Discussion and Conclusions
Sildenafil epitomizes the hallmark Rule of 5 (Ro5) properties (Lipinski, 2004; Lipinski et al. , 2001) regarding overall size and lack of polarity that allow near-complete systemic delivery into the circulation (Hong et al. , 2017; Walker et al. , 1999). Extensive structural studies of sildenafil have identified the piperazine nitrogen as amenable to modifications that affect pharmacokinetic properties without dramatically affecting pharmacological efficacy (Dunn, 2005; Kimet al. , 2001b; Kim et al. , 2001c; Yoo et al. , 2007). To develop a non-systemic PDE5i we chose to replace the methyl group of sildenafil with malonate, which is predicted to dramatically reduce permeability at post-gastric pH values between 5-7. This modification did not affect the inhibitor potency as malonyl-sildenafil was equally potent as sildenafil toward human PDE5a. The IC50 value for sildenafil determined using the assay conditions in the present study were slightly higher than previously reported (Gupta et al. , 2005). However, the results are congruent with a previous study suggesting that the carboxylate mimics the phosphate of cGMP to increase affinity (Kim et al., 2001a). We used a novel assay to assess cell permeability, whereby entry of PDE5i into colon cancer cells amplifies GC-C activity that is driven by a low level of the synthetic agonist linaclotide. We used ELISA to measure cGMP levels in our system, which demonstrated a dramatic reduction in EC50 for malonyl-sildenafil compared to the parent compound sildenafil. This observation was supported using activation of cGMP-dependent protein kinase (PKG2) as a readout for the ability of PDE5i to enter cells and increase cGMP levels. As the IC50 toward purified PDE5 was similar, these results support the idea that the carboxylic acid group impeded diffusion of malonyl-sildenafil across the plasma-membrane. However, a differential susceptibility of the compounds to efflux pumps (e.g., MDR1) might also have contributed to the exclusion but was not investigated here.
Sildenafil exhibits near-complete systemic delivery (Hong et al. , 2017), and in circulation it undergoes hepatic modification into desmethyl-sildenafil that retains some bioactivity (Kim et al., 2003). The prominent desmethyl-sildenafil peak in the MS/MS ion spectrum is due to the ionization energy as it was not detected in the liquid chromatography elution profile using pure compound, or in the plasma from malonyl-sildenafil treated mice (not shown). One of the most significant findings is that oral administration of malonyl-sildenafil even at concentrations up to 36 mg/kg exhibited minimal leakage into the circulation but was excreted at high levels in the feces. This contrasted sharply with sildenafil, which peaked at 1-2 hours in the plasma and was not detected in the feces. The typical human dose of sildenafil is 0.7-1.4 mg/kg that is equivalent to consumption of a 50 or 100 mg pill (respectively). The 36 mg/kg dose used here is 26 times the maximum human dose, and 2-4-fold the human dose with allometric scaling. This clearly demonstrates that malonyl-sildenafil has very poor permeability in vivo and is largely retained in the gut lumen. The typical transit time for these mice is between 2-4 hours (Sharman et al., 2017), so the continued excretion of malonyl-sildenafil in the feces 8 hours after oral gavage was surprising. One possibility for future exploration is that the drug is retained in the mucus layer long after excretion of the luminal contents. Regardless, this property is ostensibly ideal for a drug that targets the colon epithelium. A major drawback to the use of contemporary PDE5i for colon cancer prevention is the contraindication for patients taking nitrates to treat stable angina due to synergistic hypotensive effects (Lee et al. , 2004). It was shown here that malonyl-sildenafil did not significantly affect blood pressure in mice exposed to ISMN. This supports the pharmacokinetic data and is congruent with the notion that bioactive lipophilic metabolites were not generated from malonyl-sildenafil.
Taken together the pharmacokinetic results demonstrated that malonyl-sildenafil did not enter the circulation, but it was unclear whether it could target the colon epithelium. Previous studies have demonstrated that PDE5i treatment of mice can alter intestinal homeostasis by suppressing proliferation and apoptosis while promoting differentiation (Sharman et al., 2017; Wang et al., 2014). The suppression of proliferation in the preneoplastic epithelium has been proposed to mediate the CRC-prevention effects of PDE5i (Browning, 2019; Islam et al. , 2018) because this proliferative compartment is where replication errors and genotoxic stress drive neoplastic transformation (Chamorro-Jorganes et al. , 2010; Tomasettiet al. , 2017). Two different approaches were used to demonstrate that oral administration of malonyl-sildenafil to mice was equally effective as sildenafil at suppressing the proliferative compartment in the colon. This result was surprising because the 18-fold reduction in cell penetration for malonyl-sildenafil (EC50) would predict a lesser response in vivo . It is plausible that malonyl-sildenafil levels became concentrated in the mucosal region over time. However, a more plausible explanation is that the lengthy exposure of malonyl-sildenafil to the colon epithelium compared to transient delivery of sildenafil by systemic delivery through the vasculature, produced a larger area under the curve to compensate. These possibilities might be explored in future studies with this compound.
In summary, with the goal of creating a non-systemic PDE5i that targets the colon epithelium, we have synthesized a novel analog of sildenafil containing a carboxylic acid group derived from malonic acid. Malonyl-sildenafil remains in the colon for many hours without entering the circulation and is equally effective as sildenafil at suppressing epithelial proliferation. The results shown here are proof of principle that will enable the development of drugs for the prevention and treatment of intestinal disorders, with fewer side-effects and drug-drug interactions.
Acknowledgement: This work was supported by a grant from the Georgia Research Alliance, USA (GRA.VL21.E1 to DDB), and from the National Institutes of Health (1 R15 CA264555-01 to IL).
Conflict of interest statement: DDB and IL are cofounders of a startup biotechnology company (Targut Biotechnologies Inc.) that has the goal of developing drugs to target intestinal diseases.