Strategies for improving the solubility and metabolic stability of griseofulvin analogues

Article abstract of DOI:10.1016/j.ejmech.2016.03.071

We report two types of modifications to the natural product griseofulvin as strategies to improve solubility and metabolic stability: the conversion of aryl methyl ethers into aryl difluoromethyl ethers at metabolic hotspots and the conversion of the C-ri

Full text of DOI:10.1016/j.ejmech.2016.03.071

European Journal of Medicinal Chemistry 116 (2016) 210e215
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Strategies for improving the solubility and metabolic stability of
griseofulvin analogues
a
b
b
a
b
c
€
A.B. Petersen , G. Konotop , N.H.M. Hanafiah , P. Hammershøj , M.S. Raab , A. Kramer ,
M.H. Clausen a
,
*
a Center for Nanomedicine and Theranostics, Department of Chemistry, Technical University of Denmark, Kemitorvet 207, DK-2800, Kgs. Lyngby, Denmark
b Max-Eder Group Experimental Therapies for Hematologic Malignancies, German Cancer Research Center (DKFZ), Germany
c Clinical Cooperation Unit Molecular Hematology/Oncology, German Cancer Research Center (DKFZ) and Dept. of Internal Medicine V, University of
Heidelberg, Heidelberg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
We report two types of modifications to the natural product griseofulvin as strategies to improve sol-
ubility and metabolic stability: the conversion of aryl methyl ethers into aryl difluoromethyl ethers at
metabolic hotspots and the conversion of the C-ring ketone into polar oximes. The syntheses of the
analogues are described together with their solubility, metabolic half-life in vitro and antiproliferative
effect in two cancer cell lines. We conclude that on balance, the formation of polar oximes is the most
promising strategy for improving the properties of the analogues.
Received 18 January 2016
Received in revised form
15 March 2016
Accepted 25 March 2016
Available online 28 March 2016
© 2016 Elsevier Masson SAS. All rights reserved.
Keywords:
Griseofulvin analogues
Anticancer
Synthesis
Antifungal
Metabolic stability
1. Introduction
supernumerary centrosomes to avoid lethal multipolar cell di-
visions [8,9]. A subsequent anticancer SAR study of 34 compounds
Griseofulvin 1 was one of the first antifungal natural products
found and isolated in filamentous fungi [1]. It has since its discovery
in 1939 attracted a lot of attention due to its bioactivity. The first
report of griseofulvin's potential for the treatment of ringworm in
man was published in 1958 and griseofulvin was launched as a
human and veterinary drug for this indication under the trade
names Fulcin and Grisovin [2,3]. Since griseofulvin's discovery
more than 400 analogues have been prepared and reported in the
literature [4]. In 2001 it was demonstrated that griseofulvin could
potentiate the antitumorigenic effect of the drug Nocodazole by
inducing apoptosis in several cancer cell lines at concentrations
identified the 20-benzyl derivative 2 (GF-15) as one of the most
potent analogues [10]. A more detailed study was performed on 2
including in vivo studies in murine xenograft models of colon
cancer and multiple myeloma [11]. Here we report the in vitro
metabolic stability of 1 and 2 as well as the synthesis of two ana-
logues with a group known to block metabolism at hotspots for CYP
oxidation of griseofulvin: the 4 and the 6 positions (see Fig. 1)
[12e14].
1.1. Improving metabolic stability of griseofulvin analogues
down to 1 mM [5]. Griseofulvin has been shown to interfere with
Different major metabolites of 1 have been identified from
published in vivo studies. The major metabolites in rat, mouse, dog
and man were found to be 4-desmethyl griseofulvin (5) and 6-
desmethyl griseofulvin [15e18]. Both metabolites have previously
been shown to have no (5) or low (6) antiproliferative activity
against the cancer cell line MDA-MB-231 [19]. In order to increase
the metabolic half-life of griseofulvin analogues such as 2, we
wanted to block CYP oxidation at the two hotspots at position 4 and
6, respectively. The two aryl methyl ethers 3 and 4 were therefore
tubulin polymerization and microtubule dynamics [6,7]. In 2006, it
was shown that structural modification of griseofulvin could
improve cytotoxicity against cancer cells and in 2007 it was re-
ported that spirobenzofuranones can inhibit centrosomal clus-
tering
e a strategy of many cancer cell lines harbouring
* Corresponding author.
E-mail address: mhc@kemi.dtu.dk (M.H. Clausen).
http://dx.doi.org/10.1016/j.ejmech.2016.03.071
0223-5234/© 2016 Elsevier Masson SAS. All rights reserved.

A.B. Petersen et al. / European Journal of Medicinal Chemistry 116 (2016) 210e215
211
Analogue 16 is reported here for the first time and was designed to
increase the solubility in phosphate buffered saline (PBS) further.
Both 15 and 16 are more polar compounds than 2, so they were
expected to have lower oxidative metabolism.
2.2. Thermodynamic solubility
The thermodynamic solubility of 1e4,15 and 16 was determined
by HPLC: a reference absorption (peak area at 254 nm) linear plot
was made from a dilution series of the compounds in acetonitrile.
Saturated PBS solutions of the compounds were prepared by son-
ication and shaking and the HPLC chromatogram peak areas were
used to calculate the respective solubility from the reference
graphs.
Griseofulvin (1) is known to have low solubility (Table 1, entry 1)
and compound 2 is highly lipophilic and has lower solubility in PBS
(entry 2). As expected, modification by introducing fluorine sub-
stitution only aggravates this property (entries 3 and 4). However,
this can be improved either somewhat through the formation of a
simple oxime (entry 5) or quite dramatically by formation of the
charged oxime 16 (entry 6), where the PBS solubility is higher than
1 mg/mL. We have previously shown that oximes of griseofulvin are
hydrolysed only very slowly and as such oxime formation presents
a possible solution to issues of low solubility, provided that the
potency is not negated [19].
