Further studies on anti-invasive chemotypes: An excursion from chalcones to curcuminoids

Article abstract of DOI:10.1016/j.bmcl.2015.01.027

In our ongoing search for new anti-invasive chemotypes, we have made an excursion from previously reported potent 1,3-diarylpropenones (chalcones) to congeners bearing longer linkers between the aromatic moieties. Nine 1,ω-diarylalkenones, including curcumin and bisdemethoxycurcumin, were evaluated in the chick heart invasion assay. Unfortunately, these compounds proved less potent and more toxic than earlier evaluated chemotypes. In the 1,3-diarylpenta-2,4-dien-1-one series, fluoro and/or trimethoxy substitution caused an increase in potency. This agrees with observations made earlier for the chalcone class.

Full text of DOI:10.1016/j.bmcl.2015.01.027

Bioorganic & Medicinal Chemistry Letters 25 (2015) 1027–1031
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Further studies on anti-invasive chemotypes: An excursion from
chalcones to curcuminoids
Bart I. Roman a, , Tine De Ryck b, Sigrid Verhasselt a, Marc E. Bracke b, Christian V. Stevens a,
⇑
a Research Group SynBioC, Department of Sustainable Organic Chemistry and Technology, Faculty of Bioscience Engineering, Coupure Links 653, B-9000 Ghent, Belgium
b Laboratory of Experimental Cancer Research, Department of Radiation Oncology and Experimental Cancer Research, Ghent University, De Pintelaan 185, B-9000 Ghent, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
In our ongoing search for new anti-invasive chemotypes, we have made an excursion from previously
reported potent 1,3-diarylpropenones (chalcones) to congeners bearing longer linkers between the
aromatic moieties. Nine 1,x-diarylalkenones, including curcumin and bisdemethoxycurcumin, were
evaluated in the chick heart invasion assay. Unfortunately, these compounds proved less potent and more
toxic than earlier evaluated chemotypes. In the 1,3-diarylpenta-2,4-dien-1-one series, fluoro and/or
trimethoxy substitution caused an increase in potency. This agrees with observations made earlier for
the chalcone class.
Received 23 December 2014
Revised 13 January 2015
Accepted 15 January 2015
Available online 22 January 2015
Keywords:
Invasion
Cancer
Ó 2015 Elsevier Ltd. All rights reserved.
SAR
Chalcone
Curcuminoids
The hallmark that differentiates malignant neoplasms from
benign tumors is the ability of a subpopulation of cells to invade
neighboring tissues and cause metastases.1 These phenomena
account for 90% of all cancer deaths and currently present the most
significant hurdles in the management of malignant disease.2,3
Over the last years, our laboratories have been searching for
novel chemotypes that can suppress the invasive phenotype of
malignant cells (Chart 1).4 Starting from natural 1,3-diphenyl-
prop-2-en-1-ones,5 the main hits in our Indo-Belgian anti-invasive
screening program,6 we have developed synthetic analogues
belonging to the diarylchalcone (1b),7 stilbene8 (2,3) and diarylis-
oxazole/pyrazole (4,5) classes,9 and reported on the structure–
activity relationship of these compounds. These studies were
conducted using the phenotypic chick heart invasion (CHI) assay
as an in vitro model of invasion. The most promising molecule
identified yet is chalcone 1c, which displays strong anti-invasive
hit compounds
previous SAR work
O
O
O
chalcones
chalcones
(70 examples)
MeO
MeO
Ar1
Ar2
1a
1b
1c
F
active at
0.01 µmol.L-1
Ar= substituted phenyl,
MeO
furyl, indolyl
stilbenes
isoxazolesand pyrazoles
(10 examples)
(4 examples)
N
X
N
O
2
1
2
R
R
R
1
1
1
2
R
R
R
R
2
2
3
4
5
R
X= O, NH
present report
1,ω-diarylalkenones
curcumins
(7 examples)
(2 examples)
O
OH
O
MeO
OMe
OH
1
2
i
j
R
R
6
7
efficacy at 0.01
l
molꢀLꢁ1 in the CHI and comparable behavior in
HO
i,j= 0, 1, 2
the matrigel invasion assay.
