In vitro and in vivo antiherpetic activity of three new synthetic brassinosteroid analogues

Article abstract of DOI:10.1016/j.steroids.2004.04.011

Brassinosteroids are a novel group of steroids that appear to be ubiquitous in plants and are essential for normal plant growth and development. It has been previously reported that brassinosteroid analogues exert an antiviral activity against herpes simplex virus type 1 (HSV-1) and arenaviruses. In the present study, we report the chemical synthesis of compounds (22S,23S)-3β-bromo-5α,22,23-trihydroxystigmastan-6-one (2), (22S,23S)-5α-fluoro-3β-22,23-trihydroxystigmastan-6-one (3), (22S,23S)-3β,5α,22,23-tetrahydroxy-stigmastan-6-one (4) as well as their antiherpetic activity both in a human conjunctive cell line (IOBA-NHC) and in the murine herpetic stromal keratitis (HSK) experimental model. All compounds prevented HSV-1 multiplication in NHC cells in a dose dependent manner when added after infection with no cytotoxicity. Administration of compounds 2, 3, and 4 to the eyes of mice at 1, 2, and 3 days post-infection delayed and reduced the incidence of HSK, consisting mainly of inflammation, vascularization, and necrosis, compared to untreated, infected mice. However, viral titers of eye washes showed no differences among samples from treated and untreated mice. Since the decrease in the percentage of mice with ocular lesions occurred 5 days after treatment had ended, we suggest that brassinosteroids 2, 3, and 4 did not exert a direct antiviral effect in vivo, but rather may play a role in immune-mediated stromal inflammation, which would explain the improvement of the clinical signs of HSK observed.

Full text of DOI:10.1016/j.steroids.2004.04.011

Steroids 69 (2004) 713–720
In vitro and in vivo antiherpetic activity of three new synthetic
brassinosteroid analogues
a
b
c
´
Flavia M. Michelini , Javier A. Ramırez , Alejandro Berra ,
b
a,∗,1
´
Lydia R. Galagovsky , Laura E. Alche
a
Laboratorio de Virolog´ıa, Departamento de Qu´ımica Biolo´gica, Facultad de Ciencias Exactas y Naturales-UBA,
Ciudad Universitaria, Pabello´n II-4to. piso, 1428 Buenos Aires, Argentina
b
Departamento de Qu´ımica Orga´nica, Facultad de Ciencias Exactas y Naturales, Ciudad Universitaria,
Pabello´n II, 1428 Buenos Aires, Argentina
Laboratorio de Inmunopatolog´ıa Ocular, Facultad de Medicina Universidad de Buenos Aires,
c
J.E. Uriburu 950, E.P., 1113 Buenos Aires, Argentina
Received 25 November 2003; received in revised form 26 April 2004; accepted 29 April 2004
Available online 1 October 2004
Abstract
Brassinosteroidsareanovelgroupofsteroidsthatappeartobeubiquitousinplantsandareessentialfornormalplantgrowthanddevelopment.
It has been previously reported that brassinosteroid analogues exert an antiviral activity against herpes simplex virus type 1 (HSV-1) and
arenaviruses. In the present study, we report the chemical synthesis of compounds (22S,23S)-3␤-bromo-5␣,22,23-trihydroxystigmastan-6-one
(2), (22S,23S)-5␣-fluoro-3␤-22,23-trihydroxystigmastan-6-one (3), (22S,23S)-3␤,5␣,22,23-tetrahydroxy-stigmastan-6-one (4) as well as their
antiherpetic activity both in a human conjunctive cell line (IOBA-NHC) and in the murine herpetic stromal keratitis (HSK) experimental model.
All compounds prevented HSV-1 multiplication in NHC cells in a dose dependent manner when added after infection with no cytotoxicity.
Administration of compounds 2, 3, and 4 to the eyes of mice at 1, 2, and 3 days post-infection delayed and reduced the incidence of HSK,
consisting mainly of inflammation, vascularization, and necrosis, compared to untreated, infected mice. However, viral titers of eye washes
showed no differences among samples from treated and untreated mice. Since the decrease in the percentage of mice with ocular lesions
occurred 5 days after treatment had ended, we suggest that brassinosteroids 2, 3, and 4 did not exert a direct antiviral effect in vivo, but rather
may play a role in immune-mediated stromal inflammation, which would explain the improvement of the clinical signs of HSK observed.
© 2004 Elsevier Inc. All rights reserved.
Keywords: Brassinosteroids; HSV-1; Antiviral activity; Human conjunctive cell line; Murine herpetic stromal keratitis; Steroid
1. Introduction
We recently reported that some natural BRs and synthetic
analogues have in vitro antiviral activity against several
pathogen viruses [5,6]. Among them, natural compounds
(22R,23R)-2␣,3␣,22,23-tetrahydroxy-5␣-stigmastan-6-one
(28-homocastasterone) (1a, Fig. 1) and (22R,23R)-
2␣,3␣,22,23-tetrahydroxy-B-homo-7-oxa-stigmastan-6-one
(28-homobrassinolide) (1b, Fig. 1) showed selectivity
indices (SI) of 2 and 28, respectively, against herpes simplex
virus type 1 (HSV-1) multiplication in Vero cells. However,
synthetic analogues with 22S,23S configuration 2, 3, and 4
(Fig. 1) showed higher SI (100, 109, and 80, respectively)
[7].
