R. Alexander et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4316–4320
4319
Table 3
Potency in selected cell lines
Compound
pAKTS473 (PC3)
Proliferation (PC3)
pAKTS473 (MCF7)
Proliferation (MCF7)
Proliferation (U87-MG)
1
14h
18
2939
1555
1412
5473
2840
2066
NT
NT
199
IA
1255
862
11,070
787
3501
Values are means of two experiments. IC50 is quoted in nM. NT, not tested; IA, inactive.
lene unit between the morpholine and the phenyl group lead to a
differentiation in the activity of the enantiomers, the R, compound
14c, being inactive and the S, compound 14d, active. Replacing the
phenyl by napthyl showed a marked preference for the 1-substi-
tuted napthyl, compound 14e, over the 2-substituted, compound
14f. The 1-napthyl could be replaced by a 3-benzothienyl, 3-indo-
lyl or 3-(7-azaindolyl), compounds 14j, 14h, and 14i, respectively.
There was a clear preference for the S enantiomer as shown by
comparison of compounds 14h and 14g and supported by com-
pounds 14d and 14c. Introduction of the morpholine substituent
120
100
80
60
40
20
0
5000
4000
*
3000
2000
1000
**
***
gives a degree of selectivity for PI3Ka, d, and c over b. The in-
0
creased potency of compound 14h can be explained by Figure 3,
0
1
3
6
which shows a docking of compound 14h into the ATP binding site
Time (h) after 200 mg/kg compound 18
of PI3Kc 7 NMR data10 suggest that the 3-substituent of the mor-
.
Figure 4. Time-dependent reduction of pAKT in U87-MG human glioma tumors
after treatment with compound 18. U87-MG tumor bearing mice were treated with
vehicle (0 h), or 200 mg/kg compound 18 po. pAKT was measured by immunoassay
in tumor homogenates at 1, 3, and 6 h post-treatment (bars). Drug plasma
concentration was measured by LC–MS/MS (line). ***p < 0.001%, **p < 0.01%,
*p < 0.05% versus control, one-way ANOVA followed by Bonferroni post test.
pholine adopts an axial conformation. The lone pair of the morpho-
line oxygen of the ligand forms a hydrogen bond to the backbone
NH of Val882. The indolyl substituent of the ligand projects toward
solvent, closely fitting a space between a hydrophobic surface lar-
gely formed by the aliphatic side chain of Met953 below, and an
edge-face
Trp812 above.
p–p stacking interaction with the indolyl side chain of
pound 18 642
and its lower in vitro intrinsic clearance (compound 18 24
min/mg compared to compound 14h 35 L/min/mg, male rat
hepatocytes). The in vivo pharmacokinetics of compound 18 were
assessed in female nude mice following single po administration at
30 mg/kg. Compound 18 had a Cmax of 1455 ng/mL at 1 h post dose
and an AUClast of 2667 ng h/mL. Exposure increased at higher po
doses and was deemed sufficient to progress compound 18 into
xenograft studies.
The in vivo efficacy of compound 18 was assessed in nude mice
bearing U87-MG tumors grown subcutaneously. Ex vivo analysis of
tumor tissue after a single dose of 200 mg/kg po of compound 18
showed an inhibition of pAKT levels that was dependent on drug
concentration, Figure 4. A 65% decrease in pAKT was observed at
1 h, and was partially sustained to 6 h (32% decrease compared
to vehicle). Total AKT was also measured but stayed constant.
l
M compared to compound 14h 201
l
M, pH 7.4)
lL/
Having identified the 3-indolyl methyl group as a substituent of
interest, we sought to combine this with modification of the core
scaffold. Reaction of 2 with hydroxylamine hydrochloride in pyri-
dine followed by treatment with catalytic p-toluene sulfonic acid
in pyridine gave the lactam 15 in 65% yield, which was converted
to the bromide 16 with NBS in 81% yield. Morpholine 13h was syn-
thesized as shown in Scheme 2 from S-tryptophan and then con-
verted to the thiourea 17 which was reacted with bromide 16 to
give compound 18 in 96% yield (Scheme 3).
l
The lactam 1811 showed similar potency to the ketone 14h
against the PI3K isoforms (IC50 = 59, 1006, 18, and 31 nM for
d, and , respectively).
a, b,
c
Activation of the PI3K pathway leads to phosphorylation of
Thr308 and Ser473 of AKT and subsequently a number of down-
stream substrates, including IKK, BAD, GSK, Forkhead, and Cas-
pase-9. This activation triggers a range of responses in cell
metabolism, cell death, cell cycle progression and cell differentia-
tion. Hence, inhibiting PI3K should block phosphorylation of AKT
and inhibit cell proliferation. Compounds 1, 14h, and 18 were pro-
filed against PTEN negative PC3 prostrate tumor and U87-MG glio-
600
500
400
300
blastoma cell lines and MCF7 breast cancer cells, a mutated PI3K
cell line, results shown in Table 3.
a
Compounds 14h and 18 demonstrated the ability to inhibit the
phosphorylation of AKT and inhibit proliferation via the PI3K path-
way in PTEN negative and PI3Ka mutant cell lines and were shown
200
**
***
to be inactive against a panel of kinases that included AKT (data
not shown). Compound 1 showed comparable activity to com-
pounds 14h and 18 in inhibiting the phosphorylation of AKT in
PTEN negative cell lines. Compound 1 has similar potency to com-
pounds 14h and 18 against PI3Kb but is over twenty fold less active
100
0
0
2
4
6
8
10 12 14 16
Day from start of treatment
against PI3K
a, d, and c, suggesting a role for PI3Kb in PTEN nega-
Figure 5. In vivo efficacy of compound 18 in U87-MG xenografts. U87-MG tumor
bearing mice were treated with compound 18 po, twice per day for 14 days, at
either 100 mg/kg (open triangles) or 200 mg/kg (closed triangles). The anti-tumor
activity measured as optimum T/C % (mean volume of treated group divided by the
mean volume of vehicle treated control group (closed squares), multiplied by 100)
was 31% and 25% on day 15, respectively. **p < 0.01% versus control, ***p < 0.001%
versus control, two-way ANOVA followed by Bonferroni post test.
tive cell lines. However, compound 1 was inactive against the
MCF7 cell line which has a PI3K
activity against PI3K , d, and
Compound 18 was selected over compound 14h for in vivo
pharmacokinetic profiling based on its superior solubility (com-
a
mutation reflecting the decreased
a
c
.