Kinesin spindle protein (KSP) inhibitors. Part 7: Design and synthesis of 3,3-disubstituted dihydropyrazolobenzoxazines as potent inhibitors of the mitotic kinesin KSP

Article abstract of DOI:10.1016/j.bmcl.2007.07.067

Observations from two structurally related series of KSP inhibitors led to the proposal and discovery of dihydropyrazolobenzoxazines that possess ideal properties for cancer drug development. The synthesis and characterization of this class of inhibitors

Full text of DOI:10.1016/j.bmcl.2007.07.067

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5671–5676
Kinesin spindle protein (KSP) inhibitors. Part 7: Design
and synthesis of 3,3-disubstituted dihydropyrazolobenzoxazines
as potent inhibitors of the mitotic kinesin KSP
Robert M. Garbaccio,a,* Edward S. Tasber,a Lou Anne Neilson,a Paul J. Coleman,a
Mark E. Fraley,a Christy Olson,a Jeff Bergman,a Maricel Torrent,a Carolyn A. Buser,b
Keith Rickert,b Eileen S. Walsh,b Kelly Hamilton,b Robert B. Lobell,b Weikang Tao,b
Vicki J. South,b Ronald E. Diehl,b Joseph P. Davide,b Youwei Yan,c Lawrence C. Kuo,c
Chunze Li,d Thomayant Prueksaritanont,d Carmen Fernandez-Metzler,d
Elizabeth A. Mahan,d Donald E. Slaughter,d Joseph J. Salata,e Nancy E. Kohl,d
Hans E. Huberd and George D. Hartmana
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bDepartment of Cancer Research, Merck Research Laboratories, West Point, PA 19486, USA
cDepartment of Structural Biology, Merck Research Laboratories, West Point, PA 19486, USA
dDepartment of Drug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
eDepartment of Cellular Safety, Merck Research Laboratories, West Point, PA 19486, USA
Received 30 May 2007; revised 17 July 2007; accepted 18 July 2007
Available online 21 August 2007
Abstract—Observations from two structurally related series of KSP inhibitors led to the proposal and discovery of
dihydropyrazolobenzoxazines that possess ideal properties for cancer drug development. The synthesis and characterization of this
class of inhibitors along with relevant pharmacokinetic and in vivo data are presented. The synthesis is highlighted by a key [3+2]
cycloaddition to form the pyrazolobenzoxazine core followed by diastereospecific installation of a quaternary center.
ꢀ 2007 Elsevier Ltd. All rights reserved.
In the preceding communication, the design, synthesis,
and characterization of allosteric dihydropyrazole-based
inhibitors of KSP (kinesin spindle protein) as novel anti-
mitotics were described.1 This effort resulted in the iden-
tification of water-soluble inhibitors (e.g., 1, Fig. 1) that
demonstrated pharmacokinetics suitable for iv adminis-
tration and excellent in vivo potency. In addition, these
compounds were designed to avoid Pgp-efflux, a poten-
tial mechanism for chemotherapeutic resistance in the
clinic.
resolution of enantiomers afforded the (3S,3aS) cis
enantiomer as the active component (KSP
2
IC50 = 800 nM) (Fig. 1). Due to the poor solubility dis-
played by the chromenopyrazole series, a novel series of
dihydropyrazolobenzoxazines (e.g., 3) was envisioned
by simple rearrangement of the pyrazole nitrogens and
incorporation of the 8-fluorine substitution that was
found to enhance potency in prior efforts.1 Upon syn-
thesis, kinesin spindle protein inhibitor (KSPi) 3 was
found to have good potency (KSP IC50 = 155 nM) and
improved physical properties compared to 2.
In parallel to these developments, a structurally related
chromenopyrazole lead was identified by HTS and upon
Drawing from reported advances in KSP inhibitor
design within our group, we sought to incorporate the
weakly basic amine tethered to the dihydropyrazole2
(e.g., 1) into these dihydropyrazolobenzoxazine cores
(3 ! 4) with the goal of improving potency and solubil-
ity while maintaining activity in a P-glycoprotein (Pgp)
overexpressing cell line.
