Monoclonal antibody mediated intracellular targeting of tallysomycin S 10b

Article abstract of DOI:10.1016/j.bmcl.2004.05.089

The potency of tallysomycin S_10b (TLM S_10b) an analogue bleomycin was enhanced by up to 875-fold when it was conjugated to the internalizing antibody BR96. Attachment to the antibody is achieved via a Cathepsin B cleavable linker. The enhancement in potency is believed to be a result of cellular uptake of the conjugate upon antigen binding followed by rapid release of the drug inside the lysosome. This method provides a novel approach for increasing the potency and therapeutic index of nominally moderately-active cytotoxic agents.

Full text of DOI:10.1016/j.bmcl.2004.05.089

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4323–4327
Monoclonal antibody mediated intracellular targeting
of tallysomycin S10b
Michael A. Walker,* H. Dalton King, Richard A. Dalterio, Pamela Trail,
Raymond Firestone and Gene M. Dubowchik
Bristol–Myers Squibb Pharmaceutical Research Institute, Medicinal Chemistry, PO Box 5100, 5 Research Parkway,
Wallingford, CT 06492, USA
Received 15 September 2003; revised 26 May 2004; accepted 27 May 2004
Available online 26 June 2004
Abstract—The potency of tallysomycin S10b (TLM S10b) an analogue bleomycin was enhanced by up to 875-fold when it was
conjugated to the internalizing antibody BR96. Attachment to the antibody is achieved via a Cathepsin B cleavable linker. The
enhancement in potency is believed to be a result of cellular uptake of the conjugate upon antigen binding followed by rapid release
of the drug inside the lysosome. This method provides a novel approach for increasing the potency and therapeutic index of
nominally moderately-active cytotoxic agents.
Ó 2004 Elsevier Ltd. All rights reserved.
Monoclonal antibodies (mAbs) hold great promise as
drug delivery agents in the treatment of cancer.1 The
goals in this area are twofold, reduce exposure of sen-
sitive organs and tissue to the drug (referred to as neg-
ative targeting) and enhance the exposure of the tumor
and metastatic foci (referred to as positive targeting).
The negative and positive aspects of antibody targeting
combined yield an increase in the apparent therapeutic
index of the targeted drug. This is an especially useful
technique for reducing the toxicity associated with po-
tent cytotoxic agents. An example of the state of the art
in this area is the immunoconjugate BMS-182248. This
agent is composed of the anticancer agent doxorubicin
(Dox) linked to the tumor selective antibody BR96.
BR96 recognizes a Ley related antigen primarily asso-
ciated with carcinomas of the colon, breast, ovary, and
lung but not normal tissues.2 As such BMS-182248 is
capable of completely curing xenotransplanted tumors
in mice and rats, syngenic tumors in rats and is active
against various models of disseminated cancer.3 More
importantly, animal models show that the heart, which
is sensitive to doxorubicin, is spared the full dose of the
drug when delivered as the conjugate.
In the current disclosure we describe one additional
feature of certain antibody based antitumor agents be-
yond the negative and positive targeting aspects men-
tioned above that has yet to be exploited namely, mAb
mediated intracellular targeting.4 The idea is to use an
internalizing antibody to increase the potency of a drug,
which is poorly taken up by the cell but, which has high
intrinsic activity against an intracellular target. In this
case the drug by itself has a high maximum tolerated
dose due to poor uptake by normal cells so that negative
targeting is built in a priori provided that the carrier is
not taken up by normal cells.
The cornerstone of this approach is BR96, which is
endocytosed upon binding its antigen unlike most anti-
bodies, which normally remain bound to the cell sur-
face.5 This has allowed us to devise a method for
selective drug release at the tumor by taking advantage
of the proteolytic environment of the lysosome (or
endosome) to effect hydrolysis of the linker. Thus we
have found that dipeptide based linkers allow for
selective release of the drug in the lysosomal environ-
ment and that an optimal linker is the dipeptide, Phe-
Lys, which is recognized and cleaved by Cathepsin B.6
In order to insure that the drug does not interfere with
enzyme cleavage of the linker a p-aminobenzyl alcohol
(PABC) spacer is inserted between the peptide linker and
the drug. The PABC moiety is designed to hydrolyti-
cally decompose upon enzymatic deacylation releasing
Keywords: Immunoconjugate; Antitumor; Lysosome; Cathepsin B;
Prodrug; Antibody; BR96.
