Dipeptide-based highly potent doxorubicin antibody conjugates

Article abstract of DOI:10.1016/j.bmcl.2005.09.081

Highly potent and novel derivatives of doxorubicin were linked to monoclonal antibodies (mAbs) for site-specific drug delivery. Drug linker 5 consisted of a dipeptide linker attached directly to the daunosamine nitrogen of the n-butyldiacetate doxorubicin

Full text of DOI:10.1016/j.bmcl.2005.09.081

Bioorganic & Medicinal Chemistry Letters 16 (2006) 358–362
Dipeptide-based highly potent doxorubicin antibody conjugates
Scott C. Jeffrey,* Minh T. Nguyen, Jamie B. Andreyka, Damon L. Meyer,
Svetlana O. Doronina and Peter D. Senter
Department of Chemistry, Seattle Genetics, 21823 30th Drive SE, Bothell, WA 98021, USA
Received 2 September 2005; revised 27 September 2005; accepted 28 September 2005
Available online 3 November 2005
Abstract—Highly potent and novel derivatives of doxorubicin were linked to monoclonal antibodies (mAbs) for site-specific drug
delivery. Drug linker 5 consisted of a dipeptide linker attached directly to the daunosamine nitrogen of the n-butyldiacetate doxo-
rubicin derivative 2a. Upon hydrolysis of the peptide linker and acetate groups, the free daunosamine nitrogen is able to form the
highly potent 2-pyrrolinodoxorubicin (3a). The second approach involved the use of an oxazolidine carbamate (13) to mask an acti-
vating aldehyde group until proteolytic hydrolysis releases 3a. Both drug linkers were shown to be substrates for the lysosomal
enzyme cathepsin B. Each molecule was conjugated to the mAbs c1F6 (anti-CD70) and cAC10 (anti-CD30) to give potent drug
conjugates against renal cell carcinoma and anaplastic large cell lymphoma cell lines, respectively. The activities were immunolog-
ically selective, since antigen negative cell lines were much less sensitive to treatment with the drug conjugates. The approaches
described here for attaching highly potent doxorubicin derivatives to mAbs are novel and allow for control of drug stability while
covalently bound to the delivery agent.
Ó 2005 Elsevier Ltd. All rights reserved.
A great deal of attention has surrounded the use of
monoclonal antibodies (mAbs) for the delivery of anti-
cancer drugs to tumor cells. Several mAb–drug conju-
gates have displayed pronounced activities in
preclinical cancer models, and there are now three ap-
proved antibody–cytotoxin conjugates for cancer thera-
py: MylotargTM (gemtuzumab ozogamicin), ZevalinTM
(ibritumomab tiuxetan), and BexxarTM (tositumomab).
Research surrounding the critical parameters for thera-
peutic success has suggested that highly potent drugs
are required for mAb-based delivery strategies, and the
linker used to attach the drug to the mAb should be
highly stable in circulation. A significant body of litera-
ture describes efforts in this field using an assortment of
cytotoxic payloads including doxorubicin,1,2 taxanes,3
maytansinoids,4 minor groove binders,5–8 and others.
We have recently reported ADC technologies based on
the antimitotic agent monomethyl auristatin E9–11 and
minor groove binders8 released through proteolytic
cleavage of dipeptide linkers. This communication
describes our work with highly potent derivatives of
doxorubicin which employ a similar proteolytic release
strategy. In one example, we employed a novel masking
strategy for tethering to the mAb through a latent alde-
hyde group to result in highly potent cytotoxic activity.
Cyclic derivatives of doxorubicin involving the daunos-
amine sugar nitrogen were first reported by Acton
et al.12 The morpholino cyanide compound 1 was
reported to be intensely potent, displaying cytotoxic
activities several log units higher than doxorubicin.
