remain unclear. Here we discover that NAD(P)H oxidase 4 (NOX4), an enzyme known to catalyse the oxidation of NAD(P)H, is upregulated when p16 is inactivated by looking at gene expression profiling studies. Activation of NOX4 requires catalytic subunit p22^phox, which is upregulated following Kras activation. Both alterations are also detectable in PDAC cell lines and patient specimens. Furthermore, we show that elevated NOX4 activity accelerates oxidation of NADH and supports increased glycolysis by generating NAD⁺, a substrate for GAPDH-mediated glycolytic reaction, promoting PDAC cell growth. Mechanistically, NOX4 was induced through p16-Rb-regulated E2F and p22^phox was induced by Kras^G12V-activated NF-κB. In conclusion, we provide a biochemical explanation for the cooperation between p16 inactivation and Kras activation in PDAC development and suggest that NOX4 is a potential therapeutic target for PDAC.

Figure 1. Activated Kras or silenced p16 increased NOX4/p22^phox expression and elevated NOX activity.

(a) The expression of Kras and p16 was analysed by immunoblotting in HPNE control, HPNE/Kras and HPNE/Kras^G12V/shp16 cells. The copy numbers of Kras^WT and Kras^G12V were analysed by qPCR assay in these cells. (b) Heat map of gene expression between HPNE/Kras and HPNE/Kras^G12V/shp16 cells identified using cDNA microoarray. (c) Functional categorization of 614 genes upregulated in response to p16 suppression in HPNE/Kras^G12V cells. (d,f) The expression of NOX4 and p22^phox was analysed by qPCR and immunoblotting in indicated cells. (h) The expression of NOX4, p22^phox and p16 was analysed by immunoblotting in HPNE/Kras^G12V, Colo357 and Capan-2 cells transfected with two independent p16 siRNAs. (e,g,i) NOX activity was determined in indicated cells by measuring NADPH-dependent superoxide (O₂⁻) generation with the lucigenin-enhanced chemiluminescence assay. β-actin was used as the internal loading control. Each bar represents the mean±s.d. Data in d,e,g,i are presented as mean±s.d. (n=3). **P<0.01 for indicated comparison (one-way analysis of variance (ANOVA) with the Newman Keul's multiple comparison test).

Figure 2. NOX4 and p22^phox are overexpressed in PDAC.

(a) The mRNA levels of NOX4 and p22^phox were compared between 11 PDAC cell lines and 2 normal cell lines. (b) The mRNA levels of NOX4 and p22^phox were compared between PDAC tissues and adjacent normal tissues (N=21). (c) The expression of NOX4 and p22^phox in different PDAC cell lines was analysed by immunoblotting and compared with that in HPNE cells. β-actin was included as a loading control. (d) NOX activity was compared between 11 PDAC cell lines and 2 normal cell lines. (e) Representative IHC staining with H&E or anti-NOX4 antibody in sections of formalin-fixed tissue from wild-type or Pdx1-Cre; Kras^LSL-G12D, Ink4a^F/F mice. (f) Representative IHC staining showing no expression of NOX4 in normal pancreas acinous cells, weak and strong positive staining (10 × ) in PDAC tissues. Lower panels represent higher magnifications (× 40). NOX4 expression is considered to be significantly different between PDAC and normal tissues and higher in PDAC group (P<0.0001 analysed by Fisher's exact test). Scale bars in e, f, 100 μm. (g) NOX4 and p22^phox expression in multiple cancer microarray data sets available from Oncomine (https://www.oncomine.com//). Data in a,b,d are representative of three independent experiments and presented as mean±s.d. *P<0.05, **P<0.01 for indicated comparison (Student unpaired t-test). In g, box plot centre line and box limits represent median and interquartile range (IQR), respectively. Whisker lines represent data range (maximum and minimum) but not exceeding 1.5 × IQR from IQR. Data points beyond 1.5 × IQR from IQR are shown with circles.

