Development of albumin-binding doxorubicin pro-drugs that are cleaved by prostate-specific antigen

Article abstract of DOI:10.1002/ardp.200500130

Prostate-specific antigen (PSA) is a serine protease that is overexpressed in prostate carcinoma and represents a molecular target for selectively releasing an anticancer agent from a prodrug formulation. In this work, we developed albumin-binding prodrug

Full text of DOI:10.1002/ardp.200500130

462
DOI 10.1002/ardp.200500130
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Development of Albumin-binding Doxorubicin Pro-
drugs that are Cleaved by Prostate-specific Antigen
Felix Kratza, Ahmed Mansoura, Jens Soltaua, Andre Warneckea, Iduna Fichtnerb,
Clemens Ungera, Joachim Drevsa
a Tumor Biology Center, Freiburg, Germany
b Max-Delbrück Center, Berlin, Germany
Prostate-specific antigen (PSA) is a serine protease that is overexpressed in prostate carcinoma and
represents a molecular target for selectively releasing an anticancer agent from a prodrug formulation.
In this work, we developed albumin-binding prodrugs with the structures MT-Ser-Ser-Tyr-TyrϪ
Ser-Gly-DOXO, MT-Asn-Ser-Ser-TyrϪPhe-Gln-DOXO (MT ϭ maleimidotriethyleneglycol acid;
DOXO ϭ Doxorubicin) or EMC-Arg-Arg-Ser-Ser-Tyr-TyrϪSer-Gly-DOXO (EMC ϭ ε-maleimidoca-
proic acid; X ϭ amino acid). The maleimide Doxorubicin derivatives bound rapidly to the cysteine-
34 position of endogenous and exogenous albumin and were efficiently cleaved by PSA at the P1-PЈ1
scissile bond, releasing a respective Doxorubicin dipeptide (Ser-Gly-DOXO or Phe-Gln-DOXO). The
derivative containing arginine residues (EMC-Arg-Arg-Ser-Ser-Tyr-TyrϪSer-Gly-DOXO) exhibited ex-
cellent water solubility for intravenous administration. Subsequent biological evaluation was focused
on a PSA-negative xenograft model (PC 3) and a PSA-positive xenograft model (CWR22) in order
to assess the selectivity of our therapeutic approach. EMC-Arg-Arg-Ser-Ser-Tyr-TyrϪSer-Gly-DOXO
showed no in vivo activity in the PSA-negative PC 3 model, but good activity in the CWR22 PSA-
positive model that was comparable to Doxorubicin.
Keywords: Doxorubicin; Macromolecular prodrug; Human serum albumin; Prostate-specific antigen (PSA)
Received: May 12, 2005; Accepted: June 28, 2005
We have recently proposed a macromolecular prodrug strat-
egy for improving the therapeutic index of anticancer agents
that is based on two features: (a) rapid and selective binding
of thiol-reactive prodrugs to the cysteine-34 position of en-
dogenous albumin after intravenous administration, and (b)
acid-sensitive or enzymatic release of the albumin-bound
drug at the tumor site [1Ϫ3].
target for selectively releasing an anticancer agent from a
prodrug formulation. Based on our synthetic experience
with albumin-binding prodrugs that are cleaved by tumor-
associated proteases [3, 7], we set out to develop PSA-
specific Doxorubicin prodrugs that bind to the cysteine-34
position of human serum albumin (HSA) after intravenous
administration. The general formula of such prodrugs is
shown in Figure 1.
Several albumin-binding prodrugs with the anticancer ag-
ents Doxorubicin, Camptothecin and platinum(II) com-
plexes have shown promising antitumor activity in preclini-
cal models [1Ϫ5]. An acid-sensitive Doxorubicin prodrug,
the (6-maleimidocaproyl) hydrazone derivative of Doxorub-
icin (DOXO-EMCH) has been evaluated in a phase I clini-
cal trial and has demonstrated antitumor efficacy and a
fourfold increase in the maximum tolerated dose (MTD)
compared to the clinical standard Doxorubicin [6].
The prodrugs consist of a maleimide derivative that can
contain an additional water-soluble moiety, a peptide spacer
and the anticancer agent Doxorubicin.
When designing the prodrugs, we were aided by the work
of other groups that have investigated the sequence speci-
ficity of PSA [8Ϫ12]. These studies have revealed that dis-
tinct oligopeptides, i.e. hexa- to decapeptides, show a high
The goal of this project was to develop albumin-binding
prodrugs of Doxorubicin that are cleaved by the prostate-
specific antigen (PSA). PSA is a serine protease that is over-
expressed in prostate carcinoma and represents a molecular
Correspondence: Felix Kratz, Tumor Biology Center, Department of
Medical Oncology, Clinical Research, Breisacher Straße 117,
D-79106 Freiburg, Germany. Phone: ϩ49 761 2062176, Fax: ϩ49
761 2062905, e-mail: felix@tumorbio.uni-freiburg.de
Figure 1. General structure of albumin-binding Doxorubicin
prodrugs with incorporated peptide spacers.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Doxorubicin Prodrugs Cleaved by Prostate-specific Antigen
463
degree of substrate specificity for human PSA. For our stud-
ies we selected the three peptide sequences as shown in Fig-
ure 2.
incorporated either AsnϪSerϪSerϪTyrϪPheϪGln or
SerϪSerϪTyrϪTyrϪSerϪGly; the synthetic route for pre-
paring PSA3 and PSA4 is depicted in Scheme 1.
The hexapeptide sequence (P4-PЈ2) in these substrates is re-
ported to be cleaved efficiently by PSA at the P1-PЈ1 bond
[11, 12]. In this work, we report on the synthesis and anti-
tumor efficacy of albumin-binding Doxorubicin prodrugs
incorporating the above-mentioned peptide sequences. Bio-
logical evaluation was focused on a PSA-negative model
(PC 3) and on PSA-positive models (CWR22, LNCAP) in
order to assess the selectivity of our therapeutic approach.