* Corresponding author. Tel.: +1-203-677-6686; fax: +1-203-677-7702;
e-mail: michael.a.walker@bms.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.05.089

4324
M. A. Walker et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4323–4327
the free drug.7 Cathepsin B is a cysteine protease that is
present in all mammalian cell lysosomes, does not differ
very much from species-to-species, and is rarely found
extracellularly except in the vicinity of certain patho-
logical conditions such as metastatic sites, or in the areas
of tissue breakdown in rheumatoid arthritis.8;9 BR96
conjugates employing this linker and cytotoxic drugs
such as doxorubicin, mitomycin C, and camptothecin
are potent and selective antitumor agents.10;11
The peptide linker was assembled as shown in Scheme 1.
The commercially available succinimide ester of Fmoc-
protected phenylalanine (2) was coupled to Boc-pro-
tected lysine under basic conditions. The dipeptide 4 was
then functionalized at the C- and N-termini with the
p-aminobenzyl alcohol spacer and maleimidocaproic
acid to yield 7. Following this the linker was activated
toward coupling with TLM S10b as the corresponding
p-nitrophenyl carbonate 8.
The drug selected for this investigation is tallysomycin
S10b (TLM S10b), which belongs to the same family of
cytotoxic agents as bleomycin. The mechanism of action
for these agents involves Fe(II) mediated cleavage of
DNA.12 In preclinical models TLM S10b exhibited
activity similar to that of bleomycin against a variety of
murine tumors (P388 leukemia, B16 melanoma, Lewis
lung carcinoma, and Madison 109 lung carcinoma).13
Unfortunately, in two phase II clinical trials it yielded
no response in previously treated head and neck cancer
patients and colorectal cancer patients while at the same
time it exhibited a toxicity spectrum similar to bleomy-
cin.14 Therefore, TLM S10b represents a perfect candi-
date for intracellular targeting since an improvement in
potency will likely increase its therapeutic index.15
The Cu(II) complex of the bis-formate salt of TLM S10b
(TLM S10bÆCu(II)Æ2HCO2H) has two potential sites for
linker attachment as indicated in structure 1. The
remaining amino groups, highlighted with asterisks, are
involved in metal complexation and are thus unreactive
toward acylating agents. In fact, previous reports have
established that the amino group of the 1,4-diamino-
butane side chain of this compound is the primary
reaction site for coupling with acylating agents.17
Nonetheless, regioselective linker attachment is not an
issue for our purposes since this bond is cleaved at the
tumor.
In order to avoid over acylation linker attachment
conditions were first examined using di-tert-butyl di-
carbonate (Boc2O). Prior to reaction the formate salt
was converted to the corresponding HCl salt by treat-
ment with the ion exchange resin AGâ 2-X8.18 The
resulting salt, TLM S10bÆCu(II)Æ2HCl, was treated with
It is believed that the potency of this family of cytotoxic
agents is attenuated by poor cell penetration thus
restricting access of the drug to its target. A recent
publication provides support for this notion by showing
that the activity of bleomycin can be increased by up to
10,000 fold by using electroporation to enhance cellular
uptake.16 It is assumed that the activity of TLM S10b
would be enhanced to the same degree under similar
conditions. One additional feature of TLM S10b is that it
is water soluble, thus reducing the possible formation of
conjugate aggregation at high drug/mAb ratios.
€
1.0 equiv of Boc2O and 2.0 equiv of Hunigs base. Two
products were observed by HPLC and MS consistent
with mono- and bis-acylation compounds Boc-TLM
S10bÆCu(II)ÆHCl (9a) and (Boc)2ÆTLM S10bÆCu(II) (9b). In
addition, ion exchange chromatography (Econo-Pac S,
0.5–1.0 M HCO2NH4) showed three peaks eluting in
order and having the approximate statistical ratio one
would expect for a mixture of mono-acylated, bis-acyl-
ated and unreacted TLM S10b. Fortunately it was
found that formation of the bis-acylated product
could be suppressed by using 0.5 equiv of Boc2O.
Under these conditions the Boc-TLM S10bÆCu(II)ÆHCl
was favored by 5:1 over (Boc)2ÆTLM S10bÆCu(II)
(Scheme 2).
