Suppression of Colon Tumorigenesis in Mutant Apc Mice by a Novel PDE10 Inhibitor that Reduces Oncogenic β-Catenin

Nonsteroidal anti-inflammatory drugs (NSAID) inhibit tumorigenesis in a variety of preclinical rodent models of colorectal cancer, but long-term use in humans to treat or prevent colorectal cancer is not recommended because of potentially fatal gastrointestinal, renal, and cardiovascular toxicities resulting from COX inhibition and the suppression of physiologically important prostaglandins (1). Multiple lines of evidence suggest that the mechanism for the antineoplastic activity of NSAIDs may involve targets other than, or in addition to COX (2, 3). For example, doses of NSAIDs exceeding those needed to inhibit COX isozymes, COX-1 or COX-2, are generally required to suppress tumorigenesis in experimental animal models (4, 5). In addition, there is no apparent correlation between potency to inhibit COX and cancer cell growth among several different chemical classes of NSAIDs (6). As a nonselective COX-1/-2 inhibitor, sulindac is considered to be the most effective NSAID to inhibit adenoma formation in patients with familial adenomatous polyposis (FAP) (7–9), possibly because of high concentrations in the intestinal lumen from enterohepatic recirculation, and is among the most potent NSAIDs to inhibit cancer cell growth in vitro (10–12). Sulindac is a prodrug that requires reductive metabolism by colonic bacteria to convert it from a sulfoxide to sulindac sulfide (SS), which is the active form of the drug that inhibits COX-1 and COX-2 (13). The sulfoxide is also oxidized in the liver to a sulfone that does not inhibit COX-1 or COX-2. Despite lacking COX inhibitory activity, sulindac sulfone inhibits cancer cell growth in vitro at concentrations only slightly higher than sulindac sulfide (10, 11). Sulindac sulfone exhibited protective activity in multiple rodent models of colon tumorigenesis without affecting colonic prostaglandin levels (14, 15). Lacking toxicities associated with COX inhibition, sulindac sulfone (exisulind) was evaluated in clinical trials for its ability to inhibit adenoma formation in individuals with FAP or sporadic polyposis, but was found to be only moderately effective, most likely due to low potency and dose-limiting hepatotoxicity (16–19).

Previous studies suggest that the antineoplastic activity of sulindac and exisulind is associated with cyclic guanosine monophosphate (cGMP) phosphodiesterase (PDE) inhibition. Both drugs inhibit cancer cell growth at concentrations comparable with those that inhibit the cGMP PDE activity of cell lysates or recombinant PDE isozymes (6, 17, 18, 20–22). In addition, certain non-COX inhibitory derivatives were synthesized that exhibit appreciably greater potency to inhibit cancer cell growth and cGMP PDE activity (18, 22). However, it was unclear which cGMP degrading PDE isozyme(s) is responsible for the anticancer activity of sulindac or exisulind, as neither displayed significant selectivity for any specific GMP degrading PDE isozyme (17, 18).

Eleven distinct families of PDE isozymes have been identified that comprise over 100 different isoforms or splice variants (23). PDE5, 6, and 9 selectively hydrolyze cGMP, while PDE4, 7, and 8 selectively hydrolyze cyclic adenosine monophosphate (cAMP). PDE1, 2, 3, 10, and 11 hydrolyze both cAMP and cGMP (24). Previous studies suggest that the PDE10 isozyme is unique from other PDE isozymes because of high expression in cancer cells relative to cells from normal tissues (22, 25–27). In addition, PDE10 was found to be elevated in colon adenomas and adenocarcinomas relative to normal colonic mucosa (25, 27). An important role of PDE10 in tumorigenesis is suggested by experiments showing that PDE10 knockdown by genetic silencing or inhibition of enzymatic activity with highly specific small molecule inhibitors suppress the growth of cancer cells expressing PDE10, but not cells derived from normal tissues with low PDE10 expression (25–27).

Here we describe a novel PDE10 inhibitor, ADT 061, that was identified by synthesis and screening of a large custom library of indenes chemically related to sulindac but lacking COX inhibitory activity. ADT 061 potently and selectivity inhibited colon cancer cell growth, increased cyclic nucleotide levels, and activated cGMP-dependent protein kinase (PKG) at concentrations that inhibit recombinant PDE10. Activation of cGMP/PKG signaling in colon cancer cells by ADT 061 induced the phosphorylation of β-catenin and suppressed nuclear levels of β-catenin and TCF/LEF transcriptional activity. Oral administration of ADT 061 resulted in high local concentrations in the colonic mucosa and significantly inhibited adenoma formation without discernable toxicity in a mutant Apc mouse model of colon tumorigenesis.

The synthesis of ADT 061 [(Z)-2-(5-methoxy-2-methyl-1-(3,4,5-trimethoxybenzylidene)-1H-inden-3-yl)-N-(pyridin-3-yl) acetamide] is based on a procedure originally described in U.S. patent 6,063,818 as “Example 1” (28). The synthesis used 4-methoxy-benzaldehyde (Sigma-Aldrich) as the starting material and involved multiple reaction steps (a–f) with five key intermediates (1–5) as illustrated in Supplementary Figs. S1–S4 and described in detail in the figure legend.

