Discovery of Novel and Orally Bioavailable Inhibitors of PI3 Kinase Based on Indazole Substituted Morpholino-Triazines

Abstract

A new class of potent PI3Kα inhibitors is identified based on aryl substituted morpholino-triazine scaffold. The identified compounds showed not only a high level of enzymatic and cellular potency in nanomolar range but also high oral bioavailability. The three lead molecules (based on their in vitro potency) when evaluated further for in vitro metabolic stability as well as pharmacokinetic profile led to the identification of 26, as a candidate for further development. The IC₅₀ and EC₅₀ value of 26 is 60 and 500 nM, respectively, for PI3Kα enzyme inhibitory activity and ovarian cancer (A2780) cell line. The identified lead also showed a high level of microsomal stability and minimal inhibition activity for CYP3A4, CYP2C19, and CYP2D6 at 10 μM concentrations. The lead compound 26, demonstrated excellent oral bioavailability with an AUC of 5.2 μM at a dose of 3 mpk in mice and found to be well tolerated in mice when dosed at 30 mpk BID for 5 days.

Phosphatidyinsoitol-3-kinase (PI3K) is a lipid kinase and an important component in the PI3K/AKT/mTOR signaling pathway that plays a key role in cell proliferation, growth, survival, and apoptosis as well as protein and glucose metabolism.¹⁻³ Class 1 PI3K protein consists of four isoforms known as PI3K α, β, γ, and δ. This classification is based on their sequence homology and substrate preferences. The deregulated activation of PI3Kα and its downstream effectors including AKT and mTOR have been linked to tumor initiation and maintenance. Class 1 PI3K protein in general and PI3Kα isoform specifically plays a key role in human cancer biology.⁴⁻⁷ Aberrantly activated PI3Kα protein has been implicated in poor prognosis and survival in patients with various lymphatic tumors, as well as breast, prostate, lung, brain (glioblastoma), skin (melanoma), colon, and ovarian cancers.⁴ Therefore, inhibiting PI3Kα is an attractive strategy for cancer therapy. Clinical interest in PI3K inhibitors has been intensifying. In recent years, a large number of compounds have been reported in clinical development phase having strong antitumor activity in animal models.⁸⁻¹¹

During our efforts toward the identification and development of class 1 PI3K inhibitors, we identified hit 3 based on the triazine morpholine core.¹²⁻¹⁴ This hit was inspired by the leads reported (Figure 1, compounds 1 and 2) by Hayakawa et al.¹⁵ It was planned to develop a quick structure–activity relationship (SAR) around identified hit 3. Based on data reported by others and our internal data, we had ascertained that, the morpholine moiety is critical for activity and potency as it interacts with a hinge region backbone NH (Val882) of PI3K protein.¹⁶⁻¹⁸ Therefore, the morpholino group retained in all subsequent molecules for this study.

Figure 1.

PI3Kα IC₅₀ of identified hit 3, compared to literature leads 1 and 2.

Synthesis of the various morpholino–triazines reported in Table 1 was achieved by the general synthetic route outlined in Scheme 1 involving a combination of Suzuki coupling and nucleophilic substitution on morpholine substituted triazine ring.¹⁹

Table 1.

^*The coefficient of variation (CV %) was found to be <10% for IC₅₀.

Scheme 1.

Reagents and conditions: (a) morpholine, acetone/water (4:1), 0 °C, 3 h; (b) amino acid/ester, acetone/NaHCO₃ aq. (1:1), RT, 16 h; (c) aryl borate, Pd(PPh₃)₄, Na₂CO₃, DME/water (4:1), 120 °C, 20 min, MW.