The maleimide hexapeptide derivatives were obtained by
building up the peptide on a solid phase and introducing
maleimidotriethylene glycol acid in the final step at the
N-terminal position. The maleimide-derivatized peptides
were cleaved from the resin and purified by reverse-phase
HPLC. Subsequently, the maleimide hexapeptide derivatives
were reacted for 18 h with Doxorubicin hydrochloride in
DMF in the presence of 1-hydroxybenzotriazole, 4-methyl-
morpholine, and N,NЈ-diisopropyl carbodiimide (DIPC) as
the coupling agent. DMF was removed in high vacuo, and
the maleimide-Doxorubicin hexapeptide-derivatives PSA3
and PSA4 were isolated through chromatography on silica
gel. Identity and purity was confirmed by mass spec-
trometry and reverse-phase HPLC, respectively (see Exper-
imental).
Results
In our previous work, we have synthesized maleimide octa-
peptide derivatives with Doxorubicin containing maleimi-
dotriethyleneglycol acid that rendered excellent water solu-
bility due to the incorporation of hydrophilic ether bonds
[3]. Thus, we used this maleimide derivate for synthesizing
the first Doxorubicin derivatives PSA3 and PSA4 that
Both, PSA3 and PSA4 exhibited water solubility of
~1 mg/mL, which was considerably lower than that ob-
served with analogously constructed maleimide octapeptide
derivatives with Doxorubicin (Mal-Gly-Pro-Leu-GlyϪ
Ile-Ala-Gly-Gln-DOXO, Mal
ϭ
maleimidotriethylene-
glycol acid) that are cleaved by matrix metalloprotease 2
(MMP-2), which on the whole did not contain more polar
amino acid groups [3]. Nonetheless, we used PSA3 and
PSA4 for preparing respective albumin conjugates and pro-
bing the cleavage profile before designing prodrugs with im-
proved water solubility.
Figure 2. Peptide sequences used in the studies.
Scheme 1.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

464
Kratz et al.
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Scheme 2.
For preparing the albumin conjugates, PSA3 and PSA4
were coupled to the cysteine-34 position of HSA, thus ob-
taining A-PSA3 and A-PSA4 (Scheme 2).
PSA, and chromatograms were recorded at λ ϭ 495 nm
using reverse-phase HPLC. As shown in Figure 3A and B
after incubation of A-PSA3 or A-PSA4 with PSA for up to
24 h, the initial peak of the albumin conjugate disappears
with time and a new distinct peak (Rt ~300 or ~400 s) is
observed in each case.
Cleavage of the albumin conjugates A-PSA3 and A-PSA4
by PSA.
In order to investigate whether and how fast the albumin
conjugates A-PSA3 and A-PSA4 are cleaved by PSA, the
conjugates were incubated with enzymatically active human
LC-MS detected a mass of 711.2 (Naϩ[Ser-Gly-DOXO])
and 842.3 (Naϩ[Gln-Phe-DOXO]) for the respective peak
corresponding to the cleaved Doxorubicin dipeptide
Figure 3. (A, B) Chromatograms of cleavage studies of A-PSA3 (A) and A-PSA4 (B) with PSA (20 µg/mL) after 30 min, 3 h,
9 h, and 24 h at 37°C (pH 7.4). Concentration of the anthracycline was 50 µM. For chromatographic conditions see Experimental.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Doxorubicin Prodrugs Cleaved by Prostate-specific Antigen
465
[Ser-Gly-DOXO] for A-PSA4 and [Phe-Gln-DOXO] for
A-PSA3.
Besides the major cleavage products [Gln-Phe-DOXO] and
[Ser-Gly-DOXO], a smaller peak eluting behind the main
peak for the albumin conjugates was observed, which is un-
identified at present.
A-PSA4 demonstrated a faster rate of cleavage than
A-PSA3 (compare Figure 3A and B), and the incorporated
peptide sequence [SerϪSerϪTyrϪTyrϪSerϪGly] was thus
chosen for the development of a water-soluble albumin-
binding prodrug that could be conveniently administered in-
travenously.
Figure 5. Chromatograms of cleavage studies of A-PSA4-
Arg with PSA (20 µg/mL) after 5 min, 3 h, and 7 h at 37°C
(pH 7.4). Concentration of the anthracycline was 100 µM. For
chromatographic conditions see Experimental.
Development of the water-soluble Doxorubicin prodrug
PSA4-Arg
For developing a water-soluble analogue of PSA4, we sub-
stituted maleimidotriethylene glycol acid through 6-maleim-
idocaproic acid and introduced two adjacent arginine resi-
dues in the sequence (see Figure 4).
that free thiol groups are not present in the majority of
circulating serum proteins except for albumin and that the
HS-group of cysteine-34 of endogenous HSA is the most
reactive thiol group in human plasma [2]. In order to deter-
mine the coupling rate and selectivity of PSA4-Arg for this
sulfhydryl group, PSA4-Arg was incubated with human
blood plasma and the samples were subsequently analyzed
by reverse-phase chromatography. Chromatograms after an
incubation time of 5 and 90 min are shown in Figure 6.
Protein components were detected at 280 nm and the
anthracycline moiety simultaneously at 495 nm. Binding of
PSA4-Arg to endogenous albumin is essentially complete
after 5 min, with only traces of PSA4-Arg eluting at 750 s.
PSA4-Arg was synthesized in analogy to PSA4 (see Scheme
2 and Experimental). In contrast to PSA4, PSA4-Arg exhi-
bited excellent water solubility (~10 mg/mL) and was conse-
quently selected for further in vitro and in vivo studies.
Cleavage of the albumin conjugate A-PSA4-Arg by PSA
A further benefit of PSA4-Arg over PSA4 was a faster rate
of cleavage by PSA (see Figure 5). Cleavage to [DOXO-Gly-
Ser] is essentially complete after 7 h at 37°C under identical
conditions (compare with Figure 3B).
Albumin-binding properties of PSA4-Arg
Stability of the albumin conjugate of PSA4-Arg in human
blood plasma was assessed after PSA4-Arg had been incu-
bated at 37°C with plasma for up to 20 h. Reverse-phase
The HS-group of cysteine-34 of HSA is a unique and ac-
cessible functional group of a plasma protein, considering
Figure 4. Structure of the water-soluble Doxorubicin derivative PSA4-Arg.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

466
Kratz et al.