Fig. 1. Chemical structure of griseofulvin 1 with the IUPAC recommended numbering
and ring designation and structure of 20-benzyl modified lead compound 2. Further-
more, structures of 4-difluoro (3) and 6-difluoro (4) modified analogues of 2, proposed
to improve metabolic stability.
chosen for studying the effect of metabolic blockers. We arrived at
these specific modifications, because very few examples of CF3þ
trifluoromethylation of phenols via SN1 or SN2 mechanisms have
been reported and the potential reactivity of monofluoromethyl
derivatives towards strong nucleophiles led us to choose difluor-
omethyl ethers as metabolic blockers [20e23]. Difluoromethylation
of phenols have been reported in the literature using methyl
chlorodifluoroacetate and a suitable base such as caesium or po-
tassium carbonate [24,25].
2.3. Microsomal stability
Male mouse and male rat microsomes were purchased from a
major supplier and NADPH was used as a cofactor. The experiments
for each compound were run in triplicates, each with a control
where no cofactor was added. At the indicated times samples were
taken and the amount of non-degraded compound was quantified
by analytical HPLC (peak area). Half-life was calculated from linear
2. Results and discussion
2.1. Synthesis
For the syntheses of the 4- and 6-difluoro modified griseofulvins
(3 and 4), our starting point was a strategy first described by Arkley
et al. and Stephenson et al (see Scheme 1) [26,27]. For the position 4
modified analogue, griseofulvin was demethylated with magne-
sium iodide resulting in 5 followed by difluoromethylation with
methyl chlorodifluoroacetate and base yielding 6. Potassium car-
bonate performed best in this reaction and was therefore used
throughout. Hydrolysis of 6 gave the griseofulvic acid derivative 7,
which was chlorinated with phosphoryl chloride and lithium
chloride to give 8 (with the 40-chloro derivative as the major by-
product, not shown). The 20-chloro derivative 8 was subjected to
benzyl alcohol and DBU, which after 1,4-addition/elimination
resulted in the desired 4-difluoro modified compound 3 [19]. A
slightly different strategy was required for the 6-difluoro modified
analogue since no efficient, scalable method for direct position 6
demethylation has been reported in the literature [28]. We slightly
modified the route reported by Arkley et al. and hydrolysed gris-
eofulvin to griseofulvic acid (9), which was demethylated with
dilute sodium hydroxide yielding 10 and converted into the iso-
griseofulvin derivative 11 by acid catalysed 40-methylation with
methanol. The hereby protected 11 was difluoromethylated with
methyl chlorodifluoroacetate and base resulting in 12, which in a
sequence similar to the 4-difluoro derivative 6 was hydrolysed, 20-
chlorinated and 20-benzylated to give the desired difluoro analogue
4.
plots (see Supporting Information) and t times are given in Table 2.
½
To verify that no 20-CH3 cleavage was performed by the mouse
and rat microsomes a control experiment was performed. Deuter-
ated griseofulvin 18 was prepared as shown in Scheme 2 and
treated with mouse and rat microsomes respectively, and only
position 4/6 CH3 cleavage was observed by LC-MS.
As evident from Table 2, the introduction of metabolic blockers
at position 4 or 6 had either a negative influence on half-life (3,
entry 3) or no noticeable effect (4, entry 4). The simple oxime 15
had a microsomal profile similar to the parent analogue (entry 5),
while the more hydrophilic derivative 16 displayed an approxi-
mately 10-fold increase in half-life for both species (entry 6).
2.4. Anticancer potency
The small molecule griseofulvin has a demonstrated in vitro
activity against cancer cell lines with IC50 values ranging from
35 mM in SCC114 cells to 75 mM in HeLa cells but only little effect on
non-cancerous cell lines at these concentrations [9]. Further
chemical optimization of griseofulvin led to the isolation of ho-
mologues with much lower mean inhibitory concentrations for
proliferation and survival towards cancer cells. One of the most
potent analogues was compound 2, with more than 10-fold lower
IC50 values than griseofulvin in SCC114 cells. Additional in vitro
studies revealed that 2 was 10e30-fold more selective towards
malignant cells, including multiple myeloma, leukaemia and a va-
riety of solid tumours, when compared to healthy cells [12]. We
asked how the modifications aimed at increasing either the sta-
bility towards CYPs metabolism (3 and 4) or solubility in PBS (15
and 16) would affect the antitumour activity of these analogues in
comparison to griseofulvin (1) and compound 2. To this end, we
Two assays were used in order to investigate the compounds: 1)
IC50 against two established cancer cell lines and 2) microsomal
stabilities against male mouse and rat microsomes. In order to
obtain more SAR information, compounds 15 and 16 were included
in the investigation (Fig. 2). Compound 15 has been reported pre-
viously [19] and shown to have similar cytotoxicity as compound 2.

212
A.B. Petersen et al. / European Journal of Medicinal Chemistry 116 (2016) 210e215
determined the antiproliferative activity towards the human
epithelial breast cancer cell line MDA-MB231 and the human os-
teosarcoma U2OS cell line. The IC50 values (viable cells, luminescent
assay for ATP, see Supporting Information) are reported in Table 3.
The known compounds 2 and 15 are most potent against both
cell lines and, as previously demonstrated, both are significantly
more potent than the parent griseofulvin (1) (Table 3, entries 1, 2
and 5) [9,10,19]. Introduction of fluorine at the 4-position (entry 3)
leads to a 10-fold reduction of cytotoxicity, suggesting a significant
effect of the two fluorine atoms. For the 6-position, the reduction of
potency is less dramatic (entry 4). We have previously shown that
both demethylation and extension at positions 4 and 6 leads to
reduced anti-cancer activity for griseofulvin analogues [19]. The
fact that even the modest structural changes described here affects
potency, indicates a rather selective interaction between the target
and the methoxy-substituted arene in griseofulvins. The difluor-
omethyl group is a known hydrogen bond donor and alcohol iso-
stere, however, the HF2C hydrogen bond is considered very weak
(1.5 kcal/mol) [29e31]. Nonetheless, for the 4-position, the neigh-
bouring B-ring ketone could interact with fluorine in compound 3,
which could change the conformation of the substituent. For
analogue 4, it is harder to explain the decrease in potency, besides
from invoking the slightly larger size of the difluoromethyl group
when compared to the methyl substituent.