Keywords: KSP inhibitors; Anti-mitotics; Multidrug-resistance; Pgp;
hERG; Kinesins; Pyrazolobenzoxazines; [3+2] Cycloaddition; Quater-
nary center; Kinesin spindle protein; Cancer.
*
Corresponding author. Tel.: +1 215 652 5391; fax: +1 215 652
6345; e-mail: robert_garbaccio@merck.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.07.067

5672
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5671–5676
sodium-nitrite in the presence of ethyl 2-chloro-3-oxo-
butanoate in variable (65–92%) but good yield.
NAc
N
F
O
In the key [3+2] intramolecular cycloaddition of the
cinnamyl olefin with an in-situ formed nitrile imine,4
the major product (40% isolated yield) was found to
be the benzoxadiazene (12, Scheme 2) that arises from
an undesired competitive Claisen rearrangement of the
cinnamyl ether onto the fluoroarene, followed by attack
of the revealed phenol onto the chloroimine.4 The
desired trans-dihydropyrazolobenzoxazine 10 was iso-
lated only in 35% yield under optimized conditions. This
stands in contrast to the synthetic route that initiated
with cis-3-cinnamyl chloride and provided the desired
cis-tricycle 14 in a superior 65% yield without observa-
tion of this undesired reaction pathway (Scheme 3). Pre-
sumably, the cis-isomer cannot easily access the
transition state required for the sigmatropic rearrange-
ment. The conclusion from investigation of these two
routes was that the extra two steps required for the syn-
thesis of cis-cinnamyl chloride5 result in a doubling of
yield for the key [3+2] cycloaddition step. Both trans-
and cis-esters (10 and 14) were directly transformed to
the Weinreb amides (11 and 15) that would act as the
alkylation substrate for stereospecific installation of
the quaternary center.6
F
N
N
N
N
O
O
Me
Me
2 KSP IC50 = 800 nM
1 KSP IC50 = 3 nM
Hit to Lead
Known SAR
NAc
N
O
N
O
N
Lead Opt.
F
F
N
N
O
O
Me
Me
3 KSP IC50 = 155 nM
4 Dihydropyrrolobenzoxaxzine
Figure 1. HTS hit 2, lead 3 and related dihydropyrazole 1 that led to
dihydropyrazolobenzoxazine proposal 4.
A key observation made in the synthesis of the initial
dihydropyrazolobenzoxazine 3 was a troublesome epi-
merization seen at C-4 under basic conditions. This
observation suggested that formation of a quaternary
center through anion chemistry would provide syn-
thetic access to the proposed target. This report ad-
dresses the outcome of this proposal, and the full
characterization of the KSP inhibitors produced in
this effort.
Alkylation at C4 of either trans-11 or cis-15 gave, as ex-
pected, the same result (Scheme 4). Deprotonation with
LiHMDS occurred cleanly at ꢀ78 ꢁC and reaction of the
stabilized lithium anion with allyl bromide occurred
with high trans-diastereospecificity (>20:1) to provide
the quaternary center displayed in 16 in 69% isolated
yield. These results are consistent with approach of
the electrophile via the less-hindered face opposite to
the methylene-oxy bridge. Elaboration of the olefin to the
desired propyl amine and conversion of the Weinreb
amide to the potent methyl ketone remained as the final
synthetic hurdles. Hydroboration of the terminal olefin
yielded the primary alcohol 17 that was then treated
with methyllithium under optimized conditions. The de-
sired methyl ketone 18 was isolated in 72% yield, and the
major side product was found to be demethoxylation of
the amide (15%) as has been reported elsewhere.7 Oxida-
tion of the primary alcohol to the aldehyde 19 with
The synthetic route3 to the dihydropyrazolobenzoxazine
tricycle initiated with alkylation of 4-fluoro-2-nitrophe-
nol 5 with trans-3-cinnamyl chloride (6, Scheme 1) to
give the corresponding ether 7 in excellent yield after
recrystallization. The choice of the olefin geometry dic-
tates the relative stereochemical outcome for C4/C5,
but given the plans to deprotonate the C4 center, the
decision to employ the trans-isomer was based on its
commercial availability. Following efficient reduction
of the nitro arene 7 by zinc metal, oxidation and
conversion to the crystalline hydrazono-acetate
(Japp–Klingemann reaction) was accomplished with
9
O
CO2Et
OH
O
(a)
(d)
(b)
(3.0)
O
Cl
Cl
(c)
F
NO2
F
F
NO2
F
NH2
5
6
7
8
O
O
N
(e)
O
N
N
F
Me
H
N
N
HCl
F
NH
N
O
O
OMe
CO2Et
EtO
10
Me
N
9
(3.0)
Cl
OMe
11
Scheme 1. General synthesis of the pyrazolobenzoxazine (trans-route). Reagents and reaction conditions: (a) K2CO3, DMF, 65 ꢁC, 86%; (b) Zn
metal, NH4Cl, MeOH, 55 ꢁC, 89%; (c) NaNO2, HCl/AcOH/H2O, 65–92%; (d) Et3N, toluene, 110 ꢁC, 35%; (e) Me3Al, CH2Cl2, 88%.