The present work intends to complement the above SAR studies
by evaluating longer unsaturated spacers between the two
Chart 1. Overview of previously and presently investigated chemotypes in search
of anti-invasive lead molecules.
aromatic moieties. Hence, we have prepared and evaluated an
array of 1,x-diarylalkenones 6 (Chart 1). For completeness, also
curcumin and bisdemethoxycurcumin (7) were included in this
study.
With respect to the latter compounds, ambiguous health claims
have been put forward in both scientific and alternative literature.
⇑
Corresponding author. Tel.: +32 9 264 5957; fax: +32 9 264 6221.
E-mail addresses: Bart1.Roman@UGent.be (B.I. Roman), Tine.DeRyck@UGent.be
(T. De Ryck), Marc1.Bracke@UGent.be (M.E. Bracke), Chris.Stevens@UGent.be
(C.V. Stevens).
URL: http://www.ugent.be/bw/doct/en (C.V. Stevens).
B.I.R. is
a postdoctoral fellow of the Research Foundation—Flanders (FWO—
Vlaanderen).
http://dx.doi.org/10.1016/j.bmcl.2015.01.027
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

1028
B. I. Roman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 1027–1031
O
O
However, evidence suggests that, far from being a panaceum, cur-
1
1
2
cumin possesses liabilities such as the induction of herb–drug
interactions.10,11 A more critical attitude with respect to this puta-
tive drug is desirable, since a clear view on the risk/benefit-ratio of
curcumin and its synthetic analogues12 is yet to be established. In
the light of the above, it should be stressed that focus in the pres-
ent research was not on the supposed merits of curcuminoids as
such, but on a better grasping of the SAR of anti-invasive chalcones
through evaluation of their curcuminoid congeners.
R
R
R
8
9-16
Scheme 1. Reagents and conditions: (1) 20 mol % LiOHꢀH2O, abs EtOH, rt or 40 °C,
10 min; (2) 1 equiv cinnamaldehyde, rt or 40 °C, 72 h, see Table 1.
Table 1
Our preparation of a curcuminoid library started with the
selection and synthesis of 1,5-diarylpenta-2,4-dien-1-ones and
1,5-diarylpenta-1,4-dien-3-ones. Since the physicochemical param-
eters influencing distribution (e.g., cLogP and PSA) are very similar
for identically substituted C6–C5–C6 alkenones and chalcones, a dif-
ference in activity in the CHI assay can be attributed mainly to
altered target affinity or stability. Comparison between the two
selected C6–C5–C6 frameworks would also provide information on
the influence of the location of the carbonyl group on activity.
An array of 1,5-diarylpenta-2,4-dien-1-ones 9–16 was synthe-
sized in good yields out of suitably decorated acetophenones and
cinnamaldehydes via a LiOH-catalyzed Claisen–Schmidt condensa-
tion (Scheme 1, Table 1).7,13 This is the first dedicated report on the
use of LiOHꢀH2O for the preparation of 1,5-diarylpenta-2,4-dien-1-
ones. Only one related report has been published, in which com-
pound 9 was obtained as a side product in the asymmetric
LiOHꢀH2O-catalyzed Michael addition of acetophenone to cinna-
maldehyde, conducted in the presence of the trimethylsilyl ether
of (R)-diphenylprolinol.14
In an identical way, seven diversely substituted 1,5-diarylpen-
ta-1,4-dien-3-ones 19–25 were prepared from 4-phenylbut-3-en-
2-one 17 and a suitably decorated benzaldehyde 18. All substances
were obtained in very good yield and high purity (Scheme 2,
Table 2).
In an excursion to C6–C7–C6 curcuminoids, 1,7-diarylhepta-
1,4,6-trien-3-one 26 was prepared (Scheme 2), so as to evaluate
the behavior of molecules with a 7-atom linker. For completeness,
curcumin 27 and bisdemethoxycurcumin 28 were also included in
the library.
Preparation of 1,5-diarylpenta-2,4-dien-1-ones 9–16 (see also Scheme 1)
R1
R2
Temp (°C)
Yielda (%)
9
H
H
H
H
H
H
H
H
F
rt
rt
rt
rt
rt
rt
40
40
89
67
70
64
67
90
65
37
10
11
12
13
14
15
16
40-Me
40-OMe
40-Cl
40-F
40-NO2
30,40,50-OMe
30,40,50-OMe
a
Determined upon recrystallization in absolute EtOH.