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Figure 6. Chromatograms of incubation studies of PSA4-Arg
with human plasma at 37°C after 5 and 90 min. Concentra-
tion of the anthracycline was 100 µM. For chromatographic
conditions see Experimental.
Figure 7. Chromatograms of incubation studies of PSA4-Arg
with human plasma at 37°C over 20 h. Concentration of the
anthracycline was 100 µM. For chromatographic conditions
see Experimental.
HPLC shows that the albumin-bound form of PSA4-Arg is
highly stable over this time (see Figure 7).
2. PSA4-Arg or A-PSA4-Arg are an order of magnitude
more active than PSA4.
In vitro activity of Doxorubicin, DOXO-Gly-Ser, PSA4-Arg,
A-PSA4, and A-PSA4-Arg against LNCAP, CWR22, and
PC 3 prostate tumor cells
3. PSA4-Arg in its free form shows similar activity when
compared to its albumin-conjugated form.
4. The cleavage product DOXO-Gly-Ser generally shows
less cytotoxicity than PSA4-Arg or A-PSA4-Arg and is
considerably less active than free Doxorubicin.
The antiproliferative activity of A-PSA4, A-PSA4-Arg,
PSA4-Arg, free Doxorubicin and of the cleavage product
DOXO-Gly-Ser-OH were assessed in the following three
tumor cell lines using the MTT assay: LNCAP human pros-
tate carcinoma (PSA positive), CWR22 human prostate car-
cinoma (PSA positive), and PC 3 human prostate carci-
noma (PSA negative).
5. PSA4-Arg, A-PSA4-Arg and PSA4 are more active in
the PSA-positive cell lines, but this is also true for free
Doxorubicin.
A critical comparison of the in vitro data does not let us
conclude that PSA4-Arg, A-PSA4-Arg and PSA4 demon-
strate selectivity for PSA-positive cell lines. An explanation
for this could be the low concentrations of PSA that we
found in the cell-conditioned medium of LNCAP cells after
48 h (~28 ng/mL) and CWR22 cells (~10 ng/mL) (no PSA
The IC50 values that were obtained after a 72-h cell expo-
sure are summarized in Table 1. The cytotoxicity studies
against these three cell lines reveal some salient points:
1. Doxorubicin is the most active compound in all three
prostate carcinoma cell lines.
Table 1. IC50 values (µM)† for Doxorubicin, A-PSA-4, A-PSA4-Arg, PSA4-Arg in LNCAP, CWR22 and PC 3 prostate tumor
cells.
Compound
IC50 value in LNCAP cells
IC50 value in CWR22 cells
IC50 value in PC 3 cells
µM
µM
µM
Doxorubicin
A-PSA4
0.09 0.4
6.9 1.2
0.65 0.07
0.88 0.7
0.17 0.07
19 1.4
1.1 0.34
1.5 0.14
1.25 0.64
30 3.6
3.3 0.67
5.8 1.4
>>60
PSA4-Arg
A-PSA4-Arg
DOXO-Gly-Ser
30
4
17
2
†
IC50 values (50% inhibitory concentration) represent the means standard deviation of three independent experiments.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Doxorubicin Prodrugs Cleaved by Prostate-specific Antigen
467
was found in cell-conditioned medium of the PSA-negative
cell line, PC 3). These levels are not sufficient to cleave large
amounts of DOXO-Gly-Ser from the albumin conjugate,
which we could demonstrate through incubation studies of
A-PSA4-Arg with the cell-conditioned medium of these two
cell lines (see below). The low levels of PSA in the two PSA-
positive cell lines are in striking contrast to the levels that
we determined in tumors of CWR22 as well as in native
prostate carcinoma specimens of human patients (see be-
low).
3 cell culture supernatants and samples were analyzed over
24 h using HPLC (see Figure 8A, B).
DOXO-Gly-Ser-OH was not detected in the PSA-negative
cell line PC 3, and only small amounts were found in the
supernatants of the PSA-positive cell line LNCAP.
In vivo activity of PSA4-Arg in a PSA-negative xenograft
model (PC 3) and a PSA-positive xenograft model (CWR22)
In subsequent in vivo experiments, the antitumor efficacy of
PSA4-Arg was evaluated in nude mice in the PC 3 prostate
carcinoma model (PSA negative) and the CWR22 prostate
carcinoma model (PSA positive) in a strict comparison to
free Doxorubicin.
Incubation studies of A-PSA4-Arg with cell-conditioned
supernatants of LNCAP and PC 3 prostate tumor cells
In order to obtain a picture of the stability and cleavage
profile of A-PSA4-Arg in supernatants of these biological
samples, A-PSA4-Arg was incubated with LNCAP and PC
Preliminary toxicity studies in mice showed that the maxi-
mum tolerated dose of PSA4-Arg was approximately four-
Figure 8. (A, B) Chromatograms of incubation studies of 50 µM A-PSA4-Arg in PC 3 cell culture supernatant (A) and in LNCAP
cell culture supernatant [c(PSA) ~28 ng/mL)] (B) at 37°C. For chromatographic conditions see Experimental.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

468
Kratz et al.
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Figure 9. Curves depicting tumor growth inhibition of sub-
cutaneous PC 3 xenografts under therapy with Doxorubicin
and PSA4-Arg. Dose (i.v.); Doxorubicin: 2 ϫ 13.3 µmol/kg
(= 2 ϫ 8 mg/kg Doxorubicin); PSA4-Arg: 2 ϫ 13.3 µmol/kg
(= 2 ϫ 8 mg/kg Doxorubicin equivalents), 3 ϫ 39.9 µmol/kg
(= 3 ϫ 24 mg/kg Doxorubicin equivalents); eight mice per
group.
