Camptothecin-20-PEG ester transport forms: the effect of spacer groups on antitumor activity

Article abstract of DOI:10.1016/S0968-0896(98)00005-4

An improved synthesis of the hindered PEG-camptothecin diester transport form has been achieved using the Mukaiyama reagent. We have also assessed the effect of changing the electronic configuration of the (d-position of PEG-camptothecin transport forms on the rates of hydrolysis of the pro-moiety, and attempted to correlate these differences to efficacy in two animal models. In addition to the simple substitution of N for O, other synthetic modifications of these atoms were accomplished by employing heterobifunctional linker groups. The half lives by disappearance (rates of hydrolysis) of the transport forms in buffer and rat plasma were determined. It was established that anchimeric assistance to hydrolytic breakdown of the pro-moiety occurs in a predictable manner for some of these compounds. Results for the new derivatives in a P388 murine leukemic model and HT-29 human colorectal xenograft study are also presented. The use of a glycine linker group was found to provide similar efficacy in rodent models to that of simple camptothecin 20-PEG ester, and displayed enhanced pharmacokinetics. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00005-4

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 551±562
Camptothecin-20-PEG Ester Transport Forms: the E€ect of
Spacer Groups on Antitumor Activity
Richard B. Greenwald,* Annapurna Pendri, Charles D. Conover, Chyi Lee,
Yun H. Choe, Carl Gilbert, Anthony Martinez, Jing Xia, Dechun Wu
and Mei-mann Hsue
Enzon, Inc., 20 Kingsbridge Road, Piscataway, New Jersey 08854-3969, USA
Received 23 September 1997; accepted 5 November 1997
AbstractÐAn improved synthesis of the hindered PEG-camptothecin diester transport form has been achieved using
the Mukaiyama reagent. We have also assessed the e€ect of changing the electronic con®guration of the (d-position of
PEG-camptothecin transport forms on the rates of hydrolysis of the pro-moiety, and attempted to correlate these dif-
ferences to ecacy in two animal models. In addition to the simple substitution of N for O, other synthetic modi®ca-
tions of these atoms were accomplished by employing heterobifunctional linker groups. The half lives by disappearance
(rates of hydrolysis) of the transport forms in bu€er and rat plasma were determined. It was established that anchi-
meric assistance to hydrolytic breakdown of the pro-moiety occurs in a predictable manner for some of these com-
pounds. Results for the new derivatives in a P388 murine leukemic model and HT-29 human colorectal xenograft study
are also presented. The use of a glycine linker group was found to provide similar ecacy in rodent models to that of
simple camptothecin 20-PEG ester, and displayed enhanced pharmacokinetics. # 1998 Elsevier Science Ltd. All rights
reserved.
Introduction
rubicin and the polymers carboxymethyl dextran, alginic
acid, polyglutamate, or carboxymethyl cellulose resulted
in the conclusion that when daunorubicin was attached
irreversibly to a macromolecule, its cytotoxic properties
were lost; but attachments using a readily hydrolyzable
bond (hydrazone) resulted in sustained in vivo activity.3
This result has been interpreted to be due to hydrolysis
in the acidic milieu of the tumor; possibly after cellular
uptake. Further investigations by Ueda4 utilizing doxo-
rubicin Schi€ base conjugates of oxidized dextran
con®rmed the utility of this approach, and led to a
phase I clinical trial of the macromolecular prodrug.5
Simultaneously, extensive work on the incorporation of
doxorubicin into N-(2-hydroxypropyl)methylacrylamide
(HPMA) copolymers appeared.2,6 These conjugates
employed tetrapeptide spacers which demonstrated
controlled biodegradability (lysosomal degradation fol-
lowing endocytosis) and tumor accumulation while
exhibiting lower toxicity than the unconjugated drug. A
clinical trial of a HPMA-anthracycline conjugate has
recently been initiated.7
Synthetic polymers conjugated to cytotoxic drugs have
long been postulated as novel chemical entities which
should be capable of e€ective drug delivery.1 Thus the
greater solubility, enhanced tumor accumulation, and
longer circulatory retention attributed to the polymeric
ballast of various drug conjugates has been the focal
point of research in this particular area of drug delivery
for the past 20 years. Attesting to this fact is the litany
of review articles2 covering this essential approach to
cancer chemotherapy, notwithstanding the reality that
no clinically approved polymeric drug of this type has
yet ensued in the U.S. Currently, Maeda reports2g that
pilot clinical trials of poly(styrene-co-maleic acid n-butyl
ester) conjugated neocarzinostatin (SMANCS) have
demonstrated substantial ecacy in the treatment of
hepatocarcinoma. Earlier studies employing dauno-
*Corresponding author.