O
O
O
a
H
1
1
R
R
17
19-25
18
O
b
26 (40%)
Scheme 2. Reagents and conditions: (a) (1) 20 mol % LiOHꢀH2O, abs EtOH, rt,
10 min; (2) 1 equiv substituted benzaldehyde 18, rt, 72 h, see Table 2; (b) (1)
20 mol % LiOHꢀH2O, abs EtOH, rt, 10 min; (2) 1 equiv cinnamaldehyde,
D, 2 h.
Table 2
Preparation of 1,5-diarylpenta-1,4-dien-3-ones 19–25 (see also Scheme 2)
Having an array of 1,x-diarylalkenones in hands, we proceeded
R1
Yielda (%)
with the evaluation of their in vitro anti-invasive activity. An opti-
mal in vitro screening strategy for anti-invasive molecules has to
discern:
19
20
21
22
23
24
25
H
77
68
88
87
89
90
87
4-Me
4-OMe
4-Cl
(i) Anti-invasive effects from mere toxicity. We ultimately seek
anti-invasive agents with in vivo efficacy. The main task of
such molecules is to suppress invasion, not primary tumor
growth. Hence, these molecules need not per se be cytotoxic
towards neoplastic cells.
4-F
4-NO2
2,4,6-OMe
a
Determined upon recrystallization in absolute EtOH.
Chart 2. This illustration of the chick heart invasion assay displays 7 lm thick sections from confronting cultures between invasive human mammary carcinoma cells (MCF-
7/6) and precultured heart fragments (PHF). Panel A: scoring criteria. Panel B: from left to right: (i) results from an untreated culture, in which the confronting cells have
occupied more than half or the PHF, invasion grade IV: (ii) addition of an anti-invasive agent resulted in a clear histological separation between the confronting partners; the
invasion grade is denoted as I; (iii) outcome for a product with selective toxicity for MCF-7/6 cells with respect to the PHF. No judgment of the anti-invasive potency can be
made; this pattern is coined grade 0.

B. I. Roman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 1027–1031
1029
Table 3
In vitro anti-invasive activity data against MCF7/6 cells, obtained for nine 1,
invasion, the compound is inactive)
x-diarylalkenones in the CHI-assay (I and II = no invasion, the compound is coined active; III and IV:
Structure
Invasion grade at different concentrations (
l
molꢀLꢁ1
)
Anti-invasive activity classa
100
n.d.
10
1
0.1
0.01
n.d.
O
1a
1c
II
III
n.d.
I/II
IV
1g
O
MeO
MeO
0b/Ie
II/I
II/II–III/I
ꢁ2g
F
n.d.
OMe
O
9
T/IIb
III–IVh/III
III
n.d.
n.d.
2
3
O
12
III/II
III
II/III
III/IV
III
Cl
F
O
13
15
Id
T
II
III
III
0
1
O
MeO
T/IIb
Id/IId
III/III/IV
III
MeO
OMe
O
MeO
MeO
II/IIh
III/IIIf
III/III
n.d.
1
F
16
n.d.
OMe
O
19
26
T/Id
II/Ie
I/IIc
II
II
III
IV
IV/III
IV/III
0
0
O
IId/IIc
O
O
O
O
IV/IIf
n.d.
n.d.
1
1
HO
HO
OH
27
28
n.d.
n.d.
II
III/IV
III
OMe
OMe
IId/IId
III
OH
In all runs, solvent (negative) control gave grade III/IV. All results were obtained at least in duplo; for identical outcomes, only one number is shown.
T: toxic for PHF and MCF-7/6, to a degree that renders judgment impossible; n.d.: not determined.
Scores at the lowest active concentration are highlighted in bold.
a
Logcmin
.
b
c
d
e
f
Toxic for MCF-7/6.
Toxic for PHF.
Toxic for PHF and MCF-7/6.
Very few MCF-7/6 cells left.
PHF dedifferentiated.
See Ref. 7.
g
h
See photograph in Chart 3.
(ii) Effects on tumor cells from effects on stroma. The inherent
complexity of tumor spread mandates that the involvement
of the host tissue in the micro-ecosystem that governs
tumor behavior should not be overseen.4,6
properties, and was accordingly used to evaluate the anti-invasive
potency of our diarylalkenones. The behavior of numerous antipro-
liferative agents (e.g., vinblastine, vincristine, taxol, nocodazole,
podophyllotoxin)16 and of Hsp90 inhibitor AUY92217 in this assay
is well-documented.