Figure 10. Curves depicting tumor growth inhibition of subcu-
taneous CWR22 xenografts under therapy with Doxorubicin
and PSA4-Arg. Dose (i.v.); Doxorubicin: 2 ϫ 6.65 µmol/kg
(= 2 ϫ 4 mg/kg Doxorubicin; 2 ϫ 13.3 µmol/kg (= 2 ϫ 8 mg/
kg Doxorubicin); PSA4-Arg: 2 ϫ 13.3 µmol/kg (= 2 ϫ 8 mg/
kg Doxorubicin equivalents), 3 ϫ 39.9 µmol/kg (= 3 ϫ 24 mg/
kg Doxorubicin equivalents), 3 ϫ 59.9 µmol/kg (= 3 ϫ 36 mg/
kg Doxorubicin equivalents); four mice per group.
fold higher than for free Doxorubicin. Thus, the antitumor
efficacy of PSA4-Arg was initially compared to that of
Doxorubicin in the PC 3 model at the following doses:
Doxorubicin: 2 Ϯ 13.3 µmol/kg (ϭ 2 Ϯ 8 mg/kg Doxorub-
icin) corresponding to the maximum tolerated dose (MTD)
of Doxorubicin in nude mice that is routinely used; PSA4-
Arg: 2 Ϯ 13.3 µmol/kg (ϭ 2 Ϯ 8 mg/kg Doxorubicin equiva-
lents), 3 Ϯ 39.9 µmol/kg (ϭ 3 Ϯ 24 mg/kg Doxorubicin
equivalents). The results of this animal experiment are
shown in Figure 9 and Table 2. Doxorubicin at its optimal
dose of 2 Ϯ 13.3 µmol/kg produced a significant inhibition
in tumor growth. In contrast, therapy with PSA4-Arg at
2 Ϯ 13.3 µmol/kg and 3 Ϯ 39.9 µmol/kg was well tolerated
(no body weight change was noted) but produced a negli-
gible antitumor response in this PSA-negative model.
sponse (see Figure 10 and Table 3). A dose of 3 Ϯ 36 mg/kg
Doxorubicin equivalents was toxic, however. Doxorubicin
was tested at two doses: 2 Ϯ 8 mg/kg, which is the dose that
is used routinely in nude mice models, and a lower dose of
2 Ϯ 4 mg/kg, since for models such as CWR22, NOD/SCID
mice have to be used for successful growth. These mice are
more susceptible to cytotoxic therapies compared with nude
mice. Indeed, therapy with Doxorubicin at the dose of
2 Ϯ 8 mg/kg proved to be toxic, with all animals dying
within 21 days (see Figure 10). In contrast, 2 Ϯ 4 mg/kg
produced a good antitumor response, but also a high
reduction in body weight (see Table 3).
In order to explain the therapeutic differences of PSA4-Arg
in the PSA-negative and PSA-positive xenograft model, we
first determined the PSA concentrations in PC 3 and
CWR22 tumors as well as in three samples of native human
prostate carcinoma. In the PC 3 tumors, we were unable
In contrast, therapy with PSA4-Arg in the PSA-positive
xenograft model CWR22 at 2 Ϯ 8 mg/kg and 3 Ϯ 24 mg/kg
Doxorubicin equivalents produced a good antitumor re-
Table 2. Antitumor activity of Doxorubicin and PSA4-Arg against human PC 3 prostate xenografts.
Compound
Dose†
mg/kg
Treatment
day
Mortality
Body weight change,%
Optimum T/C
%
days 6Ϫ16
Buffer
6, 13
6, 13
6, 13, 20
6, 13
None
None
None
none
ϩ5
ϩ4
Ϫ1
Ϫ9
PSA4-Arg
PSA4-Arg
Doxorubicin
8
24
8
93
80
15‡
†
Dose refers to Doxorubicin equivalents.
Significant to buffer; statistical analysis was performed with the U-test (MannϪWhitney) with p < 0.05.
‡
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Doxorubicin Prodrugs Cleaved by Prostate-specific Antigen
469
Table 3. Antitumor activity of Doxorubicin and PSA4-Arg against human CWR22 prostate xenografts.
Compound
Dose†
mg/kg
Treatment
day
Mortality
Body weight change,%
Optimum T/C
%
days 6Ϫ16
Buffer
13, 20
13, 20
13, 20, 27
13, 20, 27
13, 20
ϩ10
Ϫ1
Ϫ1
Ϫ8
Ϫ36
Ϫ28
Ϫ
PSA4-Arg
PSA4-Arg
PSA4-Arg
Doxorubicin
Doxorubicin
8
24
36
8
28‡
24‡
Ϫ
100%
100%
Ϫ
27‡
4
13, 20
†
Dose refers to Doxorubicin equivalents.
Significant to buffer; statistical analysis was performed with the U-test (MannϪWhitney) with p < 0.05.
‡
Figure 12. Chromatograms of incubation studies of A-PSA4-
Arg with CWR22 tissue homogenate (PSA content
~2 µg/mL) at 37°C over 24 h. Concentration of the anthra-
cycline was 25 µM. For chromatographic conditions see
Experimental.
Figure 11. Chromatograms of incubation studies of A-PSA4-
Arg with PC 3 tissue homogenate at 37°C over 24 h. Con-
centration of the anthracycline was 50 µM. For chromato-
graphic conditions see Experimental.
to detect PSA using a microparticle immunoassay; in the
CWR22 tumors, a range of 0.2Ϫ2.8 µg/mL (0.8Ϫ11.2 µg/g)
was determined. The highest levels were found in human
surgical specimens of human prostate carcinoma with con-
centrations of 30Ϫ500 µg/mL (120Ϫ2000 µg/g).
Subsequently, incubation studies were performed with A-
PSA4-Arg and the respective tissue homogenates. As can be
seen in Figure 11, incubation of A-PSA4-Arg in homogen-
ates of PC 3 tumors that contained no detectable PSA re-
vealed that the dipeptide DOXO-Gly-Ser was not released
over 24 h.
Figure 13. Chromatograms of incubation studies of A-PSA4-
Arg with human prostate carcinoma tissue homogenate (PSA
content ~80 µg/mL) at 37°C over 3 h. Concentration of the
anthracycline was 50 µM. For chromatographic conditions
see Experimental.