0968-0896/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00005-4

552
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
Table 1. Rates of hydrolysis and IC50 values of various PEG-camptothecin derivatives
Compd
Linker
IC50 (nM)*
t1/2 (h)a
PBS (pH 7.4)
Rat plasma
1
Ð
-O-CO-CH2-O-PEG
7
15
16
21
7
Ð
27
Ð
2
3
10
11
16
17
24
25
28
29
-O-CO-CH2-O-CH2-CO-NH-PEG
-O-CO-CH2-O-CH2-CO-N(CH3)-PEG
-O-CO-CH2-O-CO-NH-PEG
-O-CO-CH2-O-CO-N(CH3)-PEG
-O-CO-CH2-NH-CO-CH2-O-PEG
-O-CO-CH2-N(CH3)-CO-CH2-O-PEG
-O-CO-CH2-NH-PEG
5.5
27
0.2
0.8
3
ND
5
18
12
15
24
42
28
40
6
97
12
10
3
-O-CO-CH2-N(CH3)-PEG
102
>24
*All experiments were done in duplicate: standard deviation of measurements: ‹10%.
aThese results more appropriately represent the half lives by disappearance of the transport form.
ND: Not determined.
The dramatic enhancement of circulating half life, solu-
bility and in vivo activity of the potent anti-cancer drug
camptothecin (1), which we recently reported,8 was
achieved utilizing a traditional type of prodrug strategy
in conjunction with the non-immunogenic 40 kDa poly-
meric compound, polyethylene glycol (PEG) to trans-
port the macro-molecular conjugate. Activation by the
a-alkoxy substituent provided by PEG diacid9 results in
a scissile ester bond in the polymeric conjugates (3)
which controllably releases 1 in its active lactone form.
A distinction can now be made between polymer con-
jugated drugs which undergo relatively slow rates of
internalization by endocytosis2m in order to function,
and polymer conjugated prodrugs (transport forms)
which are designed to hydrolyze in a predictable fashion
within the tumor stroma. This latter type of easily pre-
pared polymeric drug, with its controlled hydrolysis,
should be extremely potent if passive tumor accumula-
tion by the enhanced permeability and retention (EPR)
e€ect postulated by Maeda2g,2h,10 occurs in a relatively
short time, that is, t1/2 accumulation > t1/2 hydrolysis:
thus controlled release of the small hydrophobic active
drug within the neoplasm, di€usion to tumorous par-
enchymal cells, followed by cellular uptake may now be
predictable.* Presently though, the most serious draw-
back to the clinical consideration of the PEG-camp-
tothecin conjugate 3 is that it consists of a mixture of
mono (3a) and di-substituted ester (3b) prodrugs.
Therefore our e€orts were directed toward either
improving the yield of 3b or developing alternate syn-
thetic strategies in order to obtain a similar PEG diester
in a pure state. This latter approach would focus on
designing transport forms by conjugating di€erent lin-
ker groups to PEG while maintaining the integrity of
the camptothecin 20(S)- ester functionality established
for 3. In this paper we report on an improved procedure
for the preparation of 3b as well as the synthesis of
novel tripartate prodrugs11, and the results of screening
these new compounds using t1/2 hydrolysis (Table 1), in
vitro IC50 measurements (Table 1), and survival in a
P388 murine leukemic model (Table 2). From the accu-
mulated data, selections were made for more in-depth
xenograft studies. The results of these studies were
®nally compared to those of 3 in order to ®nd an ana-
logous or improved clinical candidate. We achieved the
goal of synthesizing an essentially single component
camptothecin diester with ecacy comparable to 3 by
synthesizing the glycine derivative 24. Compound 24
also demonstrated enhanced circulatory retention in the
mouse. More importantly, the hydrolytic pro®le of 24 in
human plasma showed a signi®cant increase in camp-
tothecin release compared to 3 (Figure 1).
