Discovery of matrix metalloproteases selective and activated peptide-doxorubicin prodrugs as anti-tumor agents

Article abstract of DOI:10.1016/j.bmcl.2009.12.084

To selectively target doxorubicin (Dox) to tumor tissue and thereby improve the therapeutic index and/or efficacy of Dox, matrix metalloproteinases (MMP) activated peptide-Dox prodrugs were designed and synthesized by coupling MMP-cleavable peptides to Dox. Preferred conjugates were good substrates for MMPs, poor substrates for neprilysin, an off-target proteinase, and stable in blood ex vivo. When administered to mice with HT1080 xenografts, conjugates, such as 19, preferentially released Dox in tumor relative to heart tissue and prevented tumor growth with less marrow toxicity than Dox.

Full text of DOI:10.1016/j.bmcl.2009.12.084

Bioorganic & Medicinal Chemistry Letters 20 (2010) 853–856
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of matrix metalloproteases selective and activated
peptide–doxorubicin prodrugs as anti-tumor agents q
*
Zilun Hu , Xiangjun Jiang, Charles F. Albright, Nilsa Graciani, Eddy Yue, Mingzhu Zhang, Shu-Yun Zhang,
Robert Bruckner, Melody Diamond, Randine Dowling, Maria Rafalski, Swamy Yeleswaram,
George L. Trainor, Steven P. Seitz, Wei Han
Bristol-Myers Squibb Research and Development, PO Box 5400, Princeton, NJ 08543, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
To selectively target doxorubicin (Dox) to tumor tissue and thereby improve the therapeutic index and/or
efficacy of Dox, matrix metalloproteinases (MMP) activated peptide–Dox prodrugs were designed and
synthesized by coupling MMP-cleavable peptides to Dox. Preferred conjugates were good substrates
for MMPs, poor substrates for neprilysin, an off-target proteinase, and stable in blood ex vivo. When
administered to mice with HT1080 xenografts, conjugates, such as 19, preferentially released Dox in
tumor relative to heart tissue and prevented tumor growth with less marrow toxicity than Dox.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 3 November 2009
Revised 21 December 2009
Accepted 22 December 2009
Available online 4 January 2010
Keywords:
Doxorubicin
Matrix metalloproteases
Prodrug
Doxorubicin (Dox) is an anthracycline natural product that is
widely used to treat tumors such as breast cancer, liver cancer,
soft-tissue sarcomas, and non-Hodgkin’s lymphoma. Dox has a
complex mechanism of action with some of its activity arising from
inhibition of nucleic acid synthesis within cancer cells.1 Like other
cytotoxic drugs, the therapeutic efficacy of Dox is limited by un-
wanted toxicity to non-tumor tissues, most notably myelosuppres-
sion.2 In addition to these typical chemotherapeutic toxicities, Dox
also causes cardiomyopathy which depends on the cumulative
dose of drug.
zymes are also implicated in several critical events in tumor
evolution including tumorigenesis, tumor growth, angiogenesis,
generation of reactive stroma and tumor cell metastasis. In fact,
elevated level of MMP expression in human tumors was frequently
found to correlate with disease progression.9
We reasoned that a MMP-activated prodrug of Dox might selec-
tively release Dox at the tumor sites and thereby reduce side
effects. We chose MMP-2, -9 and -14 to guide our in vitro struc-
ture–activity relationship (SAR) effort because of good expression
of these enzymes in tumors.9
There have been several attempts to develop Dox prodrugs that
increase its therapeutic index.3–5 For example, investigators used
the prostate-specific antigen to activate Dox conjugates in mice
leading to increased efficacy with reduced toxicity in mouse xeno-
grafts.6 Unfortunately, these results did not effectively target Dox
to tumors in humans.7 It was hoped that this approach could deli-
ver an efficacious concentration of Dox at tumor sites with limited
systemic exposure, and hence would significantly increase its ther-
apeutic index.