R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5671–5676
5673
40%
O
OH
O
CO2Et
F
NH
N
N
CO2Et
N
F
N
H
F
N
H
CO2Et
Cl
Cl
12
O
O
Ph
O
Ph
_
F
N
N
F
NH
N
F
N
N
35%
CO2Et
+
O
CO2Et
Cl
EtO
10
9
Scheme 2. Preferred side reaction in key [3+2] cycloaddition.
O
same cond.
Cl
Ph
CO2Et
F
NH
N
as (c)
above
65%
Cl
13
O
O
N
same cond.
F
N
F
as (e)
above
N
N
O
O
OMe
EtO
14
Me
15
N
Scheme 3. General synthesis of the pyrazolobenzoxazine (cis-route).
OH
O
N
O
N
O
F
F
(c)
F
N
N
(a)
(b)
N
N
O
O
O
N
N
N
OMe
OMe
OMe
17
11 or 15
16
NAc
O
OH
N
O
O
O
O
N
HN
(e)
(d)
F
N
N
F
N
N
F
N
N
O
O
O
Me
Me
19
Me
rac-20
18
Scheme 4. General synthesis of 20. Reagents and reaction conditions: (a) LHMDS, allylbromide, 69%, >20:1 de; (b) 9-BBN-H, H2O2, THF, 50 ꢁC,
75%; (c) MeLi, THF, 0 ꢁC, 72%; (d) Dess–Martin periodinane, CH2Cl2; (e) Na(OAc)3BH, DCE, Et3N, 80% two step yield.
Dess–Martin periodinane followed by reductive amina-
tion with N-acetyl piperazine gave the racemic final
product 20 in 80% overall yield for the two steps. The
final dihydropyrazolobenzoxazine 20 was resolved into
its constituent enantiomers using a Chiracel OD column
(5 cm column, 100% MeOH, 1 g/injection, RT = 11.6

5674
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5671–5676
and 15.5 min, respectively) with the first peak to elute
containing the potent enantiomer. Importantly, 20 was
found to have aqueous solubility in excess of 4 mg/mL
at pH 4 and its absolute stereochemistry was assigned
based on prior work, and subsequently confirmed by a
X-ray crystallographic structure bound to KSP.8
reduced basicity (20, 24–27, 30, and 31) generally
showed weaker binding, whereas strongly basic amines
(21–23, 28, and 29) engendered a stronger interaction
with this ion channel. Similar trends were seen for the
MDR ratio, wherein amines with reduced basicity, and
tertiary amines tended to have improved potency
against the Pgp-overexpressing cell line, and thus lower
‘MDR ratios’. Some direct comparisons are worth not-
ing; the comparison of 21 to 28 and 20 to 29 illustrated
that capping of the amines resulted in a dramatic reduc-
tion in the MDR ratio. These results are in line with the
observation that modulating the pKa of this basic nitro-
gen or its substitution can balance Pgp-efflux potential
with maintenance of KSP potency.
A targeted set of analogs was synthesized by this method
taking advantage of the SAR described in earlier efforts.