Figure 1. Kinetics of camptothecin (1) release from PEG-
camptothecins 3 and 24 at 37 ꢀC in human plasma: *, rate of
hydrolysis of 3; *, rate of formation of 1 from hydrolysis of 3;
!, rate of hydrolysis of 24; ~, rate of formation of 1 from
hydrolysis of 24.
*We have observed this phenomenon in a solid tumor model
with compound 24. The results of this investigation will be
published at a later date.

R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
553
Table 2. Activity of PEG-camptothecin derivatives against P388 murine leukemia in vivo
Test compound
Control
Linker
Ð
Total dosea
(mg/kg)
Mean time to
death (days)b
% ILSc
Ð
Survivors
on day 40
Ð
13.0
0/10
1
Ð
-O-CO-CH2-O-PEG
16
16
16
16
16
16
16
16
16
38.0*
38.0*
17.4y
31.6*y
23.4
192%
192%
34%
7/10
9/10
4/10
6/10
0/10
8/10
0/10
0/10
0/10
3
10
11
17
24
25
28
29
-O-CO-CH2-O-CH2-CO-NH-PEG
-O-CO-CH2-O-CH2-CO-N(CH3)-PEG
-O-CO-CH2-O-CO-N(CH3)-PEG
-O-CO-CH2-NH-CO-PEG
-O-CO-CH2-N(CH3)-CO-PEG
-O-CO-CH2-NH-PEG
143%
80%
35.0*
19.3*y
30.6*
21.4*y
169%
48%
135%
65%
-O-CO-CH2-N(CH3)-PEG
aEquivalent dose of camptothecin, mice dosed days 1±5.
bKaplan±Meier estimates with survivors censored.
cIncreased life span (ILS) is (T/C-1)Â100.
*Signi®cant ( P<0.001) compared to control (untreated).
ySigni®cant ( P<0.001) compared to compound 1.
In vivo ecacy study of the water soluble camptothecin derivatives using the P388/0 murine leukemia model. Compound 1 or prodrug
derivatives were given daily [intraperitoneal(ip)Â5], 24 h following an injection of P388/0 cells into the abdominal cavity with survival
monitored for 40 days.
Chemistry
alternative clinically relevant candidates. Using this
approach we hoped to delineate the limits of the PEG
prodrug strategy. For this reason we turned our atten-
tion to the alternative strategy outlined earlier, viz. ®rst
functionalizing the 20-OH group with a linker group
capable of further reaction (i.e. a small heterobifunc-
tional molecule which can maintain the desired place-
ment of a heteroatom in the a position). Since the
camptothecin prodrug is based on an ester, by necessity
one of the functionalities of the heterobifunctional lin-
ker must be a carboxylic acid. The other moiety, located
at the distal terminus can be designed to be situated on a
1ꢀ carbon and thus unhindered. At this stage, the
camptothecin intermediate is usually obtained in high
yield, and can be additionally puri®ed without diculty.
Further reaction of the distal functionality with a com-
plimentary functionalized PEG would then result in a
stable polymeric conjugate. In this fashion, the synthesis
of disubstituted camptothecin derivatives in high purity
can be accomplished in a facile manner. For example, in
the case of a carboxyl terminated linker one would
employ an amino PEG to form an amide linked con-
jugate. By contrast, with an amino terminated linker a
carboxy or carboxaldehydo PEG would be the appro-
priate choice leading to an amide and imine, respec-
tively; the latter can easily be reduced to a stable 2ꢀ
amine. One consequence of using linker groups is the
introduction of rate accelerations due to neighboring
group (anchimeric) participation. This enhancement to
hydrolytic breakdown of the promoiety can be incorpo-
rated into the design of the transport form. The follow-
ing examples of linkers illustrate the methodology:
Generally, when working with PEG of molecular weight
in excess of 1000 Da, it is necessary to develop functio-
nalization reactions that proceed in high yield (>90%)
since resulting mixtures of PEG-OH and PEG-X (where
X is any non OH moiety) are extremely dicult to
separate. As previously reported,8 condensation of PEG
diacid 2 with camptothecin's hindered 30 alcohol moiety
at position 20 in the presence of DIPC (EDC or DCC)
led to multicomponent mixtures of mono and diester
(35:65) under a variety of conditions (see experimental
section for a typical procedure). However, employing
the Mukaiyama reagent,12 2-chloro-1-methylpyridinium
iodide, as condensing agent, the amount of diester 3b
was increased to approximately 90%.