Structure–activity relationship studies suggested that
the potential for iminium formation through displace-
ment of the cyanide group was key to increased potency
of 1 relative to doxorubicin. More recently, Farquhar13
reported n-butyl and n-pentyl diacetate derivatives of
doxorubicin (2a,b) which displayed a similar potency
profile to 1. The basis for activity likely involves ester-
ase-mediated hydrolysis of the acyclic diacetates 2a,b.
The resulting aldehydes cyclize to form derivatives
3a,b14 (Scheme 1). These compounds alkylate double-
stranded DNA, through the corresponding cyclic
iminium, resulting in interstrand crosslinking and apop-
tosis.15 In addition, they are capable of overcoming
P-glycoprotein-mediated multidrug resistance.13 A re-
cent report demonstrated the use of the pentyl diacetate
2b tethered through a hydrazone linkage to the mAb
BR96 which displayed improved potency relative to a
doxorubicin hydrazone linker.16 We wanted to explore
Keywords: Anthracycline; Antibody–drug conjugate; Antitumor;
Cancer; Cathepsin B; Cytotoxic; DNA crosslinking; Doxorubicin;
Drug-linker; Lysosome; Monoclonal antibody; Potent; Protease;
2-Pyrollinodoxorubicin; Targeted delivery; Self-immolative spacer.
*
Corresponding author. Tel.: +1 425 527 4738; fax: +1 425 527
4109; e-mail: sjeffrey@seagen.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.081

S. C. Jeffrey et al. / Bioorg. Med. Chem. Lett. 16 (2006) 358–362
359
O
O
OH
OH
O
the geminal diacetate group and the dipeptide (resulting
in 1,6-elimination of drug through the PABC spacer),
and loss of a water molecule would generate 3a. The
second approach involved tethering the molecule
through an oxazolidine carbamate a group which would
stably mask the activating aldehyde group until proteo-
lytic cleavage from the mAb. By tethering through a
masked aldehyde, we hoped to prevent potentially
undesired chemistry involving a free aldehyde that
might compromise the drug while attached to the
mAb. Cleavage of the dipeptide would result in 1,6-
elimination to liberate an intermediate that should
rapidly cyclize to 3a. The peptide sequence val-ala
was chosen over val-cit to reduce hydrophobicity and
the potential for aggregation.
OH
OH
O
O
O
1
N
CN
HO
O
O
OH
OH
O
O
O
OH
O
OH
OH
OH
OH
esterase
O
O
O
n
O
OH
O
O
O
3a n=1
b n=2
N
NH OAc
2a n=1
b n=2
OH
OH
OAc
To construct the molecule described as Approach I
(Scheme 2), compound 2a was reacted with the p-nitro-
phenyl (PNP) carbonate of the valine-citrulline (val-cit)
dipeptide linker 418,19 (Scheme 3). This chemistry was
problematic, due to the hindered nature of daunosamine
nitrogen of 2a. Nevertheless, low yields of the desired
agent could be obtained by first protecting 2a as its
mono-TBS ether. Coupling with 4 and deprotection
afforded the drug-linker 5.
n
Scheme 1.
conjugates based on these highly potent doxorubicin
derivatives employing a more stable dipeptide linkage
for mAb-targeted delivery. We have already shown that
these molecules can be stably masked and activated by a
targeted mAb-enzyme conjugate.17
Two approaches were conceived for targeted delivery
and protease-mediated release of 3a (Scheme 2). The
key feature of both methods was that the formation
of a reactive pyrroline ring would not occur until the
drug was released from the mAb by proteolytic hydro-
lysis of the peptide linker. This is in contrast to an ap-
proach previously taken with this class of molecules in
which the drug could undergo activation while bound
to the mAb carrier.16 The first approach illustrated in
Scheme 2 involved attaching the dipeptide linker,
through a p-aminobenzyl carbamate spacer (PABC),
to the daunosamine nitrogen of 2a. Hydrolysis of both
To prepare the oxazolidine carbamate described as Ap-
proach II (Scheme 2), the benzyl alcohol 6 was activated
using diphosgene and reacted with the oxazolidine 7
formed through treatment of aldehyde 8 (prepared from
1,4-butanediol in two steps) with ethanolamine in the
˚
presence of 4 A molecular sieve powder (Scheme 4).
