Anhydride modified cantharidin analogues: Synthesis, inhibition of protein phosphatases 1 and 2A and anticancer activity

Article abstract of DOI:10.1016/S0960-894X(00)00323-1

Two series of anhydride modified cantharidin analogues were synthesised and screened for their phosphatase inhibition (PP1 and PP2A) and cytotoxicity in various cancer cell lines (Ovarian A2780, ADDP; Osteosarcoma 143B; and Colon HCT116 and HT29). One series was synthesised by a novel, high yielding one-pot hydrogenation-ring-opening-esterification procedure, the other by acid catalysed acetal formation. Analogues 5-7 and 9 displayed moderate PP2A selectivity (ca. 5- to 20-fold) and inhibition typically in the low μM range (comparable, in some cases to cantharidin). The anticancer activity of these analogues varied with the cell line under study; however, many of them showed selective cytotoxicity for the colon tumour cell lines. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00323-1

Bioorganic & Medicinal Chemistry Letters 10 (2000) 1687±1690
Anhydride Modi®ed Cantharidin Analogues: Synthesis, Inhibition
of Protein Phosphatases 1 and 2A and Anticancer Activity
Adam McCluskey,a,* Michael C. Bowyer,b Elizabeth Collins,c Alistair T. R. Sim,c
Jennette A. Sako€d and Monique L. Baldwina
aMedicinal Chemistry Group, School of Biological and Chemical Sciences, The University of Newcastle, Callaghan, NSW 2308, Australia
bSchool of Science and Technology, Central Coast Campus, The University of Newcastle, PO Box 127, NSW 2258, Australia
cDiscipline of Medical Biochemistry, School of Biomedical Science, The University of Newcastle, Callaghan, NSW 2308, Australia
dDepartment of Medical Oncology, Newcastle Mater Misericordiae Hospital, NSW 2298, Australia
Received 13 December 1999; accepted 24 May 2000
AbstractÐTwo series of anhydride modi®ed cantharidin analogues were synthesised and screened for their phosphatase inhibition
(PP1 and PP2A) and cytotoxicity in various cancer cell lines (Ovarian A2780, ADDP; Osteosarcoma 143B; and Colon HCT116 and
HT29). One series was synthesised by a novel, high yielding one-pot hydrogenation-ring-opening-esteri®cation procedure, the other by
acid catalysed acetal formation. Analogues 5±7 and 9 displayed moderate PP2A selectivity (ca. 5- to 20-fold) and inhibition typically in
the low mM range (comparable, in some cases to cantharidin). The anticancer activity of these analogues varied with the cell line
under study; however, many of them showed selective cytotoxicity for the colon tumour cell lines. # 2000 Elsevier Science Ltd. All
rights reserved.
Cantharidin (exo,exo-2,3-dimethyl-7-oxobicyclo[2.2.1]-
heptane-2,3-dicarboxylic acid anhydride) (1) (Fig. 1), in
the form of the dried body of the Chinese blister beetles:
Mylabris phalerata or M. cichorii1 has been used by the
Chinese as a natural remedy for the past 2000 years.
Western medicine decreed cantharidin to be too toxic in
the early 1900s.2 Cantharidin has important antitumour
properties and has been used as an anticancer agent by the
Chinese for the treatment of hepatoma and oesophageal
carcinoma.
There has been intense interest in developing potent and
selective inhibitors of PP1 and PP2A in recent years.
These attempts have included numerous analogues of
cantharidin (so far with little success)4 and the more com-
plex toxins such as the Microcystin analogues developed by
Chamberlin,5 which resulted in moderate PP1 selectivity.
We, too, have had an ongoing interest in cantharidin ana-
logues, an interest that was further heightened with the
discovery of fostriecin (2) (Fig. 1) as a selective and
potent PP2A inhibitor with the additional bene®t of
being a potent anticancer agent.
