662
S. V. Miert et al. / Bioorg. Med. Chem. 13 (2005) 661–669
display cytotoxic (against human tumour cell lines) and
antitubulin properties.6,7 Obviously different structure–
activity relationships will apply for interactions with
human and protozoal tubulin, and hence also for cyto-
toxicity on human cells and protozoa. Therefore it
appeared appropriate to evaluate and to compare the
antiprotozoal (antiplasmodial, antitrypanosomal and
antileishmanial) and cytotoxic properties of a series of
known and newly synthesised dihydrobenzofuran lign-
ans and benzofuran derivatives, and to apply quantita-
tive structure–activity relationship (QSAR) models
based on Quasar receptor surface modelling. These
investigations have led to the characterisation of the
dihydrobenzofuran derivative 2g as a promising anti-
leishmanial lead compound.
bulky esters lost most of there in vitro cytotoxicity:
Against the three breast cancer cell lines mentioned
above, GI50 values in the micromolar instead of the
nanomolar range were observed, and these esters were
not selected for further evaluation. Similarly, none of
the newly synthesised dihydrobenzofuran or benzofuran
derivatives described in the present work exhibited a
promising cytotoxicity in the human tumour cell line pa-
nel, as evidenced by some selected characteristic com-
pounds (Table 1).
2.2. Antiprotozoal activity
Because of the antileishmanial activity reported before
for some lignans containing two variable aromatic do-
mains (see above), a series of our dihydrobenzofuran
lignans and related benzofuran derivatives was evalu-
ated against the protozoa Leishmania donovani (axenic
amastigotes), chloroquine resistant P. falciparum (strain
K1), T.b. rhodesiense and T. cruzi, and for cytotoxicity
on L6 cells. Results are shown in Table 2 except for T.
cruzi for which no promising activities were observed.
Against P. falciparum, 2g showed an IC50 of 0.43lg/
mL, and 5j 1.06lg/mL. Most interesting activities, how-
ever, were observed against L. donovani. In a prescreen
against axenic amastigotes, three compounds showed
an IC50 around or below 0.1lg/mL (2f, 2g and 2h),
and their activity was confirmed in an antileishmanial
assay carried out in infected macrophages. In the latter
assay, 2g was the most active compound, showing an
IC50 of 0.19lg/mL. Remarkably, the lipophylic esters
2f, 2g and 2h were prepared as analogues of the methyl-
ester 2b, which was the most promising inhibitor of tub-
ulin polymerisation and the most cytotoxic compound
against human tumour cell lines, as demonstrated in
our previous investigations.6 However, the antileishma-
nial activity of 2b was almost two orders of magnitude
lower than observed for 2f or 2h. Apparently in the
dihydrobenzofuran series the introduction of larger ester
group decreases the cytotoxicity, while enhancing the
antileishmanial activity. Although also some benzofuran
derivatives such as 5j, 6d or 6j showed antileishmanial
IC50 values <1lg/mL, they did not show a pronounced
activity in the assay in infected macrophages, and the
dihydrobenzofuran derivatives were given priority for
further evaluation. Compound 2g was selected for in
vivo evaluation, although the development of more
active and less cytotoxic analogues may be necessary
before a clinically useful compound will be obtained.
2. Results and discussion
2.1. Cytotoxicity on human cancer cell lines
Evaluation of 2a–d, 3a–d, 4a–d, 5d and 6d in an in vitro
human disease oriented tumour cell line screening panel,
consisting of 60 human tumour cell lines, at the Na-
tional Cancer Institute (NCI, Bethesda, MD, USA)
has been published before.6 All other compounds re-
ported here were tested in the same assay (see Support-
ing information). More particularly the sterically
hindered esters 2f–i were prepared because the corre-
sponding methyl ester 2b was found to have a pro-
nounced cytotoxic activity, especially against the
breast cancer cell lines MB-435, MDA-N and BT-549
where GI50 values <10nM were observed (drug concen-
tration, which reduces cell growth to 50% of level ob-
tained with untreated cells). However, in the hollow
fibre assay for preliminary in vivo testing, it was not suf-
ficiently active for further in vivo evaluation.6 Because
hydrolysis of the methyl esters might explain the loss
of activity in vivo, the esters 2f–i, which are less sensitive
for hydrolysis than the methyl ester 2b, were prepared.
Unfortunately, as illustrated for 2i in Table 1, these
Table 1. Inhibition of in vitro tumour cell growth (GI50, lM)
2i
4e
5j
6j
a
Average GI50
HL-60 (TB)b
K-562
4.4
72.4
60.3
33.9
70.8
75.9
11.2
24.0
12.0
9.8
18.6
14.5
10.0
1.9
2.2
2.6
1.7
4.2
2.6
6.2
4.6
5.0
4.4
3.5
4.2
13.2
NCI-H522
HCT-15
SF-539
15.5
1.5
11.7
21.4
17.4
20.4
20.4
21.9
5.1
>100
97.7
2.3. QSAR analysis
M14
OVCAR-3
UO-31
14.8
17.4
15.1
9.1
>100
>100
>100
Since
a better quantitative understanding of the
structure–activity relationships underlying some of the
compoundsÕ selectivity with respect to cytotoxic or
antiprotozoal activity would be a prerequisite for a
rational application of structural changes in order to
arrive at more selective antiprotozoal drug leads, a 4D-
QSAR study using the concept of quasi-atomistic recep-
tor surface modelling (Quasar) was carried out. For a
detailed description of Quasar methodology see litera-
ture.8,9 Based on the assumption that the compounds
under study act specifically at a common target site
DU-145
MB-435
MDA-N
BT-549
24.0
31.6
>100
4.4
5.2
6.9
22.9
14.5
a Average logGI50 values calculated from all cell lines tested.
b HL-60 (TB), K-562, leukaemia cell lines; NCI-H522, non-small-cell
lung cancer cell line; HCT-15, colon cancer cell line; SF-539, CNS
cancer cell line; M14, melanoma cell line; OVCAR-3, ovarian cancer
cell line; UO-31, renal cancer cell line; DU-145, prostate cancer cell
line; MDA-MB-435, MDA-N, BT-549, breast cancer cell lines.
Miert, Sabine Van