At this point, even though the product purity appeared
adequate to proceed to pharmaceutical development, we
elected to further explore other methods of high yield
derivatization of camptothecin with PEG in order to
determine if modi®cation of the a position could provide

554
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
Example 1: CO2H-CO2H
Example 2: CO2H-OH
Conversion of camptothecin to the 20-acetate using
acetic anhydride was reported by Wall, Wani and
co-workers in 1966.13 By substituting diglycolic anhy-
dride (a bifunctional acylating agent)14 the slightly
water soluble camptothecin derivative 7, containing an
(a-alkoxy ester and a terminal carboxyl group, was
obtained in moderate yield (Scheme 1). It was found
more convenient to prepare 7 by ®rst opening diglycolic
anhydride with t-BuOH to give the protected ester acid
5, followed by condensation with 1 in the presence of
DIPC, and subsequent acid cleavage. The next step, as
outlined in the strategy above, was to add PEG in a
facile manner. This was accomplished by utilizing the
diamine derivative 8 which was easily prepared from the
chloride. Condensation of 7 in the presence of DIPC
provided diamide 10 in high yield and purity. In a simi-
lar way the corresponding di-N-methyl amide 11 was
also prepared from 9. From the values observed in
Table 1, this derivative gave the closest physical mea-
surements to 3, but far less in vivo activity as later
experiments demonstrated.
It was also of interest to synthesize an ester with an a-
carbamate group, which provides increased electro-
negativity, in order to observe the e€ect of this sub-
stituent on the rate of hydrolysis. To achieve this
objective we developed a simple route, as illustrated in
Scheme 1, which utilizes the a-hydroxy ester 14. This
intermediate was obtained in modest yield, and was
easily converted to the activated carbonyl imidazole
derivative 15 (see experimental). Reaction with PEG
amine 8 or 9 proceeded to give the dicarbamate deriva-
tives (16 or 17) in high yield.
Example 3: CO2H-NH2
Many amino acid ester derivatives of water insoluble
drugs have been prepared with the expressed purpose of
enhancing the aqueous solubility of the resulting pro-
drugs as the salts of strong acids.15 Anti-cancer drugs
such as paclitaxel16 and camptothecin17 have already
been modi®ed in this fashion. Generally, the problem
associated with this type of approach is the instability
Scheme 1.
Scheme 2.

