C. Sherer et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
3
Preliminary SAR of fragments
increase, and the observed anticancer activity should also increase
accordingly. Conversely, destabilising the cation should result in
lower anticancer activity since the formation of the intermediate
The 3-carbinol group
(
and hence dimer) would be retarded. Three analogues were
The 3-carbinol group of I3C is considered an important feature
for biological activity due to its role in the degradation to the active
species, DIM. Its mechanism involving the loss of water and con-
comitant stabilisation of the carbocation through conjugation can
be seen in Scheme 1. Consequently, by extending the carbon chain
from hydroxymethane to hydroxyethane, thus no longer having
the hydroxyl group conjugated with an electronegative atom on
which the positive charge can rest, should prevent the formation
of the active dimer, and biological activity would be reduced. Like-
wise, by the same argument, indole and 2-phenylindole should
produced in order to probe this hypothesis (Figs. 3, 10–12). Two
N-substituents were chosen which are both electron withdrawing
groups but to differing extents (the acetyl group being much less
electron withdrawing than the tosyl group), structural features
which should impact upon the rate of formation of cation 7, and
thus anticancer activity. Acetyl groups are known to be hydrolysa-
ble and so could be expected to be cleaved in the cell, especially in
the presence of numerous cellular proteases. However, the rate of
hydrolysis compared to the rate of dimerisation would determine
in which order these two processes would be likely to occur. For
example, if the acetyl group is removed too easily (Scheme 2,
k1Hydrol ꢀ k1Dim), then I3C would be readily formed and the rate
of dimerisation would likely be the same as the rate of dimerisa-
tion of I3C, so no difference in activity would be expected when
comparing 12 to I3C. However, if the acetyl group is sufficiently
stable, it should retard the rate of formation of the dimer (based
on the mechanism shown in Scheme 1), which should also result
in a reduced activity. These competing pathways are shown in
Scheme 2.
also be less active since they have no 3-substituent and thus a
0
3
,3 -diindolylmethane metabolic product cannot be formed.
In relation to two established glioblastoma cell lines (U87,
U251) and three short-term glioblastoma cultures (IN1472,
IN1528, IN1760) against which the compounds were tested, indole
itself and 2-phenylindole (4) had EC50 values that were too high to
calculate based on the concentration ranges used (entries 1 and 4),
suggesting a lack of activity in glioblastoma for these compounds.
However, compound 2 was shown to have EC50 values in the range
of 10–1430
l
M, demonstrating a large increase in activity upon
To ascertain if the inclusion of a hydrolysable group is signifi-
cant in these compounds, an analogue containing a non-hydrolysa-
ble N-substituent should be included for comparison. This non-
hydrolysable group should be as chemically similar as possible to
the acetyl group in order to keep the analogue as otherwise similar.
Therefore, the group should have electron withdrawing mesomeric
properties, be relatively small, and be a hydrogen bond acceptor. A
substituent that fulfils all of these criteria is the tosyl group. As
such, the N-tosyl and N-acetyl analogues of I3C (compounds 11
and 12 respectively, shown in Fig. 3) were synthesised and tested
as direct comparisons to I3C, in addition to the N-acetyl analogue
of 2, namely compound 10.
When screened against the glioblastoma cell cultures (U251,
IN1528 and IN1760), the effect of the N-acetyl group shows a clear
negative effect on anticancer activity, with both I3C and compound
2 (entries 2 and 3) having better activity than their N-acetyl ana-
logues 10 and 12 (entries 8 and 10). Based on the established pro-
drug nature of I3C to convert to its active DIM form in biological
assays, these results imply that the acetyl group is not effectively
removed, either by hydrolysis or by enzymatic activity, and the
formation of the dimer, through a mechanism similar to that
shown in Scheme 1, is prevented. That said, it cannot be ruled
out that the role of the NH group is important for binding, or that,
through N-substitution, the ability of the compounds to directly
bind to cellular targets in their ‘‘prodrug” form has been altered.
inclusion of the 3-carbinol group (compare entries 2 and 4), results
which were supported by the data for indole vs. I3C (EC50(I3C)
=
119–1580
lM, entries 1 and 3) (see Table 1).
Extension of the 3-carbinol group (entries 6 and 7) yields less
conclusive results. Comparing I3C with its hydroxyethane ana-
logue 8 (entries 3 and 6) shows that 8 has consistently worse activ-
ity across all five cell lines and cultures on which they were tested.
However, the 2-phenyl analogues (2 and 9, entries 2 and 7) show
much less consistency, with compound 2 having better activity
against the two established cell lines U87 and U251, whereas its
hydroxyethyl analogue (9) has better activity against the three
short term cultures IN1472, IN1528 and IN1760, and is particularly
active against IN1528 (EC50 = 4 lM). That said, when comparing 2
with 3, and 8 with 9, with respect to the 2-phenyl group, it is
evident that the phenyl ring increases efficacy across all cells lines
and cultures studied.
The effect of N-substituents
Since a cationic intermediate (7, Scheme 1) is involved in the
conversion of I3C to its active form, DIM, the stability of this inter-
mediate presumably impacts on the amount of DIM produced
in vitro. Therefore, by increasing the stability of intermediate 7 suf-
ficiently to enable dimerisation to occur, the yield of DIM should
Table 1
Comparison of EC50 values (lM).
Entry
Compound
U87 (lM)
U251 (lM)
IN1472 (
l
M)
IN1528 (lM)
IN1760 (lM)
a,b
1
2
3
4
5
6
7
8
9
Indole
2
–
–
–
–
–
b
30 ± 8
290 ± 19
–
–
580 ± 71
110 ± 9
–
–
–
–
–
130 ± 15
390 ± 5
–
480 ± 7
950 ± 232
–
–
1950 ± 268
350 ± 27
–
–
–
–
–
10 ± 8
119 ± 6
–
440 ± 153
1740 ± 198
4 ± 1
1430 ± 394
1580 ± 438
–
b
I3C, 3
a,b
4
b,c
DIM, 6
280 ± 5
1220 ± 82
250 ± 37
600 ± 280
140 ± 13
1760 ± 10
1240 ± 172
50 ± 8
230 ± 98
4570 ± 767
330 ± 62
>10,000
650 ± 79
>10,000
1080 ± 139
50 ± 17
b
8
9
10
11
12
13
14
c
>10,000
c
440 ± 60
5820 ± 260
3990 ± 526
270 ± 35
c
1
1
1
0
1
2
c
c
a
b
c
EC50 not reached.
Commercially available from Sigma-Aldrich (UK).
U87 and IN1472 not tested. Errors are standard deviations of experiments run at least in triplicate.