The ®rst recorded use as an antitumour agent was in
1264.1 A unique feature of cantharidin observed during
clinical trials was the stimulation of the bone marrow
production of white cells, which is in contrast to most
other anticancer drugs that readily induce myelosup-
pression. Despite such qualities, the nephrotoxicity of
cantharidin has prevented it from entering mainstream
oncology. Norcantharidin (11), the demethylated analogue
of cantharidin, also possesses anticancer activity and
stimulates the bone marrow, but without the urinary
toxicity.1 Both agents are known protein phosphatase 1
(PP1) and 2A (PP2A) inhibitors.3
The anticancer activity of fostriecin (2) has been linked
to its selective inhibition of PP2A. Fostriecin exhibits
>40 000-fold selectivity for PP2A (IC50 3.4 nM) over
PP1.6 As with other protein phosphatase inhibitors
(okadaic acid, calyculin A), fostriecin abrogates the
G2 checkpoint of the cell cycle, which forces the cell
prematurely into mitosis with inadequate spindle
*Corresponding author. Tel.: +61-2492-16486; fax: +61-2492-15472;
e-mail: amcclusk@mail.newcastle.edu.au
Figure 1.
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00323-1

1688
A. McCluskey et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1687±1690
formation and poorly replicated DNA.6 Fostriecin is
active against leukemia (L1210, IC50 0.46 mM), lung,
breast, and ovarian cancer cells and displays ecacious
in vivo antitumour activity against L1210 leukemia in
mice and is currently undergoing phase I clinical trials at
NCI as an anticancer agent. In addition to its phospha-
tase inhibition fostriecin also inhibits topoisomerase II
(IC50 40 mM) in vitro through a novel, non-DNA-strand
cleaving mechanism, but does not induce G2 arrest like
other topoisomerase II inhibitors.7 Fostriecin, however,
is a more potent inhibitor of protein phosphatase than
topoisomerase II.
As can be seen from Figure 1 there is a super®cial struc-
tural commonality between 1 and 2. Intrigued by this,
we set out to synthesise a series of fostriecin/cantharidin
`cross' analogues. We believed that the synthesis of such
analogues would allow the development of oxidatively
stable fostriecin analogues possessing the stability and
bone marrow stimulatory properties of 1 (and 11), whilst
possessing the potent phosphatases inhibition and anti-
cancer properties of fostriecin. Additionally, we felt that
these modi®cations may allow the introduction of a
hydrophobic tail, more in line with the more potent
okadaic acid class of compounds.5 In the ®rst instance,
we were primarily interested in the addition of an ali-
phatic side chain within the ®ve-membered ring of can-
tharidin, followed by both phosphatase (PP1 and PP2A)
and anticancer screening in an attempt to de®ne the
biological action of these agents. We have recently
observed that the lactone (4) (and other cantharidin
analogues, not shown) showed selective cytotoxicity
towards colon tumour cell lines compared with leuke-
mia, ovarian, and osteosarcoma cell lines.8 We were
thus encouraged to develop a series of cantharidin/fos-
triecin analogues as potential anticancer agents. Herein
we wish to report some of our preliminary results in the
areas of protein phosphatase inhibition and anticancer
activity.
Scheme 1. (i) 10% Pd-C, H2 4 atm, `super-dry ROH';11 (ii) 10% Pd-C,
H2 4 atm, acetone, 24 h; (iii) 10% Pd-C, H2 4 atm, wet EtOH; (iv)
ROH, cat p-TosOH, re¯ux 2 h.
4 with catalytic p-TosOH and methanol (ethanol or
propanol) at re¯ux for 2 h followed by standard work
up, resulted in the clean generation of a series of acetal-
lactones 8±10.
Biology
The cantharidin analogues, 5±10, were screened for
their ability to inhibit protein phosphatase 1 and protein
phosphatase 2A.13 Cantharidin (1) and norcantharidin
(11) were included as internal standards to ensure the
relative validity of our protocol. The results of the phos-
phatase inhibition study are shown in Table 1. We note
that analogues 5±7 show excellent potency compared
with 1 and 11, as well as improving (albeit slightly) on
PP2A selectivity.