Matrix metalloproteases (MMPs) are a family of structurally re-
lated zinc-containing proteases containing more than 20 members.8
Under normal conditions, these enzymes play an important role in
the maintenance and remodeling of connective tissues. These en-
The desired conjugates of Dox should possess several proper-
ties. In particular, they should be good substrates for MMP en-
zymes to allow efficient activation in the tumor. Conjugates
should be poor substrates for other enzymes, including enzymes
typically found in the plasma compartment. Of particular concern
for MMP-activated prodrugs was neprilysin since this cell-surface
protease is expressed outside tumor tissue and may therefore lead
to non-tumor activation of the prodrugs.10 In addition, the pro-
drugs should not be cytotoxic prior to activation11 and should have
aqueous solubility compatible with intra venous administration.
When properly designed, the resulting prodrugs have the potential
to efficiently and preferentially deposit Dox in tumor tissue rela-
tive to non-tumor tissue leading to an improved therapeutic index
and improved tumor growth inhibition.
Based on these considerations, our medicinal chemistry ap-
proach was to design and link an MMP peptide substrate,
ꢀ ꢀ ꢀP3P2P1–P0 P0 P0 ꢀ ꢀ ꢀ, to Dox to form a peptide–Dox conjugate. The
q
Presented in part at the 231st National Meeting of the American Chemical
Society, Atlanta, GA, March 2006, MEDI-271, 272.
* Corresponding author. Tel.: +1 609 818 5290.
E-mail address: zilun.hu@bms.com (Z. Hu).
1
2 3
COOH terminus of a peptide was conjugated by an amide bond
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.084

854
Z. Hu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 853–856
COOH terminal residue linked directly to Dox because L-Dox was
O
Cap-P3-P2-P1~P1'-P2'-Leu-Dox
HO
OH O
reported to be more efficiently converted to Dox than other conju-
gates.13–15 The N-termini of the conjugates were capped to prevent
aminopeptidase degradation before MMP cleavage. Based on these
considerations, N-terminus capped peptide conjugates with Gly as
P1 and Leu linked to Dox as shown in Figure 1 were designed and
optimized for MMP cleavage and selectivity. Incorporation of polar
groups either within the N-terminal caps or within the side chains
of amino acid residues was used to improve solubility of the con-
jugates. A target solubility of 1 mg/mL was chosen to guide com-
pound design.
HO
MMPs
(elevated level in tumors)
O
OH O OMe
P1'-P2'-Leu-Dox
O
H2N
Aminopeptidase
OH
Doxorubicin (Dox),its amino group is
Dox
linked to the COOH-terminus of peptitdes
Figure 1. Activation of peptide–Dox prodrug.
The synthesis of an example of the peptide–Dox conjugates is
outlined in Scheme 1. Compound preparation was performed on
a peptide synthesizer following a standard Fmoc solid phase proto-
col starting from Fmoc-Leu–Wang resin using HBTU as the cou-
pling reagent.16 The N-terminus of the completed peptide on
resin was capped and the peptide was then cleaved from the resin
with 90% trifluoroacetic acid in dichloromethane. The resulting
peptide was then coupled to Dox to form the conjugate using the
BOP coupling reagent.
SPPS
PLGS(OBn)Y(OtBu)L-Wang resin
Fmoc-L-Wang resin
peptide synthesizer
1) Capping
Dox, BOP
Cap-PLGS(OBn)YL-Dox
Cap-PLGS(OBn)YL-OH
2) 90% TFA in DCM
DIPEA, DMF
Scheme 1. Synthesis of peptide–Dox conjugates.
The initial SAR was based on simple collagen-like peptide con-
jugates containing the sequence PLG ꢁ L, which is cleaved by most
MMPs.14 We first determined the preferred peptide length. As we
previously reported17 and show in the set of conjugates 1–6 in
Table 1, the optimized conjugate length was a hexapeptide with
with the amino group of Dox. In our design of the peptide se-
quence, Glycine (Gly, G) was chosen as the P1 group because it
was found to be optimal for MMP cleavage from our initial findings
and also a literature report.12 Leucine (Leu, L) was chosen as the
Table 1
In vitro profiles of peptide–Dox conjugates
No.