As previously observed, formation of the simple oxime (15,
entry 5) does not change potency [19]. However, when the size of
the 40 substituent increases, the cytotoxicity decreases (entry 6).
This could be due to an unfavourable steric or charge interaction
from the carboxymethyl group on the oxime. The highly increased
polarity of 16 compared to 2 or 15 could also lead to decreased
permeability and/or a lower affinity for the target.
3. Discussion
Metabolism of griseofulvin is known to go through position 4
and 6 demethylation. Our attempts to block metabolism at one site
appears to have simply re-routed the main metabolic reaction to
the non-blocked position. In order to increase metabolic half-life
using this strategy, it would appear that one would need to
modify both the 4- and the 6-position. However, given the lengthy
synthetic routes involved and the decreased anti-cancer activity
observed for 3 and 4 when compared to 2, this is not an attractive
strategy. A more viable solution comes from rendering the com-
pounds more polar, which both increases solubility (Table 1, entries
5 and 6) and half-life (Table 2, entry 6), with unchanged or only
moderately decreased potency (Table 3, entries 5 and 6).
4. Conclusions
Griseofulvin analogues have interesting properties related to
cancer biology and potential as small molecule therapeutics in
oncology. However, the physico-chemical properties of these
compounds are not germane to further drug development as they
suffer from low aqueous solubility and modest microsomal stabil-
ity. We have investigated two strategies aimed at ameliorating one
or both of these problems: the introduction of metabolic blockers at
two hotspots and the introduction of oximes to increase polarity.
Results show that between these two approaches, increasing po-
larity has the greatest potential for providing candidates for further
development. Future efforts should aim at improving potency
while retaining good solubility and long half-lives.
Scheme 1. Synthetic route for preparation of 4-difluoro (3) and 6-difluoro (4) modified
20-benzyl griseofulvin analogues from griseofulvin (1). (a) MgI2, Et2O, PhH, reflux; (b)
K2CO3, ClF2CCO2Me, DMF, 70 ꢀC; (c) H2SO4 (2 M), AcOH, 80 ꢀC; (d) LiCl, POCl3, 1,4-
dioxane, 85 ꢀC; (e) BnOH, DBU, 1,4-dioxane, 60 ꢀC; (f) H2SO4 (1 M), AcOH, 100 ꢀC;
(g) NaOH (0.7 M), reflux; (h) CSA (cat.), MeOH, reflux; (i) K2CO3, ClF2CCO2Me, DMF,
75 ꢀC.

A.B. Petersen et al. / European Journal of Medicinal Chemistry 116 (2016) 210e215
213
temperature under argon before use. All reagents were purchased
from major suppliers and used without further purification, unless
otherwise noted. All solvents were analytical grade and dried when
necessary over 4 Å activated molecular sieves and water content
was determined to be less than 20 ppm by Karl Fischer titration on
a Metrohm 899 e Coulometer. Thin Layer Chromatography (TLC)
was carried out on commercially available pre-coated plates (Merck
60, F254). Dry Column Vacuum Chromatography (DCVC) was per-
formed according to a published procedure [32]. When concen-
trating on Celite in vacuo for DCVC purification, high vacuum was
also used. 1H, 13C, HSQC, and HMBC NMR spectra were recorded on
a Bruker Ascend 400 Fourier transform spectrometer using an in-
ternal deuterium lock. 19F NMR Spectra were recorded on a Varian
Mercury 300 B Fourier transform spectrometer. Solvents were used
Fig. 2. Oxime analogues 15 and 16.
Table 1
Measured PBS solubility of the compounds investigated.
Entry
Compound
Solubility (
mg/mL)
1
2
3
4
5
6
1
2
3
4
15
16
7.7
0.3
<0.3a
<0.3a
3.0
as internal standards when assigning 1H and 13C NMR spectra (
CHCl3 7.26 ppm, DMSO-d5 2.50; C: CDCl3 77.16 ppm, DMSO-d6
dH:
d
39.52 ppm). For E/Z isomer mixtures, chemical shifts for the minor
isomer are given in brackets. Purity by HPLC was performed on a
Waters e2629 separation module fitted with a Waters 2998 photo
>1000.0
a
Below the detection limit of the HPLC setup using the described assay.
diode array detector using a Waters Symmetry 3.5 mm C18 column
(4.6 ꢁ 75 mm) employing a 2 solvent system consisting of solvent
A: water and solvent B: acetonitrile (both solvents containing 0.1%
formic acid). HRMS was performed on a Bruker Daltonics maXis 3G
QTOF-MS fitted with a Dionex Ultimate 3000 UHPLC. A Stuart
Advanced Digital Melting Point SMP30 was used to determine
melting point and they are reported uncorrected. Optical rotation
was measured on a PerkineElmer 241 Polarimeter, and the values
are given in degrees and concentrations are given as g/100 mL.
Compounds 2, 5, 9 and 15 (E/Z mixture) were synthesised according
to literature procedures while 10 and 11 were synthesised using
modified literature procedures [11,27,29]. For 10: 0.7 M NaOH was
used instead of 0.5 M NaOH and for 11: CSA and MeOH was used
instead of (MeO)2CHMe2 and p-TSA followed by HCl (conc.) for
acetal cleavage.
Table 2
Microsomal stability.
Entry
Compound
t½ (Mouse)b
t½ (Rat)b
1
2
3
4
5
6
1
2
3
4
15
16
37
27
11
31
45
58
60a
26
53
57
339
449
a
Duplicate experiment.
