Synthesis and pharmacological characterization of potent, selective, and orally bioavailable isoindoline class dipeptidyl peptidase IV inhibitors

Focused structure-activity relationships of isoindoline class DPP-IV inhibitors have led to the discovery of 4b as a highly selective, potent inhibitor of DPP-IV. In vivo studies in Wistar/ST rats showed that 4b was converted into the strongly active metabolite 4l in high yield, resulting in good in vivo efficacy for antihyperglycemic activity.


Background
With the advent of sitagliptin (MK-0431) and vildagliptin (LAF-237), doubt no longer exists regarding the potential of dipeptidyl peptidase IV (DPP-IV; CD26; E.C. 3.4.14.5) inhibitors for the treatment of type 2 diabetes [1][2][3][4]. Hence, intensive research efforts are being continued, and have led to the discovery of a number of potent DPP-IV inhibitors ( Figure 1) [5][6][7][8][9]. Research on second-generation DPP-IV inhibitors has focused on selectivity for DPP-IV over other proline-specific dipeptidyl peptidases, especially DPP8/9, since it has been suggested that inhibition of DDP-8/9 is associated with severe toxicity [10,11]. In addition, the results of recent clinical trials have indicated that prolonged and marked inhibition of DPP-IV would be beneficial for severely diabetic patients [12,13]. The requirement for prolonged, high exposure in humans imposes stringent requirements on the safety profiles and ADME properties of back-up compounds. In this article, we describe our preliminary results with potent and selective isoindoline class DPP-IV inhibitors with respect to CYP, cytochrome P450, induction, and rodent PK, studies as well as inhibition of DPP-IV activity.

Results and discussion
Very recently, Jiaang and co-workers reported that prolinenitrile-based inhibitors with heterocyclic rings showed high selectivity and potency for DPP-IV as well as in vivo efficacy compared to vildagliptin [14]. We had also pursued the possibility of isoindoline class DDP-IV inhibitors and found their high potency and excellent in vivo efficacy [15]. Thus, isoindolines were synthesized as shown in Figure 1 and evaluated in vitro for their ability to inhibit human recombinant DPP-IV and were also screened for their selectivity over DPP-8/9 by a fluorescence assay using glycyl-proline 7-amino-4-methylcoumarin (H-Gly-Pro-AMC). The inhibitory potency is reported as the IC 50 value (Table 1). All the compounds had excellent selectivity for DPP-IV over the other related peptidases. Monosubstitution at positions around the benzene ring of 4a was well tolerated, while retaining a high level of selectivity. Disubstitution, however, led to a slight decrease in potency (4j, k). Disappointingly, most compounds showed CYP induction of either or both of the two enzymes. Eventually, 4b was subjected to further investigation.
In vivo PK studies on 4b showed a short plasma halflife and reduced AUC when dosed intravenously (Table  2). Apparently, the reduction in AUC was partly due to a very high clearance. On the other hand, oral administration of 4b showed an improved half-life and a dosedependent increase in AUC. As it was estimated from the PK profiles that 3-10 mg/kg doses of 4b would provide > 50% inhibition of DPP-IV for several hours, we tried to briefly examine the potency of 4b in oral glucose tolerance tests (OGTT).
Fasted male Wistar/ST rats received either vehicle or 4b at different oral doses ( Figure 2). After 30 min (t = 0), oral glucose challenges (1 g/kg) were conducted and then plasma DPP-IV activities and blood glucose levels were monitored at various intervals over a 2 h period.
Selected data are shown in Figure 2. To our surprise, the 1 mg/kg dose of 4b resulted in 95% inhibition of plasma DPP-IV activity within 30 min post-dose and inhibition of greater than 90% was maintained throughout the study. The inhibitory effect was dose-dependent, and even the 0.1 mg/kg dose produced 30% inhibition. Similarly, reduction of glucose levels paralleled DPP-IV inhibition and a reduction of 18% was observed at a  dose of 1 mg/kg. In addition, increased insulin levels at 10 min post-challenge strongly suggested preservation of active GLP-1.
The high clearance of 4b suggested that the unexpected in vivo efficacy might be explained by the presence of active metabolites. Therefore, further PK studies were conducted. A major metabolite was detected by LC-MS analysis and its structure was determined by comparison of the LC retention time and MS/MS fragmentation pattern with synthetic standards. Consequently, 4l was identified as a very active metabolite, which showed a reasonable degree of systemic exposure by virtue of in vivo conversion as high as 60%.