Conjugate
Enzyme cleavage kcat/Km (mMꢂ1 sꢂ1
)
Stabilitya (%)
Solubilityb (mg/mL)
MMP-2
MMP-9
MMP-14
Neprilysin
1
2
3
Ac-PLG–L-Dox
Ac-PLG–LL-Dox
Ac-LG–LL-Dox
<1
18
<1
<1
>120
<1
—
4
<1
—
22
5
—
—
—
—
—
—
4
Ac-LG–LYL-Dox
6
1
24
8
—
—
5
6
7
8
Ac-PLG–LYL-Dox
Ac-PLG–LYAL-Dox
88
>120
24
7
11
>120
116
21
55
31
20
25
29
21
31
40
>120
69
47
390
>120
79
34
<1
34
>120
43
73
43
52
38
120
64
55
107
>120
>120
68
85
25
>120
>120
69
25
19
>120
>120
58
>120
62
89
56
72
48
83
97
>120
>120
97
22
>120
2.1
<1
1
<1
<1
<1
<1
<1
<1
<1
8
<1
<1
<1
<1
<1
<1
<1
<1
<1
<1
—
—
31
—
—
—
—
0.04
—
—
0.01
0.001
0.38
0.13
1.27
1.4
>2.3
1.9
2.1
>2.6
>2.2
1.5
>3.9
>2.2
1.8
Ac-PLG–S(OMe)cYL-Dox
Ac-PLG–S(OBn)dYL-Dox
Ac-PLG–HofeAL-Dox
Ac-PLG–HofYL-Dox
Ac-PLG–HofHoyfL-Dox
Ac-PLG-HoagYL-dox
Ac-PLG–HofGmphL-Dox
Ac-PLG–HofK(NMe2)L-Dox
Cap1i-PLG–S(OBn)YL-Dox
Cap2j-PCitk G–S(OBn)YL-Dox
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
47
13
59
100
90
100
29
70
83
88
90
100
95
81
87
90
90
100
Ac-
Ac-
Ac-
Ac-
Ac-
c
c
c
c
c
E-PQG–S(OBn)YL-Dox
E-PCitG–S(OBn)YL-Dox
E-PLG–S(OBn)YL-Dox
E-PLG–C(SBn)YL-Dox
E-PLG–HoyYL-Dox
Ac-bD-PLG–S(OBn)YL-Dox
Cap3l-PQG–S(OBn)YL-Dox
Cap3-PLG–S(Bn)YL-Dox
Cap2-PLG–S(OBn)YL-Dox
Cap3-PCitG-T(OBn)YL-Dox
Cap4m-PSG-T(OBn)YL-Dox
30
37
59
114
76
111
>120
>120
>2.4(2.9)n
(2.9)
(2.5)
68
79
a
b
c
% Remaining after 6 h in blood.
In pH 7.4 buffer solution.
O-Methylserine.
O-Benzylserine.
Homophenylalanine.
d
e
f
Homotyrosine.
g
h
i
2-Amino-4-(pyridine-4-yl)butanoic acid.
N-Methylpiperazinepropylglycine.
Cap1: succinic acid.
Cap2:2-sulfoacetic acid.
Citrulline.
Cap3:3-sulfobenzoic acid.
Cap4: 3,5-disulfobenzoic acid.
Value in bracket was the solubility in 5% dextrose solution.
j
k
l
m
n

Z. Hu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 853–856
855
Table 3
(A) After administration of Dox(1.4µmol/kg)
Tumor/heart distribution ratio of Dox in neu mice 24 h after iv dose
10000
1000
100
10
Compound
Dox
1
17
19
20
23
24
8.2
tumor
heart
Dox ratioa tumor/heart
Tumor Dox concentration (pmol/g) 2600 310 500 104 400 290
6.7
7.2
11
14
a
Normalized ratio.
saline
Dox 7um/k q4dx3
Dox 14um/k q4dx3
conjugate19,14um/k, qdx10
10000
1
0
500
1000
1500
Time (min)
1000
100
10
(B) After administration of conjugate 19 (1.4µmol/kg)
10000
tumor
heart
1000
100
10
1
0
10
20
30
40
50
60
time (day)
Figure 3. Tumor growth inhibition in HT1080 mice.
such as succinic acid (15), 2-sulfoacetic acid (16 and 25) or Ac-c-
1
glutamic acid (17–21), both the solubility and stability of the con-
jugates improved. Further exploration found capping groups such
as Ac-b-aspartic acid (22), 3-sulfobenzoic acid (23, 24 and 26) or
3,5-disulfobenzoic acid (27) were tolerated with increased solubil-
ity. This SAR exercise resulted in a series of peptide–Dox conju-
gates with excellent profiles in terms of enzyme selectivity,
stability, solubility, and lack of acute toxicity that met our program
criteria.