In contrast, incubation studies with homogenates from
CWR22 tumors showed a gradual release of DOXO-Gly-
Ser over 24 h (Figure 12).
a) IC50 values (50% inhibitory concentration) represent the
means
ments.
standard deviation of three independent experi-
Release of DOXO-Gly-Ser from A-PSA4-Arg was consider-
ably faster in the homogenate of a native human prostate
tumor that showed levels of ~80 µg/mL, with cleavage being
essentially complete within 3 h (see Figure 13). However,
further degradation of DOXO-Gly-Ser to Doxorubicin
could not be demonstrated in these studies with any of the
PSA-positive tumor homogenates.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

470
Kratz et al.
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Discussion
Recently, we were able to demonstrate that a MMP-2
specific albumin-binding prodrug initially released
a
In recent years a number of low- and high-molecular-weight
prodrugs with Doxorubicin have been developed that were
designed to be cleaved by tumor-associated enzymes such
as β-glucuronidase (HMR 1826) [13], PSA [8, 10, 14Ϫ17], to
date unidentified peptidases (CPI-0004Na) [18], cathepsin B
(PK1 Ϫ Doxorubicin conjugated with N-(2-hydroxypropyl)
methacrylamide copolymer) [19], or MMP-2 (albumin-bind-
ing Doxorubicin prodrug) [3, 20] (reviewed in [20]).
Doxorubicin tetrapeptide [Ile-Ala-Gly-Gln-DOXO] in mela-
noma tumors and was then degraded to free Doxorubicin
by further proteases [3, 16, 17]. This prodrug exhibited en-
hanced efficacy over Doxorubicin in a melanoma xenog-
raft model.
Thus, our future goal will be to optimize the C-terminal
amino acid sequence in PSA4-Arg (EMC-Arg-Ser-Ser-Tyr-
TyrϪSer-Gly-DOXO) by substituting Gly for other amino
acids that allow free Doxorubicin to be released in PSA-
positive prostate tumors as the final cleavage product.
PSA is especially attractive as a target protease because it is
primarily expressed in prostate tissue and prostate carci-
noma. Besides the albumin-binding prodrug PSA4-Arg pre-
sented in this paper, two other low-molecular-weight
Doxorubicin prodrugs have been developed that aim to ex-
ploit PSA as a protease target [10, 14Ϫ17]. These Doxorub-
icin derivatives contain the peptide sequences Mu-His-Ser-
Experimental
Ser-Lys-Leu-Gln-Leu-OH (Mu
ϭ morpholinocarbonyl)
Chemicals, materials and spectroscopy
[LϪ377,202] and N-glutaryl-(hydoxypropyl)-Ala-Ser-cyclo-
hexaglycyl-Gln-Ser-Leu-OH bound to the amino position
of Doxorubicin. Both low-molecular-weight prodrugs were
designed to release DOXO-Leu following cleavage of PSA.
The dose dependency of L-377,202 and PSA4-Arg in vivo is
similar; i.e. a 5Ϫ7-fold increase in the maximum tolerated
dose (MTD) in mice is noted when compared to Doxorub-
icin. On the whole, all of the prodrugs have demonstrated
good antitumor activity in PSA-positive animal models.
Both L-377,202 and PSA4-Arg have additionally been
tested in PSA-negative models (DuPro-1 and PC 3, respec-
tively) and only marginal activity was seen for the prodrugs
in these studies, confirming a lack of cleavage of the incor-
porated PSA-specific sequence in the prodrugs.
Doxorubicin hydrochloride (MW 580.0) was purchased from Hande
Tech Development Co. USA, Inc.; Mal-Ser-Ser-Tyr-Tyr-Ser-Gly-
OH, Mal-Asn-Ser-Ser-Tyr-Phe-Gln-OH (Mal ϭ maleimidotriethy-
leneglycol acid) and 6-maleimidohexanoyl-Arg-Arg-Ser-Ser-Tyr-
Tyr-Ser-Gly-OH was custom-made by BACHEM AG (Bubendorf,
Switzerland); maleimidotriethylene glycol acid was prepared ac-
cording to our published procedure[5]; organic solvents: HPLC
grade (Merck, Darmstadt, Germany). All other chemicals used were
at least reagent grade and obtained from Sigma-Aldrich (Deisen-
hofen, Germany) or Merck and used without further purification.
HSA (20% solution) was purchased from Dessau Pharma, Ger-
many, and contained approximately 30% free thiol groups as as-
sessed with the Ellman·s test. Enzymatically active PSA was pur-
chased from Calbiochem (Bad Soden, Germany). The buffers used
were vacuum-filtered through a 0.2-µm membrane (Sartorius, Ger-
many) and thoroughly degassed with argon or nitrogen prior to use.
Cell culture media, supplements and fetal calf serum (FCS) were
purchased from Bio Whittaker (Serva, Heidelberg, Germany). All
culture flasks were obtained from Greiner Labortechnik (Fricken-
hausen, Germany). ESI-MS spectra were obtained on a Finnigan
MAT 312 with associated MAT SS 200 data system, using electron
spray ionization at 4.0 kV.ESI. UV/VIS-spectrophotometry was car-
ried out with a double-beam spectrophotometer U-2000 from Hita-
chi. LC-ESI-MS was performed by bioproof AG (Munich, Ger-
many). PC 3, LNCAP, CWR22 tumor cell lines were obtained from
American Type Culture Collection (ATCC). Samples of surgical
prostate tumor specimens were supplied by the Department of Urol-
ogy, University of Freiburg, Germany.
L-377,202 has been evaluated in a phase I study. Nineteen
patients with advanced hormone-refractory prostate cancer
were treated intravenously with L-377,202 at escalating dose
levels from 20 to 315 mg/m2 of L-377,202 [21]. Dose-limit-
ing grade 4 neutropenia was noted in two of two patients
that were treated with a dose of 315 mg/m2. The re-
commended dose for phase II studies was 225 mg/m2, which
corresponds to approximately 90 mg/m2 Doxorubicin
equivalents. PK studies demonstrated that L-377,202 was
cleaved to N-l-leucyldoxorubicin and Doxorubicin.