In the current publication we illustrate the synthesis of a
TLM S10b BR96 conjugate and demonstrate that the
resulting conjugate yields a substantial increase in the
potency for the attached drug presumably by mAb
mediated intracellular targeting (Fig. 1).
O
*
NH2
O
NH2
NH
NH2 • HCO2H
N
H
N
4
NH2
S
O
•Cu(II)
N
N
N
O
*
S
H
N
HO
O
H2N
HO
O
N
H
O
HN
O
O
N
H
HO
*
O
N
*
OH
O
OHOH
N
H
HCO2H • NH2
OH
OH
O
O
OH
O
OH
O
NH2
1 Tallysomycin S10b, Cu(II), (HCO2H)2
Figure 1. Tallysomycin S10b Cu(II)(CHCO2H)2 (*Cu(II) binding atoms).

M. A. Walker et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4323–4327
4325
BocHN
O
BocHN
BocHN
O
OH
+
H
N
a
b
N
FmocHN
O
O
FmocHN
N
H
FmocHN
N
N
H
CO2H
H2N
CO2H
O
OH
O
5
2
3
4
BocHN
O
O
O
H
O
O
O
H
c
N
N
N
N
N
H
O
5
6
O
O
O
O
R
O
7 R = H
8 R =
d
O
O
NO2
Scheme 1. Regents and conditions: (a) NaHCO3, 1,2-dimethoxyethane, H2O, (b) di-tert-butyl dicarbonate, p-aminobenzyl alcohol, pyridine, DMF,
(c) (1) Et2NH, CH2Cl2, (2) 6, i-Pr2NEt, CH2Cl2, (d) bis-(p-nitrophenyl)-carbonate, i-Pr2NEt, DMF.
AG2-X8 (Cl-)
.
.
.
.
TLM S10b Cu(II) (HCl)2
TLM S10b Cu(II) (HCO2H)2
a,b,c or d
.
.
.
.
.
.
TLM S10b Cu(II) (HCl)2
TLM S10b Cu(II) (-C(O)R)
TLM S10b Cu(II) (-C(O)R)2
+
a
b
BocHN
O
O
H
N
H
N
N
N
H
5
R =
R =
O
O
O
O
O
9
10
Scheme 2. Reagents and conditions: (a) 1.0 equiv (Boc)2O, 2.0 equiv i-Pr2NEt, (b) 0.5 equiv (Boc)2O, 2.0 equiv i-Pr2NEt, (c) 1.0 equiv 8, 2.0 equiv
i-Pr2NEt, (d) 0.5 equiv 8, 2.0 equiv i-Pr2NEt, 1 h.
When peptide 8 was used in place of Boc2O the results
were qualitatively the same. Two products, 10a (mono-
acylated) and 10b (bis-acylated), were obtained when
1.0 equiv of peptide was used in the reaction while mainly
10a was produced when 0.5 equiv of 8 were used. It is
important that the reaction time be restricted to 1 h, since
longer reaction times lead to formation of a third prod-
uct having a molecular weight consistent with the addi-
tion of two equiv of TLM S10bÆCu(II). Nonetheless, after
selective acylation, the crude reaction mixture consisted
mainly of the mono-acylated product and unreacted
TLM S10bÆCu(II) as indicated by HPLC and MS. The
desired product was easily isolated by preparative HPLC
(C18, 60:40 CH3OH/H2O 0.05 M HCO2NH4). Following
isolation, 10a was treated with TFA to effect protecting
group removal to yield compound 11.
step prior to carrying out the conjugation to BR96.