Stability against male liver microsomes, t½ (min.).
b
5.1.1. (2S,60R)-20-(Benzyloxy)-7-chloro-4-(difluoromethoxy)-6-
methoxy-60-methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-
3,40-dione (3)
To a stirred solution of 8 (0.20 g, 0.50 mmol) in 1,4-dioxane
(3 mL) under argon was added benzyl alcohol (0.11 mL, 1.0 mmol)
followed by DBU (0.19 mL, 1.25 mmol). The solution was heated to
60 ꢀC, after 20 h the reaction was cooled to 22 ꢀC, celite was added
and the mixture was concentrated in vacuo. Purification by DCVC
(⌀ ¼ 20 mm) gradient eluting from 100% toluene / 75/25 toluene/
acetonitrile with 2.5% increment/20 mL fraction and recrystalliza-
tion from toluene/heptane yielded 0.16 g (50%) 3 as white crystals.
Scheme 2. Synthetic route for preparation of the deuterated compound 18. (a) I)
CD3OD, DBU, 1,4-dioxane, RT, II) MeOH, DBU, reflux.
Table 3
Cytotoxicity of griseofulvin and analogues towards two cancer cell lines.
Entry
Compound
MDA-MB231 (
m
M)
U2OS (mM)
Rf
¼
0.31 (90/10, toluene/acetonitrile); Mp. 113e114 ꢀC;
¼ þ177ꢀ (c ¼ 0.25, CHCl3); 1H NMR (400 MHz, CDCl3)
a 2D0
ꢂ
1
2
3
4
5
6
1
2
3
4
15
16
25.33 3.33
1.62 0.54
11.77 1.30
5.34 1.38
1.29 0.22
10.10 2.59
22.44 1.44
1.45 0.19
14.32 3.83
5.56 0.13
1.61 0.40
14.13 3.83
½
d
7.33e7.24 (3H, m), 7.15 (2H, dd, J ¼ 7.4 & 1.8 Hz), 6.91 (1H, t,
2JHF ¼ 73.5 Hz), 6.47 (1H, s), 5.62 (1H, s), 4.89 (2H, AB, J ¼ 12.3 Hz,
D
¼ 33.7 Hz), 4.01 (3H, s), 2.98 (1H, dd, J ¼ 16.1 & 13.3 Hz),
2.94e2.84 (1H, m), 2.46 (1H, dd, J ¼ 16.1 & 4.0 Hz), 0.98 (3H, d,
J ¼ 6.5 Hz); 13C NMR (101 MHz, CDCl3)
d 196.6, 192.6, 169.5, 168.9,
3
164.2, 146.9 (t, JCF ¼ 3.7 Hz), 134.6, 128.8, 128.5, 126.9, 115.6 (t,
1JCF ¼ 265.4 Hz), 107.2, 106.4, 102.4, 98.9, 91.3, 71.0, 57.5, 40.1, 36.5,
5. Experimental
14.4; 19F NMR (282 MHz, CDCl3)
d
ꢃ83.93 (2F, ABX, 2JFF ¼ 163.8 Hz &
2JFH ¼ 73.5 Hz); HRMS (ESIþ) 465.0916 calcd. for [M þ Hþ] 465.0911.
5.1. General
5.1.2. (2S,60R)-20-(Benzyloxy)-7-chloro-6-(difluoromethoxy)-4-
methoxy-60-methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-
3,40-dione (4)
To a stirred solution of 14 (0.20 g, 0.50 mmol) in 1,4-dioxane
(3 mL) under argon was added benzyl alcohol (0.11 mL, 1.0 mmol)
followed by DBU (0.19 mL, 1.25 mmol). The solution was heated to
60 ꢀC, after 20 h the reaction was cooled to 22 ꢀC, celite was added
For reactions carried out under argon the glassware was dried
before use, employing the following procedure: heating was
applied with a heat gun while the glassware was evacuated
(16 Torr), the evacuation was followed by argon flushing. Evacua-
tion and flushing were applied several times under continuous
heating, and finally the glassware was allowed to cool to room

214
A.B. Petersen et al. / European Journal of Medicinal Chemistry 116 (2016) 210e215
and the mixture was concentrated in vacuo. Purification by DCVC
(⌀ ¼ 20 mm) gradient eluting from 100% toluene / 75/25, toluene/
acetonitrile with 2.5% increment/20 mL fraction yielded 0.065 g
(28%) 4 as white crystals. Rf ¼ 0.16 (90/10, toluene/acetonitrile); Mp.
Purification by DCVC (⌀ ¼ 40 mm) gradient eluting from 100%
toluene / 90/10, toluene/acetonitrile with 1% increment/50 mL
fraction (each gradient run with 2 ꢁ 50 mL) yielded 0.58 g (55%) 8
as an uncoloured oil (the major side product was the 40-chloro
substituted analogue, not illustrated, of which was isolated 0.45 g
109e112 ꢀC; ½a 2D0
ꢂ
¼ þ147ꢀ (c ¼ 0.5, CHCl3); 1H NMR (400 MHz,
CDCl3)
d
7.35e7.24 (3H, m), 7.20e7.11 (2H, m), 6.66 (1H, t,
(43%)). Rf ¼ 0.46 (90/10, toluene/acetonitrile); ½a D20
ꢂ
¼ þ301ꢀ
2JHF ¼ 72.1 Hz), 6.39 (1H, s), 5.61 (1H, s), 4.88 (2H, AB, J ¼ 12.3 &
39.8 Hz), 3.95 (3H, s), 3.03 (1H, dd, J ¼ 16.5 & 13.3 Hz), 2.95e2.83
(c ¼ 0.25, CHCl3); 1H NMR (400 MHz, CDCl3)
d 6.91 (1H, t,
2JHF ¼ 73.2 Hz), 6.53 (1H, s), 6.46 (1H, s), 4.04 (3H, s), 3.08 (1H, dd,
J ¼ 16.6 & 13.9 Hz), 3.01e2.88 (1H, m), 2.50 (1H, dd, J 16.6 & 4.2 Hz),
(1H, m), 2.47 (1H, dd, J ¼ 16.5 & 4.4 Hz), 1.00 (3H, d, J ¼ 6.7 Hz); 13
C
NMR (101 MHz, CDCl3)
d
196.7, 193.3, 170.2, 168.9, 157.0, 155.9 (t,
0.99 (3H, d, 6.6 Hz); 13C NMR (101 MHz, CDCl3)
d 194.2, 191.1, 169.0,
3JCF ¼ 2.6 Hz), 134.6, 128.8, 128.5, 126.7, 115.33 (t, JCF ¼ 265 Hz),
164.6, 151.8, 147.0 (t, 3JCF ¼ 3.7 Hz), 131.7, 115.5 (t, 1JCF ¼ 265.7 Hz),
1
108.8, 106.3, 101.8, 97.5, 91.4, 71.0, 56.8, 40.1, 36.4, 14.5; 19F NMR
107.0, 102.5, 99.1, 91.9, 57.6, 40.0, 37.6, 14.9; 19F NMR (282 MHz,
2
(282 MHz, CDCl3)
d
ꢃ82.14 (2F, d, JFH ¼ 72.1 Hz); HRMS (ESIþ)
CDCl3)
d
ꢃ83.95 (2F, ABX, 2JFF ¼ 163.5 Hz & 2JFH ¼ 72.1 Hz); HRMS
465.0928 calcd. for [M þ Hþ] 465.0911.