Human HT-29, HCT 116, SW480, and Caco2 colon carcinoma cells were purchased from the ATCC (HTB-38, CCL-247, CCL-228, and HTB-37, respectively) and cultured according to the manufacturer's recommendations. Cancer cell line identity was confirmed by short tandem repeat profiling at LabCorp (Genetica). Cells were maintained in culture for 20 passages or less, and tested regularly to ensure absence of Mycoplasma contamination. The human normal colonic mucosal epithelial cell line, NCM460, was purchased from INCELL Corporation, LLC, and grown according to the manufacturer's recommendations, and maintained in culture for less than 12 passages (29). Cell growth inhibition assays were performed as described previously (26, 27). PDE10 siRNA knockdown cells were generated previously (25).

PDE assays were performed using recombinant PDE isozymes (BPS Bioscience) in conjunction with the IMAP FP Screening Express Kit (Molecular Devices Inc,), as described previously (27). To determine enzyme kinetics, the concentrations of enzyme, ADT 061, and fluorescently labeled cyclic nucleotide were held constant, while unlabeled cyclic nucleotides were added in increasing concentrations to the reaction. The PDE10 specific inhibitor, Pf2545920 (Selleck Chemicals, LLC), served as a positive control for assessing PDE10 inhibitory activity of ADT 061.

Cell lysates were incubated with 10 μL of Protein A agarose beads (Cell Signaling Technology) per 100 μL of lysate for 30 minutes on a rotator at 4°C. Lysates were spun at 6,000 rpm for 10 minutes and beads discarded. PDE10 antibody (C2C3, Genetex, Inc.) and IgG control (Cell Signaling Technology) were incubated overnight at 4°C on a rotator. Protein A agarose beads were added and incubated for 1 hour on a rotator at 4°C, centrifuged and the supernatant discarded. Beads were washed in lysis buffer and prepared for Western blot analysis. For the PDE assay, an additional wash was performed in PDE reaction buffer (Molecular Devices) and enzymatic activity was measured by fluorescence polarization using a Biotek Synergy H4 plate reader.

COX inhibition was determined using the COX Fluorescent Inhibitor Screening Assay Kit (Cayman Chemical) according to the manufacturer's recommendation. SS (Sigma-Aldrich; 10 μmol/L) served as a positive control for assessing the COX-1 and COX-2 inhibitory activity of ADT 061. Celecoxib (Selleck Chemicals; 100 nmol/L) and indomethacin (Sigma-Aldrich; 100 nmol/L) served as additional controls as COX-2 and COX-1 specific inhibitors, respectively. The protocol was modified to include a 20-minute incubation step prior to the addition of arachidonic acid. Fluorescent readings were obtained using a Biotek Synergy H4 plate reader.

Intracellular cyclic nucleotide concentrations were assayed using Cyclic GMP/AMP ELISA Kits (Cayman Chemical) according to the manufacturer's recommendations and as described previously (27).

Western blots were performed as described previously (27). All antibodies were purchased from Cell Signaling Technology except PDE10 (C2C3, Genetex, Inc. GTX118886), total β-catenin (Clone 14, BD Biosciences 610153), and β-actin (Sigma-Aldrich A1978). Cell Signaling antibodies are as follows: p-VASP Ser239 (Cell Signaling Technology, 3114), p-VASP Ser157 (Cell Signaling Technology, 3111), VASP (Cell Signaling Technology, 3132), p-β-catenin Ser33/37/Thr41 (Cell Signaling Technology, 9561), β-catenin Ser552 (Cell Signaling Technology, 5651), Survivin (Cell Signaling Technology, 2808), Cl Caspase-3 (Cell Signaling Technology, 9664), p-CREB (Cell Signaling Technology, 9198), GAPDH (Cell Signaling Technology, 5174).

TCF/LEF transcriptional activity was determined as described previously (22).

Fluorescent images were acquired using a Nikon A1 Laser Scanning Confocal microscope. SW480 cells were analyzed for nuclear translocation of β-catenin following Wnt stimulation as described previously (27). Samples were co-stained with antibodies against total β-catenin (Clone 14, BD Biosciences) and non-phospho serine 45-β-catenin (D2U8Y, Cell Signaling Technology, 19807) before counterstaining with Alexa-Fluor 488 and 647, respectively. For quantification, the region of interest generator in Nikon Elements was used to automatically select nuclei positive for DAPI staining. Staining intensity was then exported for each channel of interest on a minimum of 100 cells per treatment. Hyperspectral imaging coupled with linear unmixing was performed to determine the localization of PDE10 in colon tissue. Spectral imaging and analysis was used to separate the autofluorescence signal from the AlexaFluor-488 labeling of PDE10, as described previously (30, 31). A library containing pure spectra of AlexaFluor-488 (PDE10), AlexaFluor-647 (β-catenin), and DAPI (nuclei) was constructed using single labeled controls of HT-29 cells labeled with only AlexaFluor-488, AlexaFluor-647, and DAPI, respectively. Normal colon tissue section was used to obtain the autofluorescence spectral signature. Images were scaled similarly and false colored for visualization as indicated in the text.