Brassinosteroids (BRs) are recognized as a new class
of plant steroidal hormones that control many of the same
developmental and physiological processes as their insect
and mammalian counterparts, including regulation of gene
expression, cell division, differentiation, and homeostasis
[1–4].
∗
Corresponding author. Tel.: +54 11 4576 3334; fax: +54 11 4576 3342.
E-mail address: lalche@qb.fcen.uba.ar (L.E. Alche´).
Research Member of the National Research Council of Argentina (CON-
1
ICET).
0039-128X/$ – see front matter © 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2004.04.011

714
F.M. Michelini et al. / Steroids 69 (2004) 713–720
Fig. 1.
HSV-1 induces an ocular disease in humans named her-
2.1.2. (22E)-3␤-Bromo-5␤,6␤-epoxystigmast-22-ene (6)
A mixture of potassium permanganate (2 g, 12.65 mmol)
and ferric nitrate nonahydrate (1 g, 2.47 mmol) was ground to
a fine powder, and water (0.1 ml) was added. To a stirred sus-
pension of this mixture in methylene chloride (20 ml), (22E)-
3␤-bromostigmasta-5,22-diene [12] (5, 1.94 g, 4.08 mmol)
and t-butyl alcohol (2 ml) were added. After 2 h at room tem-
perature, the reaction was completed, and the product was
separated from the inorganic residue by adding ether (50 ml),
stirring for 5 min, and filtering through Celite. After evapo-
ration of the solvent, the crude product was purified by col-
umn chromatography (CH2Cl2/ethyl acetate 1:1) yielding 6
(73%). mp (acetone): 93 ◦C.
petic stromal keratitis (HSK), which is a significant cause
of ocular morbidity around the world leading to vision im-
pairment and blindness [8,9]. Studies of animal models have
been very useful for the characterization of the evolution of
HSK. Within one to two weeks, HSV-1 corneal infection in
mice initiates a series of events that culminate in a stromal
necrotizing keratitis, histologically similar to that of HSK
[10]. Furthermore, it resembles many characteristics of the
recrudescent disease in humans, both in the duration of virus
shedding and with respect to the effect of antiviral drugs [11].
In this paper, we present the chemical synthesis of com-
pounds 2-4 and investigate their antiherpetic activity both in
a human conjunctive cell line and in the murine HSK exper-
imental model.
1H NMR (CDCl3): 0.67 (18-H3, 3H, s), 0.88 (27-H3, 3H,
d, J = 6 Hz), 0.95 (26-H3, 3H, d, J = 6 Hz), 0.96 (29-H3,
3H, d, J = 7 Hz), 1.00 (19-H3, 3H, s), 1.02 (21-H3, 3H, d, J
= 7 Hz), 3.10 (6␣-H, 1H, d, J = 2 Hz), 4.66 (3␣-H, 1H, m),
5.02 (23-H, 1H, dd, J = 15 Hz, 8 Hz), 5.15 (22-H, 1H, dd, J
= 15 Hz, 8 Hz).
2. Experimental
2.1. Synthesis of compounds
13C NMR (CDCl3): 12.1 (C18), 12.1 (C29), 16.9 (C19),
19.1 (C26), 20.9 (C11), 21.2 (C27), 22.0 (C21), 24.3 (C15),
25.3 (C28), 29.7 (C16), 30.0 (C8), 31.9 (C25), 32.8 (C7),
33.6 (C2), 35.4 (C10), 39.9 (C1), 40.1 (C12), 40.3 (C20),
42.5 (C13), 44.3 (C4), 51.3 (C9), 51.8 (C24), 53.4 (C3), 56.4
(C14), 56.5 (C17), 63.4 (C6), 68.2 (C5), 129.7 (C23), 137.9
(C22).
2.1.1. General
All compounds were recrystallized until constant melt-
ing point (mp). Melting point were determined on a Fisher
Johns apparatus and are uncorrected. H and 13C spectra
1
were recorded on a Bruker AM-500 spectrometer at 500 and
125 MHz, respectively. Chemical shifts (δ) are given in ppm
downfield from TMS as the internal standard. 19F NMR spec-
tra were recorded on a Bruker AM-500 at 470.4 MHz, and
chemical shifts (δ) are given in ppm upfield from CFCl3 as the
internal standard. Coupling constant (J) values are in Hz. Un-
less otherwise indicated, all solvents and reagents used were
of commercial grade. FAB spectra (low and high resolution)
were recorded on a VG ZAB-T instrument (VG-Micromass)
using 3-NBA as matrix. Low resolution mass spectra (EI)
were recorded on a Shimatzu QP-5000 at 70 eV. Reactions
were monitored by TLC on plates precoated with silicagel
F254 0.2 mm (Merck). Column chromatography was carried
out on silica gel 60, 0.04–0.063 mm (Merck).
MS (EI) m/z (%): 492 ([M]+, 81Br, 4); 490 ([M]+, 79Br, 4);
411 (4); 269 (9); 55 (100).
HRMS (EI): Calculated for C29H47BrO: 490.2810; found:
490.2808.