When mercaptoethanol was used as the nucleophilic
reagent, thiol attack at the maleimide to form the cor-
responding succinimide could be selectively carried out,
to yield 12 without interference from Cu(II). Conjuga-
tion to BR96 11 was carried out using our standard
method. Selective disulfide bond reduction of BR96
yielded 8 conjugatable thiol groups. Subsequent reaction
with 8 equiv of peptide 11 (pH 6.0 PBS) gave the desired
conjugate 13 having a TLM S10bÆCu(II)/BR96 ratio of
9.5 as determined by second derivative UV-spectro-
scopy. Size exclusion chromatography (SEC) of the
conjugate showed only one peak eluting at 9.7 min
(1.0 mL/min, 0.1 M KH2PO4, SEC-250 (Biorad)) and
having the same time as free BR96. UV analysis of this
peak indicated an increase in the absorption at
k ¼ 293 nm relative to k ¼ 280 for the conjugate com-
pared to the same 293 nm/280 nm absorption ratio for
the free mAb. Co-injection of BR96 and 11 gave two
Since Cu(II) is capable of oxidizing thiols to disulfides
we decided to model the thioether forming conjugation

4326
M. A. Walker et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4323–4327
Table 1. Activity of TLM S10b conjugates against cell lines with
varying BR96 expression
NH2
O
Cell line
IC50 lM equivalent of TLM S10bÆCu(II)
O
H
N
H
N
1
13
14
N
N
H
LS174T
L2987
2
0.04
0.04
3
5
.
O
O
O
TLM S10b Cu(II)
0.5
7
4
O
H3396
MCF7
0.008
0.006
2
0.8
O
2
11
a or b
LS174t, L2987, H3396, and MCF7, respectively), which
was used as a reference agent. In contrast the potency of
the drug is enhanced by up to 875-fold when delivered as
the corresponding BR96 conjugate. The data indicate
that the potency enhancement is antigen dependent since
the extent of the potency enhancement and the net IC50
are related to the antigen expression level. In addition,
the corresponding IgG conjugate does not yield an in-
crease in activity suggesting the antigen binding and
internalization into the cell are required as well.
NH2
O
O
H
N
H
N
R
S
N
N
5
H
.
O
O
O
TLM S10b Cu(II)
O
O
It is interesting to note that for TLM S10b conjugation to
the nonbonding mAb (IgG) yielded activity equivalent
to the free drug. This is in contrast to our previous work
with other drugs, which showed a significant reduction
in activity when bound to IgG. In order to explain this
apparent discrepancy we offer three possible explana-
tions. First, the conjugate might contain free drug, but
this seems unlikely since the conjugates were extensively
purified in order to remove free TLM S10b and no free
drug was detected prior to assaying the conjugate. Sec-
ond, the activity of the IgG-conjugate and TLM S10b (at
2 h) might be due to extracellular activation of the redox
pathway associated with the activity of the parent drug,
which somehow harms the cellular membrane. Litera-
ture reports have provided evidence for cell membrane
destruction mediated by BLM conjugated to albumin or
nonbinding IgG.19 Last, the conjugate might be inter-
nalized into the cell through a TLM S10b binding site
located on the outside of the cell. Some evidence exists
to suggest that such a site exists for bleomycin.20
Regardless, of the mechanism the results achieved in the
current study show that the activity of BR96-TLM S10b
compared to IgG-TLM S10b is sufficiently large to yield a
desirable therapeuetic index in animals.
12 (R = -SCH2CH2OH)
13 (R = BR96)
14 (R = IgG)
Scheme 3. Reagents and conditions: (a) 1.2 equiv HS(CH2)2OH,
CH3OH, (b) (1) 8 equiv DTT, 1 equiv BR96 (or IgG), (2) 8 equiv 11.
peaks one at 9.7 min and the other at 14.9 min suggest-
ing that the conjugate, which shows only one peak is free
of unreacted peptide 11. Further analysis of the conju-
gate using size exclusion chromatography indicated
6 1% protein aggregation. Peptide 11 was likewise
conjugated with a nontumor cell recognizing IgG to
provide a reference agent (14) for cell culture assays
(Scheme 3).
Prior to evaluating the potential utility of 13 as an
antitumor agent we set out to determine the Cathepsin B
cleavage rate for the release of TLM S10b from the Phe-
Lys linker. Incubation of 12 with Cathepsin B (pH 5,
37 °C) resulted in rapid release of the drug (t1=2 ¼ 10 min)
while the no release of drug was observed when 13 was
incubated at pH 5 (37 °C) for up to 24 h in the absence of
Cathepsin B. This suggests that the resulting conjugate
should be highly selective for Cathepsin B mediated
drug release.
Table 2 compares the key activity parameters of BR96-
TLM S10b with our previously published BR96 based
immunoconjugates. As can be seen, the TLM S10b con-
jugate is an order of magnitude more potent than BMS-
182248 and 2–3-times more active than our previous
peptide linked immunoconjugates despite the fact that
the apparent activity of parent drug is an order a mag-
nitude weaker than doxorubicin and camptothecin
(CPT). This suggests that the intrinsic activity of the
TLM S10b is actually higher than doxorubicin and
camptothecin and is consistent with its proposed
mechanism of cytotoxicity, which involves catalytic
generation of radical species at the site of action.