Note that with analogues 5±7, we have been able to
maintain the low micromolar inhibition of PP1 and
PP2A previously observed with cantharidin, and an
increase, albeit a modest one, over PP2A selectivity.
This is in keeping with other results, which indicated that
``a ring-opened anhydride derivative'' was necessary to
promote ecient binding with PP2A.14 Analogues 8±10
showed marginal or no protein phosphatase activity.
However, 9 displayed greater PP2A selectivity than any
of the analogues generated herein. We are currently
investigating the origin of the PP2A selectivity of 9.
Chemistry
We based our original synthesis of 4 on the work of
Egglette, starting from the room temperature Diels±
Alder addition of furan and maleic anhydride, a€ording
anhydride 3 in excellent yield.9 However, we found that
using the literature protocol failed to generate even the
moderate quantities of 4 required for subsequent synthetic
manipulations. Closer examination of the products
obtained indicated that the major product isolated was
not the expected lactol 4,10 but the hydrogenated, ring-
opened, esteri®ed analogue 6 (see Scheme 1).11 Subse-
quently, we examined the generality of this reaction and
found that: The reaction proceeds smoothly in the cases
of methanol, ethanol and propanol. Results with exten-
ded-chain alcohols will be reported in due course.
Additionally, the reaction only proceeds in this manner
if `super-dry' alcohols are used (doubly distilled from
Mg/I2).11,12
Table 1. Inhibition of protein phosphatases 1 and 2A by compounds
1 and 5±11,a
Compound
PP1 inhibition
IC50 (mM)
PP2A inhibition
IC50 (mM)
PP2A
selectivity
1
5
6
7
8
9
10
11
1.78
4.71
2.96
4.82
>1000
746
>1000
1.98
0.26
0.41
0.45
0.47
>1000
55
>1000
0.37
6.8
11.5
6.6
10.3
ND
13.6
ND
5.5
In cases where rigorous drying conditions were not
enforced, 4 was the sole product isolated. Treatment of
aAverage of three experiments in triplicate.

A. McCluskey et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1687±1690
1689
Table 2. IC50(mM)a values of tumour cell lines after 72 h continuous exposure to test compounds 1, 5±11
Tumour type/
cell line
1
5
6
7
8
9
10
11
A2780b
ADDPb
143Bc
HCT116d
HT29d
10Æ2
11Æ1.2
10Æ1.2
9Æ1
100Æ10
180Æ8
118Æ8
76Æ14
105Æ5
185Æ5
110Æ10
315Æ65
450Æ50
75Æ5
538Æ83
323Æ40
>1000
143Æ23
28Æ1
100Æ20
107Æ9
103Æ17
80Æ17
26Æ8
375Æ176
535Æ120
>1000
50Æ0
47Æ3
43Æ9
24Æ4
33Æ7
330Æ39
437Æ37
266Æ9
195Æ5
6.4Æ0.7
243Æ39
15Æ4.5
41Æ11
aIC50 is the concentration that induces 50% growth inhibition compared with untreated control cells.
bOvarian.
cOsteosarcoma.
dColon.
Conclusions
In view of the anticancer activity and clinical use of can-
tharidin, norcantharidin, and fostriecin we also conducted
cytotoxicity studies in a number of tumour cell lines,
and the results from these studies are shown in Table 2.
We have developed a novel one-pot reaction for the
synthesis of mono-esteri®ed analogues of cantharidin
which displayed both potency and levels of selectivity at
PP2A comparable to cantharidin. The corresponding
acetal lactones exhibited almost no protein phosphatase
inhibition, with the exception of 9 reinforcing the belief
that the ring-opened form of cantharidin is the active spe-
cies. Additionally, the analogues tested showed promise as
selective anticolon-cancer agents. Those agents exhibit-
ing greater inhibition of PP2A also exhibit greater cyto-
toxicity. This study does not rule out any other possible
mode of action, but it is intriguing to think that selective
inhibition of PP2A may give rise to more potent and
organ/tissue speci®c anticancer agents.