2.1.3. (22E)-3␤-Bromo-5-hydroxy-5␣-stigmast-22-en-6-ona
(7)
Jones reagent (0.5 ml) was added dropwise to a stirred
solution of compound 6 (359 mg, 0.73 mmol) in acetone
(25 ml). After 1 h, isopropyl alcohol was added to stop the
reaction. The mixture was neutralized with solid sodium bi-
carbonate, filtered, and the solvent was evaporated under re-
duced pressure. The crude residue was purified by column

F.M. Michelini et al. / Steroids 69 (2004) 713–720
715
chromatography (hexane/ethyl acetate 9:1) to afford com-
pound 7 [13] in 83% yield. mp (methanol): 200 ◦C.
1H NMR (CDCl3): 0.68 (18-H3, 3H, s), 0.86 (19-H3, 3H,
s), 0.88 (27-H3, 3H, d, J = 6 Hz), 0.95 (26-H3, 3H, d, J =
6 Hz), 0.96 (29-H3, 3H, d, J = 7 Hz), 1.02 (21-H3, 3H, d, J
= 7 Hz), 4.32 (3␣-H, 1H, m), 5.02 (23-H, 1H, dd, J = 15 Hz,
8 Hz), 5.15 (22-H, 1H, dd, J = 15 Hz, 8 Hz).
ride/ethyl acetate gradient) afforded 3 in 83% yield. mp
(methanol/water): 137 ◦C.
1H NMR (CDCl3/CD3OD 9:1): 0.68 (18-H3, 3H, s), 0.88
(27-H3, 3H, d, J = 6.7 Hz), 0.90 (19-H3, 3H, s), 0.95 (26-
H3, 3H, d, J = 6.7 Hz), 0.96 (29-H3, 3H, d, J = 6.8 Hz), 1.02
(21-H3, 3H, d, J = 6.8 Hz), 2.58 (7␣-H, 1H, dd, J = 12.5 Hz,
12.5 Hz), 3.62 (22-H and 23-H, 2H, m), 3.88 (3␣-H, 1H, m).
13C NMR (CDCl3/CD3OD 9:1): 12.2 (C18), 13.7 (C21),
13.8 (C19, d, JCF = 6.2 Hz), 14.1 (C29), 17.3 (C26), 18.3
(C28), 21.1 (C27), 21.2 (C11), 24.0 (C15), 26.6 (C25), 28.6
(C16), 30.0 (C1), 30.8 (C2), 33.9 (C4, d, JCF = 22.2 Hz),
37.8 (C8), 39.4 (C12), 42.1 (C20), 42.7 (C7), 42.9 (C13),
43.2 (C10, d, JCF = 14.7 Hz), 45.6 (C9), 48.4 (C24), 56.0
(C14), 56.5 (C17), 66.9 (C3), 70.1 (C23), 71.7 (C22), 101.3
(C5, d, JCF = 176.9 Hz), 207.2 (C6, d, JCF = 27.2 Hz).
13C NMR (CDCl3): 12.3 and 12.3 (C18 and C29), 14.1
(C19), 19.1 (C26), 21.1 (C11), 21.2 and 21.3 (C21 and C27),
24.0 (C15), 25.4 (C28), 28.7 (C16) 32.0 (C25), 32.2, 32.9
(C1 and C2), 37.4 (C8), 38.8 (C12), 39.5 (C4), 40.4 (C20),
42.0 (C7), 42.3 and 43.1 (C10 and C13), 44.7 (C9), 48.0 (C3),
51.3 (C24), 56.0 and 56.5 (C14 and C17), 81.0 (C5), 129.6
(C23), 138.9 (C22), 211.1 (C6).
MS (EI) m/z (%): 508 ([M]+, 81Br, 2); 506 ([M]+, 79Br, 2);
427 (1); 43 (100).
HRMS (EI): Calculated for C29H47BrO2: 506.2759;
found: 506.2754.
19F NMR (CDCl3/CD3OD 9:1): −161.1 (d, JFH
=
45.0 Hz).
MS (FAB) m/z (%): 481 ([M + H]+, 19); 463 (20); 305
(21); 154 (100).
2.1.4. (22S,23S)-3␤-Bromo-5,22,23-trihydroxy-5␣-
stigmastan-6-one (2)
HRMS (FAB): Calculated for C29H50FO4 [M + H]+:
481.3694; found: 481.3693.
A mixture of 7 (1.63 g, 3.20 mmol), THF (20 ml), t-
butanol/water (1:1) (70 ml), pyridine (25 ␮l), potassium fer-
ricyanide (3.2 g, 9.7 mmol), potassium carbonate (1.34 g,
10.5 mmol), and potassium osmate dihydrate (50 mg,
0.13 mmol) was stirred at room temperature for 9 days [14].
An excess of solid sodium bisulfite was added until no evolu-
tion of bubbles was observed. Layers were separated, and
the aqueous phase was thoroughly extracted with methy-
lene chloride/methanol (5%). The combined organic layers
were dried and evaporated, and the crude solid was puri-
fied by column chromatography (hexane/ethyl acetate gradi-
ent) to give compound 2 (89% yield). mp (methanol/water):
132 ◦C.