(ESIþ) 393.0099 calcd. for [M þ Hþ] 393.0103.
5.1.3. (2S,60R)-7-Chloro-4-(difluoromethoxy)-20,6-dimethoxy-60-
methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-3,40-dione (6)
To a stirred solution of 5 (5.80 g, 17 mmol) in DMF (50 mL) under
argon was added potassium carbonate (2.35 g, 17 mmol) followed
by methyl chlorodifluoroacetate (2.46 mL, 23 mmol). After 18 h at
70 ꢀC the reaction was cooled to 22 ꢀC before being poured into
100 mL ¼ saturated sodium carbonate. The resulting precipitate
was isolated, dissolved in dichloromethane and concentrated on
celite in vacuo. Purification by DCVC (⌀ ¼ 60 mm) gradient eluting
from 100% heptane / 100% ethyl acetate with 10% increment/
100 mL fraction yielded 5.1 g (76%) 6 as white powder. Rf ¼ 0.60
5.1.6. (2S,60R)-7-Chloro-6-(difluoromethoxy)-4,40-dimethoxy-60-
methyl-3H-spiro[benzofuran-2,10-cyclohexan]-30-ene-20,3-dione
(12)
To a stirred solution of 11 (2.0 g, 6.0 mmol) in DMF (50 mL)
under argon was added potassium carbonate (0.83 g, 6.0 mmol)
followed by methyl chlorodifluoroacetate (0.96 mL, 9.1 mmol).
After 6½ h. at 75 ꢀC the reaction was cooled to 22 ꢀC before being
poured into 50 mL ¼ saturated sodium carbonate. The resulting
precipitate was isolated, dissolved in dichloromethane and
concentrated on celite in vacuo. Purification by DCVC (⌀ ¼ 40 mm)
gradient eluting from 100% heptane / 100% ethyl acetate with 10%
increment/50 mL fraction yielded 1.9 g (82%) 12 as white powder, a
sample was recrystallised from dichloromethane/heptane.
(ethyl acetate); ½a D20
ꢂ
¼ þ275ꢀ (c ¼ 0.25, CHCl3); 1H NMR (400 MHz,
2
CDCl3)
d
6.89 (1H, t, JHF ¼ 73.3 Hz), 6.48 (1H, s), 5.54(1H, s), 4.00
(3H, s), 3.62 (3H, s), 2.93 (1H, dd, J ¼ 16.1 &13.3 Hz), 2.89e2.78 (1H,
m), 2.43 (1H, dd, J ¼ 16.1 & 4.0 Hz), 0.92 (3H, d, J ¼ 6.5 Hz); 13C NMR
Rf ¼ 0.64 (ethyl acetate); Mp. 160.0e161.0 ꢀC; ½a D20
ꢂ
¼ þ160ꢀ
(101 MHz, CDCl3)
d 196.7, 192.5, 170.3, 169.4, 164.2, 147.0 (t,
(c ¼ 0.25, CHCl3); 1H NMR (400 MHz, CDCl3)
d 6.65 (1H, dd,
3JCF ¼ 3.7 Hz), 115.5 (t, 1JCF ¼ 265.2 Hz), 106.9, 105.1, 102.3, 98.6, 91.1,
2JHF ¼ 72.8 & 72.0 Hz), 6.37 (1H, s), 5.44 (1H, d, J ¼ 1.4 Hz), 3.92 (3H,
s), 3.78 (3H, s), 3.18 (1H, ddd, J ¼ 17.6, 12.0 & 1.4 Hz), 2.92e2.81 (1H,
m), 2.49 (1H, dd, J ¼ 17.6 & 5.7 Hz), 1.03 (3H, d, 6.7 Hz); 13C NMR
57.5, 56.9, 39.9, 36.5, 14.2; 19F NMR (282 MHz, CDCl3)
d
ꢃ83.90 (2F,
ABX, 2JFF ¼ 164.0 Hz & 2JFH ¼ 73.5 Hz); HRMS (ESIþ) 389.0594 calcd.
for [M þ Hþ] 389.0598.