Cytotoxicity was evaluated using the MTT assay.15
From Table 2 it was apparent that cantharidin (1) and
norcantharidin (11) were potent, but nonselective anti-
cancer agents. In the ®rst series of modi®cations that we
have carried out, compounds 5±7 had the e€ect of an
overall decrease in the cytotoxic potency against the cell
lines illustrated. The only notable exception being 7,
which showed both moderate potency and selectivity for
the colon cell lines: HCT116 and HT29 at 75 and 15
mM, respectively.
In the second series of analogues 8±10, the cytotoxic
potency of 8 and 10 in A2780, ADDP, and 143B cell
lines was lower than that observed for the ®rst series of
analogues. However, the potency of 9 was typically
comparable or better than that of the ®rst series of ana-
logues (A2780 100 Æ 20 mM; ADDP 107 Æ 9 mM; 143B
103 Æ 17 mM). Interestingly, as observed for analogue 7,
analogues 8±10 showed a marked increase in cytotoxi-
city in the colon cell lines particularly the HT29 cells,
where cytotoxicity resembled that observed for nor-
cantharidin (11). It is noteworthy that in HT29 cells, the
presence of the alkyl chain (methyl, ethyl or propyl)
appeared to have little e€ect on cytotoxicity. Again with
compounds 8±10, irrespective of the magnitude of their
cytotoxicity the trend observed (low-high-low) parallelled
the observed ability of these compounds to inhibit protein
phosphatases. Most interestingly, compound 9 showed
reasonable potency across the cell lines screened, increas-
ing in the case of colon cells, and displayed moderate
phosphatase inhibition with moderate PP2A selectivity.
Acknowledgements
We gratefully acknowledge ®nancial support from Hunter
Medical Research and the NH and MRC Australia. We
also thank Dr. Ian van Altena for useful discussions.
References and Notes
1. Wang, G.-S. J. Ethnopharmacol. 1989, 26, 147.
2. Goldfarb, M. T.; Gupta, A. K.; Sawchuk, W. S. Dermato-
logic Clinics 1991, 9, 287.
3. Li, Y.-M.; Casida, J. E. Proc. Natl. Acad. Sci. USA 1992,
89, 11867.
4. (a) McCluskey, A.; Taylor, C.; Quinn, R. J.; Suganuma,
M.; Fujiki, H. Bioorg. Med. Chem. Lett. 1996, 6, 1025. (b)
Enz, A.; Zenke, G.; Pombo-Villar, E. Bioorg. Med. Chem.
Lett. 1997, 7, 2513; (c) Sodeoka, M.; Baba, Y.; Kobayashi, S.;
Hirukawa, Bioorg. Med. Chem. Lett. 1997, 7, 1833; (d) Tat-
lock, J. H.; Linton, M. A.; Hou, X. J.; Kissinger, C. R.; Pel-
letier, L. A.; Showalter, R. E.; Tempczyk, A.; Villafranca, J. E.
Bioorg. Med. Chem. Lett. 1997, 7, 1007.
Finally, it was apparent that in both series of compounds,
the highest cytotoxicity was observed for those analogues
showing the greatest selectivity for PP2A over PP1. This
observation notwithstanding, it is inherently dicult to
correlate in vitro protein phosphatase inhibition with cel-
lular cytotoxicity, as the latter will potentially be in¯uenced
by not only the degree of protein phosphatase inhibition
but also by intracellular drug uptake, intracellular drug
metabolism, and variations in intracellular protein
phosphatase activity. Furthermore, we cannot preclude
the involvement of other intracellular targets.