2.1.6. (22S,23S)-3␤,5,22,23-Tetrahydroxy-5␣-
stigmastan-6-one (4)
Compound 12 [15] (150 mg, 0.34 mmol) was dihydrox-
ylated in a similar way as described for compound 7.
Purification by column chromatography (methylene chlo-
ride/ethyl acetate gradient) afforded 4 in 86% yield. mp
(methanol/water): 168 ◦C.
1H NMR (CDCl3/CD3OD 9:1): 0.67 (18-H3, 3H, s), 0.82
(19-H3, 3H, s), 0.88 (27-H3, 3H, d, J = 6.7 Hz), 0.95 (26-
H3, 3H, d, J = 6.7 Hz), 0.96 (29-H3, 3H, d, J = 6.8 Hz), 1.02
(21-H3, 3H, d, J = 6.8 Hz), 2.58 (7␣-H, 1H, dd, J = 12.6 Hz,
12.5 Hz), 3.60 (22-H and 23-H, 2H, m), 3.95 (3␣-H, 1H, m).
13C NMR (CDCl3/CD3OD 9:1): 11.8 (C18), 13.6 (C19),
13.6 (C21), 14.0 (C29), 17.3 (C26), 18.3 (C28), 20.7 (C11),
21.0 (C27), 23.6 (C15), 26.6 (C25), 28.3 (C16), 29.5 (C2),
30.1 (C1), 35.7 (C4), 37.0 (C8), 39.1 (C12), 41.6 (C7), 42.0
(C13), 42.0 (C20), 2.6 (C10), 43.9 (C9), 48.4 (C24), 55.5
(C17), 55.8 (C14), 66.1 (C3), 70.0 (C23), 71.6 (C22), 79.7
(C5), 212.8 (C6).
1H NMR (CDCl3/CD3OD 9:1): 0.68 (18-H3, 3H, s), 0.84
(19-H3, 3H, s), 0.88 (27-H3, 3H, d, J = 6.7 Hz), 0.95 (26-H3,
3H, d, J = 6.7 Hz), 0.96 (29-H3, 3H, d, J = 6.8 Hz), 1.02 (21-
H3, 3H, d, J = 6.8 Hz), 2.71 (7␣-H, 1H, t, J = 12.6 Hz), 3.60
(22-H and 23-H, 2H, m), 3.34 (3␣-H, 1H, m).
13C NMR (CDCl3/CD3OD 9:1): 11.3 (C18), 13.6 (C19),
13.6 (C21), 14.1 (C29), 17.3 (C26), 18.3 (C28), 21.0 (C27),
21.6 (C11), 23.9 (C15), 26.6 (C25), 27.7 (C16), 31.9 (C2),
32.7 (C1), 37.2 (C8), 38.2 (C4), 39.3 (C12), 41.5 (C7), 42.0
(C10), 42.0 (C20), 43.4 (C13), 44.1 (C9), 48.4 (C24), 49.4
(C3), 52.3 (C17), 55.8 (C14), 70.0 (C23), 71.7 (C22), 80.1
(C5), 213.3 (C6).
MS (FAB) m/z (%): 479 ([M + H]+, 40); 461 (100); 447
(17).
HRMS (FAB): Calculated for C29H51O5 [M + H]+:
479.3737; found: 479.3720.
MS (FAB) m/z (%): 541 ([M + H]+, 4); 523 (9); 505 (2);
154 (100).
2.2. Antiviral assays
HRMS (FAB): Calculated for C29H50BrO4 [M + H]+:
541.2892; found: 541.2885.
2.2.1. Cells, viruses, and treatment solutions
Vero and IOBA-NHC cells [16] were grown in Eagle’s
minimal essential medium supplemented with 5% inactivated
calf serum (MEM 5%) and 50 ␮g/ml gentamycin, and main-
tained in monolayer formation in MEM supplemented with
1.5% inactivated calf serum (MEM 1.5%).
2.1.5. (22S,23S)-5-Fluoro-3␤,22,23-trihydroxy-5␣-
stigmastan-6-one (3)
Compound 10 [15] (198 mg, 0.44 mmol) was dihydrox-
ylated in a similar way as described for compound 7.
Purification by column chromatography (methylene chlo-
HSV-1 Cgal+ (HSV-1 Cgal) KOS strain was propagated
at low multiplicity, plaque-assayed on Vero cell monolayers,

716
F.M. Michelini et al. / Steroids 69 (2004) 713–720
and used for in vitro assays. This system contains the Es-
cherichia coli lacZ reporter gene encoding ␤-galactosidase,
which is under the control of a strong heterologous promoter
from the human cytomegalovirus immediate-early (IE) reg-
ulatory region [17], and was kindly provided by Dr. Alberto
cells were fixed with formaldehide 10%, stained with Crystal
Violet, and plaque forming units were counted.
2.2.4. Measurement of ␤-gal activity
IOBA-NHC cell monolayers grown in coverslips inside
24-well plates were infected with HSV-1 Cgal (m.o.i. = 0.2).