In such a case, a phenotypic, tissue-based assay constitutes the
In the model, precultured heart tissue fragments (PHFs), dis-
sected from 9-day-old chick embryos, are confronted with aggre-
gates of human invasive MCF-7/6 mammary carcinoma cells in
the presence of a certain concentration of a test substance. After
most informatory test setup. The chick heart invasion (CHI)
model,15 developed at the Laboratory of Experimental Cancer
Research (Ghent University, Belgium), possesses all of these

1030
B. I. Roman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 1027–1031
eight days, the interaction between the partners is evaluated histo-
logically and classified along a 5-grades scale (Chart 2).
Compounds are designated ‘active’ or ‘anti-invasive’ when the
invasion grade is I or II, and ‘inactive’ when the grade is III or IV.
For substances that exhibit selective toxicity versus MCF-7/6 cells,
a judgment of the anti-invasive potency cannot be made (invasion
grade 0).
wealth of reports on the anti-invasive activity of these compounds.
The efficacy documented in literature is mostly observed at high
concentrations. For example, in one of the most relevant recent
reports, using MCF-7 cells in the matrigel assay, curcumin was
used at a concentration of 30
l
molꢀLꢁ1 18
.
We also observe efficacy
at 10
l
molꢀLꢁ1 for curcumin, but not at lower concentrations.
Therefore, we must conclude that the potential of curcumin is infe-
rior to that of some of our earlier investigated compound classes
(i.e., chalcones and stilbenes).
Because the CHI assay is not high throughput, nine 1,x-diary-
lalkenones were chosen for evaluation. A test set was compiled
that would enable an assessment of the influence of the different
linkers (keeping the substitution pattern unchanged), as well as
the role of the aromatic substituents (keeping the linker
unchanged). For these nine compounds, the invasion grade in the
CHI assay was determined at concentrations ranging from
Overall, the evaluated 1,x-diarylalkenones also seem to pos-
sess a low safety margin: the majority of the evaluated mole-
cules display cytotoxic effects against the heart tissue at the
lowest anti-invasive concentration or one decade above. In all,
the most attractive profile is exhibited by compound 19: this
molecule is toxic for both PHF and MCF-7/6 cells at
100 down to 0.01
into one score per compound (Eq. 1): ‘Logcmin’, being the decadic
logarithm of the lowest concentration (expressed in
which a substance exhibits an anti-invasive behavior (invasion
l
molꢀLꢁ1 (Table 3). The assay data was translated
100
l
molꢀLꢁ1, selectively toxic for the malignant cells at a ten-
l
molꢀLꢁ1) at
fold lower concentration, and still mediating anti-invasive effects
another decade lower.
grades I or II). Six activity levels exist, ranging from class ꢁ2 for
Although no exceptional anti-invasive behavior was observed
for the assessed molecules, the obtained results bear value from
a SAR point of view. The main goal of this investigation was to gain
insight into the influence of the type of spacer between the aro-
matic moieties on activity. Such an evaluation is best conducted
for the unsubstituted scaffolds (series 1, Chart 4) and for the ana-
logs of the most promising molecule in our anti-invasive library,
4-fluoro-30,40,50-trimethoxychalcone 1c (series 2). An overview of
the potency and a number of relevant properties of these two ser-
ies is presented in Chart 4.
the most active compounds (invasion grade
I
or II at the
0.01
l
molꢀLꢁ1 level) to class 3 (compounds with no apparent effect
at a concentration of 100
l
molꢀLꢁ1):
Anti-invasive activity class ¼ log cmin
ð1Þ
From the results displayed in Table 3, it is clear that alkenones
9, 12, 15, 16, 27 and 28 display a very low anti-invasive potency.
Molecules 13, 19 and 26 proved active down to 1
l
molꢀLꢁ1. Repre-
sentative images of these results are presented in Chart 3.
The disappointing results obtained for curcumin (27) and bisde-
methoxycurcumin (28) in the CHI model do not contradict the
Judged by activity class alone (Logcmin), our data indicate that,
within both series, linker modifications are not well-tolerated.