Compounds 4, 5, and 6 illustrate the importance of the phenolic group, as protection, removal, or replacement results in complete loss of activity. This finding is in line with literature reports^17,18 that highlighted the importance of hydrogen bond donor moieties for potency in PI3K inhibitors having aryl morpholine based structural motifs. Further, compound 3 was coupled with different amines to produce their corresponding amides (entry 9–12, Table 1) and evaluated for their in vitro PI3Kα inhibition activity. The PI3Kα enzyme inhibitory activity of the compounds was determined using a homogeneous time resolved fluorescence (HTRF) assay. The percent inhibition was calculated and plotted against the concentration of the inhibitor to calculate IC₅₀.¹⁹ Interestingly, and as illustrated by compounds 7 through 13, variations of the glycine moiety of 3 does not result in any significant change in activity. Compound 13 even suggests that the glycine moiety may not be interacting with the protein and may potentially be oriented outside the protein into the solvent. This is complementary to literature findings.²⁰ The activity of compound 8 suggests that this loss of activity may be due to steric interaction between the ligand and the protein or the lack of hydrogen bonding interaction between the protein and the ligand. The marginal increase in activity of compound 13 over compound 3 may be due to the either less steric hindrance or less encumbered ability for the formation of the hydrogen bond between the ligand and the kinase.

From this series of molecules, compound 11 was selected for further optimization, primarily based on its potency against PI3Kα. Though, we had initially established the importance of the phenolic group, we needed to find an alternate of phenol. This is due to the generally accepted observation of the poor PK/ADME characteristics imparted to a compound with the presence of a phenol moiety. A small set of new molecules were synthesized as analogues of 11 to find a replacement for phenol (Table 2). A two/three step synthesis, involving a Suzuki coupling with different boronic acids as a key step for diversification was exploited to generate these analogues (Scheme 2).¹⁹ The data from this new set of molecules led to the selection of compound 16 over compound 11 for further optimizations, even though it was 2-fold lower in in vitro potency, primarily due to the anticipated issues with the phenol.

Table 2.

^*The coefficient of variation (CV %) was found to be <10% for IC₅₀.

Scheme 2.

Reagents and conditions: (b) glycine amide, acetone/NaHCO₃ (aq.) (1:1), RT, 16 h; (c) aryl borate, Pd(PPh₃)₄, Na₂CO₃, DME/water (4:1), 120 °C, 20 min, MW.

Further optimization was planned targeting enhancement of the in vitro enzymatic potency by varying the substituents on the glycine amide side chain on compound 16 (Table 2). Scheme 3 illustrates the synthetic route deployed for the synthesis of this series of molecules based on the indazole-morpholino-triazine scaffold. The new molecules in this series were screened for their cellular potency on the human ovarian cell line (A2780) along with enzymatic activity. The cellular potency was determined by the MTT assay using human ovarian cancer cell line (A2780). The results of this screening are documented in Table 3. Compound 18, synthesized from homoglycine amide, was found to be slightly more potent than the original lead 16. Intriguingly, conformationally restricted analogues of glycine amide, i.e., 6- and 7-membered cyclic amides (compound 19 and 20) as a replacement, were found to be more potent compared to original lead 16 and compound 18. The absence of a potential hydrogen bonding interaction by the amino group attached to the triazine and the steric size of the cyclic lactams negates our explanation for the loss of activity of compound 8 over compound 3. The addition of methyl on N-linker may have conformational impact leading to steric clashes. The small but favorable shift in in vitro potency for compound 18 to 20, relative to the original lead 16, highlights the role of the spacer and the spatial orientation of the terminal amide. To explore this further, several other analogues were synthesized using an aromatic ring spacer between the triazine and terminal amide (compounds 21–24). All of the new analogues demonstrated improved enzymatic and cellular potency. Replacement of the bridging nitrogen (compound 22) with oxygen (compound 25) led to a 4-fold loss in enzymatic as well as cellular potency, but this loss was recovered by moving the amide group from the meta- to the para-position on the phenyl ring (compound 26) resulting in the best compound in this series. Replacing the aromatic ring spacer of 26 with a cycohexyl ring (compound 28 and 29) resulted in compounds equipotent to 26. Hence, compounds 26, 28, and 29 were identified as the optimized leads based on their enzymatic and cellular potencies. These three leads were further evaluated for their in vitro metabolic stability with human liver microsome (HLM), dog liver microsome (DLM), and mouse liver microsomes (MLM). The three leads were incubated with isolated liver microsomes of the three species for 2 h. The concentration of the parent compound was monitored by HPLC at four time points (0, 0.5, 1, and 2 h, respectively). All three molecules (26, 28, and 29) showed encouraging level of microsomal stability particularly in HLM. The data of 2 h time point is shown in Table 4 and a detailed study report is provided as Supporting Information.