Figure 3. NOX4 plays a critical role in regulation of glycolysis by generating NAD⁺.

(a) OCR was determined in HPNE, HPNE/Kras^G12V and HPNE/Kras^G12V/shp16 cells. (b) Glucose uptake and lactate production, NAD⁺/NADH and NADP⁺/NADPH ratios were measured in HPNE, HPNE/Kras^G12V and HPNE/Kras^G12V/shp16 cells. (c) Immunoblotting analysis showed the NOX4 and p22^phox knockdown efficiency. Scrambled siRNA (sc) was used as a negative control here and in d–f. (d) NOX activity, NAD⁺/NADH, NADP⁺/NADPH ratios and glycolytic activity (glucose uptake, lactate production and ATP levels) were measured in NOX4-silenced HPNE/Kras^G12V/shp16 cells. (e) NOX activity, glucose uptake and lactate production levels were measured in p22^phox-silenced HPNE/Kras^G12V/shp16 cells. (f) Glucose uptake was measured in AsPc-1 and Panc-28 cells after siRNA depletion of NOX4 or p22^phox or co-expressing siRNA resistant NOX4 (NOX4-R). (g) The expression levels of NOX4 were analysed by qPCR and immunoblotting in NOX4-overexpressing HPNE cells. Data are presented as mean±s.d. (n=3). **P<0.01 for indicated comparison (Student unpaired t-test). (h) The pyruvate and lactate levels were measured in HPNE/NOX4 or HPNE/Kras^G12V/shp16 cells compared with parental HPNE cells using metabolite isotope tracing experiments with 13 carbon labelled glucose (U-¹³C₆ Glu). Data in b,d–f,h are presented as mean±s.d. (n=3). *P<0.05, **P<0.01 for indicated comparison (one-way analysis of variance (ANOVA) with the Newman Keul's multiple comparison test).

Figure 4. Kras upregulates p22^phox expression via the Tak1-NF-κB pathway.

(a) The expression of p-Tak1, Tak1, p-NF-κB/p65 and NF-κB/p65 was analysed by immunoblotting in HPNE, HPNE/Kras^G12V and HPNE/Kras^G12V/shp16 cells. (b) The expression of Tak1, p-NF-κB/p65, NF-κB/p65 and p22^phox was analysed by immunoblotting in AsPc-1/i-Tak1shRNA cells with Dox induction. (c) The expression of Kras, p-NF-κB/p65, NF-κB/p65 and p22^phox was analysed by immunoblotting in mPDAC/iKras cells with Dox induction. (d) The NF-κB/p65 and p22^phox mRNA expression was analysed by qPCR in HPNE/Kras^G12V/shp16 cells transfected with NF-κB/p65 siRNAs. (e) The expression of phosphorylation-defective mutant IκBα, p-NF-κB/p65, NF-κB/p65 and p22^phox was analysed by immunoblotting in wild-type (WT) and IκBα-mutant (Mu) AsPc-1 and Panc-28 cells. (f) NOX activity was compared in IκBα wild-type and mutant AsPc-1 and Panc-28 cells. (g) The sequences of human p22^phox promoter regions are presented with NF-κB binding sites (BS-1 and BS-2) indicated. (h) The activities of NF-κB binding sites BS-1 and BS-2 in p22^phox promoter were analysed in indicated cells by ChIP and qPCR assay. IgG was used as a negative control. (i) The activities of NF-κB binding sites BS-1 in p22^phox promoter were analysed in indicated cells by ChIP assay and qPCR. (j) The expression of p-NF-κB/p65, NF-κB/p65 and p22^−phox was analysed by immunoblotting in AsPc-1 cells treated with TNF-α (10 ng ml⁻¹) for 48 h. β-actin was used as the internal loading control. Data in d,f,h,i are presented as mean±s.d. (n=3). **P<0.01 for indicated comparison (Student unpaired t-test).