The resulting product was N-deprotected with piperi-
dine to give 9. Coupling with Fmoc-Val-OSu and depro-
tection of both the TBDPS and Fmoc groups with
TBAF afforded 10. This was followed by coupling of
the free amine to maleimidocaproyl N-hydroxysuccina-
mide ester (MC-OSu) and oxidation of the alcohol
O
O
OH
O
O
O
OH
O
OH
OH
OH
OH
O
OH
O
O
R1
O
O
OH
O
H
H
i. esterase
N
N
mAb
O
N
O
ii. peptidase
H
R2
N
O
O
O
N
OH
OH
AcO
2-pyrrolinodoxorubicin (3a)
AcO
NH2
HO
O
O
OH
O
O
O
OH
OH
O
OH
OH
OH
O
R1
O
OH
peptidase
H
H
N
N
mAb
O
OH O
N
O
O
H
R2
O
O
O
O
O
NH
N
HN
OH
OH
NH
O
O
Scheme 2. Approach I: n-butyl diacetate carbamate. Approach II: oxazolidine carbamate.

360
S. C. Jeffrey et al. / Bioorg. Med. Chem. Lett. 16 (2006) 358–362
O
O
OH
OH
O
(a), (b)
NHFmoc
NH2
O
OH
HO
TBDPSO
TBDPSO
OH
O
8
H
OH
(a), (b), (c)
(c)
O
O
H
NH2
H
N
O
N
HN
NH OAc
HO
O
OH
O
7
OAc
6
2a
O
H
N
H
(d), (e)
N
N
H
H
N
O
O
O
O
O
H
N
O
NH2
–O
O
N
H
N+
N
4
O
O
O
O
10
O
O
O
(f), (g)
N
O
OH
O
O
N
O
O
HN
NH2
(h), (i)
OH
O
O
OH
OTBDPS
OH
O
H
N
H
N
O
O
O
OH
O
O
9
H
H
N
N
N
H
O
O
N
H
O
(j)
N
O
O
O
OH
N
O
O
O
5
11
AcO
N
O
N
O
O
O
OH
O
AcO
O
OH
O
OH
O
Scheme 3. Reagents and conditions: (a) TBSOTf, 2,6-DTBP, CH2Cl2,
81%; (b) 4, pyridine, DMF, rt, 48 h, 36%; (c) HF–pyridine complex,
CH3CN, 47%.
O
H
N
H
N
O
O
O
OH
O
N
H
O
O
NH
group to the corresponding aldehyde 11. Finally, reduc-
tive alkylation with doxorubicin hydrochloride (12)
using sodium cyanoborohydride afforded the desired
drug linker 13 with the aldehyde stably masked as the
oxazolidine carbamate.
N
OH
O
13
O
N
O
Scheme 4. Reagents and conditions: (a) TBDPSCl, NaH, 56%; (b)
˚
Dess–Martin, CH2Cl2, 97%; (c) ethanolamine, benzene, 4 A powdered
To determine if the drug-linkers 5 and 13 were likely to
release free drug once transported by the mAb to the
lysosomes of target cells, the reactive maleimide groups
were quenched with cysteine (cys) and the resulting com-
pounds were exposed to the human lysosomal enzyme
cathepsin B.19 Both drug-linkers were substrates for
cathepsin B, with the val-cit compound cys-5 being
cleaved at approximately twice the rate of the val-ala
compound cys-13 (Table 1). The linker cys-5 cleanly
released the butyl diacetate 2a, while the linker cys-13
released 2-pyrrolinodoxorubicin 3a directly as deter-
mined by LC–MS analysis of the sample digests. Both
reactions went to completion. In the absence of cathep-
sin B, both cysteine-quenched drug-linkers were highly
stable.