ADT 061 oral bioavailability was evaluated using 7–8 weeks old, female C57BL/6 mice. ADT 061 was administered at a dosage of 100 mg/kg by oral gavage once as a suspension in 0.5% carboxymethyl cellulose/0.25% tween 80 (T80) in sterile water. Each mouse had blood collected twice: the first timepoint was a survival bleed (0.5, 1, or 4 hours after administration) and the second timepoint was a terminal bleed (2, 8, or 24 hours after administration). There were 4 mice per timepoint. Blood was processed for plasma and frozen. In addition to blood collection, lungs, colonic mucosa, ovaries, uterus, and brain were collected 2, 8, or 24 hours after treatment and frozen. Drug levels were measured using LC/MS-MS methodology. Results were presented as mean ± SD. This study was approved by the Institutional Animal Care and Use Committee of the University of South Alabama (protocol #755414).

Mouse tissue samples were homogenized either 1:5 or 1:10 with 5 mmol/L ammonium acetate buffer using an Omni THQ homogenizer. Calibration standards, blanks and quality control samples (QCs) were prepared by spiking naïve mouse plasma or tissue homogenate (50 μL) with the appropriate amount of ADT 061 to achieve concentrations in tissue homogenate or plasma ranging from 5–10,000 ng/mL standards, blanks, QCs and samples were spiked with internal standard (500 μL of 10 ng/mL terfenadine in acetonitrile) and vortexed for 10 seconds to precipitate the proteins. After centrifugation for 5 minutes at 21,000 × g, the supernatant was removed and transferred to an autosampler vials and analyzed in positive ion mode by LC/MS-MS. The LC/MS-MS system consisted of Shimadzu Prominence system equipped with LC20-AD dual HLPC pumps, an SIL20-AC HT autosampler, and a DGU-20A2 in-line degasser. Detection was performed using an Applied BioSystems 4000 QTRAP (Applied Biosystems) triple quadrupole mass spectrometer operated in the positive ion mode. Mass calibration, data acqusition, and quantitaion were performed using Applied Biosystem Analyst 1.6.2 software (Applied Biosystems). Separation of the ADT 061 and the internal standard from the plasma or homogenate matrix was achieved using a Phenomenex Luna C18, 100×2 mm 5 μm particle column equipped with a C18 Security Guard cartridge (Phenomenex). The mobile phase was delivered at a flow rate of 400 μL/minute using a gradient elution profile consisting of deionized (DI) water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B). ADT 061 was analyzed using an elution profile in which mobile phase B was held at 20% for 1 minute, then increased linearly to 90% over 5 minutes, held at 90% for 1.5 minutes, returned to 20% and equilibrated for 2.5 minutes. The analyte and internal standard were detected after a 10 μL injection using multiple reaction monitoring for the following the following transitions: ADT 061 (m/z 473.3.4→121.0), Terfenadine (m/z 472.4→ 436.3).

Male Apc^+/Min-FCCC mice (6–8 weeks of age, N = 40) were randomized to treatment groups and treated for 6 weeks. The activity of the liver enzymes aspartate aminotransferase (AST) and alanine aminotransferase (ALT) was measured in plasma samples collected at necropsy (n = 9/group) using commercial ELISA kits, according to manufacturer's instructions (Biovision). This study was approved by the Institutional Animal Care and Use Committee of Fox Chase Cancer Center (protocol #14-07).

A colony of Apc^+/Min-FCCC mice has been established at Fox Chase Cancer Center (FCCC, Philadelphia, PA) and maintained for over 80 generations (32). Because male APC^+/Min-FCCC mice develop twice as many colon adenomas and have a higher incidence of colon tumors than femail APC^+/Min-FCCC mice, only male mice were used in the current study. Mice were maintained on a Harlan Teklad 2018SX diet (Envigo) and had access to food and water ad libitum. Prior to ADT 061 treatment, colonoscopies were performed on all Apc^+/Min-FCCC mice (ref. 33; 6–7 weeks of age) to ensure the absence of colorectal adenomas. At 7 to 8 weeks of age, mice were assigned to groups of 19–21 mice receiving either unsupplemented chow (control), or chow supplemented with either 1,000 or 1,500 ppm ADT 061 for 14 weeks. Body weights were recorded weekly. At the time of euthanasia, the entire intestine was examined grossly, fixed in 10% buffered formalin, and submitted for histopathologic review. This study was approved by the Institutional Animal Care and Use Committee of FCCC (Philadephia, PA; protocol #14-07).

Formalin-fixed paraffin-embedded (FFPE) tissues were sectioned, stained with hematoxylin and eosin, and reviewed in a blinded manner by a gastrointestinal pathologist with extensive experience with the Apc^+/Min-FCCC mouse model. Tumors were classified as adenomas (>4 dysplastic crypts) or microadenomas (1–4 dysplastic crypts). Adenomas were further categorized as polypoid or flat. Polypoid adenomas projected above the mucosa, while flat adenomas were less than twice the height of the adjacent normal mucosa. This definition is consistent with the criteria for adenomas established by a leading panel of intestinal pathologists (34).