5. Aggen, J. B.; Humphrey, J. M.; Gauss, C.-M.; Huang, H.-
B.; Nairn, A. A. C.; Chamberlin, A. R. Bioorg. Med. Chem.
1999, 7, 543.
6. Murray, R. W. Nature 1992, 359, 599.
7. (a) Chen, G. L.; Yang, L.; Rowe, T. C.; Halligan, B. D.;
Tewey, K. M.; Liu, L. F. J. Biol. Chem. 1984, 259, 13560. (b)
Tewey, K. M.; Rowe, T. C.; Yang, L.; Hallighan, B. D.; Liu,
L. Science 1984, 266, 466.
8. Sako€, J.; Ackland, S.; Baldwin, M. L.; Keane, M. A.;
McCluskey, A. Invest. New Drugs 2000, submitted.

1690
A. McCluskey et al. / Bioorg. Med. Chem. Lett. 10 (2000) 1687±1690
9. Eggelte, T. A.; de Koning, H.; Huisman, H. O. Tetrahedron
1973, 29, 2445.
10. It is known that this hydrogenation is solvent dependent,
with the ole®n only being reduced if the hydrogenation is carried
out in either acetone or ethyl acetate as the solvent.
on ice the sample was centrifuged and a 100 mL aliquot of the
supernatant was removed for scintillation counting of the [32P]
released during the reaction. Data is expressed as the IC50 con-
centration of the compound, which represents the concentration
of compound required to produce 50% inhibition of protein
phosphatase activity relative to a control (absence of inhibitor)
incubation (100% activity).
11. A typical synthesis of 6 is as follows: anhydride 11 (5.028 g,
0.03 mol) was dissolved in dry ethanol (doubly distilled after
re¯ux over Mg/I2), 10%-Pd/C (500 mg) added and the mixture
was shaken under 4 atmos of hydrogen at room temperature
overnight (typically 17 h). The catalyst was ®ltered o€ and the
solvent removed in vacuo. The crude solid was recrystallised
from ethyl acetate yielding a white solid (4.089 g, 63%). Mp
14. McCluskey, A.; Keane, M. A.; Mudgee, L.-M.; Sim, A. T.
R.; Sako€, J. A.; Quinn, R. J. Eur. J. Med. Chem. 2000, in press.
15. Cell culture and stock solutions. Stock solutions were pre-
pared as follows and stored at 20 ꢀC: cantharidin (Biomol,
USA) and cantharidin analogues as 10 mM solutions in
phosphate bu€ered saline (PBS). All cell lines were cultured at
37 ꢀC, under 5% CO2 in air. The cell lines A2780 (human
ovarian carcinoma) and 143B (human osteocarcinoma) were
maintained in DMEM (Trace Biosciences, Australia) supple-
mented with 5% foetal bovine serum and 10 mM sodium
bicarbonate. HT29 (human colon carcinoma) cells were main-
tained in DMEM supplemented with 10% foetal bovine serum
and 10 mM sodium bicarbonate. HCT116 (human colon car-
cinoma) cells were maintained in RPMI 1640 (Trace Bios-
ciences, Australia) supplemented with 10% foetal bovine
serum. ADDP (cisplatin resistant A2780) cells were maintained
in RPMI supplemented with 5% foetal bovine serum. All culture
media was further supplemented with penicillin (100 IU/mL),
streptomycin (100 mg/mL), and glutamine (4 mM). The doubling
time for each cell line was 14 h for HCT116, 20±24 h for HT29
cells, 13±15 h 143B cells, 20±24 h for A2780, and 14±16 h for
ADDP cells. Cells were passaged every 3±7 days and all cell
lines were routinely tested and found to be mycoplasma free.