sieves; (d) diphosgene (2 equiv), pyr (8 equiv), CH2Cl2, À78 °C, 2 h,
then 7, À78 °C to rt, 61%; (e) DMF, piperidine (4:1); (f) Fmoc-Val-
OSu, DMF, 94% two steps; (g) TBAF, THF, 82%; (h) MC-OSu,
DMF, 88%; (i) Dess–Martin, CH2Cl2, 91%; (j) doxorubicin–HCl (12),
CH3CN, H2O (2:1), NaCNBH3, 0 °C to rt, 26%.
to the drug linkers 5 and 13 resulted in conjugates that
were largely monomeric (Table 2) as determined by size
exclusion chromatography. Achieving a high level of
loading with 5 (>4 drugs/mAb) was complicated by pre-
cipitation and aggregation of the resulting ADCs. In
contrast, 13 could be loaded with 7.4 drugs/mAb with-
out exceeding an acceptable level of aggregation
(>10%). Drug loading was determined using spectromet-
ric methods by comparing the absorbance of the
Compounds 5 and 13 were conjugated to the chimeric
mAbs 1F6 and AC10. Chimeric 1F6 binds to CD70
(TNF receptor superfamily) which has a high relative
expression profile on both hematologic and renal cell
carcinoma lines compared to normal tissues.20–23
cAC10 binds to the CD30 antigen found on malignant
B cell lines.24 For the preparation of conjugates, both
mAbs were treated with dithiothreitol (DTT) to reduc-
tively cleave the heavy–heavy and heavy–light chain
disulfide linkages. Full reduction provides a mAb with
approximately eight free sulfhydryl groups. Exposure
Table 1. Cathepsin B digest of cysteine-quenched 5 and 13
Compound
Human cathepsin B hydrolysis (t1/2 min)
Cys-5
Cys-13
72
142
Results from a single experiment. Substrate:enzyme ratio = 20,000.
Final substrate concentration of 2.5 mM. Half-life was determined by
LC measuring percent remaining starting material at 254 nm.

S. C. Jeffrey et al. / Bioorg. Med. Chem. Lett. 16 (2006) 358–362
Table 2. Characterization and in vitro cytotoxicity of the conjugatesa
361
Compound
Target antigen
Drug loading
% Aggregation
Caki-1 cellsb
786-O cellsb
Karpas 299 cellsb
(CD70+, CD30À)
0.039
112
(CD70+, CD30À)
0.006
64.5
0.4
(CD30+, CD70À)
0.099
29.2
—
2a
Doxorubicin (12)
c1F6-5
—
—
—
—
—
—
<1
3
CD70
CD30
CD70
CD30
2.4
3.3
4.9
7.4
1.6
cAC10-5
c1F6-13
cAC10-13
65
0.65
>50
108
1.42
>50
0.57
<1
10
2.3
>50
a The activities of the ADCs were compared to that of doxorubicin, and a highly potent doxorubicin derivative (2a) that leads to the same molecule
released from the conjugates.
b Cells were treated with the test agents for 96 h and viability was determined by reduction of resazurin. The IC50 values indicated are the
concentrations (nM) of the drug component of the conjutates.
doxorubicin chromophore (490 nm) to the protein
absorbance (280 nm).
>40-fold specificity when evaluated against an antigen
negative cell line and non-binding control ADCs. Both
drug-linkers gave conjugates that were >70-fold more
potent than doxorubicin when evaluated on the basis
of drug concentration. Results from the in vivo evalu-
ation of ADCs based on 5 and 13 will be reported
elsewhere.
Compound 2a showed exquisite cytotoxic activity on
the three cell lines evaluated: the renal cell carcinoma
(RCC) lines Caki-1 (CD70+) and 786-O (CD70+), and
ALCL line Karpas 299 (CD30+) with potencies rang-
ing from 6 to 99 pM (Table 2). Doxorubicin (12) was
2.5–4 orders of magnitude less active on the same cell
lines.