Growth assays as well as all enzymatic assays (PDE and COX) were performed with a minimum of three replicates and repeated at least twice. All Western blots and immunofluorescence experiments were repeated with at least two biological replicates. Statistical analysis was performed using GraphPad Prism software using ANOVA for two or more groups. Statistical comparisons of colon tumor data were performed using Fisher exact test for tumor incidence and the Wilcoxon test for tumor multiplicity.

ADT 061 contains the same indene scaffold as sulindac but with several chemical modifications (Fig. 1A) aimed at blocking COX binding to reduce toxicity while improving potency and selectivity to inhibit PDE10 that we hypothesized could enhance anticancer efficacy. SS nonselectively inhibited COX-1 and COX-2 by 68%–70% at a concentration of 10 μmol/L, while ADT 061 completely lacked COX-1 or COX-2 inhibitory activity at concentrations up to 50 μmol/L (Fig. 1B). As controls, the COX-1 inhibitor, indomethacin, selectively inhibited COX-1, while the COX-2 inhibitor, celecoxib selectively inhibited COX-2. Despite lacking COX inhibitory activity, ADT 061 potently inhibited the growth of colon cancer cell lines, HCT116, HT-29, and SW480, with IC₅₀ values of 0.31, 0.33, and 0.37 μmol/L, respectively, but did not appreciably affect the growth of normal colon mucosal (NCM460) cells grown under similar conditions at concentrations up to 50 μmol/L (Fig. 1C). SS was appreciably less potent than ADT 061 to inhibit the growth of colon cancer cells with IC₅₀ values 110-fold to190-fold higher and was also less selective for cancer cells as compared with NCM460 cells (Fig. 1D). PDE10 was expressed in all colon cancer cell lines examined but was essentially undetectable in NCM460 cells (Fig. 1E) as published previously (25).

The PDE isozyme selectivity of ADT 061 was determined using recombinant PDE isozymes, PDE1–11, with cAMP and/or cGMP as substrates based on the known substrate specificity of each PDE isozyme. At a concentration of 1 μmol/L, ADT 061 inhibited PDE10 by 75% and 50% with cGMP and cAMP as a substrate, respectively, without appreciably affecting the activity of all other PDE isozymes (Fig. 2A). Testing a wide range of concentrations revealed that ADT 061 inhibited recombinant PDE10 using cGMP as a substrate with an IC₅₀ of 0.3 μmol/L. Slightly higher concentrations of ADT 061 were required to inhibit PDE10 when cAMP was used as the substrate with an IC₅₀ of 1.0 μmol/L (Fig. 2B). By comparison, SS inhibited PDE10 with an IC₅₀ of 70 μmol/L that was comparable with its potency to inhibit cancer cell growth but was not isozyme selective as reported previously (22).

Experiments were also conducted to determine whether ADT 061 binds PDE10 in a competitive manner by measuring the enzymatic activity of recombinant PDE10 in the presence of varying cGMP concentrations. As calculated by Michaelis-Menten kinetics, ADT 061 increased the K_m of PDE10 for cGMP to reach half maximal velocity from 5.4 to 17.9 μmol/L, while not significantly affecting maximal velocity (V_max; Fig. 2C). The competitive nature by which ADT 061 inhibits PDE10 was also apparent using the Eadie–Hofstee method by plotting velocity as a function of velocity/substrate concentration showing non-parallel intersecting lines for ADT 061 versus control (Supplementary Fig. S5).

To further determine whether PDE10 inhibition is responsible for the cancer cell growth inhibitory activity of ADT 061, PDE10 levels were suppressed in HT-29 cells by siRNA knockdown. This resulted in reduced sensitivity to ADT 061 from an IC₅₀ of 0.32 μmol/L in vector control cells to 1.9 μmol/L in PDE10 knockdown cells (Fig. 2D). The activity of ADT 061 remaining despite treatment was likely attributed to incomplete knockdown of PDE10. In addition, the sensitivity of PDE10, extracted from HT-29 cells by immunoprecipitation, to ADT 061 was determined to be comparable to that of recombinant PDE10 (Fig. 2E).