The composite data presented in Table 1 include po-
tency against the KSP enzyme,9 mitotic arrest activity
(Cell EC50) in A2780 human ovarian carcinoma cells,10
hERG binding,11 and an MDR ratio. The MDR ratio
is expressed as the ratio of cell potency between paired
cell lines that have high and low expression of PGP,
respectively, and these data have been used extensively
in this program as a surrogate measure of Pgp suscepti-
bility.12 A compound with an MDR ratio < 10 is desir-
able in our view, and a value of ꢁ1 is ideal.2
Representative pharmacokinetic data from dog are pro-
vided for several compounds in Table 2. In general,
these pyrazolobenzoxazines were found to have moder-
ate to high clearance in dog using a cassette dosing strat-
egy. Given its favorable balance of properties (potency,
MDR ratio, hERG, and PK), compound 20 was selected
for further study.
Consistent with earlier SAR in related series, substitu-
tion at R1 with F, Cl or Me resulted in similar potency
and MDR ratios (compounds 20–27). In addition, po-
tency was not greatly altered by the identity of the R2
amine tethered by the propyl chain at C4. However, sub-
stitution at R2 had significant impact on hERG binding
and the MDR ratio, and was thus used to optimize these
properties. With respect to hERG binding, amines with
In the design of these novel KSP inhibitors, efforts have
been made to minimize binding to hERG with the hope
of avoiding acute cardiotoxicity that would be dose-lim-
iting. While hERG binding is an imperfect predictor for
in vivo QT prolongation, it is one of the few available
Table 1. Potency, MDR ratio, and hERG binding
R2
O
N
R1
N
O
R1
R2
MDRa ratio
a
KSP IC50 (nM)
a,b
Cell EC50 (nM)
a
hERG IC50 (nM)
Compound
21c
22c
23c
F
1.6
1.0
1.8
7.4
4.8
6.4
2.3
2.5
2.3
5940
2000
2400
Cl
Me
NMe2
24
F
0.4
1.8
1.0
4.4
13
0.8
1.5
1.4
6100
3300
6300
25c
26
Cl
Me
N
O
11
20
27
F
Cl
1.6
1.0
5.0
9.0
1.0
1.6
10,000
5600
N
NAc
28
18
F
F
0.5
5.2
2.5
11
1.7
0.6
940
NH2
OH
18,000
29
30
31
F
F
F
4.3
1.5
0.9
5
40
4600
17,000
9600
N
N
N
NH
O
6.9
4.8
3.3
2.6
NH
O
NMe
9–12
a See Refs.
b EC50 for mitotic arrest (G2/M block) in A2780 cell line.
c Tested as a racemate; All values are reported for n = 3 or greater with a standard deviation within 25–50% of the reported value.
for the details of these assays; all compounds tested as single enantiomers unless noted.

R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5671–5676
5675
Table 2. Dog cassette data
R2
O
R1
N
N
O
R1
R2
Dog pharmacokineticsa
CL (mL/min/kg)
a
KSP IC50 (nM)
Compound
t1/2 (h)
21b
23b
F
1.6
1.8
5.7
52
37
NMe2
Me
8.3
20
29
30
31
F
F
F
F
1.6
5.2
1.5
0.9
5.6
14
40
15
11
N
N
N
N
NAc
NH
3.5
2.4
3.9
O
NH
O
NMe
a Dosed iv as a cassette DMSO solution. Average of two dogs dosed at each of 0.25 mg/kg iv. The individual determinations are within 25% of the mean.
b Tested as a racemate.
Table 3. Summarized data for 20
Compound 20 was evaluated in our in vivo mouse xeno-
graft tumor model of mitotic arrest.14 In this pharmaco-
dynamic model, an aqueous solution of 20 was
administered to nude mice bearing A2780 xenograft
implants via Alzet osmotic pumps for 24 h. The sectioned
O
N
N
O
tumors were stained for phospho-histone H3, an induc-
ible marker of mitotic arrest. KSPi 20 was found to induce
F
N
maximal mitotic block with a plasma exposure of 100 nM.
Based on the similar free fraction of 20 in mouse plasma
(3%) and human plasma (3%), this plasma concentration
may also produce a similar level of pharmacodynamic
potency in human. Intra-tumor levels of compound 20
were not measured in this experiment.