After incubation for 1 h at 37 ◦C, inocula were eliminated,
and cells were covered with MEM 1.5%. At 24 h p.i., super-
natants were harvested and stored at −70 ◦C for titration, and
cells were stained in situ for ␤-gal. A solution containing 1%
formaldehyde and 0.2% glutaraldehyde in PBS (200 ␮l/well)
was added for 30 min at 4 ◦C. Subsequently, supernatants
were removed, and cells were washed twice with PBS. A ␤-
gal activity staining solution (50 mM potassium ferrocyanide,
50 mM potassium ferricyanide, 20 mM MgCl2, and 1 mg/ml
X-gal) was added to the cells for 3 h at 37 ◦C. Coverslips
were mounted, and ‘blue’ cells were observed with an Olym-
pus BX61 microscope, with a slider slot for Normarski DIC,
and photographed.
´
Epstein (Universite Claude Bernard, Lyon, France).
The HSV-1 KOS strain (wt) was also propagated at low
multiplicity and was used for in vivo experiments. Com-
pounds 2-4 were dissolved in ethanol and diluted with PBS
or culture medium for testing. The maximum concentration
of ethanol tested (1%) exhibited no toxicity under the in vitro
or in vivo conditions.
2.2.2. Cytotoxicity assay
Cell viability in the presence of the compounds was
determined using the cleavage of the tetrazolium salt
MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide) (Sigma) by the mitochondrial enzyme succinate
dehydrogenase to give a blue product (formazan) [18]. The
absorbance of each well was measured on an Eurogenet-
ics MPR-A 4i microplate reader using a test wavelength of
570 nm and a reference wavelength of 630 nm. Results were
expressed as a percentage of absorbance of treated cell cul-
tures with respect to untreated ones. The CC50 was defined as
the concentration of compound that caused a 50% reduction
in absorbance.
2.2.5. Animals
Five to seven-week old male Balb/c mice were purchased
from the I.N.T.A. (Castelar, Buenos Aires). All mice were
handled according to the Animal Care Guidelines from the
National Institute of Health (USA) and the Association for
Research in Vision and Ophthalmology (ARVO, USA) reso-
lution on the use of animals in research.
2.2.3. Virus titration
2.2.6. Corneal inoculation of HSV-1
Verocellsgrownin24-wellplateswereinfectedwithserial
10-fold dilutions of viral yields and incubated for 1 h at 37 ◦C.
Residual inocula were replaced by maintenance medium with
0.7% of methylcellulose. After 72 h of incubation at 37 ◦C,
Mice were anaesthetized with 2 mg intraperitoneal ke-
tamine (Ketalar, Parke-Davis, Morris Plains, NJ) and xy-
lazine (Rompun, Mobay, Shanee, KA). The right cornea of
Scheme 1. Synthesis of compound 2. Reagents and conditions: (a) KMnO4/Fe(NO3)3·9H2O/t-BuOH/H2O/CH2Cl2, 2 h, r.t. (b) Jones’ reagent/acetone, 1 h, r.t.
(c) K2OsO4/K4Fe(CN)6/K2CO3/pyridine/t-BuOH/H2O/THF, 9 days, r.t.

F.M. Michelini et al. / Steroids 69 (2004) 713–720
717
each mouse was scratched eight times in a criss-cross pattern
with a 27 gauge needle, and 5 ␮l of an HSV-1 suspension
containing 2.5 × 104 PFU of virus were instilled in the cul-
de-sac. Brassinosteroids were topically administered three
times a day at 1, 2, and 3 days post-infection (p.i.) (volume
= 5 ␮l). The control mice received PBS.
The severity ofHSKwasdetermined undera binocular mi-
croscope at 8 and 10 days p.i. The diagnosis of disease was
based on the density of the inflammatory infiltrates, neovas-
cularization, and the presence of superficial or deep ulcers.
Thus, mice exhibiting puffy eyelids with moderate to severe
crusting (blepharitis), vessel enlargement with corneal inva-
sion (10% or more) (neovascularization), and slight to severe
corneal turbidity/opacity, with or without signs of corneal
necrosis or perforation, were considered as ill (stromal ker-
atitis).
of the oxirane ring of 6 with Jones’ reagent gave (22E)-
3␤-bromo-5-hydroxy-5␣-stigmast-22-en-6-one (7) [13]. The
side chain double bond of 7 was dihydroxylated with cat-
alytic osmium (VIII) using ferric ferricyanide as cooxidant
[14]. No chiral ligand was necessary due to the presence
of the C-24 ethyl group, which induced the desired stere-
oselectivity, yielding 89% of (22S,23S)-3␤-bromo-5,22,23-
trihydroxy-5␣-stigmastan-6-one (2).
3. Results and discussion
Synthesis of compound 2 is summarized in Scheme 1.
Compound 5 ((22E)-3␤-bromostigmasta-5,22-diene) was
obtained in three steps from stigmasterol [12]. Regio
and stereoselective epoxidation of compound 5 with
KMnO4/Fe(NO3)3 [19] yielded 73% of the corresponding
5␤,6␤ epoxide 6. Hydrolytic opening and further oxidation
Fig. 2. Cytotoxicity and antiviral activity of compounds 2, 3, and 4 in vitro.