As predicted for a PDE10 inhibitor, treatment of HT-29 colon cancer cells with ADT 061 significantly increased both intracellular cGMP and cAMP levels at concentrations that inhibited cancer cell growth (Fig. 3A). To determine whether the magnitude of the effect of ADT 061 treatment on elevating intracellular cGMP and/or cAMP levels was sufficient to induce downstream signaling, PKG and PKA activity were measured by Western blot analysis using phospho-specific antibodies against VASP and CREB. VASP is phosphorylated preferentially by PKG on Ser²³⁹ and by PKA on Ser¹⁵⁷ residues (25, 27, 35, 36). ADT 061 at a concentration of 0.3 μmol/L increased Ser²³⁹ phosphorylation within 30 minutes of treatment without affecting the phosphorylation of VASP at Ser¹⁵⁷ (Fig. 3B). Testing a wider range of concentrations confirmed that ADT 061 increased Ser²³⁹ phosphorylation at concentrations effective for inhibiting colon cancer cell growth, while appreciably higher concentrations were required to increase Ser¹⁵⁷ phosphorylation (Fig. 3C). In addition, ADT 061 did not induce CREB phosphorylation at Ser¹³³, which is exclusively phosphorylated by PKA (37). These results suggest that the increase in intracellular cGMP levels by ADT 061 treatment is sufficient to activate PKG and inhibit colon cancer cell growth, while the increase in cAMP levels is insufficient to activate PKA and inhibit growth.

Previous studies have reported that PKG can phosphorylate β-catenin to induce proteasomal degradation of β-catenin (24, 26, 38). To determine whether PDE10 inhibition and activation of cGMP/PKG signaling by ADT 061 can phosphorylate and reduce levels of β-catenin, HT-29, and SW480 colon cancer cells were treated with ADT 061 at concentrations that inhibit colon cancer cell growth and extracts were probed by Western blot analysis using phospho-specific antibodies against specific β-catenin residues. ADT 061 treatment reduced the non-phosphorylated (stable) form of β-catenin at Ser⁴⁵, which represents a reduction of the oncogenic pool of β-catenin (Fig. 3D). Conversely, ADT 061 treatment increased levels of phosphorylated β-catenin at Ser^{33, 37}/Thr⁴¹ residues, which are known to induce ubiquitination and proteasomal degradation of β-catenin (24). Consistent with a mechanism involving the activation of cGMP/PKG signaling, ADT 061 treatment caused an increase in phosphorylation of β-catenin at Ser⁵⁵², a residue known to be phosphorylated by PKG (26). ADT 061 treatment of HT-29 and SW480 colon cancer cells also reduced levels of survivin (Fig. 3F), a protein that is regulated by TCF/LEF transcription and essential for cancer cell proliferation and survival (39). This effect occurred within the same concentration range of ADT 061 that was effective for inducing apoptosis as detected by increased levels of cleaved caspase 3. ADT 061 treatment of HCT 116 colon cancer cells also reduced the activity of the TCF/LEF transcription factor (Fig. 3E). These effects occurred within the same concentration range of ADT 061 that activated cGMP/PKG signaling in colon cancer cells and inhibited the growth of colon cancer cells.

The effect of ADT 061 on Wnt-induced nuclear translocation of β-catenin was determined in SW480 colon cancer cells harboring mutant APC using Wnt-conditioned media as described previously (27, 40). Confocal immunofluorescence microscopy, utilizing antibodies against total β-catenin and its non-phosphorylated (stable, oncogenic) form at Ser⁴⁵, revealed that Wnt-conditioned media stimulated nuclear translocation of β-catenin, while control media failed to induce nuclear accumulation of β-catenin (Fig. 4A). Nuclear translocation of β-catenin was confirmed by co-staining cells with antibodies against total and oncogenic β-catenin along with the nuclear marker, DAPI (Fig. 4B). Treatment with ADT 061 at a concentration of 0.5 μmol/L completely blocked Wnt-induced β-catenin translocation to the nucleus (Fig. 4C). Nuclear localization of β-catenin was quantified for total (Fig. 4D) as well as active oncogenic non-phosphorylated forms (Fig. 4E). While staining intensity for total β-catenin did not differ significantly among the treatment groups, the staining intensity for non-phosphorylated active β-catenin increased significantly over control with Wnt stimulation. The effect of Wnt stimulation was completely inhibited by ADT 061.

Consistent with the overexpression of PDE10 mRNA and protein levels during early stages of colon tumorigenesis as reported previously (25), levels of PDE10 protein were found to be elevated in colon adenomas, as well as in adjacent uninvolved mucosa of Apc^+/Min-FCCC mice relative to normal colon mucosa from wildtype (WT) mice when measured by Western blot analysis (Fig. 5A). While PDE10 levels were not appreciably different in uninvolved mucosa and adenomas, cGMP levels as measured by ELISA were significantly reduced in adenomas compared with uninvolved mucosa, suggesting that the enzymatic activity of PDE10 is higher in adenomas relative to uninvolved mucosa (Fig. 5B), despite comparable levels of PDE10 protein (Fig. 5A). Detection of PDE10 in FFPE colon tissue from Apc^+/min-FCCC mice by immunolabeling confirmed high levels of PDE10 expression in dysplastic regions (Fig. 5C, red label) that co-localized (white arrows) with β-catenin (Fig. 5C, green label).