0
200
400
600
800
1000
1200
1400
1600
Time (min)
Figure 2. Dox deposition in HT1080 xenografts and heart tissue after administra-
tion of Dox (A) and conjugate 19 (B).
three prime side and three non-prime side residues to achieve
optimal MMP-2, -9 and -14 cleavage and selectivity over
neprilysin.
To determine how the peptide–Dox conjugate prodrugs were
metabolized in animals, HT1080 cells were used as a model sys-
tem. MMP-selective conjugates were injected into mice with
HT1080 xenografts to determine the tissue distribution of Dox fol-
lowing the procedure described by us previously.17 As shown in
Figure 2A, B and Table 2, when the peptide–Dox conjugates were
administered, more Dox was deposited in HT1080 tumor than in
heart with tumor to heart Dox ratio ranging from 6.5 to 17, consis-
tent with our previous results. Further metabolic studies showed
conjugates disappeared rapidly from the plasma compartment
after dosing. Dox levels were below detectable limits in plasma.
In heart tissue, L-Dox levels (data not shown) were lower than
the levels of conjugates, and heart tissue accumulated Dox pre-
sumably by metabolism of L-Dox. Unlike plasma and heart, L-
Dox levels in HT1080 tumor were higher than those of conjugates.
Since L-Dox did not preferentially accumulate in HT1080 tumor,
these findings support the idea that the conjugates were preferen-
tially metabolized in HT1080 tumor relative to heart and plasma.
This was further confirmed by injecting a non-MMP cleavable con-
jugate into mice resulting in no detectable Dox in HT1080 xeno-
grafts. Similar preferential deposition of Dox in tumors (tumor/
heart ratio 6.7–14) was observed when selected conjugates (Table
3) were administered to neu mice but with less Dox exposures than
in HT1080 mice.
Further SAR efforts kept P1 Gly and P0 Leu constant and focused
3
on hexapeptide conjugates with three prime side and three non-
prime side residues. When P01 was changed from Leu and O-methyl
serine to bulky hydrophobic residues such as homophenylalanine
(Hof), O-benzyl serine (S(OBn)) and 2-amino-4-(pyridine-4-
yl)butanoic acid (Hoa), the selectivity over neprilysin increased
(5-12). Replacing the small P02 residue Ala (9) with amino acids
containing larger side groups (Y in 10 and Hoy in 11) further in-
creased selectivity and improved MMP cleavage. It was also found
that P2 and P0 could tolerate a variety of amino acids (natural and
2
non-natural) of different sizes and properties such as N-methyl
piperazinepropylglycine (13), N,N-dimethyllysine (14), citrul-
line(16, 18) and glutamine (17). This discovery provided an oppor-
tunity to incorporate amino acids with polar side groups such as
basic amines to make conjugates with increased aqueous solubil-
ity. Unfortunately, it was observed that conjugates with positively
charged functional groups such as basic amine derivatives incorpo-
rated at P0 , P2 positions or as capping groups such as in 12, 13 and
2
14, frequently caused acute toxicity when administered to mice.
Capping groups of the conjugates also affected the stability and
solubility properties. Simple acetyl (Ac) capped conjugates (9–12)
were quite stable in blood but had limited solubility in pH 7.4 buf-
fer solution. With the introduction of water soluble capping groups
Table 2
Tumor/heart distribution ratio of Dox in HT1080 mice 24 h after iv dose
Compd
Dox
17
18
19
21
23
24
25
26
27
Dox ratioa Tumor/heart
1
16
16
17
10
17
17
16
9.7
13
Dox Conc. in tumor (pmol/g)
1600
993
800
900
1450
870
727
1500
790
600
a
Normalized ratio.