IOBA-NHC cells grown in 96-well plates were treated with 1–314 ␮M of 2
(ꢀ), 3 (ꢁ), and 4 (᭹) for cytotoxicity assays (right axis). Following 24 h of
incubation at 37 ◦C, a MTT colorimetric assay was performed. IOBA-NHC
cells grown in 24-well plates were infected with HSV-1 Cgal (m.o.i. = 0.2)
and mock-treated or treated with 0.1–20 ␮M of 2 (ꢂ), 3 (ꢀ), and 4 (ꢁ) (left
axis). After 24 h of incubation at 37 ◦C, free and cell-associated virus was
collected and titrated in Vero cells.
Scheme 2. Synthesis of compounds 3 and 4. Reagents and conditions:
(a) Jones’ reagent/acetone, 30 min, r.t. (b) K2CO3/MeOH, 4 h, r.t. (c)
K2OsO4/K4Fe(CN)6/K2CO3/pyridine/t-BuOH/H2O/THF, 9 days, r.t. (d)
HF-Py/CH2Cl2, 10 min, 0 ◦C. (e) PCC/HPy+CF3COO−, 2 h, r.t.

718
F.M. Michelini et al. / Steroids 69 (2004) 713–720
The synthesis of compounds 3 and 4 is shown in Scheme 2.
ent non-growing cells in concentrations ranging from 1 to
314 ␮M, and after 24 h of incubation at 37 ◦C, a MTT col-
orimetric assay was performed. As shown in Fig. 2, brassi-
nosteroids 2, 3, and 4 exhibited CC50 values of >277, >314,
and 146 ␮M, respectively, showing cytotoxicity levels for hu-
man conjunctival cells even lower than those corresponding
to Vero cells, except for compound 4, as previously reported
[7].
To evaluate the antiviral activity of these compounds,
IOBA-NHC monolayers grown in 24-well plates were in-
fected with HSV-1 Cgal (m.o.i. = 0.2) and then, treated with
different concentrations of 2, 3, and 4 (from 0.1 to 20 ␮M) or
control media. After 24 h of incubation at 37 ◦C, free and cell-
associated virus were collected and titrated in Vero cells. A
dose-dependent inhibition of viral replication was observed
in drug-treated cultures, regardless of the compound assayed
(Fig. 2). We established that all compounds prevented HSV-1
Cgal multiplication in human conjunctival cells when added
after infection, exhibiting 50% inhibition of virus yield at
concentrations of 1, 4, and 5 ␮M for 2, 3, and 4 (Fig. 2),
respectively, with SI of >277, >78.5, and 29.2, respectively.
Stigmasteryl acetate was converted into its 5␤,6␤ epoxide 8
as previously described [19]. Nucleophylic opening of this
epoxide with a hydrofluoric acid–pyridine complex (Olah’s
reagent) afforded compound 9; the proper stereochemistry
at C-5 was achieved by trans-diaxial acid catalyzed open-
ing of the oxirane ring. Subsequent oxidation at C-6 and
hydrolysis of the acetate in C-3 yielded (22E)-5-fluoro-3␤-
hydroxy-5␣-stigmast-22-en-6-one (10) [15]. Dihydroxyla-
tion of the ꢀ22 double bond with osmium (VIII) allowed
for the synthesis of (22S,23S)-5-fluoro-3␤,22,23-trihydroxy-
5␣-stigmastan-6-one (3) from stigmasterol with a 45% yield.
Alternatively, epoxide 8 was treated with Jones’ reagent to af-
ford the 5␣-hydroxy-6-keto compound 11 [15]. Hydrolysis of
11yieldedcompound12. Dihydroxylationofthesidechainof
12 gave (22S,23S)-3␤,5,22,23-tetrahydroxy-5␣-stigmastan-
6-one (4).
Anti-HSV-1 activity of brassinosteroid analogues 2, 3, and
4 were tested in human conjunctival cells. First, the 50%
cytotoxic concentration (CC50) for IOBA-NHC cells was
determined. Compounds 2, 3, and 4 were added to conflu-
Fig. 3. Effect of brassinosteroids 2, 3, and 4 on HSV-1 propagation. IOBA-NHC cell monolayers grown in coverslips inside 24-well plates were infected with
HSV-1 Cgal (m.o.i. = 0.2) and incubated for 1 h at 37 ◦C. Then, cells were covered with MEM 1.5% or media containing 20 ␮M of compounds 2, 3, and 4 per
well. After incubation for 24 h at 37 ◦C, cells were stained in situ for ␤-gal. (A) HSV-1 infected -untreated cells; (B) HSV-1 infected 2-treated cells; (C) HSV-1
infected 3-treated cells; and (D) HSV-1 infected 4-treated cells. Magnification: 40×.

F.M. Michelini et al. / Steroids 69 (2004) 713–720
719
Only brassinosteroid 2 exhibited a higher SI than that ob-
tained in Vero cells [7].
of 10 mice each were infected with HSV-1 KOS strain in the
right eye. One group was mock-treated with PBS (control),
whereas the others were treated topically with a concentra-
tion of 40 ␮M of compounds 2, 3, and 4 three times a day at
1, 2, and 3 days p.i.