Moreover the potency enhancing effects of conjugation
are higher for TLM S10b than for doxorubicin and CPT
as expected for a drug, which poorly internalized in the
free state. The immunoselective activity of BR96-TLM
The activity of the BR96- and IgG-conjugates and TLM
S10b were evaluated against a panel of cell lines having
varying levels of BR96 antigen expression with the
corresponding results shown in Table 1. LS174T
(human colon adenocarcinoma) and L2987 (human lung
adenocarcinoma) show moderate expression of the
BR96 antigen while H3396 (human breast adenocarci-
noma) and MCF7 (human breast adenocarcinoma) are
high expressers (cf. Ref. 2). The IC50’s reported for the
conjugates in the table are calculated based on the
number of moles of TLM S10b delivered in order to
allow evaluation of the potency enhancing properties of
BR96 conjugation. As can be seen, free TLM S10b is only
modestly active against the cell lines compared to
doxorubicin (IC50 ¼ 0.3, 0.3, 0.2, and 0.5 lM against

M. A. Walker et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4323–4327
4327
Table 2. Summary of BR96-TLM S10b conjugate (13) compared to conjugates of Dox and CPT
Conjugateb
Experimenta
BMS-182248
BR96-F-K-Dox
BR96-F-K-CPT
BR96-F-K-TLM S10b (13)
IC50 L2987 (lM)
IC50 drug/conj.
IC50 IgG/BR96
0.63
0.1
0.15
1
0.1
4
0.04
13
17
80
30
100
a Experiments: IC50 L2987; IC50 of the immunoconjugate against the L2987 cell line; IC50 drug/conjugate; ratio of the parent drug IC50 divided by the
IC50 of the immunoconjugate; IC50 IgG/BR96; ratio of the IC50 obtained for the drug conjugated to the nonspecific IgG antibody divided by the
IC50 of the drug conjugated to BR96.
b Conjugates: BMS-182248; doxorubicin conjugated to BR96 via a hydrazone linker; BR96-F-K-Dox; doxorubicin conjugated to BR96 via the
cathepsin B cleavable Phe-Lys-PABC peptide linker; BR96-F-K-96; camptothecin conjugated to BR96 via the cathepsin B cleavable Phe-Lys-PABC
peptide linker; BR96-F-K-TLM S10b; compound 13 described in the text.
8. Sinh, A. A.; Jamuar, M. P.; Wilson, M. J.; Rozhin, J.;
Sloane, B. F. Prostate 2001, 49, 172.
9. Hashimoto, Y.; Kakegawa, H.; Narita, Y.; Hachiya, Y.;
Hayakawa, T.; Kos, J.; Turk, V.; Katunuma, N. Biochem.
Biophys. Res. Commun. 2001, 283, 334.
10. (a) Dubowchik, G. M.; Firestone, R. A; Willner, D;
Hofstead, S. J.; Trail, P. A.; Lasch, S. J.; Henderson, A. J.;
Jure, M.; Mosure, K. W.; Knipe, J. O. Peptides 1996,
Proceedings of the European Peptide Symposium, 24th,
Edinburgh, Sept. 8–13, 1996; pp 347–348; (b) Dubowchik,
G. M.; Firestone, R. A.; Padilla, L.; Willner, D.; Hofstead,
S. J.; Mosure, K.; Knipe, J. O.; Lasch, S. J.; Trail, P. A.
Bioconjugate Chem. 2002, 13, 855.
11. Walker, M. A.; Dubowchik, G. M.; Hofstead, S. J.; Trail,
P. A.; Firestone, R. A. Bioorg. Med. Chem. Lett. 2002, 12,
217–219.
12. Stubbe, J.; Kozarich, J. W. Chem. Rev. 1987, 87, 1107.
13. Schurig, J. E.; Rose, W. C.; Hirth, R. S.; Schein, A.;
Huftalen, J. B.; Florezyk, A. P.; Bradner, W. T. Cancer
Chemother. Pharmacol. 1984, 13, 164.