856
Z. Hu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 853–856
2. Von Hoff, D.; Layard, M. W.; Basa, P.; Davis, H. L.; Von Hoff, A. L.; Rozencweig,
Table 4
M.; Muggia, F. M. Ann. Intern. Med. 1979, 91, 710.
Bone marrow toxicity in mice
3. Yoneda, Y.; Steiniger, S. C. J.; Capkova, K.; Mee, J. M.; Liu, Y.; Kaufmann, G. F.;
Janda, K. D. Bioorg. Med. Chem. Lett. 2008, 18, 1632.
Compound
Dose schedule
Reticulocyte% of RBCa
4. Ravel, D.; Dubois, V.; Quinonero, J.; Meyer-Losic, F.; Delord, J. P.; Rochaix, P.;
Nicolazzi, C.; Ribes, F.; Mazerolles, C.; Assouly, E.; Vialatte, K.; Hor, I.; Kearsey,
J.; Trouet, A. Clin. Cancer Res. 2008, 14, 1258.
5. Meyer-Losic, F.; Quinonero, J.; Dubois, V.; Alluis, B.; Dechambre, M.; Michel, M.;
Cailler, F.; Fernandez, A.-M.; Trouet, A.; Kearsey, J. J. Med. Chem. 2006, 49, 6908.
6. Garsky, V. M.; Lumma, P.; Feng, D.-M.; Wai, J.; Ramjit, H. G.; Sardana, M. K.;
Oliff, A.; Jones, R. E.; DeFeo-Jones, D.; Freidinger, R. M. J. Med. Chem. 2001, 44,
4216.
7. DiPaola, R. S.; Rinehart, J.; Nemunaitis, J.; Ebbinghaus, S.; Rubin, E.; Capanna, T.;
Ciardella, M.; Doyle-Lindrud, S.; Goodwin, S.; Fontaine, M.; Adams, N.;
Williams, A.; Schwartz, M.; Winchell, G.; Wickersham, K.; Deutsch, P.; Yao,
S.-L. J. Clin. Oncol. 2002, 20, 1874.
8. Woessner, J.; Nagase, H. Matrix Metalloproteinases and TIMPs, 1st ed.; Oxford
University Press: New York, 2000.
Saline
Dox
Dox
Dox
Dox
3.7
0.4
0.1
0.3
0.2
3.5
2
4
6
14
14
l
l
l
l
l
mol/kg qdx10
mol/kg qdx10
mol/kg qdx10
mol/kg qdx10
mol/kg qdx10
Conjugate 19
a
Reticulocytes measured 3 days after last dose.
The efficacies of selected conjugates were determined by mea-
suring tumor growth inhibition in mice implanted with HT1080
and treated with Dox or conjugates. Shown in Figure 3 as a repre-
sentative example, conjugate 19 was more effective than Dox. At
9. Egeblad, M.; Werb, Z. Nat. Rev. Cancer 2002, 2, 161.
10. Turner, A. In Handbook of Proteolytic Enzymes; Barrett, A., Rawlings, N.,
Woessner, J., Eds.; Academic Press: San Diego, 1998; p 1080.
11. Trouet, A.; Passioukov, A.; Van Derpoorten, K.; Fernandez, A.-M.; Abarca-
Quinones, J.; Baurain, R.; Lobl, T. J.; Oliyai, C.; Shochat, D.; Dubois, V. Cancer Res.
2001, 61, 2843.
14
48 days. It was superior to Dox at its maximum tolerated doses
(14 mol/kg, q4dx3). It was also observed that mice treated with
lmol/kg qdx10, 19 reduced tumor size to baseline for up to
l
12. Netzel-Arnett, S.; Sang, Q. X.; Moore, W. G. I.; Navre, M.; Birkedal-Hansen, H.;
Van Wart, H. E. Biochemistry 1993, 32, 6427.
conjugate 19 did not show signs of toxicity.
To further investigate the toxicity of the conjugates, marrow
toxicity was determined by quantification of reticulocytes, the
short-lived precursors of red blood cells (RBC). As shown in Table
4, there was a dramatic drop in reticulocyte count for mice treated
13. Grant, G.; Giambernardi, T.; Grant, A.; Klebe, R. Matrix Biol. 1999, 18, 145.
14. Nagase, H.; Fields, G. Biopolymers 1996, 40, 399.
15. Berman, G.; Green, M.; Sugg, E.; Andereg, R.; Millington, D. S.; Norwood, D. L.;
McGreehan, J.; Wiseman, J. J. Biol. Chem. 1992, 267, 1434.