In summary, PSA4-Arg demonstrated good selectivity when
comparing the significant difference in antitumor activity
observed in the PSA-negative (PC 3) and the PSA-positive
(CWR22) model. Our investigations also emphasize the
need to elucidate the cleavage profile of enzymatically cleav-
able prodrugs in native tumor tissue at physiologically rel-
evant protease concentrations of PSA.
Methods
HPLC for binding and cleavage studies with PSA3, PSA4, PSA4-
Arg and the respective albumin conjugates was performed with a
BioLogic Duo-Flow System from Biorad (Munich, Germany),
which was connected with a Lambda 1000 visible monitor from
Bischoff (at λ ϭ 495 nm); UV-detection at 280 nm; column: Waters,
˚
300 A, Symmetry C18 (4.6 Ϯ 250 mm) with pre-column; chromato-
graphic conditions: flow: 1.2 mL/min, mobile phase: 27.5%
CH3CN, 72.5% 20 mM potassium phosphate (pH 7.0), mobile
phase B: CH3CN, gradient: 0Ϫ25 min 100% mobile phase; 25Ϫ40
min increase to 70% CH3CN, 30% 20 mM potassium phosphate;
40Ϫ50 min 70% CH3CN, 30% 20 mM potassium phosphate;
50Ϫ60 min decrease to initial mobile phase; injection volume:
50 µL.
Incubation studies with tumor homogenates reveal, how-
ever, that the full potential of PSA4-Arg has not been ex-
ploited, considering that a less active Doxorubicin dipeptide
DOXO-Gly-Ser and not Doxorubicin was released in PSA-
positive prostate carcinoma tissues.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
Doxorubicin Prodrugs Cleaved by Prostate-specific Antigen
471
PSA concentration in cell culture supernatants and tumor tissue
homogenates was determined nephelometrically with a micropar-
ticle immunoassay (MEIA) from Abbott.
Synthesis of the albumin conjugates of PSA3, PSA4 and PSA4-Arg
The albumin conjugates of PSA3, PSA4 and PSA4-Arg were pre-
pared by reacting them with commercially available albumin, which
contains approximately 30% mercaptalbumin, and isolating the
conjugate through size-exclusion chromatography: 10 mL HSA was
incubated at 37°C for 1 h with 12.1 mg PSA3, with 13.2 mg PSA4
or with 13.4 mg PSA4-Arg. The albumin conjugate was obtained
by subsequent size-exclusion chromatography (Sephacryl HR100;
buffer 0.004 M sodium phosphate, 0.15 M NaCl, pH 6.5). The con-
tent of anthracycline in the sample was determined using the ε-
value for Doxorubicin (ε495 (pH 7.4) ϭ 10 650 MϪ1 cmϪ1) and
Synthesis and characterization of PSA3 and PSA4
PSA3 or PSA4 was prepared by reacting the maleimidotriethylene
glycol hexapeptide derivative [Mal-Asn-Ser-Ser-Tyr-Phe-Gln-OH]
or [Mal-Ser-Ser-Tyr-Tyr-Ser-Gly-OH] with Doxorubicin hydrochlo-
ride in DMF using a standard coupling procedure: 58 mg (0.1
mmol) Doxorubicin hydrochloride, 102.8 mg (0.1 mmol) Mal-Asn-
Ser-Ser-Tyr-Phe-Gln-OH or 94.6 mg (0.1 mmol) Mal-Ser-Ser-Tyr-
Tyr-Ser-Gly-OH, 13.5 mg (0.1 mmol) 1-hydroxybenzotriazole hy-
drate and 33 µL (30.3 mg, 0.3 mmol) 4-methylmorpholine were dis-
solved in 20 mL anhydrous DMF; after stirring at ϩ5°C for 15
min, 46.5 µL (37.9 mg, 0.3 mmol) N,NЈ-diisopropyl carbodiimide
was added as the coupling agent. After stirring at ϩ5°C for 4 days,
DMF was removed by evaporation in high vacuum and the residue
dissolved in a minimal amount of chloroform/methanol 3 : 1 (for
PSA3) or 4 : 1 (for PSA4), and the product was purified twice on
a silica gel column using chloroform/methanol 3 : 1 or 4 : 1 to afford
either 50 mg PSA4 or 60 mg PSA4, respectively, as a red powder
after precipitating the combined fractions containing the product
with diethyl ether. PSA3: Mass (ESI-MS, MW 1553.5): m/z 1576
[Naϩ salt adduct], HPLC (495 nm): >98%; PSA4: Mass (ESI-MS,
MW 1471.5): m/z 1494 [Naϩ salt adduct], HPLC (495 nm): > 98%.
was adjusted to 400
50 µM by concentrating the sample with
CENTRIPREP-10 concentrators from Amicon, Germany (4°C
and 4500 rpm). Samples were kept frozen at Ϫ78°C and thawed
prior to use.
Incubation studies of the albumin conjugates of PSA3, PSA4 and
PSA4-Arg with PSA
Samples (50 or 100 µM) of the albumin conjugates of PSA3, PSA4
or PSA4-Arg were incubated with human PSA (20 µg/mL) and
chromatograms recorded at λ ϭ 495 nm using reverse-phase HPLC
at the time points stated in the figures.
Incubation studies with human plasma
Synthesis and characterization of PSA4-Arg
PSA4-Arg was added to human blood plasma (EDTA stabilized)
pre-incubated at 37°C at a final concentration of 100 µM, and the
samples were incubated for the time periods stated in the figures at
37°C; a 50-µL sample was analyzed by HPLC.
Doxorubicin hydrochloride (50 mg, 0.086 mmol), 100.7 mg
(0.086 mmol) 6-maleimidohexanoyl-Arg-Arg-Ser-Ser-Tyr-Tyr-Ser-
Gly-OH, 11.6 mg (0.086 mmol) 1-hydroxybenzotriazole hydrate and
37.8 µL (37.8 mg, 0.34 mmol) 4-methylmorpholine were dissolved
in 20 mL anhydrous DMF; after stirring at ϩ5°C for 15 min,
39.8 µL (32.5 mg, 0.21 mmol) N,NЈ-diisopropyl carbodiimide was
added as the coupling agent. After stirring at ϩ5°C for 4 days, the
product was precipitated with diethyl ether, washed three times with
20 mL diethyl ether to afford 130 mg PSA4-Arg as a red powder.