Chart 3. Representative images of notable results from the CHI assay (see also Table 3).
C6-C2-C6
C6-C3-C6
C6-C5-C6
C6-C7-C6
stilbenes
chalcones
1,ω-diarylalkenones
curcuminoids
O
O
O
O
s
e
r
1
2
i
29
1a
9
Log cmin= 2
cLogP= 4.1
19
26
e
s
Log cmin= -1
Log cmin= 1
Log cmin= 0
Log cmin= 0
cLogP= 4.54
cLogP= 4.22
cLogP= 3.59
cLogP= 4.02
Activity versus toxicity: legend
OMe
MeO
OMe
O
O
CHI result at indicated concentrations:
MeO
MeO
s
e
r
i
MeO
MeO
MeO
toxic for PHF
non-toxic -1
100 10
1
0.1 0.01 µmol.L
MeO
OMe
F
MeO
F
not assessed
F
F
OMe
30
31
1c
16
e
s
toxic for MCF-7/6, not for PHF
Log cmin= 0
Log cmin= 1
Log cmin= -2
Log cmin= 1
cLogP= 4.0
cLogP= 4.0
cLogP= 3.36
cLogP= 3.88
lowest active concentration
Chart 4. Comparison of the potency and toxicity of homologous stilbenes, chalcones and 1,
x
-diarylalkenones in the CHI assay.

B. I. Roman et al. / Bioorg. Med. Chem. Lett. 25 (2015) 1027–1031
1031
Besides, potency comparison across the two series within a given
Acknowledgements
compound class (stilbene, chalcone, curcuminoid) shows that opti-
mal linker size and geometry are dependent on the decoration pat-
tern of the aromatic moieties. From a potency point of view,
optimization of the aromatic substitution pattern and the linker
should thus not be considered separate exercises.
The authors thank the Research Foundation—Flanders (Fonds
voor Wetenschappelijk Onderzoek—Vlaanderen) for financial sup-
port for this research, and would like to express their sincere
appreciation to Marleen De Meulemeester for meticulously con-
ducting the CHI-assay on all compounds investigated in this study.
When taking into account other properties (lipophilic efficiency,
toxicity), it is evident that the curcuminoid class performs poorer
than the stilbene and chalcone classes. Overall, chalcone 1c
remains the molecule with the most attractive profile: reasonable
LogP, high anti-invasive potency but no cytotoxicity at nanomolar
concentrations, and selective toxicity towards MCF-7/6 cells at
Supplementary data
Supplementary data (experimental details, compound charac-
terization and 1H, 13C and 19F NMR spectra of selected compounds)
associated with this article can be found, in the online version, at
http://dx.doi.org/10.1016/j.bmcl.2015.01.027.
10
l
molꢀLꢁ1
.
For completeness, a short analysis can be made of the influence
of the substitution pattern of 1,5-diarylpenta-2,4-dien-1-ones on
activity and toxicity. Compared to the unsubstituted scaffold (9),
introduction of a chlorine atom at the 40-position (12) lowers tox-
icity but does not strongly influence activity. Similar to what has
been observed for the chalcone class, trimethoxy- and/or fluoro
substitution is attractive from a potency point of view. However,
the gain in potency observed for the 30,40,50-trimethoxy pattern
(15 and 16) comes with an increase in undesired effects (toxicity,
dedifferentiation). Similarly, while the presence of a fluorine atom
at the 40-position in curcuminoid 13 has a positive effect on activ-
References and notes
1. Hanahan, D.; Weinberg, R. A. Cell 2011, 144, 646.
2. Weigelt, B.; Peterse, J. L.; Van’t Veer, L. J. Nat. Rev. Cancer 2005, 5, 591.
3. Mehlen, P.; Puisieux, A. Nat. Rev. Cancer 2006, 6, 449.
4. Vanhoecke, B. W.; Depypere, H. T.; De Beyter, A.; Sharma, S. K.; Parmar, V. S.; De
Keukeleire, D.; Bracke, M. E. Pure Appl. Chem. 2005, 77, 65.
5. Roman, B. I.; Heugebaert, T. S. A.; Bracke, M. E.; Stevens, C. V. Curr. Med. Chem.
2013, 20, 186.