References and notes
When the active drug 3a was conjugated to the mAbs
c1F6 or cAC10 either via drug-linker 5 or 13, significant
levels of cytotoxic activity were obtained and the effects
were immunologically specific. The cells ranged from 41-
to 270-fold more sensitive to binding conjugates than to
non-binding conjugates. In addition, the IC50 values of
conjugates with these two drugs were similar, suggesting
that both linker constructs delivered the active drug 3a
with equal effectiveness. However, the potency of the
conjugates was significantly less than that of the free
drugs 2a. This might be due to the fact that passive cel-
lular uptake of free drug may lead to higher intracellular
concentrations compared to those obtained through
mAb-mediated delivery.
1. Trail, P. A.; Willner, D.; Lasch, S. J.; Henderson, A. J.;
Hofstead, S.; Casazza, A. M.; Firestone, R. A.; Hellstrom,
I.; Hellstrom, K. E. Science 1993, 261, 212.
2. Saleh, M. N.; Sugarman, S.; Murray, J.; Ostroff, J. B.;
Healey, D.; Jones, D.; Daniel, C. R.; LeBherz, D.; Brewer,
H.; Onetto, N.; LoBuglio, A. F. J. Clin. Oncol. 2000, 18,
2282.
3. Ojima, I.; Geng, X.; Wu, X.; Qu, C.; Borella, C. P.; Xie,
H.; Wilhelm, S. D.; Leece, B. A.; Bartle, L. M.; Goldm-
acher, V. S.; Chari, R. V. J. Med. Chem. 2002, 45, 5620.
4. Liu, C.; Tadayoni, B. M.; Bourret, L. A.; Mattocks, K.
M.; Derr, S. M.; Widdison, W. C.; Kedersha, N. L.;
Ariniello, P. D.; Goldmacher, V. S.; Lambert, J. M.;
Blattler, W. A.; Chari, R. V. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 8618.
5. Chari, R. V.; Jackel, K. A.; Bourret, L. A.; Derr, S. M.;
Tadayoni, B. M.; Mattocks, K. M.; Shah, S. A.; Liu, C.;
Blattler, W. A.; Goldmacher, V. S. Cancer Res. 1995, 55,
4079.
6. Suzawa, T.; Nagamura, S.; Saito, H.; Ohta, S.; Hanai, N.;
Yamasaki, M. Bioorg. Med. Chem. 2000, 8, 2175.
7. Suzawa, T.; Nagamura, S.; Saito, H.; Ohta, S.; Hanai, N.;
Kanazawa, J.; Okabe, M.; Yamasaki, M. J. Controlled
Release 2002, 79, 229.
8. Jeffrey, S. C.; Torgov, M. Y.; Andreyka, J. B.; Bodding-
ton, L.; Cerveny, C. G.; Denny, W. A.; Gordon, K. A.;
Gustin, D.; Haugen, J.; Kline, T.; Nguyen, M. T.; Senter,
P. D. J. Med. Chem. 2005, 48, 1344.
9. Francisco, J. A.; Cerveny, C. G.; Meyer, D. L.; Mixan, B.
J.; Klussman, K.; Chace, D. F.; Rejniak, S. X.; Gordon,
K. A.; DeBlanc, R.; Toki, B. E.; Law, C. L.; Doronina, S.
O.; Siegall, C. B.; Senter, P. D.; Wahl, A. F. Blood 2003,
102, 1458.
We compared the potency of our ADCs (relative to
doxorubicin) to that of the hydrazone-based BR96
conjugate of 2b previously described by King
et al.16 For the hydrazone of 2b on BR96, an in-
creased potency of 10-fold was seen over the free
drug doxorubicin. The increased potency of ADCs
based on 5 and 13 ranged from 70- to 170-fold, rel-
ative to doxorubicin. We believe the increased relative
potency of our ADCs stems from the potency of the
released drug. In our hands, compound 2a was
approximately 10-fold more potent than 2b on a ser-
ies of cell lines (data not included).