(101 MHz, CDCl3)
d 192.5, 187.9, 179.1, 170.4, 157.0, 155.8 (t,
3JCF ¼ 2.7 Hz), 115.4 (dd, 1JCF ¼ 265 Hz, 264 Hz), 108.8, 102.1, 99.6,
5.1.4. (2S,60R)-7-Chloro-4-(difluoromethoxy)-40-hydroxy-6-
methoxy-60-methyl-3H-spiro[benzofuran-2,10-cyclohexan]-30-ene-
20,3-dione (7)
To a stirred solution of 6 (2.9 g, 7.5 mmol) in acetic acid (20 mL)
was added sulphuric acid (6 mL, 2 M), the solution was heated to
80 ꢀC. After 45 min the solution was cooled to 22 ꢀC before being
poured into water (150 mL). The resulting precipitate was isolated
and extracted extensively with water yielding 2.7 g (97%) 7 as light
yellow powder. A sample was recrystallised from ethyl acetate/
heptane yielding light yellow crystals. Mp. 205e208 ꢀC (dec.);
97.4, 95.8, 56.7, 56.5, 35.3, 33.0, 14.7; 19F NMR (282 MHz, CDCl3)
2
2
d
ꢃ82.03 (2F, ABX, JFF ¼ 164.4 Hz & JFH ¼ 72.4 Hz); HRMS (ESIþ)
389.0600 calcd. for [M þ Hþ] 389.0598.
5.1.7. (2S,60R)-7-Chloro-6-(difluoromethoxy)-40-hydroxy-4-
methoxy-60-methyl-3H-spiro[benzofuran-2,10-cyclohexan]-30-ene-
20,3-dione (13)
To a stirred solution of 12 (1.7 g, 4.4 mmol) in acetic acid (15 mL)
was added sulphuric acid (4 mL, 2 M), the solution was heated to
80 ꢀC. After 45 min the solution was cooled to 22 ꢀC before being
poured into water (50 mL). The mixture was extracted ethyl acetate
(3 ꢁ 40 mL) and the combined organic extracts was dried with
sodium sulphate before celite was added and it was concentrated in
vacuo. Purification by DCVC (⌀ ¼ 40 mm) gradient eluting from
100% heptane / 100% ethyl acetate with 10% increment/50 mL
fraction followed by ethyl acetate (4 ꢁ 50 mL) (all solvents con-
taining 0.2 vol% acetic acid) yielded 1.3 g (79%) 13 as white powder,
a sample was recrystallised from ethyl acetate/heptane. Rf ¼ 0.32
(0.2% acetic acid in ethyl acetate); Mp. 209e210 ꢀC (dec.);
½
a 2D0
d
ꢂ
¼ þ234ꢀ (c ¼ 0.25, MeOH); 1H NMR (400 MHz, DMSO-d6)
12.00 (1H, bs), 7.46 (1H, t, 2JHF ¼ 72.8 Hz), 6.74 (1H, s), 5.33 (1H, s),
4.03 (3H, s), 2.89e2.75 (2H, m), 2.60e2.51 (1H, m), 0.88 (3H, d,
J ¼ 5.9 Hz); 13C NMR (101 MHz, DMSO-d6)
d 191.2, 187.0, 179.8,
168.9, 163.7, 147.4 (t, 3JCF ¼ 4.1 Hz), 115.9 (t, 1JCF ¼ 259.9 Hz), 105.7,
101.1, 99.9, 96.8, 95.2, 58.0, 34.4, 32.7, 14.2; 19F NMR (282 MHz,
2
DMSO-d6)
d
ꢃ84.64 (2F, d, JFH ¼ 72.8 Hz); HRMS (ESIþ) 375.0444
calcd. for [M þ Hþ] 375.0441.
5.1.5. (2R,60R)-20,7-Dichloro-4-(difluoromethoxy)-6-methoxy-60-
methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-3,40-dione (8)
To a stirred solution of 7 (1.0 g, 2.7 mmol) in 1,4-dioxane (10 mL)
was added lithium chloride (1.0 g, 24 mmol) and phosphoryl tri-
chloride (5 mL). The mixture was heated to 85 ꢀC for 40 min before
cooling with ice bath (carefully, not frozen). Saturated sodium
carbonate was added (slowly) until the pH reached 8 and the
mixture was extracted with dichloromethane (3 ꢁ 30 mL). The
combined organic extracts were dried with sodium sulphate before
celite was added and the mixture was concentrated in vacuo.
½
a 2D0
d
ꢂ
¼ þ191ꢀ (c ¼ 0.25, MeOH); 1H NMR (400 MHz, DMSO-d6)
11.98 (1H, bs), 7.56 (1H, t, 2JHF ¼ 72.3 Hz), 6.67 (1H, s), 5.31 (1H, s),
3.90 (3H, s), 2.90e2.75 (2H, m), 2.59e2.50 (1H, m), 0.86 (3H, d,
J ¼ 5.7 Hz); 13C NMR (101 MHz, DMSO-d6)
d 191.9, 186.8, 179.8,
169.4, 156.8, 155.8 (t, 3JCF ¼ 3.5 Hz), 116.1 (t, 1JCF ¼ 260.4 Hz), 107.4,
101.2, 98.6, 96.0, 94.9, 56.9, 34.3, 32.9, 14.2; 19F NMR (282 MHz,
2
DMSO-d6)
d
ꢃ83.86 (2F, d, JFH ¼ 72.3 Hz); HRMS (ESIþ) 375.0461
calcd. for [M þ Hþ] 375.0441.

A.B. Petersen et al. / European Journal of Medicinal Chemistry 116 (2016) 210e215
215
5.1.8. (2R,60R)-20,7-Dichloro-6-(difluoromethoxy)-4-methoxy-60-
methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-3,40-dione
(14)
left for 2 days at 22 ꢀC were upon crystals of 18 was formed.
Filtration yielded 170 mg (85%) 18 as a white powder, Rf ¼ 0.65
(ethyl acetate); Mp. 217e219 ꢀC; 1H NMR (400 MHz, DMSO-d6)
To a stirred solution of 13 (1.2 g, 3.2 mmol) in 1,4-dioxane
(12 mL) was added lithium chloride (1.2 g, 28 mmol) and phos-
phoryl trichloride (6 mL). The mixture was heated to 85 ꢀC for
50 min before cooling with ice bath (carefully, not frozen). Satu-
rated sodium carbonate was added (slowly) until the pH reached 8
and the mixture was extracted with dichloromethane (3 ꢁ 30 mL).