16. Chan, W. C.; White, P. D. Fmoc Solid Phase Peptide Synthesis:
A Practical
Approach; Oxford: Oxford University Press, 2000. A representative procedure
for the synthesis of conjugate 25: Fmoc-Leu–Wang resin (0.9 mmol/g load,
0.25 mmol scale) was loaded on a peptide synthesizer and peptide synthesis
was performed under HBTU/HOBt/DIPEA in dimethylformamide (DMF)
coupling condition and piperidine (20% in DMF) for Fmoc deprotection to
give PLGS(OBn)Y(OtBu)L-Wang resin. To this resin swollen in DMF were added
2-sulfoacetic acid (5 equiv), Bop reagent (3 equiv) and DIPEA (5 equiv). After
shaken for 2 h, the resin was washed alternatively with water, DMF and DCM
for three times and dried in vacuo. Peptide cleavage from the resin was
performed by adding 90% TFA/10% DCM (10 mL) and triisopropylsilane
(0.2 mL). The resin was removed by filtration and further washed with DCM.
Solvent was removed from filtrate to leave a crude product. Purification by
reverse phase HPLC (acetonitrile–water with 0.05% TFA) provided the
corresponding capped peptide 140 mg (yield 65%, MS: 861.5 [M+1]). The
peptide was dissolved in DMF (5 mL), to which were added Dox hydrochloride
(Aldrich, 94 mg, 0.16 mmol), BOP reagent (108 mg, 0.24 mmol) and DIPEA
(0.083 mL, 0.48 mmol). The reaction was stirred at 0 °C for 2 h. Solvent was
removed in vacuo. The sample was dissolved in H2O:CH3CN and purified using
a reverse phase HPLC (Rainin, C18 reverse phase column, 41.4 ꢃ 250 mm) with
a linear gradient from 30% to 80% acetonitrile, 0.05% ammonium acetate over
20 min with a flow rate of 45 mL/min. Fractions were pooled and freeze dried
to afford the purified peptide–Dox conjugate 25 122 mg (55%). ES MS(Mꢂ1):
1384.5.
with Dox even at a dose as low as 2
mice treated with conjugate 19 at its efficacious dose of
14 mol/kg qdx10 showed no difference in the levels of reticulo-
lmol/kg qdx10. In contrast,
l
cytes (3.5%) from the vehicle-treated mice (3.7%).
In a mouse pharmacokinetic study, conjugate 19 demonstrated
a clearance 0.4 L/h/kg, a volume of distribution of 0.4 L/kg and a
half life of 0.6 h after iv administration.
In summary, a series of soluble and stable peptide–Dox conju-
gates was discovered and optimized for selectivity, solubility, effi-
cacy and toxicity profiles. These conjugates were selectively
cleaved by MMPs and preferentially delivered Dox to tumor rela-
tive to heart tissues. Conjugate 19 was shown to be more effective
than Dox with less toxicity in HT1080 mouse model.
Acknowledgment
We thank Dr. James A. Johnson for proof reading the Letter.
References and notes
17. Albright, C. F.; Graciani, N.; Han, W.; Yue, E.; Stein, R.; Lai, Z.; Diamond, M.;
Dowling, R.; Grimminger, L.; Zhang, S.; Behrens, D.; Musselman, A.; Bruckner,
R.; Zhang, M.; Jiang, X.; Hu, Z.; Higley, A.; DiMeo, S.; Rafalski, M.; Mandlekar, S.;
Car, B.; Yeleswaram, S.; Stern, A.; Copeland, R. A.; Combs, A.; Seitz, S. P.;
Trainor, G. L.; Taub, R.; Huang, P.; Oliff, A. Mol. Cancer Ther. 2005, 4, 751.
1. Hardman, J.; Limbird, L.; Gilman, A. The Pharmacologic Basis of Therapeutics,
10th ed.; McGraw-Hill: New York, 2001.