N
20
O
Property
20
KSP IC50 (nM)
Cell EC50 (nM)
1.6
5.0
>4
Through the combination of known SAR in two struc-
turally related series, dihydropyrazolobenzoxazines were
proposed, synthesized, and found to be highly potent,
water-soluble inhibitors of KSP that retained full
potency against a Pgp-overexpressing cell line. By utiliz-
ing existing SAR and discovering a robust synthetic
route highlighted by a diastereoselective installation of
a quaternary center, compound 20 was identified as
having appropriate potency and pharmacokinetics to
further characterize. In a key finding, 20 was shown to
have diminished functional interaction with the hERG
channel as recorded in a patch-clamp assay relative to
another optimized compound (1) despite demonstrating
tighter binding to the channel. These data prompted
examination of 20 in the cardiovascular dog model
and were found not to cause QTc prolongation at a
plasma concentration 350-fold above the concentration
that produced a maximal in vivo pharmacodynamic
endpoint in a mouse xenograft model. Based on these
Solubility (pH 4, mg/mL)
MDR ratio
In vivo mitotic arrest EC90 (nM)
1.0
100
high throughput assay techniques.11 Interestingly, we
noted that structurally similar compound 1 was found
to induce a QTc prolongation in a canine cardiovascular
model that was not predicted by hERG binding
(>30,000 nM), but corroborated only by measurement
of IKr current blockade in a classical patch-clamp assay.
Importantly, while 20 demonstrated measurable affinity
to hERG (10,000 nM), it did not produce functional
blockade in the same patch-clamp study.13
As a critical follow-up to these measurements, com-
pound 20 was evaluated for its ability to prolong QTc
in vivo in anesthetized dogs. Compound 20 showed no
effect on QTc interval when dosed up to 35 lM plasma
levels.

5676
R. M. Garbaccio et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5671–5676
EC50 are reported as the average of at least two indepen-
dent determinations; standard deviations are within 25–
50% of EC50 values.
properties, 20 represents a highly optimized lead with
ideal properties for a cancer drug (Table 3).15 In the next
letter, efforts to explore a potent structurally related
series of KSP inhibitors are described.
11. The hERG IC50 values are reported as the average of at
least two independent determinations and were acquired
by radioligand binding competition experiments using
membrane preparations from human embryonic kidney
cells that stably express hERG; standard deviations are
within 25–50% of reported values. For assay details, see:
Bilodeau, M. T. et al. J. Med. Chem. 2004, 47, 6363, and
references therein.
Acknowledgments
We thank Dr. Chuck Ross for high resolution mass
spectral analysis, Dr. Sandor Varga for NMR identifica-
tion of 12, and Mr. Carl Homnick for chiral separations.
12. Relative to the parental KB-3-1 cells, KB-V-1 cells,
originally derived by culturing KB-3-1 cells in the presence
of the Pgp substrate vinblastine (J. Biol. Chem. 1986, 261,
7762), express >500-fold higher levels of Pgp mRNA and
protein. The compound potency (IC50) for induction of
mitotic arrest was determined by evaluating the levels of
the mitotic marker phospho-nucleolin after 16 h incuba-
tion with the test compound in an 11-point half-log
dilution series. The ratio of IC50 obtained in KB-V-1 cells
vs. that in KB-3-1 cells is defined as the MDR ratio. As a
general guideline, we considered compounds with MDR
ratios < 10 to be of interest for their ability to enter and
kill cells that overexpress Pgp. Verapamil, a competitive
inhibitor of Pgp, restores the activity of Taxol and our
KSP inhibitors in the KB-V-1 cell line to nearly that
observed in the parental KB-3-1 line, confirming that Pgp-
efflux is responsible for the observed resistance to drug-
mediated mitotic arrest. The absolute values of potency
between the A2780 and KB cell lines were similar (within
2· of each other).
Supplementary data
Full experimental details for the synthesis of 20 are pro-
vided as an attachment to this manuscript. Supplemen-
tary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2007.07.067.