In summary, the drug linkers 5 and 13, based on the
potent doxorubicin derivative 3a, were designed, syn-
thesized, and linked to mAbs that recognize tumor
associated antigens on RCC and ALCL tumors. The
cysteine-quenched adducts of 5 and 13 were substrates
for the lysosomal enzyme cathepsin B and were
cleaved at comparable rates. Conjugation to the mAbs
c1F6 and cAC10 gave ADCs with single-digit
nanomolar to sub-nanomolar cytotoxicity IC50s, and
10. Doronina, S. O.; Toki, B. E.; Torgov, M. Y.; Mendelsohn,
B. A.; Cerveny, C. G.; Chace, D. F.; DeBlanc, R. L.;
Gearing, R. P.; Bovee, T. D.; Siegall, C. B.; Francisco, J.
A.; Wahl, A. F.; Meyer, D. L.; Senter, P. D. Nat.
Biotechnol. 2003, 21, 778.
11. Hamblett, K. J.; Senter, P. D.; Chace, D. F.; Sun, M. M.;
Lenox, J.; Cerveny, C. G.; Kissler, K. M.; Bernhardt, S.

362
S. C. Jeffrey et al. / Bioorg. Med. Chem. Lett. 16 (2006) 358–362
X.; Kopcha, A. K.; Zabinski, R. F.; Meyer, D. L.;
Francisco, J. A. Clin. Cancer Res. 2004, 10, 7063.
12. Acton, E. M.; Tong, G. L.; Mosher, C. W.; Wolgemuth,
R. L. J. Med. Chem. 1984, 27, 638.
13. Cherif, A.; Farquhar, D. J. Med. Chem. 1992, 35, 3208.
14. Farquhar, D.; Cherif, A.; Bakina, E.; Nelson, J. A. J.
Med. Chem. 1998, 41, 965.
15. Zwelling, L. A.; Altschuler, E.; Cherif, A.; Farquhar, D.
Cancer Res. 1991, 51, 6704.
19. Dubowchik, G. M.; Firestone, R. A.; Padilla, L.; Willner,
D.; Hofstead, S. J.; Mosure, K.; Knipe, J. O.; Lasch, S. J.;
Trail, P. A. Bioconjug. Chem. 2002, 13, 855.
20. Diegmann, J.; Junker, K.; Gerstmayer, B.; Bosio, A.;
Hindermann, W.; Rosenhahn, J.; Von Eggeling, F. Eur. J.
Cancer 2005.
21. Junker, K.; Hindermann, W.; Von Eggeling, F.; Dieg-
mann, J.; Haessler, K.; Schubert, J. J. Urol. 2005, 173,
2150.
16. King, H. D.; Staab, A. J.; Pham-Kaplita, K.; Yurgaitis,
D.; Firestone, R. A.; Lasch, S. J.; Trail, P. A. Bioorg. Med.
Chem. Lett. 2003, 13, 2119.
17. Torgov, M. Y.; Alley, S. C.; Cerveny, C. G.; Farquhar, D.;
Senter, P. D. Bioconjug. Chem. 2005, 16, 717.
18. Dubowchik, G. M.; Walker, M. A. Pharmacol. Ther. 1999,
83, 67.
22. Held-Feindt, J.; Mentlein, R. Int. J. Cancer 2002, 98, 352.
23. Tesselaar, K.; Gravestein, L. A.; Van Schijndel, G.
M.; Borst, J.; Van Lier, R. A. J. Immunol. 1997, 159,
4959.
24. Wahl, A. F.; Klussman, K.; Thompson, J. D.; Chen, J. H.;
Francisco, L. V.; Risdon, G.; Chace, D. F.; Siegall, C. B.;
Francisco, J. A. Cancer Res. 2002, 62, 3736.