The combined organic extracts was dried with sodium sulphate
before celite was added and the mixture was concentrated in vacuo.
Purification by DCVC (⌀ ¼ 40 mm) gradient eluting from 100%
toluene / 95/5, toluene/acetonitrile with 1% increment/50 mL
fraction (each gradient run with 2 ꢁ 50 mL) yielded 0.61 g (48%) 14
as an uncoloured oil (the major side product was the 40-chloro
substituted analogue, not illustrated, of which was isolated 0.53 g
d 6.51 (1H, s), 5.61 (1H, s), 4.05 (3H, s), 3.95 (3H, s), 2.88e2.74 (1H,
m), 2.68 (1H, dd, J ¼ 16.4 & 13.3 Hz), 2.36 (1H, dd, J ¼ 16.2 & 4.7 Hz),
0.81 (3H, d, J ¼ 6.7 Hz); 13C NMR (101 MHz, DMSO-d6)
d 196.0,191.6,
170.7, 169.0, 164.9, 158.1, 105.1, 104.5, 95.7, 91.8, 90.6, 58.1, 57.1, 35.9,
14.3; HRMS (ESIþ) 356.0979 calcd. for [M þ Hþ] 356.0975.
Acknowledgements
This work was supported by the Novo Nordisk Foundation,
Biotechnology-based Synthesis and Production Research and Pre-
seed Grants, as well as through an instrument grant from the
Carlsberg Foundation.
(42%)). Rf ¼ 0.34 (90/10, toluene/acetonitrile); ½a D20
ꢂ
¼ þ302ꢀ
Appendix A. Supplementary data
(c ¼ 0.25, CHCl3); 1H NMR (400 MHz, CDCl3)
d 6.70 (1H, t,
2JHF ¼ 71.9 Hz), 6.45 (1H, s), 6.44 (1H, s), 3.97 (3H, s), 3.09 (1H, dd,
J ¼ 16.8 & 13.9 Hz), 2.99e2.87 (1H, m), 2.49 (1H, dd, J ¼ 16.6 &
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2016.03.071.
4.3 Hz), 0.99 (3H, d, J ¼ 6.7 Hz); 13C NMR (101 MHz, CDCl3)
d 194.4,
3
191.8, 169.7, 157.1, 156.4 (t, JCF ¼ 2.6 Hz), 151.8, 131.7, 115.3 (t,
References
1JCF ¼ 266 Hz), 108.7, 101.8, 97.6, 91.9, 56.9, 40.1, 37.7, 15.1; 19F NMR
2
(282 MHz, CDCl3)
d
ꢃ82.27 (2F, ABX, JFF
¼
163.8 Hz &
[1] A.E. Oxford, H. Raistrick, P. Simonart, Biochem. J. 33 (1939) 240e248.
[2] D.I. Williams, R.H. Marten, I. Sarkany, Lancet 272 (1958) 1212e1213.
[3] R.P.T. Davenport-Hines, J. Slinn, Glaxo: A History to 1962, Cambridge Uni-
versity Press, Cambridge, 1992, p. 219.
[4] A.B. Petersen, M.H. Rønnest, T.O. Larsen, M.H. Clausen, Chem. Rev. 114 (2014)
12088e12107.
2JFH ¼ 72.0 Hz); HRMS (ESIþ) 393.0104 calcd. for [M þ Hþ] 393.0103.
5.1.9. 2-((((2S,60R)-7-Chloro-4,6-dimethoxy-60-methyl-3-oxo-20-
phenoxy-3H-spiro[benzofuran-2,10-cyclohexan]-20-en-40-ylidene)
amino)oxy)acetic acid (16) (E/Z isomeric mixture)
[5] Y.-S. Ho, J.-S. Duh, J.-H. Jeng, Y.-J. Wang, Y.-C. Liang, C.-H. Lin, C.-J. Tseng, C.-
F. Yu, R.-J. Chen, J.-K. Lin, Int. J. Cancer 91 (2001) 393e401.
To a stirred solution of 2 (0.215 g, 0.50 mmol) in ethanol
(12.5 mL) and DMSO (6.7 mL) under argon, was added sodium
acetate (0.176 g, 2.15 mmol) and (aminooxy)acetic acid hemi-
hydrochloride (0.191 g, 1.75 mmol). The mixture was heated to
65 ꢀC for 20 h before cooling to 22 ꢀC. Dichloromethane was added
(40 mL) and the mixture was washed with water (3 ꢁ 25 mL). The
organic phase was dried with sodium sulphate before celite was
added and the mixture was concentrated in vacuo. Purification by
DCVC (⌀ ¼ 40 mm) gradient eluting from 50% toluene / 60/40,
toluene/acetonitrile (all solvents containing 0.5% acetic acid) with
4% increment/50 mL fraction yielded 0.225 g (90%) 16 as white
crystals. Rf ¼ 0.23 (59/39/2, toluene/acetonitrile/acetic acid); 1H
~
[6] A.R. Chaudhuri, R.F. Luduena, Drug Dev. Res. 53 (2001) 44e49.
[7] R.A.B. Keates, Biochem. Biophys. Res. Commun. 102 (1981) 746e752.
[8] T. Oda, J. Antibiot. 59 (2006) 114e116.
€
[9] B. Rebacz, T.O. Larsen, M.H. Clausen, M.H. Rønnest, H. Loffler, A.D. Ho,
€
A. Kramer, Cancer Res. 67 (2007) 6342e6350.
[10] M.H. Rønnest, B. Rebacz, L. Markworth, A.H. Terp, T.O. Larsen, A. Kramer,
€
M.H. Clausen, J. Med. Chem. 52 (2009) 3342e3347.