Considering that brassinosteroids inhibited HSV-1 Cgal
multiplication in human conjunctival cells, we investigated
whether viral propagation was restrained by these com-
pounds. For this purpose, IOBA-NHC cell monolayers grown
in coverslips inside 24-well plates were infected with HSV-
1 Cgal (m.o.i. = 0.2) and incubated for 1 h at 37 ◦C. Then,
inocula were eliminated, and cells were covered with MEM
alone or MEM 1.5% containing 20 ␮M of each compound per
well. After incubation for 24 h at 37 ◦C, supernatants were
collected and titrated on Vero cells, and human conjunctival
cells were stained in situ for ␤-gal. In the absence of brassi-
nosteroids, ‘blue’ cells clustered in characteristic HSV-1 foci
were observed (Fig. 3A), whereas cell cultures treated with
any of the three compounds exhibited only scattered ‘blue’
cells (Fig. 3B–D). Therefore, brassinosteroids impeded HSV-
1 Cgal propagation in IOBA-NHC cells. Likewise, the inhibi-
tion of HSV-1 Cgal multiplication in supernatants from cells
treated with a 20 ␮M concentration of 2, 3, and 4 were 92.6,
92.6, and 99.9%, respectively, when titrated in Vero cells,
with respect to untreated infected cells.
Prompted by the anti-HSV-1 activity of compounds 2, 3,
and 4 in cultured cells from conjunctival tissue, we decided
to evaluate their ability to heal HSK in a murine experimental
model. First, a toxicity assay was performed. Daily observa-
tion of groups of five mice each that received compounds 2, 3,
and 4 three times a day during 3 consecutive days confirmed
no toxic effect at a concentration of 40 ␮M. To determine
if treatment with compounds 2, 3, and 4 after viral inocula-
tion affected the clinical signs of murine HSK, four groups
Signs of ocular disease that gradually developed from 6 to
7 days of infection onwards were observed until day 14 p.i. in
all groups (data not shown). The percentage of animals with
HSK at 8 and 10 days p.i. was registered (Fig. 4), and the
percentages were compared using the χ2-test, with 90% con-
fidence intervals. At the onset of disease (day 8 p.i.), 70% of
untreated, infected mice exhibited lesions of keratitis, which
increased to 80% after 10 days of infection. By day 8 p.i.,
compound 2 significantly restrained the evolution of HSK,
since only 30% of animals had signs of keratitis. A healing
effect of compounds 3 and 4 was less evident. Although the
percentage of sick animals decreased to 50% at 10 days p.i.,
no significant differences with respect to untreated infected
animals were observed (Fig. 4). Among mice with keratitis,
those evolving to necrosis showed no significant differences
in the incidence of this lesion between untreated and treated
infected animals, regardless of the compound administered
(data not shown).
To determine whether compounds 2, 3, and 4 inhibited
HSV-1 multiplication in vivo, the eyes of untreated and
treated infected mice were gently washed with a 10 ␮l drop
of MEM 1.5%, at 1, 3, 6, and 9 days after infection. Each
sample was diluted 1:10 and titrated in Vero cells. No sig-
nificant differences in viral titers among samples from un-
treated and treated animals were detected, indicating that
none of the compounds exerted an antiviral effect in-vivo
(Fig. 5).
It is remarkable that the decrease in the percentage of mice
with ocular lesions occurred 5 days after the treatment with
Fig. 4. Effect of brassinosteroids 2, 3, and 4 on the incidence of HSV-1
induced-ocular disease in mice. Forty adult Balb/c mice were inoculated
with 105 UFP/5 ␮l of HSV-1 KOS strain in their corneas. Three groups of 10
mice each received 40 ␮M of 2, 3, or 4 three times a day for 3 consecutive
days, beginning at 1 day p.i. The remaining animals (10) were used as a
control for HSV-1 infection. Signs of HSK (blepharitis, neovascularization,
corneal clouding, edema, irritation and necrosis) were assessed for 10 days.
Fig. 5. Effect of brassinosteroids 2, 3, and 4 on HSV-1 multiplication in the
eye. The eyes of infected adult Balb/c mice untreated and treated with 40 ␮M
of 2 (ꢀ), 3 (ꢀ), and 4 (ꢁ) were gently washed with a 10 ␮l drop of MEM
1.5% at 1, 3, 6, and 9 days after infection. To quantify viral infectivity, each
sample diluted 1:10 was titrated in Vero cells. HSV-1 (ꢂ).

720
F.M. Michelini et al. / Steroids 69 (2004) 713–720
2, 3, and 4 had ended, suggesting that brassinosteroids did
not exert a direct antiviral effect, as shown in Fig. 5. It is sus-
pected that multiple immunopathological mechanisms are at
play during the development of HSK and that early lesions
are principally the consequence of CD4+ T cells recogniz-
ing virally derived peptides [20]. Since intact HSV is rarely
found in the cornea beyond day 6 p.i., it seems likely that the
persistence of incomplete viral antigens and the viral immune
alteration of cell membranes in the corneal stroma incite in-
flammatory reactions [21]. In fact, keratocytes in the corneal
stroma of the mouse can function as antigen-presenting cells
[22]. Despite the lack of anti-HSV activity in vivo, the fact
thatcompounds2, 3, and4improvedtheclinicalsignsofHSK
led us to consider that these compounds may play a role in
immune-mediated stromal inflammation. Further studies to
elucidate an eventual antiinflammatory effect of compounds
2, 3, and 4 are in progress.