Pharmacokinetic analysis in mice following a single oral dose of ADT 061 (100 mg/kg) resulted in plasma levels (C_max = 4.9 μmol/L) that exceeded growth IC₅₀ values by greater than 10-fold for at least 8 hours following administration with a biphasic profile indicative of enterohepatic recirculation (Supplementary Fig. S6). Measurement of ADT 061 levels in various tissues revealed that ADT 061 achieved high levels in colon mucosa (C_max = 62.6 nmol/g) that were greater than 10- and 30-fold higher relative to levels in plasma and other tissues (C_max = 1–2 nmol/g), respectively. Unlike known PDE10 inhibitors that achieve high concentrations in the brain and have associated side effects (e.g., sedation), oral administration of ADT 061 resulted in nearly undetectable (0.1 and 0.2 nmol/g) levels in the brain and did not cause sedation.

ADT 061 was evaluated for its ability to inhibit the formation of colon adenomas in Apc^+/Min-FCCC mice. Mice used for this study were identified as tumor free by colonoscopy prior to treatment. After 14 weeks of treatment with control (unsupplemented) chow or chow supplemented with either 1,000 or 1,500 ppm ADT 061, no significant difference in body weights was observed between control and treatment groups (Fig. 6A). A dose of 1,500 ppm yielded plasma levels of 0.5 μmol/L, exceeding colon tumor cell growth inhibition IC₅₀ values (0.3 μmol/L) in cell culture (see Fig. 2D). Of note, no hepatotoxicity was observed in an independent study when Apc^+/Min-FCCC mice were fed chow supplemented with higher doses of ADT 061 (up to 2,400 ppm). Circulating levels of AST and ALT, biomarkers for liver damage, were unremarkable in mice treated with ADT 061 even at doses higher than those used in the current study (Supplementary Fig. S7).

The activity of ADT 061 against colon tumorigenesis was dose dependent. Treatment of Apc^+/Min-FCCC mice with ADT 061 reduced the incidence of colon adenomas by 18.6% at 1,000 ppm (P = 0.186) and 36.9% at 1,500 ppm (P = 0.019; Fig. 6B). The multiplicity of total adenomas (sum of flat, polypoid, and microadenomas) was reduced 27.5% in mice treated with 1,000 ppm ADT 061 (P = 0.31) and 51.3% in mice treated with 1,500 ppm ADT 061 (P = 0.028; Fig. 6C). The multiplicity of adenomas (> 4 dysplastic crypts) was reduced 31.2% in mice treated with 1,000 ppm ADT 061 (P = 0.11) and 46.0% in mice treated with 1,500 ppm ADT 061 (P = 0.026; Fig. 6D). Although not statistically significant, the multiplicity of microadenomas (≤ 4 crypts) was also decreased in mice treated with 1,500 ppm ADT 061 (Fig. 6E). These data indicated that the higher dose of ADT 061 was needed to inhibit colon tumorigenesis in Apc^+/Min-FCCC mice.

The results from the present study demonstrate that ADT 061 significantly inhibits colon tumorigenesis when administered orally to Apc^+/Min-FCCC mice without discernable toxicity. The mechanism, as studied using colon cancer cells, appears to involve direct inhibition of neoplastic cell growth by targeting PDE10, a PDE isozyme recently found to be overexpressed during early stages of colon tumorigenesis and essential for the proliferation and survival of cancer cells (22, 25, 27). Consistent with a mechanism involving PDE10 inhibition, ADT 061 increased intracellular cyclic nucleotide levels and activated PKG at concentrations that inhibit recombinant PDE10 and cancer cell growth in vitro. ADT 061 activation of cGMP/PKG signaling was associated with reduced levels of the oncogenic pool of β-catenin and the suppression of TCF/LEF transcriptional activity.

As further evidence that PDE10 inhibition is responsible for the cancer cell growth inhibitory activity of ADT 061, siRNA knockdown of PDE10 in colon cancer cells reduced sensitivity to ADT 061 in growth assays. ADT 061 also inhibited the activity of PDE10 extracted from colon cancer cells with an IC₅₀ value achieved when using recombinant PDE10. In addition, treatment of colon cancer cells with ADT 061 increased both cAMP and cGMP levels as would be expected for a PDE10 inhibitor. The increase in intracellular cGMP levels by ADT 061 was sufficient to activate PKG signaling as evidenced by an increase in VASP phosphorylation at residues known to be phosphorylated by PKG, while the role of cAMP signaling requires further study. The evidence that PDE10 inhibition is responsible for the growth inhibitory activity of ADT 061 agrees with previous studies reporting that genetic knockdown of PDE10 or inhibition of enzymatic activity small molecule inhibitors, such as Pf2545920, can also inhibit colon cancer cell growth (25, 27).

An important distinction between ADT 061 and existing PDE10 inhibitors is that the latter were developed for central nervous system (CNS) disorders such as schizophrenia and Huntington's disease and were designed to cross the blood–brain barrier (41, 42). Pf2545920 and likely other PDE10 inhibitors developed for CNS disorders do not appear to be suitable for cancer indications because of rapid liver metabolism, as well as associated side effects (sedation; refs. 43, 44). ADT 061 did not result in appreciable levels in brain following oral administration and did not cause sedation, but achieved high local concentrations in colon mucosa, likely because of enterohepatic recirculation. These properties of ADT 061, along with its ability to target PDE10 that is overexpressed during early stages of colon tumorigenesis would appear to be consistent with low toxicity as observed in these studies and essential for colorectal cancer chemoprevention.