Conclusions
In summary, the focused, small SARs of the isoindoline derivatives have led to the discovery of 4b as a highly selective, potent inhibitor of DPP-IV. The in vivo studies showed that the active metabolite 4l had a very high inhibitory potency with respect to DPP-IV. Consequently, we abandoned further development of compounds in this series. However, on the basis of the results described here, we found anagliptin, which has advanced into PIII trials, to have improved safety profiles and PK parameters. This article is also intended to provide information on the scope and limitations of isoindoline-based DPP-IV inhibitors and to facilitate research on the new generation DPP-IV inhibitors.

Compound synthesis 4.1.1. General
All commercially available reagents and solvents were used as-received. All reactions were carried out using oven-dried flasks or glassware, and mixtures were stirred with stirring bars and concentrated using a standard rotary evaporator unless otherwise noted.
Procedures for preparation of all intermediates 2 and 3 were described previously [15]. The 1 H NMR spectra were recorded by a JEOL JNM-ECP400 spectrometer operating at 400 MHz in DMSO-d 6 at 25°C with tetramethylsilane as the internal standard. The data are reported as follows: chemical shift in ppm (δ), integration, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br = broad singlet, m = multiplet), and coupling constant (Hz). LC/MS spectra were determined on a Waters ZMD2000 equipped with a Waters 2690 injector and a PDA detector operating at 210-400 nm and interfaced with a Micromass ZMD mass spectrometer.

Biological evaluation 4.2.1. In vitro assay for DPP-IV inhibition
Inhibition of DPP-IV activity was determined by measuring the rate of hydrolysis of a surrogate substrate, H-Gly-Pro-7-amino-4-methylcoumarin (H-Gly-Pro-AMC). Human recombinant DPP-IV was purchased from R&D Systems, Minneapolis, MN. 10 μL of appropriately diluted solutions of the test compounds in water was added to 96-well microtiter plates, followed by the addition of 40 μL of DPP-IV diluted in assay buffer (25 mM HEPES, 140 mM NaC1, 0.1 mg/mL BSA, pH 7.8). After a 30-min preincubation at room temperature, the reaction was initiated by the addition of 50 μL of the assay buffer containing 0.2 mM H-Gly-Pro-AMC. After incubation at room temperature for 20 min, the reaction was stopped by the addition of 100 μL of 25% aqueous acetic acid and fluorescence was measured using an excitation wavelength of 390 nm and an emission wavelength of 460 nm. A standard curve for AMC was generated by adding 0.2-20 μmol of AMC to buffer solutions containing 12.5% aqueous acetic acid. The inhibitory rate relative to the control without inhibitor was calculated and IC 50 values were determined by nonlinear regression (GraphPad Prism 4, ver. 4.03 software).

In vitro assays for inhibition of DPP-8 and DPP-9
Human DPP-8 and DPP-9 were expressed in baculovirus-infected Sf9 insect cells and purified using Histagged protein purification resins. Inhibition of DPP-8 and -9 activities was determined as described above. 10 μL of appropriately diluted aqueous solutions of the test compounds was added to 96-well microtiter plates, followed by the addition of 50 μL of 1.0 mM H-Gly-Pro-AMC in buffer solution (50 mM HEPES, 0.1 mg/ mL BSA, pH 8.0). The reaction was initiated by the addition of 40 μL of the enzyme solution diluted in the assay buffer. After incubation at room temperature for 30 min, the reaction was stopped by the addition of 100 μL of 25% aqueous acetic acid and fluorescence was measured using an excitation wavelength of 390 nm and an emission wavelength of 460 nm.