Mass (ESI-MS, MW 1693.7): m/z 1807.6 [TFA salt adduct], HPLC
(495 nm): > 97%.
Preparation of prostate tumor tissue homogenates
For obtaining prostate carcinoma tissue homogenates, all steps were
carried out on ice where possible. Tissues of PC 3 and CWR22
xenograft tumors as well as tissues from three native human prostate
tumors (received from the Department of Urology, Freiburg) were
cut into small pieces, and 200-mg samples were transferred in a 2-
mL Eppendorf tube to which was added 800 µL homogenate buffer
(50 mM Tris-HCl buffer, pH 7.4, containing 1 mM monothiogly-
cerol). Homogenization was carried out with a micro-dismem-
berator at 3000 rpm for 3 min with the aid of glass balls, and the
samples were then centrifuged at 5000 rpm for 10 min and kept
frozen at Ϫ78°C until use.
Synthesis of DOXO-Gly-Ser-OH
DOXO-Gly-Ser-OH was prepared by reacting Ser-Gly-OH with tri-
methylsilyl chloride, diisopropylethylamine and 4-methoxytrityl
chloride in DMF using the following procedure: 400 mg (2.47
mmol) Ser-Gly-OH, 653.6 µL (2.1 mmol) trimethylsilyl chloride and
431.1 µL (1.05 mmol) diisopropylamine in 200 mL DMF were
heated at 40°C for 4 h until the solution became clear. The solution
was left to cool to room temperature and 1273 µL diisopropylamine
was added, as well as 800 mg 4-methoxytrityl chloride. Stirring was
continued at room temperature for 2 days. The mixture was then
kept at ϩ5°C overnight and the formed precipitate was isolated and
washed with diethyl ether and dried in vacuum (yield of protected
4-methoxytrityl-Ser-Gly was 400 mg. In a second reaction, 50 mg
(0.086 mmol) Doxorubicin hydrochloride, 200 mg 4-methoxytrityl-
Ser-Gly, 11.6 mg (0.086 mmol) 1-hydroxybenzotriazole hydrate, and
37.8 µL (0.34 mmol) 4-methylmorpholine were dissolved in 15 mL
anhydrous DMF. After stirring at ϩ5°C for 15 min, 39.8 µL (0.21
mmol) N,NЈ-diisopropyl carbodiimide was added as the coupling
agent. After stirring at ϩ5°C for 4 days, the product was precipi-
tated with 200 mL diethylether and isolated by centrifugation. The
red product was dissolved in 4 mL methanol, and the protective
group was removed by adding 1.6 mL dichloroacetic acid for 5 min
and subsequent precipitation with 30 mL diethylether (yield 122
mg). Mass [ESI-MS, MW 687.6): m/z 686.8 ([MϩϪ1]), HPLC
(495-nm): > 95%.
Incubation studies of the albumin conjugate of A-PSA4-Arg
Incubation studies of the albumin conjugate of A-PSA4-Arg were
performed with supernatants and tissue homogenates of PSA-nega-
tive (PC 3) and PSA-positive prostate carcinoma (CWR22, LNCAP,
human prostate cancer samples). The albumin conjugate of PSA4-
Arg was incubated with the biological samples at 37°C at a final
concentration of 25 or 50 µM, and chromatograms were recorded
at 495 nm using reverse-phase HPLC at the time points stated in
the figures.
Cell culture
PC 3 (PSA negative), LNCAP (PSA positive) and CWR22 (PSA
positive) prostate carcinoma cells were grown as monolayer cultures
in cell culture flasks in RPMI 1640 culture medium without phenol
red, supplemented with 10% heat-inactivated FCS, 100 µg/mL glu-
tamine, 100 U/mL penicillin and 100 µg/mL streptomycin. Cells
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de

472
Kratz et al.
Arch. Pharm. Chem. Life Sci. 2005, 338, 462−472
were cultured in a humidified atmosphere of 95% air and 5% car-
bon dioxide at 37°C. Media were routinely changed every 3 days.
For subculture or experiments, cells growing as monolayer cultures
were released from the tissue flasks by treatment with 0.05%
trypsine/EDTA, and viability was monitored using the cell analyzer
system Casy 1 from Schärfe Systems (Reutlingen, Germany). For
the experiments, cells were used during the logarithmic growth
phase.
Acknowledgments
We thank the Deutsche Forschungsgemeinschaft and the
Hopp Stiftung for their financial support.
References
[1] F. Kratz, R. Mueller-Driver, I. Hofmann, J. Drevs, C. Unger,
J. Med. Chem. 2000, 43, 1253Ϫ1256.
MTT assay
[2] F. Kratz, A. Warnecke, K. Scheuermann, C. Stockmar, J.
Schwab, J. Lazar, P. Druckes et al., J. Med. Chem. 2002, 45,
5523Ϫ5533.
The MTT [3-(4,5-dimethylthiazolyl-2)-2,5-diphenyltetrazolium
bromide] assay was performed in analogy to a literature protocol
[21]. Briefly, approximately 5000 cells were plated in each well of a
96-well tissue culture plate. Medium supplemented with 10% FCS
was added and cells were allowed to adhere for 24 h. Subsequently,
cells were pre-incubated with various drug concentrations
(0.001Ϫ60 µM) for 72 h and then labeled by adding sterile-filtered
MTT (5 mg/mL PBS). Cultures were incubated at 37°C for 2 h.
Subsequently, the supernatant was removed and 100 µL DMSO ad-
ded to each well. After an incubation time of 10 min, the extinction
of the samples was quantified with an ELISA reader (Dynatech
Laboratories Inc., Sullyfield, UK) at a wavelength of 570 nm. Four
separate cultures were determined per concentration. Results are
shown as means standard deviation of three independent experi-
ments (n ϭ 3).