Mutations in the APC gene, encoding for aberrant adenomatous polyposis coli (APC) protein, occur in over 80% of patients with colorectal cancer and result in loss of tumor suppressor function and activation of Wnt/β-catenin signaling (45). In the absence of Wnt ligand bound to its receptors, β-catenin not bound by cadherin junctions becomes associated with APC, axin, glycogen synthase kinase 3β (GSK3β), and casein kinase 1 (CK1) referred to as the “β-catenin destruction complex,” is degraded by the proteasome (46). CK1 phosphorylation of β-catenin at Ser⁴⁵ is believed to act as a primer for the phosphorylation by GSK3β at Ser^33/37, and Thr⁴¹, to induce ubiquitination by β-transducin repeat containing protein (β-TrCP) and proteasomal degradation (47, 48). Wnt binding to Frizzled phosphorylates Dishevelled (Dvl), allowing Dvl to bind axin. Dissociation of the β-catenin destruction complex occurs during colorectal cancer, resulting in the accumulation of β-catenin in the cytoplasm and nuclear translocation, which drives TCF/LEF-mediated transcription of genes encoding for proteins that are essential for proliferation and survival of colon tumor cells (49–53). Individuals with FAP who harbor germline mutations in the APC gene have aberrantly high levels of oncogenic (non-phosphorylated/stabilized) β-catenin, leading to unregulated growth and the formation of adenomas (45, 54). Individuals with sporadic adenomas have similar mutations in APC or β-catenin within dysplastic lesions that give rise to adenomas (55).

PKG activation by ADT 061 was found to counter the effects of mutant APC by reducing β-catenin-dependent TCF/LEF transcriptional activity as shown by a decrease in the phosphorylated (stabilized) form of β-catenin at Ser⁴⁵ with a corresponding increase in phosphorylation of β-catenin on Ser^33/37/Thr⁴¹, phosphorylation sites known to induce ubiquitination and degradation (24, 46, 53). ADT 061 treatment of cancer cells also decreased levels of survivin, a protein known to be regulated by TCF/LEF transcription and essential for the proliferation and survival of cancer cells (39), which coincided with increased cleaved caspase 3, a marker of apoptosis. The ability of ADT 061 to block Wnt-induced nuclear localization of β-catenin can be attributed to reduced cytoplasmic levels of β-catenin, resulting from proteasomal degradation. Together, these results show that ADT 061 decreases TCF/LEF transcriptional activity and inhibits colon cancer cell growth by selectively reducing the oncogenic pool of β-catenin leading to induction of apoptosis. These results support previous studies reporting that ectopic expression of PKG can induce apoptosis of colon tumor cells by reducing β-catenin levels (56, 57), as well as reports that PDE10 inhibition by genetic silencing or small molecule inhibitors can suppress β-catenin/TCF-mediated transcriptional activity by activating cGMP/PKG signaling (25). Importantly, we previously published evidence showing that the known PDE10 inhibitor, Pf-2545920, inhibited the growth of colon tumor cell lines through the activation of PKG. However, clinical use of this compound as well as other PDE10 inhibitors developed for CNS diseases is not suitable for colorectal cancer chemoprevention given that such compounds readily cross the blood-brain barrier and cause sedation (27, 41).

The significance of these findings cannot be overstated with respect to the ability of ADT 061 to inhibit PDE10 and activate cGMP/PKG signaling as a novel mechanism for the suppression of the oncogenic pool of β-catenin in cancer cells without affecting normal stem cells that require Wnt/β-catenin signaling for stemness. The lack of effect of ADT 061 on normal stem cell function is supported by experiments showing potent inhibition of colon cancer cell growth expressing PDE10 by ADT 061 without affecting PDE10-null NCM460 cells that were isolated from normal colon mucosa and reported to contain a normal stem cell population (29). Such selectivity may not be shared by other inhibitors of Wnt/β-catenin signaling that presumably failed clinical trials because of toxicities associated with the disruption of normal stem cell function.

Apc^+/Min-FCCC mice were used to evaluate the antitumor activity of ADT 061. This spontaneous mouse model of colorectal cancer displays a high incidence and multiplicity of colon adenomas and has been previously used to assess the efficacy of sulindac (32, 58). In addition, tumors and uninvolved colon mucosa from this model express high PDE10 levels relative to normal mucosa from WT mice, findings that agree with previous reports of low PDE10 levels in normal tissues outside of the CNS (59). High levels of PDE10 in dysplastic cells of the colon are consistent with results from other investigators reporting high levels of PDE10 mRNA in adenomas from Apc^Min mice (38). In the current study, ADT 061 treatment reduced the multiplicity and incidence of total adenomas (microadenomas, flat, and polypoid adenomas) in a dose-dependent manner. Although treatment with 1,000 ppm ADT 061 decreased colon tumor incidence and multiplicity in Apc^+/Min-FCCC mice, the result did not achieve statistical significance. Only the 1,500 ppm dose of ADT 061 inhibited colon tumorigenesis significantly. The lack of toxicity at these doses provides an additional opportunity to explore the antitumor activity of ADT 061 at higher doses.