R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
555
of the charged a-ammonio substituted ester which
undergoes rapid hydrolysis.18 Bifunctional amino acid
groups employed in a PEG prodrug strategy will, how-
ever, result in a water soluble nonionic a-amidoester
prodrug that should be much less susceptible to hydro-
lysis. The route taken is shown in Scheme 2 where ester
formation was easily accomplished leaving a reactive
amino group, usually in a protected form, but easily
unblocked. Attachment of PEG to the a-amino group of
camptothecin 20-glycinate and sarcosinate was accom-
plished in standard fashion by utilizing PEG diacid 2
(Scheme 2). Although DIPC and the Mukaiyama
reagent worked well in dichloromethane solvent to
a€ord 24, we preferred the use of EEDQ since it could
be employed with the more environmentally acceptable
solvent, toluene. Acylated amino acids appear to have a
pKa in the range of about 3.0±3.6 (N-acetyl gly=3.6,
glygly=3.12)19 which is similar to that reported for a-
alkoxy acids (3.48). We felt that since an a-amido sub-
stituent possesses similar electronegativity to a-alkoxy
groups as evidenced by pKa, a similar hydrolytic pro®le
to that of the e€ective a-alkoxy ester 3 should result for
this novel series of esters. As Table 1 shows, the t1/2
hydrolysis for 24 was only about 1.4 times greater than
that of 3 in PBS bu€er (pH 7.4). However, this was not
the case for the N-substituted amide 25 derived from
sarcosine which hydrolyzed at a rate of 3.5 times slower
than 3 in bu€er and 2.5 times slower in rat plasma. The
rate of hydrolysis provided by the glycine linker, seren-
dipitously, gave rise to a compound (24) with excellent
ecacy in the biological models. For purposes of com-
parison, PEG was attached directly to the slightly less
electronegative amino group (pKa of neutral gly=4.3)19
by reaction of the bromo ester 27 with PEG amine 8 to
give 28 (Scheme 2). The N-methyl derivative 29, was
synthesized in a similar way. In this case, anchimeric
participation of the basic NH of amine 28, probably
via an a-lactam intermediate,20 was clearly observed:
compared to the N-methylamine derivative 29, t1/2 in
PBS bu€er was almost nine times faster and 16 times
faster in plasma.
Results
In vitro cytotoxicity
The in vitro biological ecacy of the PEG-camptothecin
transport forms were tested using the P388/0 murine
leukemia cell lines. The cytotoxicities of 1, 3 and the
various PEG transport forms synthesized in this study
are shown in Table 1.
In vivo murine leukemia screen
The relative in vivo equivalency of compound 1 to the
new prodrugs synthesized in this study was assessed by
monitoring survival in a P388/0 murine leukemia model.
Mice were injected with P388/0 cells and then treated for
®ve consecutive days. This was followed by daily survi-
val monitoring for another 35 days. The results of this
in vivo study are shown in Table 2. The mean time to
death for native camptothecin (1, administrated within
an intralipid suspension delivered ip) group at 16 mg/kg
was 38 days which resulted in an increased life span
(ILS) of 192% with a 70% cure rate. Similarly, the
mean time to death for an equivalent dose of the pro-
drug, compound 3 (administered ip as an aqueous solu-
tion) was 38 days (ILS=192%) with a cure rate of 90%.
Both 1 and 3 showed no signs of overt acute toxicity.
Thus, equivalency ( P ˆ 0:66) was demonstrated for
compound 1 and its transport form, compound 3, in this
murine leukemia model. By comparison, compounds 11,
24, and 28 demonstrated an ILS similar to 3, although
compound 28's group was void of long term survivors.
In vivo colorectal xenograft
Prodrug congeners that exhibited a comparable ILS to
3 in the P388 model were subsequently evaluated in
Table 3. Antitumor activity against human colorectal xenografts (HT-29) in vivo
Test compound
Dose (mg/kg)a
Week 5 (end of treatment)
BW change (%)c
Week 7 (two weeks post)
BW change (%)
Total mortality
Ecacy (%)b
Ecacy (%)
Control
Topotecan
-
+727
+78
20
+10
+1047
+300
+62
98
+18
+6
0/10
0/10
5/10
0/10
4/10
0/10
2.5
2.5
2.5
2.5
2.5
8
3
1
3d
11d
24d
+9
89
60
1
+13
+13
+25
+1
0
73
96
80
aMice dosed ®ve days a week for ®ve weeks.
bEcacy was expressed as per cent tumor volume change from initial.
cBW, body weight.
dEquivalent dose of compound 1.
In vivo ecacy study of the water soluble camptothecin derivatives using the HT-29 human colorectal xenograft. A subcutaneous
injection of HT-29 cells was allowed to reach an average tumor volume of 250 mm3, prior to treatments. Mouse weight and tumor size
were measured at the beginning of study and weekly through week 7.