In vivo assay methods
All procedures were approved by the Sanwa Kagaku Kenkyusho Institutional Animal Care and Use Committee. 7-week old Wistar/ST rats were housed under standard conditions and allowed free access to water and a commercial diet for at least 5 days. The rats were fasted overnight prior to dosing and then received 4b orally at doses of 0.1-1 mg/kg or vehicle as a 5 mL/kg aqueous solution 30 min before glucose challenge. After an oral glucose challenge (5 mL/kg of an aqueous solution of 20% glucose), blood samples were collected from the tail vein of each animal into heparin-containing tubes at serial time points for 2 h. Plasma was prepared immediately, frozen, and stored at -20°C prior to analysis.

Inhibition of rat plasma DPP-IV ex vivo
Plasma DPP-IV activity was determined as described above. A 20 μL plasma sample was mixed with 5 μL of reaction buffer (140 mM NaCl, and 10 mM KCl, 25 mM Tris-HCl, pH 7.4, 1% bovine serum albumin) and 10 μL of buffer containing 60 μM H-Gly-Pro-AMC. After incubation at room temperature for 30 min, the reaction was stopped by the addition of 20 μL of 25% aqueous acetic acid and fluorescence was measured using an excitation wavelength of 360 nm and an emission wavelength of 460 nm.

Measurement of plasma glucose and insulin concentrations
Plasma glucose and insulin were determined with a glucometer (Glutest Pro; SKK, Japan) and a rat insulin ELISA kit (Shibayagi, Japan), respectively, according to the manufacturer's instructions. Statistical analyses were performed using Microsoft Excel. Individual comparisons among more than two experimental groups were assessed using ANOVA, with Fisher's Least Significant Difference post hoc test. Differences were considered significant at P values < 0.05. Analysis of dose-response data was performed by Dunnett's test.

Pharmacokinetics (PK) in rats
Sprague-Dawley (SD) rats were housed under standard conditions and allowed free access to water and a commercial diet. On the day before the experiment, rats were fasted overnight and for the first 12 h of the experiment. Compounds 4b were prepared in a saline/ ethanol vehicle (50/50 v/v) at appropriate concentrations of 4b as an intravenous (iv) injection of 1 mL/kg via the femoral vein and as a suspension in 5% gum arabic solution for oral (po) administration. Blood samples were collected from the jugular vein of each animal with a heparinized syringe under diethyl ether anesthesia at serial time points for 24 h after drug administration. Plasma was obtained by centrifugation at 4°C and stored at -70°C until analysis. Protein precipitation was carried out by the addition of the internal standard solution (70% CH 3 CN with 0.2% acetic acid) to samples. The tubes underwent vigorous shaking and centrifugation for 5 min; then the supernatant was subjected to LC/MS/ MS analysis. Peak areas were determined using Xcalibur ® software (Thermo Electron Corporation, UK) and AUC values were calculated by the trapezoidal rule.

Metabolic stability
The incubation mixture containing 0.25 mg of rat or human liver microsomes was preincubated with an NADPH-generating system for 5 min at 37°C. The reaction was started by the addition of 5 μL of a DMSO solution containing the test compound (5 μM). At t = 0 and at two additional time points between 0 and 30 min, aliquots (100 μL) were removed and added to termination mixtures (CH 3 CN). Proteins were sedimented by centrifugation and an aliquot of the supernatant was analyzed by LC/MS/MS. In determinations of the in vitrot 1/2 , the analyte/ISTD peak area ratio was converted to percentage of drug remaining by assigning a value of 100% to the peak area ratio at t = 0. The slope of the regression line fitted to the log (percentage remaining) versus incubation time relationship (-k) was used in the conversion of raw data to the in vitrot 1/2 value. In vitro CL int was calculated using the following formula.