6. Bracke, M. E.; Vanhoecke, B. W. A.; Derycke, L.; Bolca, S.; Possemiers, S.;
Heyerick, A.; Stevens, C. V.; De Keukeleire, D.; Depypere, H. T.; Verstraete, W.;
Williams, C. A.; McKenna, S. T.; Tomar, S.; Sharma, D.; Prasad, A. K.; DePass, A.
L.; Parmar, V. S. Anticancer Agents Med. Chem. 2008, 8, 171.
7. Bracke, M.; Stevens, C.; De Ryck, T.; Roman, B.; Vanhoecke, B. WO2013/113722
A1, 2013.
8. Roman, B. I.; De Coen, L. M.; Mortier, S. T. F. C.; De Ryck, T.; Vanhoecke, B. W. A.;
Katritzky, A. R.; Bracke, M. E.; Stevens, C. V. Bioorg. Med. Chem. 2013, 21, 5054.
9. Roman, B. I.; De Ryck, T.; Dierickx, L.; Vanhoecke, B. W. A.; Katritzky, A. R.;
Bracke, M. E.; Stevens, C. V. Bioorg. Med. Chem. 2012, 20, 4812.
10. Esatbeyoglu, T.; Huebbe, P.; Ernst, I. M. A.; Chin, D.; Wagner, A. E.; Rimbach, G.
Angew. Chem., Int. Ed. 2012, 51, 5308.
11. Mancuso, C.; Barone, E. Trends Pharmacol. Sci. 2009, 30, 333.
12. Roman, B. I.; Monbaliu, J.-C.; De Coen, L. M.; Verhasselt, S.; Schuddinck, B.; Van
Hoeylandt, E.; Stevens, C. V. Eur. J. Org. Chem. 2014, 2594.
13. Bhagat, S.; Sharma, R.; Sawant, D. M.; Sharma, L.; Chakraborti, A. K. J. Mol. Catal.
A 2006, 244, 20.
14. Li, W.; Wu, W.; Yang, J.; Liang, X.; Ye, J. Synthesis 2011, 1085.
15. Bracke, M. E.; Parmar, V. S.; Depass, A. L.; Stevens, C. V.; Vanhoecke, B. W.;
Mareel, M. M. Methods Mol. Biol. 2014, 1070, 93.
16. Mareel, M. M.; De Mets, M. Int. Rev. Cytol. 1984, 90, 125.
17. Jensen, M. R.; Schoepfer, J.; Radimerski, T.; Massey, A.; Guy, C. T.; Brueggen, J.;
Quadt, C.; Buckler, A.; Cozens, R.; Drysdale, M. J.; Garcia-Echeverria, C.; Chène,
P. Breast Cancer Res. 2008, 10, R33.
18. Kim, J. M.; Noh, E. M.; Kwon, K. B.; Kim, J. S.; You, Y. O.; Hwang, J. K.; Hwang, B.
M.; Kim, B. S.; Lee, S. H.; Lee, S. J.; Jung, S. H.; Youn, H. J.; Lee, Y. R. Phytomedicine
2012, 19, 1085.
ity, resulting in anti-invasive efficacy down to 1
l
molꢀLꢁ1 (class 0),
this molecule is particularly toxic at 10
l
molꢀLꢁ1. Remarkably,
none of our earlier evaluated fluoro chalcones proved toxic for
the PHF.7 Taken together, the aromatic substitution pattern of
the 1,5-diarylpenta-2,4-dien-1-ones has a strong influence on both
toxicity and potency.
In conclusion, the evaluated 1,x-diarylalkenones, including cur-
cumin and bisdemethoxycurcumin, lack nanomolar potency and
display higher toxicity than their chalcone and stilbene analogs.
Still, 1,5-diphenylpenta-1,4-dien-3-one 19 exhibits an interesting
cytotoxicity profile. Analogous to what was observed earlier for
the chalcone class, trimethoxy and/or fluoro substitution yields
more potent compounds (13, 15 and 16). However, in the case of
the diarylpentenones and -heptenones, this modification entails
undesired effects (toxicity, dedifferentiation).
Overall, compared to the curcuminoid and stilbene series, the
chalcone class remains the most interesting source of anti-invasive
compounds in this research program, with 4-fluoro-30,40,50-tri-
methoxychalcone as the exponent.