[11] M.S. Raab, I. Breitkreutz, S. Anderhub, M.H. Rønnest, B. Leber, T.O. Larsen,
L. Weiz, G. Konotop, P.J. Hayden, K. Podar, J. Fruehauf, F. Nissen, W. Mier,
U. Haberkorn, A.D. Ho, H. Goldschmidt, K.C. Anderson, M.H. Clausen,
€
A. Kramer, Cancer Res. 72 (2012) 5374e5385.
[12] W.E. Barnette, CRC Crit. Rev. Biochem. 15 (1984) 201e235.
[13] M. van Heek, C.F. France, D.S. Compton, R.L. McLeod, N.P. Yumibe, K.B. Alton,
E.J. Sybertz, H.R. Davies Jr., J. Pharmacol. Exp. Ther. 283 (1997) 157e163.
[14] J.W. Clader, J. Med. Chem. 47 (2004) 1e9.
[15] S. Symchowicz, K.K. Wong, Biochem. Pharmacol. 15 (1966) 1595e1600.
[16] P.A. Harris, S. Riegelman, J. Pharm. Sci. 58 (1969) 93e96.
[17] C. Lin, R. Chang, J. Magat, S. Symchowicz, J. Pharm. Pharmacol. 24 (1972)
911e913.
NMR (400 MHz, CDCl3) d 7.32e7.20 (3H, m), 7.20e7.12 (2H, m), 6.08
(1H, s), 5.65 (6.28) (1H, s), 4.88 (4.96) (1H, d, J ¼ 12.3 (12.2) Hz), 4.77
(4.82) (1H, d, J ¼ 12.3 (12.2) Hz), 4.66e4.60 (2H, m), 4.01 (3H, s),
3.95 (3H, s), 2.78 (3.05) (1H, dd, J ¼ 17.0 (15.2) & 13.0 (13.4) Hz),
2.70e2.54 (1H, m), 2.43 (3.12) (1H, dd, J ¼ 15.2 (17.0) & 4.2 (4.9) Hz),
0.97 (0.98) (3H, d, J ¼ 6.7 (6.8) Hz); 13C NMR (101 MHz, CDCl3)
[18] C.-C. Lin, J. Magat, R. Chang, J. McGlotten, S. Symchowicz, J. Pharmacol. Exp.
Ther. 187 (1973) 415e422.
€
[19] M.H. Rønnest, M.S. Raab, S. Anderhub, S. Boesen, A. Kramer, T.O. Larsen,
M.H. Clausen, J. Med. Chem. 55 (2012) 652e660.
d
194.0 (193.8), 174.4 (174.3), 169.7, 164.5, 158.7 (161.3), 157.7
[20] T. Umemoto, K. Adachi, S. Ishihara, J. Org. Chem. 72 (2007) 6905e6917.
[21] K. Stanek, R. Koller, A. Togni, J. Org. Chem. 73 (2008) 7678e7685.
[22] R. Koller, K. Stanek, D. Stolz, R. Aardoom, K. Niedermann, A. Togni, Angew.
Chem. Int. Ed. 48 (2009) 4332e4336.
[23] R. Ringom, T. Benneche, Acta Chem. Scand. 53 (1999) 41e47.
[24] S.B. Christensen IV, S. Dabbs, J.M. Karpinski, PCT International Application
WO/1996/023754, 1996.
(157.6), 157.1 (153.9), 135.7 (135.5), 128.5, 127.9 (128.0), 126.7
(126.8), 105.8 (105.7), 99.7 (94.1), 97.2 (97.2), (91.4) 91.4, (89.4) 89.4,
70.3 (70.2), (70.1) 70.1, (57.0) 57.0, 56.4, (36.5) 35.4, 26.4 (30.8), 14.5
(14.4); HRMS (ESIþ) 502.1259 calcd. for [M þ Hþ] 502.1263.
[25] S. Zheng, G. Kaur, H. Wang, M. Li, M. Yang, X. Macnaughtan, S. Reid,
J. Prestegard, B. Wang, H. Ke, J. Med. Chem. 51 (2008) 7673e7688.
[26] V. Arkley, J. Attenburrow, G.I. Gregory, T. Walker, J. Chem. Soc. (1962)
1260e1268.
[27] L. Stephenson, T. Walker, W.K. Warburton, G.B. Webb, J. Chem. Soc. (1962)
1282e1292.
[28] For an example of a small-scale protocol for direct position-6 demethylation,
see: M.H. Rønnest, P. Harris, C.H. Gotfredsen, T.O. Larsen, M.H. Clausen Tet-
rahedron Lett. 51 (2010) 5881e5882.
[29] J.A. Erickson, J.I. McLoughlin, J. Org. Chem. 60 (1995) 1626e1631.
[30] F. Narjes, K.F. Koehler, U. Koch, B. Gerlach, S. Colarusso, C. Steinkühler,
M. Brunetti, S. Altamura, R. De Francesco, V.G. Matassa, Bioorg. Med. Chem.
Lett. 12 (2002) 701e704.
5.1.10. (2S,60R)-7-Chloro-4,6-dimethoxy-20-(methoxy-d3)-60-
methyl-3H-spiro[benzofuran-2,10-cyclohexan]-20-ene-3,40-dione
(18)
To a stirred solution of 17 (0.20 g, 0.56 mmol) in 1,4-dioxane
(3 mL) under argon was added deuterated methanol (0.4 mL) fol-
lowed by DBU (0.4 mL). The solution was stirred at 22 ꢀC for 3 days.
Celite was added and the mixture was concentrated in vacuo. Pu-
rification by DCVC (⌀ ¼ 20 mm) gradient eluting from 100%
heptane / 100% ethyl acetate with 10% increment/25 mL fraction.
The isolated compound (Rf ¼ 0.65 (ethyl acetate)) and DBU (0.4 mL)
were refluxed in MeOH (50 mL) for 5 h. The reaction mixture was
[31] N.A. Meanwell, J. Med. Chem. 54 (2011) 2529e2591.
[32] D.S. Pedersen, C. Rosenbohm, Synthesis 16 (2001) 2431e2434.