Figure 5. p16 upregulates NOX4 expression via the Rb-E2F pathway.

(a) The expression of p-Rb (Ser780), p-Rb (Ser795), p-Rb (Ser807/811), Rb and E2F1 in HPNE/Kras^G12V and HPNE/Kras^G12V/shp16 cells was analysed by immunoblotting. β-actin and Lamin A/C was used as loading control. (b) The mRNA expression of E2F family members (E2F1, E2F2, E2F3, E2F4 and E2F5) in HPNE/Kras^G12V/shp16, AsPc-1 and Panc-28 cells was analysed by RT-PCR. (c) The expression of NOX4 was analysed by qPCR in HPNE/Kras^G12V/shp16 and Panc-28 cells transfected with siRNAs targeting different E2F family members. (d) The sequences of human NOX4 promoter regions are presented with E2F binding sites (BS-1 and BS-2) indicated. (e) The activities of E2F1 binding sites BS-1 and BS-2 in the NOX4 promoter were analysed in indicated cells by ChIP and qPCR assay. (f) NOX activity was determined in HPNE/Kras^G12V/shp16, AsPc-1 and Panc-28 cells transfected with E2F1 siRNA. (g) The activities of E2F1 binding sites BS-1 in the NOX4 promoter were analysed in indicated cells by ChIP and qPCR assay. IgG was used as a negative control. Data in c,e,f,g are presented as mean±s.d. (n=3). **P<0.01 for indicated comparison (Student unpaired t-test).

Figure 6. Suppression of NOX4 by shRNA or DPI inhibits PDAC growth in vitro and in vivo.

(a,b) Growth curves and clonogenic growth were measured in NOX4-silenced HPNE/Kras^G12V/shp16 cells. (c) Weights of PDAC tumours removed on day 60 from mice (N=8) injected orthotopically with NOX4-silenced and control HPNE/Kras^G12V/shp16 cells. (d) The nude mice were inoculated subcutaneously with indicated cells (N=5). The tumour sizes were measured throughout the experiment to evaluate NOX4 knockdown effects. (e) Tumour weight derived from indicated group was measured. (f) Photograph and comparison of excised tumour size. (g) Sizes and weights of PDAC tumours (N=10) removed on day 42 from mice injected orthotopically with NOX4-silenced AsPc-1 cells or control AsPc-1 cells treated with NOX inhibitor DPI (1.5 mg kg⁻¹ per mouse, i.p., 5 days per week). (h) Paraffin-embedded tumour sections were stained with H&E (P: PDAC; S: Spleen) or anti-Ki67 antibody (Scale bars, 100 μm); apoptotic cells were visualized by TUNEL staining (green) and counterstained with DAPI (blue) (Scale bars, 10 μm). Quantification of proliferation index and apoptotic index in PDAC tumours was shown. (i) Sizes and weights of tumour tissues (N=5) removed on day 42 from mice injected orthotopically with AsPc-1/i-shNOX4 cells. On: mice were fed with doxy-containing water from 2 weeks after inoculation, and continued for 4 weeks. **P<0.01 for indicated comparison (Student unpaired t-test). (j) Kaplan–Meier overall survival analysis for mice treated with DPI as indicated (Kaplan–Meier analysis with the log-rank test). On: mice were fed with doxy-containing water from 3 weeks of age. DPI: mice were treated with DPI (1.5 mg kg⁻¹ per mouse, i.p., 5 days per week) from 8 weeks of age. Data in a,b are presented as mean±s.d. (n=3). Data in c–e,g,i are representative of two independent experiments and presented as mean±s.d. **P<0.01 for indicated comparison (one-way analysis of variance (ANOVA) with the Newman Keul's multiple comparison test).

Figure 7. Mutant Kras and inactivated p16 orchestrate metabolic reprogramming in PDAC.

The proposed working model of the function and mechanisms of NOX4 in reprogramming aerobic glycolysis initiated by activated Kras and inactivated p16 in PDAC.