Quinone-induced activation of Keap1/Nrf2 signaling by aspirin prodrugs masquerading as nitric oxide

Abstract

The promising therapeutic potential of the NO-donating hybrid aspirin prodrugs (NO-ASA) includes induction of chemopreventive mechanisms and has been reported in almost 100 publications. One example, NCX-4040 (pNO-ASA), is bioactivated by esterase to a quinone methide (QM) electrophile. In cell cultures, pNO-ASA and QM-donating X-ASA prodrugs that cannot release NO rapidly depleted intracellular GSH and caused DNA damage; however, induction of Nrf2 signaling elicited cellular defense mechanisms including upregulation of NAD(P)H:quinone oxidoreductase-1 (NQO1) and glutamate-cysteine ligase (GCL). In HepG2 cells, the "NO-specific" 4,5-diaminofluorescein reporter, DAF-DA, responded to NO-ASA and X-ASA, with QM-induced oxidative stress masquerading as NO. LC-MS/MS analysis demonstrated efficient alkylation of Cys residues of proteins including glutathione-S-transferase-P1 (GST-P1) and Kelch-like ECH-associated protein 1 (Keap1). Evidence was obtained for alkylation of Keap1 Cys residues associated with Nrf2 translocation to the nucleus, nuclear translocation of Nrf2, activation of antioxidant response element (ARE), and upregulation of cytoprotective target genes. At least in cell culture, pNO-ASA acts as a QM donor, bioactivated by cellular esterase activity to release salicylates, NO(3)(-), and an electrophilic QM. Finally, two novel aspirin prodrugs were synthesized, both potent activators of ARE, designed to release only the QM and salicylates on bioactivation. Current interest in electrophilic drugs acting via Nrf2 signaling suggests that QM-donating hybrid drugs can be designed as informative chemical probes in drug discovery.

Figure 1

Chemical structures of NO-ASA isomers, NCX-4016 (mNO-ASA) and NCX-4040 (pNO-ASA), and novel X-ASA and (ASA)₂ derivatives. Esterase-mediated bioactivation liberates SA, ASA, and HBN. The quinone methide is formed from o- and p- isomers, but not from m-isomers. The highly reactive electrophilic QM depletes GSH and modifies Cys residues of proteins including GST-P1 and Keap1.

Figure 2

Effect of NO-ASA and X-ASA analogues on intracellular GSH concentrations and DNA damage in HepG2 and Hepa 1c1c7 cells, respectively. (A) Total intracellular GSH concentrations measured after treatment of HepG2 cells with mBr-ASA (25 μM circle; orange), pBr-ASA (25 μM circle; green), or pMs-ASA (25 μM square; blue), data show mean and s.e.m. analyzed by one way ANOVA with Tukey's post test relative to data at the first time point (*** P < 0.001; ** P < 0.01); (B) Expression of GCLC mRNA in HepG2 cells after 4 h treatment with pMs-ASA, mBr-ASA, pBr-ASA, and pNO-ASA (10-25 μM). Data show mean and s.e.m. analyzed by one way ANOVA with Tukey's post test (*** P < 0.001) with respect to the DMSO-treated controls. (C) Single strand DNA damage, measured by Comet assay in Hepa 1c1c7 cells, after one hour incubation with pNO-ASA, mBr-ASA, pBr-ASA, and pMs-ASA (10 μM, 25 μM). Data show mean and s.e.m: significant differences with DMSO vehicle control were analyzed (** P < 0.01) by ANOVA with Tukey's post test. (D) Figurative representation of the relationship between flux of QM from bioactivation and biological consequences.

Figure 3

QM release from X-ASA analogue, pBr-ASA, and pNO-ASA induces Nrf2 translocation in HepG2 cells. Immunofluorescence images of fluorescein Nrf2 staining and DAPI nuclear staining with overlapped images showing translocation of Nrf2 from cytoplasm to cell nucleus. HepG2 cells were treated: (A) for 4 h with DMSO, pNO-ASA (6.25 μM, 25 μM), 4'-bromoflavone (200 ng/mL), and pNO-ASA (25 μM) with LY29402; (B) DMSO control, or pBr-ASA (6.25 μM and 25 μM) after 4 or 18 h, respectively.

Figure 4

Effect of NAC pretreatment on NQO1 induction by QM-releasing X-ASA. NQO1 induction in Hepa 1c1c7 cells was measured after 48 h incubation with X-ASA compounds and after 10 min pre-incubation with (solid bars) or without (open bars) addition of NAC (1 mM): (A) pMs-ASA; (B) oBr-ASA; (C) pBr-ASA. Fold-induction is calculated relative to in-plate DMSO vehicle controls; for comparison 4-BF as a positive control caused approximately sevenfold induction. Data show mean and s.e.m: ** significantly different (P < 0.05) with respect to the DMSO-treated controls by ANOVA with Tukey's post test.

Figure 5

The effect of X-ASA analogue, pBr-ASA, and pNO-ASA on DAF fluorescence in HepG2 cells. HepG2 cells were treated with DMSO, pNO-ASA (25 μM) and pBr-ASA (25 μM) for 24 h or spermine NONOate (25 μM) for 20 min. After treatment cells were labeled with DAF-DA and Hoechst nuclear stain. In similar experiments, the effect of NOS inhibition on pBr-ASA treatment showed no effect relative to pBr-ASA alone (Figure S1 in the Supporting Information).

Figure 6

LC-MS/MS analysis of GST-P1 modification by QM at Cys47 and Cys101. GST-P1 (15 μM) was treated with pNO-ASA or pBr-ASA (30 μM) for 30 min and unreacted Cys blocked with NEM (20 mM). The protein after in-gel digest was analyzed by LC-MS/MS, yielding extracted ion chromatograms of: a) QM modified Cys47 peptide, and; b) QM modified Cys101 peptide. The MS/MS spectra with b and y ion fragmentations are shown for: c) QM modified Cys47 peptide, and; d) QM modified Cys101 peptide.

Figure 7

Attenuated QM modification at Cys47 of mutant GST-P1(C101A) by addition of free Cys. GST-P1(C101A) (4 μM) was treated with pNO-ASA (25 μM) for 60 min in the presence of different Cys concentrations (1-500 μM) followed by blocking the unreacted Cys47 with NEM (20 mM). After in-gel digest, analysis by LC-MS/MS of the signal of Cys47 containing peptide (45-54: ASCLYGQLPK, 593+2) relative to the adjacent peptide peak (55-70: FQDGDLTLYQSNTILR, 943+2) yielded a relative measure of QM modification. Data shown are normalized mean and SD. (n=4).

Figure 8

Effect of novel QM releasing isomers, o(ASA)₂ and p(ASA)₂, on ARE-luciferase gene reporter induction. ASA derivatives at varying concentrations were incubated with HepG2 cells for 8 h. Results are shown as fold-induction relative to control as measured by enzyme activity and normalized to protein concentrations. Data show mean and SEM from at least triplicate measures.

Scheme 1