The greater potency and selectivity of ADT 061 to inhibit PDE10 and cancer cell growth may account for its greater antitumor efficacy in vivo, as compared with sulindac. The ability of sulindac to inhibit intestinal tumors in Apc^+/Min-FCCC mice was evaluated in a prior study by this group, using mice that were endoscopically confirmed to be tumor free at the time of study enrollment (same design as the current study; ref. 58). Although administration of sulindac to Apc^+/Min-FCCC mice reduced the multiplicity of small intestinal tumors, the incidence and multiplicity of colon tumors in sulindac-treated mice remained comparable with that of untreated controls (58). Other attributes, such as its ability to achieve high local concentrations in colon mucosa as compared with plasma and other tissues following oral administration and lack of COX inhibitory activity may result in low toxicity which is essential for colorectal cancer chemoprevention.

In summary, ADT 061 is a novel non-COX inhibitory derivative of sulindac with PDE10 inhibitory activity that potently and selectively inhibits colon cancer cell growth. The mechanism of action involves competitive inhibition of PDE10 that is overexpressed during early stages of tumorigenesis. The PDE10 inhibitory activity of ADT 061 induced cGMP/PKG signaling to reduce the oncogenic pool of β-catenin driving the transcription of a variety of genes essential for colon cancer cell proliferation and survival. ADT 061 treatment by oral administration achieved high local concentrations in the colon mucosa and significantly inhibited the formation of colon adenomas in Apc^+/Min-FCCC mice without discernable toxicity. Although further investigation of efficacy and safety is needed, we conclude that ADT 061 has the potential to prevent or treat adenomas in individuals with FAP or sporadic polyposis at high risk of developing colorectal cancer.

X. Chen reports grants from NCI and other support from ADT Pharmaceutical LLC during the conduct of the study; other support from ADT Pharmaceutical LLC outside the submitted work; in addition, X. Chen has a patent for US Patent App No. 15/537,283) pending. S.J. Leavesley has formed a startup company, SpectraCyte LLC, to commercialize novel spectral imaging technologies that have been developed at the University. However, all of the work performed in this article was performed using a commercially-available Nikon microscope. G. Gorman reports other support from University of South Alabama during the conduct of the study. L. Coward reports other support from University of South Alabama during the conduct of the study. A.B. Keeton reports grants from NCI during the conduct of the study; other support from ADT Pharmaceuticals, LLC outside the submitted work; in addition, A.B. Keeton has a patent for ADT-061 pending. M.L. Clapper reports grants from NIH/NCI during the conduct of the study. G.A. Piazza reports other support from ADT Pharmaceuticals LLC during the conduct of the study; in addition, G.A. Piazza has a patent for U.S. Patent Application No. 15/537,283 pending. No disclosures were reported by the other authors.

K.J. Lee: Conceptualization, data curation, formal analysis, validation, investigation, methodology, writing–original draft, writing–review and editing. W.-C.L. Chang: Resources, data curation, formal analysis, investigation, methodology, writing–review and editing. X. Chen: Data curation, formal analysis, methodology, writing–review and editing. J. Valiyaveettil: Data curation, formal analysis, validation. V. Ramirez-Alcantara: Writing–review and editing. E. Gavin: Data curation, writing–review and editing. A. Musiyenko: Data curation. L. Madeira da Silva: Data curation, formal analysis, writing–review and editing. N.S. Annamdevula: Data curation, formal analysis. S.J. Leavesley: Resources. A. Ward: Data curation. T. Mattox: Writing–review and editing. A.S. Lindsey: Data curation, writing–review and editing. J. Andrews: Resources, software, writing–review and editing. B. Zhu: Writing–review and editing. C. Wood: Data curation. A. Neese: Data curation. A. Nguyen: Data curation. K. Berry: Resources, writing–review and editing. Y. Maxuitenko: Writing–review and editing. M.P. Moyer: Resources. E. Nurmemmedov: Resources, writing–review and editing. G. Gorman: Resources, data curation, formal analysis, writing–review and editing. L. Coward: Data curation, formal analysis, writing–review and editing. G. Zhou: Resources, writing–review and editing. A.B. Keeton: Data curation, writing–review and editing. H.S. Cooper: Resources, formal analysis, writing–review and editing. M.L. Clapper: Resources, formal analysis, writing–review and editing. G.A. Piazza: Conceptualization, resources, supervision, funding acquisition, project administration, writing–review and editing.

The animal work was performed with the technical support of the following core facilities at Fox Chase Cancer Center: Laboratory Animal Facility, Genotyping Components of the Genomics Facility, Small Animal Imaging Component of the Biological Imaging Facility, and the Biostatistics and Bioinformatics Facility.

Funding provided by the NCI, NIH under Award Numbers 1R01CA131378, 1RO1CA148817, 1RO1CA197147, and 1RO1CA155638, to G.A. Piazza and P30 CA006927, to Fox Chase Cancer Center. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.