[6] Wachsman MB, L
o´pez E, Ram´ırez JA, Galagovsky LR, Coto CE.
Antiviral effect of brassinosteroids against herpes virus and are-
naviruses. Antiviral Chem Chemother 2000;11:71–7.
[7] Talarico LB, Ram´ırez JA, Galagovsky LR, Wachsman MB.
Structure–activity relationship studies in a set of new brassinosteroid
derivatives assayed against herpes simplex virus type 1 and 2 in cell
cultures. Med Chem Res 2002;11:434–44.
[8] Zhao Z-S, Granucci F, Yeh L, Schaffer PA, Cantor H. Molecular
mimicry by herpes simplex virus-type 1: autoimmune disease after
viral infection. Science 1998;279:1344–7.
[9] Carr JJD, Ha¨rle P, Gebhardt BM. The immune response to ocular her-
pes simplex virus type 1 infection. Exp Biol Med 2001;226:353–66.
[10] Liu T, Tang Q, Hendricks RL. Inflammatory infiltration of the trigem-
inal ganglion after herpes simplex virus type 1 corneal infection. J
Virol 1996;70:264–71.
[11] Hendricks RL. An immunologist’s view of herpes simplex keratitis:
Thygeson Lecture. Presented at the Ocular Microbiology and Im-
munology Group Meeting October 26, 1996. Cornea 1997;16:503–6.
[12] Ponce MA, Ram
´ırez JA, Galagovsky LR, Gros EG, Erra-Balsells R.
Photooxygenation of different 3␤-substituted stigmastanes by using
singlet oxygen. A mechanistic approach. J Chem Soc Perkin Trans
2000;2:2351–7.
[13] Ram´ırez JA, Teme Centurio´n OM, Gros EG, Galagovsky LR. Syn-
thesis and bioactivity evaluation of brassinosteroid analogs. Steroids
2000;65:329–37.
Acknowledgements
[14] Kolb HC, VanNieuwenhze MS, Sharpless KB. Catalytic asymmetric
dihydroxylation. Chem Rev 1994;94:2483–547.
[15] Ram´ırez JA, Gros EG, Galagovsky LR. Effects on bioactivity due
to C-5 heteroatom substituents on synthetic 28-homobrassinosteroid
analogs. Tetrahedron 2000;56/34:6171–80.
This work was supported by grants from the Universi-
dad de Buenos Aires, UBA X-051 and Agencia Nacional de
´
´
´
Promocion Cientıfica y Tecnica (ANPCYT) PICT 6594/99.
Mass spectrometry was provided by the Washington Uni-
versity Mass Spectrometry Resource, an NIH Research Re-
source (Grant No. P41RR0954).
[16] Diebold Y, Calonge M, Enr
´ıquez de Salamanca A, Callejo S, Cor-
rales R, Sa´ez V, et al. Characterization of a spontaneously immor-
talized cell line [IOBA-NHC cells] from normal human conjunctiva.
Invest Ophthalmol Vis Sci 2003;44:4263–74.
[17] Johnson PA, Best MG, Friedmann T, Parris DS. Isolation of a herpes
simplex virus type 1 mutant deleted for the essential UL42 gene and
characterization of its null phenotype. J Virol 1991;65:700–10.
[18] Denizot F, Lang R. Rapid colorimetric assay for cell growth and
survival. J Immunol Methods 1986;89:271–7.
[19] Syamala MS, Das J, Baskaran S, Chandrasekaran S. A novel and
highly selective ␤-epoxidation of ꢀ5- unsaturated steroids with per-
manganate ion. J Org Chem 1992;57:1928–30.
References
[1] Grove MD, Spencer GF, Rohwedder WK, Mandava N, Worley JF,
Warthen Jr JD, et al. Brassinolide, a plant growth-promoting steroid
isolated from Brassica napus pollen. Nature 1979;281:216–8.
[2] Bajguz A, Tretyn A. The chemical characteristic and distribution of
brassinosteroids in plants. Phytochemistry 2003;62:1027–46.
[3] Khripach VA, Zhabinskii VN, de Groot AE, editors. Brassinos-
teroids: a new class of plant hormones. San Diego: Academic Press;
1999.
[20] Deshpande SP, Zheng M, Lee S, Rouse BT. Mechanisms of patho-
genesis in herpetic immunoinflammatory ocular lesions. Vet Micro-
biol 2002;86:17–26.
[4] Thummel C, Chory J. Steroid signalling in plants and
insects—common themes, different pathways. Genes Dev
2002;16:3113–29.
[5] Wachsman MB, Ramirez JA, Galagovsky LR, Coto CE. Antiviral
activity of brassinosteroids derivatives against measles virus in cell
cultures. Antiviral Chem Chemother 2002;13:61–6.
[21] Holbach L, Font R, Naumann G. Herpes simplex stromal and en-
dothelial keratitis. Ophthalmology 1990;97:722–8.
[22] Seo SK, Gebhardt BM, Lim HY, Kang SW, Higaki S, Varnell ED,
et al. Murine keratocytes function as antigen-presenting cells. Eur J
Immunol 2001;31:3318–28.