556
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
HT-29 human colorectal xenografts which were mon-
itored for tumor growth kinetics after a ®ve week dosing
period. All treatment groups were monitored an addi-
tional two weeks in order to assess the short-term
regrowth kinetics of the treated tumors. Compound 1
appeared more toxic (50% mortality) and less e€ective
than equivalent doses of the prodrugs (Table 3). Both 3
and 24 showed considerable antitumor activity which
resulted in a striking 95% or greater reduction in initial
tumor burden by the end of the study without any sig-
ni®cant weight loss. Topotecan treatments resulted only
in tumor growth inhibition following the ®ve weeks of
treatment. Following drug treatment withdrawal, con-
trol, topotecan and compound 1 treated mice all dis-
played tumor regrowth (Table 3). In contrast, mice in all
prodrug groups showed continued tumor regression,
with all the groups displaying approximately a 10%
further reduction in tumor volume during this period.
(Table 1). Anchimeric assistance emerges as a primary
mechanism in the hydrolytic process for those structures
where an NH or NHC=O functionality is present, and
a 3, 5, or 6 membered cyclic transition state can be
formed with the terminal ester. Precedent for such rate
enhancements are known for amides,23 carbamates,24 and
amines.25 Thus, 10 hydrolyzes at a rate ®ve times faster
than 11 in bu€er, and three times faster in rat plasma.
The P388 model re¯ects this in the mean time to death
which is only minimally better than control, and would
seem to indicate that a plasma circulating half life closer
to that of 3 (2 h), is desirable. The rate enhancement is
so pronounced for 16 (t1/2=0.2 h, pH 7.4) that no
further experimentation could be justi®ed for this com-
pound since t1/2 is less than that reported for campto-
thecin 20-glycinate TFA salt which releases 1 very
rapidly, and has been shown to be less potent than 3.8
On the other hand, by replacing the N±H group of 16 by
N±Me, the resulting derivative 17 has a rate of hydro-
lysis which is 140 times slower. Compound 17, while
exhibiting a similar hydrolytic pro®le to 3, produces a
result in the P388 model that is only half as ecacious.
Similarly, the N-methyl derivative 25 with a t1/2=97 h at
pH 7.4 is almost four times slower than 3, and does not
show a signi®cant % ILS. We attribute this ®nding to the
inability of 25 to release camptothecin in a timely fashion
to the model. The N-unsubstituted analog 24 hydrolyzes
2.5 times faster than 25 at pH 7.4, and 1.5 times faster in
rat plasma. Compared to 3, compound 24 hydrolyzes at a
rate (t1/2=40 h) that is only about 35% slower than 3 in
bu€er, but 3 times slower in plasma. Fortunately, in this
case the di€erence in the timed release of 1 from 24 still
results in a compound that demonstrates very similar
P388 activity to 3. Based on the data provided in Table 1,
a circulating t1/2 range of approximately 2±6 h in rat
plasma is required to provide ecacy in the ascites P388
murine leukemic model. Rates either on the high or low
side of this range results in a diminution of ILS.
In vivo circulatory retention
Both compounds 3 and 24, which showed the greatest
ecacy in the solid tumor model, were tested for blood
circulatory retention. The blood t1/2a of compound 3
was estimated to be 5 min, and the t1/2b was 3.5 h with
the area under the curve (AUC) of 17.7 mg/mL.h.
Similarly, compound 24 displayed t1/2a=5 min, and a
t1/2b= 5.3 h with an AUC of 15.9 mg/mL.h.
Discussion
It is evident from the IC50 values shown in Table 1 that
all of the camptothecin transport forms prepared in this
study demonstrate in vitro activity similar to native drug
(1). Examination of the in vivo data presented in
Table 2, clearly shows that signi®cant di€erences in
biological activity result from the structural changes
made by the incorporation of a linker group. However,
predictive changes, ultimately, could only be broadly
characterized, and it appears that each linker must be
evaluated on an individual basis.
On the basis of their performance in the simple P388
model, compounds 11 and 24 were chosen for more
in-depth evaluation in a solid tumor (HT-29 colorectal
xenograft) study which would compare them to 1, 3,
and the water soluble camptothecin derivative, topote-
can. The results of the study are presented in Table 3.26
It is not clear at this time why a lesser tumor reduction
and greater degree of toxicity exists for 11: it may pos-
sibly be due to a more rapid release of 1 in vivo caused
by an initial proteolytic attack on the 30 amide tripartate
prodrug. Compounds 10, 16, 17, 24 and 25 can also be
classi®ed as tripartate prodrugs11 with two enzymati-
cally mediated pathways available for hydrolysis.
