Colon tumor cell growth-inhibitory activity of sulindac sulfide and other nonsteroidal anti-inflammatory drugs is associated with phosphodiesterase 5 inhibition

Abstract

Nonsteroidal anti-inflammatory drugs (NSAID) display promising antineoplastic activity, but toxicity resulting from cyclooxygenase (COX) inhibition limits their clinical use for chemoprevention. Studies suggest that the mechanism may be COX independent, although alternative targets have not been well defined. Here, we show that the NSAID sulindac sulfide (SS) inhibits cyclic guanosine 3',5'-monophosphate (cGMP) phosphodiesterase (PDE) activity in colon tumor cell lysates at concentrations that inhibit colon tumor cell growth in vitro and in vivo. A series of chemically diverse NSAIDs also inhibited cGMP hydrolysis at concentrations that correlate with their potency to inhibit colon tumor cell growth, whereas no correlation was observed with COX-2 inhibition. Consistent with its selectivity for inhibiting cGMP hydrolysis compared with cyclic AMP hydrolysis, SS inhibited the cGMP-specific PDE5 isozyme and increased cGMP levels in colon tumor cells. Of numerous PDE isozyme-specific inhibitors evaluated, only the PDE5-selective inhibitor MY5445 inhibited colon tumor cell growth. The effects of SS and MY5445 on cell growth were associated with inhibition of β-catenin-mediated transcriptional activity to suppress the synthesis of cyclin D and survivin, which regulate tumor cell proliferation and apoptosis, respectively. SS had minimal effects on cGMP PDE activity in normal colonocytes, which displayed reduced sensitivity to SS and did not express PDE5. PDE5 was found to be overexpressed in colon tumor cell lines as well as in colon adenomas and adenocarcinomas compared with normal colonic mucosa. These results suggest that PDE5 inhibition, cGMP elevation, and inhibition of β-catenin transcriptional activity may contribute to the chemopreventive properties of certain NSAIDs.

Figure 1

In vitro association between colon tumor cell growth inhibition and other molecular mechanisms for a panel of NSAIDs and COX-2 inhibitors. A, positive association between the potencies of for HT-29 growth inhibition and inhibition of cGMP hydrolysis in HT-29 cell lysates. B, no association between the potencies for HT-29 growth inhibition and inhibition of COX-2.

Figure 2

Colon tumor cell growth- and cGMP PDE-inhibitory activity of SS. A, dose-dependent growth inhibitory activity of SS after 72 hours of treatment. B, dose-dependent increase in apoptotic (TUNEL positive) HT-29 cells after 48 hours of SS treatment. C, dose-dependent cAMP (▲, ○) and cGMP (●, ◆) PDE inhibitory activity of SS in HT-29 cell lysates (solid lines) compared to FHC lysates (dashed lines). D, dose-dependent increase in intracellular cGMP levels in HT-29 and SW480 cells after 30 minutes of SS treatment. * represents that the treatment is significantly different than vehicle control with p<0.05.

Figure 3

Effects of cyclic nucleotide elevating agents on growth of human colon cancer cells. A, dose-dependent growth inhibitory activity of the NO donor and guanylyl cyclase activator NOR-3 after 72 hours of treatment. B, non-growth inhibitory activity of the adenylyl cyclase activator forskolin after 72 hours of treatment. C, dose-dependent growth inhibitory activity of the PDE5 selective inhibitor MY5445. D, dose-dependent cAMP (■) and cGMP (●) PDE inhibitory activity of MY5445 in HT-29 cell lysates.

Figure 4

PDE isozyme expression and activity. A, inhibition of cGMP hydrolytic activity of recombinant PDE2, 3, and 5 by SS. B, relative levels of PDE2, 3, and 5 isozyme expression in HT-29 and FHC as measured by Western blotting with PDE family-specific antibodies. C, relative activities of PDE2, 3, and 5 families in HT-29 cell lysates as measured by overall inhibition of cGMP hydrolysis by the PDE2 selective inhibitor EHNA, the PDE3 selective inhibitor milrinone, and the PDE5 selective inhibitor sildenafil. D, PDE5 expression in normal colon mucosa, adenoma, and adenocarcinomas from human clinical specimens as determined by immunohistochemistry following avidin-biotin peroxidase labeling. E, effects of SS (top graph) or MY5445 (bottom graph) on Tcf/Lef transcriptional activity as measured by a luciferase reporter assay in HT-29 (hatched bars), HCT-116 (vertically stripped bars), and SW480 (horizontally stripped bars) colon cancer cells. F, time-dependent decrease in the expression of cyclin D1 and survivin in SW480 cells after treatment with SS or MY5445, with β-actin as a loading control.

Figure 5

In vivo antitumor efficacy of sulindac. A, selectivity of SS for inhibition of cGMP hydrolysis in colon tumor homogenates (solid line) compared to homogenates of normal colonic mucosa (dashed line). B, selectivity of MY5445 for inhibition of cGMP hydrolysis in colon tumor homogenates (solid line) compared to homogenates of normal colonic mucosa (dashed line). C, MTD determination of sulindac in male athymic mice. D, dose-dependent inhibition of tumor growth in HT-29 xenograft mice treated with 50mg/kg/day compared to vehicle treated controls.