MED
which were obtained in reasonable yields, were subjected to
smooth chlorination with N-chlorosuccinimide to afford the
chloro[1,2]oxazoles 13a–g in moderate yields (40–50%).
All the synthesized compounds were submitted to the US
National Cancer Institute (NCI) in Bethesda for cell-based anti-
proliferative screens. With the exception of the [1,2]oxazolo-
[4,5-g]indoles 6a,b, all submitted [1,2]oxazolo indoles were se-
lected for the one-dose (10ꢀ5 m) prescreening on the full panel
of 60 human tumor cell lines divided into nine subpanels (leu-
kemia, non-small-cell lung, colon, central nervous system, mel-
anoma, ovarian, renal, prostate, and breast cancers).
series (compare 6n with 13e), the chloro derivative main-
tained biological activity, whereas in the 3,5-dimethoxy series
(compare 6t with 13g), a decrease in biological activity by
more than one order of magnitude was observed.
Analysis of the GI50 values listed in Table 3 indicates that the
most active compound, 6o, was particularly effective against
the leukemia subpanel. In fact, the calculated MG_MID value
for the leukemia subpanel was higher than the overall cell
lines MG_MID value (DMG_MID=0.05 mm). The most sensitive
cell lines were SR and K-562, having GI50 values in the nanomo-
lar range, 50 and 60 nm, respectively. Compound 6o was also
highly selective against the colon cancer, ovarian cancer and
CNS cancer subpanels, having subpanel MG_MID values at
sub-micromolar levels, 0.25, 0.31 and 0.36 mm, respectively.
Also in these subpanels the most sensitive cell lines showed
GI50 values in the nanomolar range: KM12 of the colon cancer
subpanel (90 nm), OVCAR-3 and NCI/ADR-RES of the ovarian
cancer subpanel (40 and 80 nm, respectively), and SF-295 of
the CNS cancer subpanel (60 nm). The same compound, how-
ever, got responses in the nanomolar range in other cell lines
belonging to different subpanels: NCI-H522 of the non-small-
cell lung cancer subpanel (50 nm), MDA-MB-435 of the melano-
ma subpanel (30 nm), A498 of the renal cancer subpanel
(80 nm), and MCF-7 of the breast cancer subpanel (50 nm).
Compound 6t, which was slightly less active than 6o, revealed
selectivity against leukemia, colon cancer, and ovarian cancer
subpanels, with MG_MID values at sub-micromolar levels (0.33,
0.35, and 0.44 mm, respectively), but in none of the cell lines
reached the nanomolar range. In fact, it scored sub-micromolar
figures in 51 out of 56 tested cell lines (93%) and in the re-
maining four cell lines, with the exception of the T-47D cell
line, the response was at the low micromolar level.
Among these, eight compounds (6h,k,n,o,t,u, and 13e,g)
were selected for evaluation on the full panel at five doses
(10ꢀ4–10ꢀ8 m). All selected compounds gave positive GI50
values against all tested human cell lines with the exception of
6k, 6t, and 13g, each of which were not responsive against
one cell line (Table 2), respectively MCF7 and T47D of the
breast cancer subpanel, and NCI-H322M of the non-small-cell
lung cancer subpanel (Table 3).
Table 2. Overview of the antitumor screenings of 6h,k,n,o,t,u and 13e,g.
Compd
No. of cell lines[a]
screened positive
GI50 range
[mm][b]
MG_MID[c]
6h
6k
6n
6o
6t
6u
13e
13g
59
59
59
54
56
59
59
59
59
58
59
54
55
59
59
58
0.37–16.4
0.88–52.4
0.16–39.6
0.03–31.1
0.14–2.89
0.25–46.5
0.30–10.1
2.15–50.0
3.89
5.13
1.41
0.25
0.47
1.35
1.99
7.08
[a] The number of cell lines screened and the number of cell lines in
which the test compound gave a positive GI50 value. [b] The GI50 value is
defined as the compound concentration causing 50% growth inhibition
of tumor cells. [c] Mean graph mid-points (MG_MID).
The less active compounds 6u, 6n, and 13e, however, ex-
hibited MG_MID values of 1.35, 1.41, and 1.99 mm, respectively,
showing in 30–50% of the tested cell lines GI50 values in the
sub-micromolar range, and in the remaining cell lines, with
some rare exceptions, the GI50 values remained in the low
micromolar range.
An evaluation of the data listed in Table 3 indicates that the
presence of a 3,5-dimethoxy-substituted benzyl moiety at the
indole nitrogen (position 8) appears crucial for conferring
good activity to the [1,2]oxazoloindole derivatives. In fact, the
most active compounds were 6o and 6t, having mean graph
mid-points (MG_MID) of 0.25 and 0.47 mm, respectively. Within
the 3,5-dimethoxybenzyl series, when the ethoxycarbonyl sub-
stituent at position 7 (6o, MG_MID=0.25 mm) was replaced by
a phenyl group (6t), a slight decrease in activity was observed
(MG_MID=0.47 mm), whereas the corresponding unsubstituted
derivative (6h) was significantly less active (MG_MID=
3.89 mm). Removal of the methoxy group from position 5 of
the benzyl portion led to a substantial decrease in activity
(compare 6o, MG_MID=0.25 mm, with 6n, MG_MID=1.41 mm,
and 6h, MG_MID=3.89 mm, with 6g, inactive). Moving the me-
thoxy moiety from position 3 to 4 led to compounds devoid of
activity. The introduction of a 4-methoxy group to generate
a 3,5-dimethoxybenzyl derivative also led to a decrease in bio-
logical activity (compare 6t and 7u). Substitution at position 6
with a chloro group led to mixed results: in the 3-methoxy
In conclusion, reported a method for the synthesis of deriva-
tives of the new ring system [1,2]oxazolo[4,5-g]indole. The anti-
proliferative activity exhibited by derivatives 6h,k,n,o,t,u and
13e,g against the totality of the NCI full panel of human
tumor cell lines makes this class of compounds interesting for
further studies. In particular, the potent activity of compounds
6o and 6t encourages the synthesis of new derivatives and
makes them lead compounds for a new class of 3,5-dimethoxy-
benzyl[1,2]oxazolo[4,5-g]indoles with the aim of obtaining
more potent antiproliferative agents.
Experimental Section
Full details of the protocols used to synthesize the compounds
described and a brief summary of the US NCI cell-based antiproli-
ferative screens are given in the Supporting Information.
ChemMedChem 0000, 00, 1 – 4
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
3
&
ÞÞ
These are not the final page numbers!