14. (a) Nicaise, C.; Ajani, J.; Goudeau, P.; Rozencweig, M.;
Levin, B.; Krakoff Cancer Chemother. Pharmacol. 1990,
26, 221; (b) Nicaise, C.; Hong, W. K.; Dimery, W.;
Usakewicz, J.; Rozencweig, M.; Krakoff, I. Invest. New
Drug. 1990, 8, 325.
S10b is higher than that of BMS-182248 and is approxi-
mately equivalent to that observed for our other peptide
linked immunoconjugates. This window is large enough
to expect that the immunoconjugate could be dosed at
concentrations much higher than the IC50 while main-
taining TLM S10b exposure below toxic levels.
We have successfully synthesized an immunoconjugate
of TLM S10b employing the mAb BR96 and a dipeptide
linker designed to release the drug by Cathepsin B cat-
alyzed proteolysis. The parent drug is only moderately
active against a panel of tumor cell lines in contrast to
the conjugate (13), which results in a 13–875-fold in-
crease in potency. This is presumably due to the inter-
nalizing properties of BR96, which allows for more
efficient uptake of the drug and exposure to its primary
target. This provides support for using internalizing
mAbs as a vehicle to increase the potency of drugs,
which exhibit poor cellular penetration in the free
solution.
References and notes
15. Bleomycin immunoconjugates have previously been re-
ported. (a) Manabe, Y.; Tsubota, T.; Haruta, Y.; Okazaki,
M.; Haisa, S.; Nakamura, K.; Kimura, I. Biochem.
Biophys. Res. Commun. 1983, 115, 1009; (b) Peng, Z.;
Zhen, Y. S. Yaoxue Xuebao 1991, 26, 331; (c) Deng, Y. C.;
Zhen, Y. S.; Zheng, S.; Jiang, M. Yaoxue Xuebao 1993, 28,
410; (d) Wang, W. G.; Wang, S. H.; Xue, Y. C.; Zhen,
Y. S. Yaoxue Xuebao 1995, 30, 583.
1. Dubowchik, G. M.; Walker, M. A. Pharmacol. Ther. 1999,
83, 67.
€
€
2. Hellstrom, I.; Garrigues, H. J.; Garrigues, U.; Hellstrom,
K. E. Cancer Res. 1990, 50, 2183.
3. (a) Trail, P. A.; Willner, D.; Lasch, S. J.; Henderson, A. J.;
€
Hofsted, S.; Casazza, A. M.; Firestone, R. A.; Hellstrom,
I.; Hellstrom, K. E. Science 1993, 261, 212; (b) Sjogren, H.
€
€
€
€
O.; Isaksson, M.; Willner, D.; Hellstrom, I.; Hellstrom, K.
E.; Trail, P. A. Cancer Res. 1997, 57, 4530.
16. Poddevin, B.; Orlowski, S.; Belehradek, J., Jr.; Mir, L. M.
Biochem. Pharmacol. 1991, 42, 67.
4. Reviews of intracellular delivery of mAbs: (a) Kato, Y.;
Sugiyama, Y. Crit. Rev. Ther. Drug 1997, 14, 287; (b)
Chen, B.-X.; Erlanger, B. F. Immunol. Lett. 2002, 84, 63.
17. Mistry, J. S.; Jani, J. P.; Morris, G.; Mujumdar, R. B.;
Reynolds, I. J.; Sebti, S. M.; Lazo, J. S. Cancer Res. 1992,
52, 709.
€
5. Garrigues, J.; Garrigues, U.; Hellstrom, I.; Hellstrom,
€
18. Bio-Rad Laboratories, Hercules CA.
K. E. Am. J. Pathol. 1993, 142, 607.
6. Dubowchik, G. M.; Firestone, R. A. Bioorg. Med. Chem.
Lett. 1998, 8, 3341.
19. Voznesenskii, A. I.; Galanova, Y. V.; Shkrob, A. M.;
Mathanov, I. E.; Archakov, A. I. Arch. Biochem. Biophys.
1990, 283, 519.
7. Carl, P. L.; Chakravarty, P. K.; Katzenellbogen, J. A.
J. Med. Chem. 1981, 24, 479.
20. Pron, G.; Belehedrak, J., Jr.; Mir, L. M. Biochem. Biophys.
Res. Commun. 1993, 194, 333.