[3] A. M. Mansour, J. Drevs, N. Esser, F. M. Hamada, O. A. Bad-
ary, C. Unger, I. Fichtner et al., Cancer Res. 2003, 63,
4062Ϫ4066.
[4] A. Warnecke, I. Fichtner, D. Garmann, U. Jaehde, F. Kratz,
Bioconjugate Chem. 2004, 15, 1349Ϫ1359.
[5] A. Warnecke, F. Kratz, Bioconjugate Chem. 2003, 14, 377Ϫ387.
[6] D. Gmehling, M. Medinger, K. Mross, B. Haering, F. Kratz,
C. Unger, J. Drevs et al., Proc. Am. Assoc. Cancer Res. 2005,
46, 3987.
[7] F. Kratz, J. Drevs, G. Bing, C. Stockmar, K. Scheuermann, P.
Lazar, C. Unger, Bioorg. Med. Chem. Lett. 2001, 11,
2001Ϫ2006.
[8] S. R. Denmeade, A. Nagy, J. Gao, H. Lilja, A. V. Schally, J. T.
Isaacs, Cancer Res. 1998, 58, 2537Ϫ2540.
Xenograft experiments
[9] S. R. Khan, S. R. Denmeade, Prostate 2000, 45, 80Ϫ83.
[10] S. R. Denmeade, J. T. Isaacs, Cancer J. Sci. Am. 1998, 4,
15Ϫ21.
The antitumor efficacy of PSA4-Arg and Doxorubicin was investi-
gated in a PSA-negative model (PC 3) and in a PSA-positive xenog-
raft model (CWR22); male Ncr nude mice were used for the PC 3
model and male NOD/SCID mice for the CWR22 transplants.
NOD/SCID mice were necessary because of the insufficient growth
of CWR22 in nude mice. The mice were held in laminar flow shelves
under sterile and standardized environmental conditions (25 2°C
[11] G. S. Coombs, R. C. Bergstrom, J. L. Pellequer, S. I. Baker, M.
Navre, M. M. Smith, J. A. Tainer, E. L. Madison, D. R. Corey,
Chem. Biol. 1998, 5, 475Ϫ488.
[12] C. F. Yang, E. S. Porter, J. Boths, D. Kanyi, M. Hsieh, B. S.
Cooperman, J. Peptide Res. 1999, 54, 444Ϫ448.
room temperature, 50
10% relative humidity, 12-h light-dark
[13] T. E. Murdter, B. Sperker, K. T. Kivisto, M. McClellan, P.
Fritz, G. Friedel, A. Linder et al., Cancer Res. 1997, 57,
2440Ϫ2445.
rhythm). They received autoclaved food and bedding (ssniff, Soest,
Germany) and drinking water ad libitum.
PC 3 and CWR22 prostate carcinomas were xenografted subcutane-
ously as fragments according to a literature protocol [22]. Mice were
randomly distributed to the experimental groups. When the tumors
had reached a palpable size, treatment was initiated. Mice were
treated intravenously in a weekly schedule with 10 mM sodium
phosphate and 5% d-glucose (pH 6.4), PSA4-Arg or Doxorubicin
(2.0 mg/mL). PSA4-Arg was dissolved in a sterile isotonic buffer
containing 10 mM sodium phosphate and 5% d-glucose (pH 6.4)
at a concentration of 6.0 mg/mL; for doses and schedules see Figs.
8 and 9 and Tables 2 and 3. Four to eight mice were used per group
for the experiments.
[14] S. R. Denmeade, C. M. Jakobsen, S. Janssen, S. R. Khan, E.
S. Garrett , H. Lilja, S. B. Christensen, J. T. Isaacs, J. Natl.
Cancer Inst. 2003, 95, 990Ϫ1000.
[15] S. R. Denmeade, W. Lou, J. Lovgren, J. Malm, H. Lilja, J. T.
Isaacs, Cancer Res. 1997, 57, 4924Ϫ4930.
[16] D. DeFeo-Jones, S. F. Brady, D. M. Feng, B. K. Wong, T.
Bolyar, K. Haskell, D. M. Kiefer et al., Mol. Cancer Ther. 2002,
1, 451Ϫ459.
[17] D. DeFeo-Jones, V. M. Garsky, B. K. Wong, D. M. Feng, T.
Bolyar, K. Haskell, D. M. Kiefer et al., Nat. Med. 2000, 6,
1248Ϫ1252.
Tumor size was measured twice weekly with a calliper-like instru-
ment in two dimensions. Individual tumor volumes (V) were calcu-
lated by the formula V ϭ (length Ϯ [width]2)/2 and related to the
values on the first day of treatment (relative tumor volume, RTV).
At each measurement day, treated/control values (T/C) were calcu-
lated as percentage for each experimental group; the optimum (low-
est) values obtained within 4 wk after treatment were used for evalu-
ating the efficacy of the compounds, and optimum T/C-values are
presented in Tables 2 and 3. The body weight of mice was deter-
mined twice weekly and related to the body weight on the first day
of treatment (body weight change, BWC). Statistical analysis was
performed with the U-test (MannϪWhitney) with p < 0.05.
[18] V. Dubois, L. Dasnois, K. Lebtahi, F. Collot, N. Heylen, N.
Havaux, A. M. Fernandez et al., Cancer Res. 2002, 62,
2327Ϫ2331.
[19] P. M. Loadman, M. C. Bibby, J. A. Double, W. M. Al-Shakhaa,
R. Duncan, Clin. Cancer Res. 1999, 5, 3682Ϫ3688.
[20] F. Kratz, A. Warnecke, B. Schmid, D. Chung, M. Gitzel, Curr.
Med. Chem. Anti-Cancer Agents 2005, in press.
[21] S. P. Cole, Cancer Chemother. Pharmacol. 1986, 17, 259Ϫ263.
[22] M. A. Wainstein, F. He, D. Robinson, H. J. Kung, S. Schwartz,
J. M. Giaconia, N. L. Edgehouse et al., Cancer Res. 1994, 54,
6049Ϫ6052.
 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.archpharm.de