Scheme 3.

Reagents and conditions: (d) benzyl mercaptan, DIPEA, morpholine, THF, 0 °C, 5 h; (e) aryl borate, Pd(PPh₃)₄, Na₂CO₃, DME/water (4:1), 90 °C, 18 h; (f) oxone, THF/water (1:1), 0 °C–RT, 25 h; (g) RX, DMF, K₂CO₃, RT, 24 h; (h) methanesulfonic acid, methanol/water (2:1), 55 °C, 1 h.

Table 3.

^*The coefficient of variation (CV %) was found to be <10% for IC₅₀ and <5% for EC₅₀.

Table 4.

Compounds 26, 28, and 29 were screened further for their cytochrome P450 inhibition activity. All three leads showed no or minimal inhibitions for activity of CYP3A4, CYP2C19, and CYP2D6 at 10 μM concentrations. All three molecules were also evaluated for oral bioavailability in mice. Each was dosed orally at 3 mpk to Swiss mice to evaluate their pharmacokinetic profile and determine the key parameters such as C_max, AUC, t_max, and t_1/2. The results are summarized in Table 4. Compounds 26 and 28 showed better PK profile compared to 29 with higher plasma exposure of >5 μM at 3 mpk dosing. On the basis of its potency and pharmacokinetic profile, 26 was selected as the molecule of choice for further evaluation

Compound 26 was evaluated in vivo for dose ranging pharmacokinetic study in mice at three different doses of 1, 3, and 10 mpk. A summary of the pharmacokinetic parameters are shown in Table 5. The plasma exposure of 26 demonstrated a reasonable dose response exposure with escalating dose from 1 to 3 to 10 mpk.

Table 5. Dose Ranging Pharmacokinetics Study of 26 by Oral Administration in Mice.

	dose (oral)
PK parameters	1 mpk	3 mpk	10 mpk
tmax (h)	1.00	1.00	2.33
Cmax (nM)	360	1243	1798
AUC (nM·h)	2117	5246	9791
elimination t1/2 (h)	6.93	4.86	6.87

The lead 26 was also screened for hERG activity using the HERG-lite assay from Chantest Corporation.¹⁸ The objective of the study was to estimate in vitro effects of 26 on surface expression of the hERG potassium channel. Both wild type (WT) and the single-point mutant (SM) channel G601S were examined. Predicted hERG liability was indicated as channel block and/or trafficking inhibition. On the basis of the hERG lite assay data, the IC₅₀ of the hERG liability determined by the patch clamp method was predicted. Overall, the predicted hERG liability of 26 was found to be >10 μM implying that 26 is not expected to have hERG liability.

The in vivo multiday dose tolerability of 26 was evaluated in Swiss mice.¹⁹ The lead 26 was dosed at 30 mpk BID for 5 consecutive days and monitored for effect and changes in body weight, body temperature, body coat, and level of physical activity. The animals did not show any significant body weight loss (Figure 2), hyper- or hypothermia, change in body coat or lowered physical activity. This clearly indicates that 26 is well tolerated by mice.

Figure 2.

Effect of 26 on mice bodyweight when dosed at 30mpk BID oral for 5 days.

In conclusion, our efforts toward development of class I PI3K inhibitors led to identification of a small molecule 26 as a potent and orally bioavailable inhibitor of PI3Kα protein. The lead 26 was found to have excellent metabolic stability, no hERG liability with appreciable oral exposure, and tolerability. Efforts for the development of 26 are in progress and will be summarized in a future publication.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.5b00322.

• Experimental details of chemical synthesis as well as characterization and in vitro and in vivo screening of all compounds (PDF)

The authors declare no competing financial interest.