References and notes
1. Preceding communication.
2. Cox, C. D.; Breslin, M. J.; Whitman, D. B.; Coleman, P.
J.; Garbaccio, R. M.; Fraley, M. E.; Zrada, M. M.; Buser,
C. A.; Walsh, E. S.; Hamilton, K.; Lobell, R. B.; Tao, W.;
Abrams, M. T.; South, V. J.; Huber, H. E.; Kohl, N. E.;
Hartman, G. D. Bioorg. Med. Chem. Lett. 2007, 17, 2697.
3. All compounds were characterized by H NMR and high
resolution mass spectrometry. For detailed experimental
procedures, see the supplementary material.
1
13. Wang, J.; Della Penna, K.; Wang, H.; Karczewski, J.;
Connolly, T. M.; Koblan, K. S.; Bennett, P. B.; Salata, J.
J. Am. J. Physiol. 2003, 284, H256.
14. Athymic nude mice (nu/nu) were xenografted subcuta-
neously with the human ovarian carcinoma cell line
A2780 and the resulting tumors were allowed to reach
200–300 mm3 before mice were surgically implanted with
Alzet mini-pumps (Durect Corporation) filled with the
appropriate KSP inhibitors according to manufacturer’s
recommendations. Prior to the implant, pumps contain-
ing KSP inhibitors were primed by incubation in 37 ꢁC
water bath for 3 h, so that the pumps would discharge
4. Garanti, L.; Sala, A.; Zecchi, G. J. Org. Chem. 1977, 42,
1389, The undesired Claisen-rearrangement was also
observed in this report as a side product.
5. Synthesized in two steps from 3-phenyl-2-propyn-1-ol.
Lindlar reduction of the alkyne provided the cis-olefin.
Conversion of the alcohol to the chloride was accom-
plished using methanesulfonylchloride, 2,6-lutidine, and
LiCl.
6. Levitt, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun.
1982, 12, 989.
KSP inhibitors at
a constant rate of 8 ll/h after
subcutaneous implantation. Mice were euthanized 22 h
post pump implant and blood and tumors were
harvested. Blood was collected in EDTA Vacutainers
and processed for plasma to determine pharmacokinet-
ics. Tumors were fixed in 10% neutral buffered formalin
and then processed and embedded in paraffin. Paraffin
embedded tumors were sectioned 5 lm thick and used in
7. Sibi, M. P.; Sharma, R.; Paulson, K. L. Tetrahedron Lett.
1992, 33, 1941.
8. PDB coordinates for 20: 2Q2Y.
9. KSP inhibitory activity was measured using a standard
ATPase assay. IC50 values are reported as averages of at
least two determinations; standard deviations are 25–50%.
10. Mitotic arrest was measured by assessing the mitosis-
specific phosphorylation of nucleolin using an antibody
coated, bead-based assay. In this assay, total nucleolin is
captured on a streptavidin-coated paramagnetic bead
coupled with biotinylated nucleolin monoclonal IgG1
antibody 4E2 (Research Diagnostics, Inc.). Nucleolin
phosphorylation is detected by an antibody complex
consisting of a phospho-specific nucleolin IgM monoclo-
nal antibody, TG3 (Applied NeuroSolutions, Inc.) and a
goat anti-mouse IgM labeled with a ruthenium Tris–
bipyridyl complex (BV-TAG Technology, BioVeris
Corp.). The complex is quantitated via electrochemilumi-
nescence based on the excitation/emission properties of the
Tris–bipyridyl complex using a BioVeris M-series ana-
lyzer. EC50 for KSP-induced nucleolin phosphorylation
was determined after treatment with a compound in a
7-point, half-log dilution series for 16 h. The values of
a
phospho-histone H3 immunohistochemistry assay
designed to determine the percentage of cells blocked
in mitosis compared to control treated tumors. After
paraffin removal, re-hydration, and antigen retrieval,
sections were incubated with anti-phospho-histone H3
(ser10; Upstate). Following incubation with a biotinyl-
ated secondary antibody, staining of antigen positive
nuclei was accomplished using avidin:biotin complexed
horseradish peroxidase and diaminobenzidine reagent.
Sections were imaged using an Olympus BX51 micro-
scope with a motorized stage and Image-Pro Analysis
software. Quantization of the sections was accomplished
by measuring the percentage of positively stained nuclei
(black) per unit area.
15. Compound 20 and its analogs are also described in
WO2006023083.