Cytotoxicity assay. Cells in logarithmic growth were transferred
to 96-well plates. Cytotoxicity was determined by plating cells in
triplicate in 100 mL medium at a density of 2500±3500 cells/
well for all cell lines. On day 0 (24 h after plating) when the cells
were in logarithmic growth, 100 mL medium with or without the
test agent was added to each well. After drug exposure growth
inhibitory e€ects were evaluated using the MTT (3-[4,5-dime-
thyltiazol-2-yl] 2,5-diphenyl-tetrazolium bromide) assay and
absorbance read at 540 nm (Alley, M. C.; Scudiero, D. A.;
Monks, A.; Hursey, M. L.; Czerwinski, M. J.; Fine, D. L.;
Abbott, B. J.; Mayo, J. G.; Shoemaker, R. H.; Boyd, M. R.
Cancer Res. 1998, 48, 589). The IC50 was the drug concentration
at which cell growth is 50% inhibited based on the di€erence
between the optical density values on day 0 and those at the
end of drug exposure (Bergman, A. M.; Ruiz van Haperen, V.
W.; Veerman, G.; Kuiper, C. M.; Peters, G. J. Clin. Cancer
Res. 1996, 2, 521).
108 110 ꢀC. H NMR (CDCl3) d 1.2 (t, 3H), 1.5 (m, 2H), 1.7
1
(m, 2H), 2.97 (q, 4H), 4.0 (q, 2H), 4.8 (d, 1H), 4.9 (d, 1H); 13
C
NMR (CDCl3) d 14.75, 29.64, 29.68, 52.97, 61.83, 79.06, 79.3,
171.86, 176.86. A typical synthesis of 8 is as follows: hydroxy-
lactone (4) (0.85 g, 0.005 mol) was dissolved in ethanol (10
mL), and a catalytic quantity of p-TosOH (ca. 20 mg) was
added and the mixture was then re¯uxed for 2 h. After cooling,
the solvent was removed in vacuo, and the resulting oil taken up
in chloroform (20 mL) and washed successively with saturated
NaHCO3 (2Â10 mL); water (10 mL), dried over Na2SO4.
Removal of the solvent (in vacuo) yielded upon standing an
o€-white solid (0.723 g, 73% as a mixture of diasteromers).
Mp 50±52 ꢀC. H NMR (CDCl3) d 1.2 (m, 3H), 1.5 (m, 2H),
1
1.7 (m, 2H), 2.4 (d, 1H), 2.8 (d, 1H), 3.5±3.8 (m, 2H), 4.6 (d,
1H), 4.8 (d, 1H), 5.2 (s, 1H). 13C NMR (CDCl3) d 15.0, 28.5,
29.3, 51.0, 51.7, 66.0, 80.2, 80.9, 107.0, 176.0.
12. The use of distilled or reagent grade alcohols result in the
formation of 4. Additionally, 4 appears to be the major product
isolated when the reaction is conducted in higher alcohols; e.g.,
hexanol.
13. Protein phosphatase assays were carried out essentially as
described (Collins, E.; Sim, ATR. Methods Mol. Biol. 1998,
93, 79±102) using [32P]-glycogen phosphorylase a as substrate
and recombinant PP1 (Bernt, N. Methods Mol. Biol. 1998, 93,
67±78) or partially puri®ed (chicken skeletal muscle) PP2A
catalytic subunits (Mackintosh, C. In Protein Phosphorylation:
A Practical Approach; Hardie, D. G., Ed.; IRL, 1993). Brie¯y,
enzyme activity was measured at 30 ꢀC in a bu€er (®nal volume
of 30 mL) containing 50 mM Tris±HCl (pH 7.5), 1 mM EGTA,
0.1 mM EDTA, 5 mM ca€eine, 0.1% mercaptoethanol, 0.3 mg/
mL BSA. The concentration of PP1 or PP2A used was such
that the reaction was limited to 15% dephosphorylation to
ensure linearity. The reaction was started with the addition of
30 mg [32P]-glycogen phosphorylase a and terminated after 20
min by the addition of 100 mL ice-cold 70% TCA. After 10 min