Kinetic di€erentiation between the two possibilities are
beyond the scope of this investigation, but one would
anticipate that biological results would be re¯ective of
the pathways. In compound 3 (a bipartate prodrug)
Of the seven new compounds presented, three (11, 24,
and 28) can be immediately distinguished as possessing
an ILS greater than the other candidates in the P388
murine leukemic model. It is generally recognized that a
per cent ILS > 120 is signi®cant,22 and 11, 24, and 28
all have values in excess of 130 albeit slightly less than
that observed for compound 3. Compound 28 was
eliminated from further consideration since dosing with
this compound provided no survivors on day 40 of the
study compared to 11 and 24 where the survival rate
was 60 and 80%, respectively. Some degree of correla-
tion of physical parameters with P388 activity can be
made from comparisons of t1/2 hydrolysis in bu€er

R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
557
only pH or esterase mediated release is possible. There-
fore other tripartate prodrug candidates which mimic
the overall biological activity of 3 might be considered
to hydrolyze by a similar mechanism; those that man-
ifest di€erent biological parameters such as greater
toxicity or lesser ecacy can be viewed as hydrolyzing
by a di€erent, or combination of, mechanisms including
anchimeric assistance. While this interpretation is useful
when explaining in vivo di€erences, the weakness in this
simplistic view of the data is that alternate mechanisms
can lead to the same result in a single model, and thus
cannot be adequately distinguished from each other.
Structure±activity correlations between di and tripartate
prodrugs must therefore be regarded as only approx-
imations of in vivo mechanisms. While the ester hydro-
lysis model for 24 in bu€er and rat plasma is attractive
to support its similarity to 3, examination of rates of
hydrolysis in human plasma clearly show that 24 can
indeed behave quite di€erently when models are chan-
ged.28 The t1/2 hydrolysis of 24 in fresh human plasma is
shown in Figure 2. The amount of 1 released from the
transport form is approximately 65%. For comparison
the curve for 3 is superimposed, where only about 25%
of 1 becomes available. This decrease in recovery of 1
has been attributed to a species speci®c enzymatic
interaction.8 Additionally, a circulatory retention study
for 24 in the mouse was carried out and yielded a t1/2b of
5.3 h; about 30% longer than that of 3.21 Overall, the ease
of synthesis, long circulatory retention, and high ecacy
makes 24 an ideal candidate for future development.
glycine linker group yielded a PEG transport form that
was minimally as e€ective as the lead compound 3, but
provided enhanced pharmacokinetics, and demon-
strated superior ecacy. It is noteworthy, that all the
PEG-camptothecin transport forms tested appeared to
demonstrate less toxicity and more ecacy than equal
levels of free camptothecin. More detailed reports of
this important ®nding will be published shortly. We are
currently exploring the e€ect of substituting other
amino acids for glycine in the PEG camptothecin tri-
partate delivery system, as well as examining di€erent
PEG-amino acid-drug transport forms in several in vivo
systems.
Experimental
Biological assays
Materials. All PEG-camptothecin compounds were dis-
solved in sterile water for injection (WFI) prior to in vivo
drug treatments. With the exception of the circulatory
retention study, all PEG-camptothecin dosages were
given as their camptothecin equivalents (absolute amount
of camptothecin given). For in vivo administration,
compound 1 was dispersed in intralipid (Liposyn1 III
10%, Abbott Laboratories, North Chicago, IL) by
sonication. Topotecan was synthesized according to
published procedures.29 All animals received humane care
in compliance with the `Principles of Laboratory Animal
Care' formulated by the National Society of Medical
Research and the `Guide for the Care and Use of
Laboratory Animals' published by the National Institute
of Health. These experimental protocols were approved
by the Institutional Animal Care and Use Committee of
UMDNJ-Robert Wood Johnson Medical School.
Conclusion
In the present study we have investigated various high
yield synthetic approaches leading to PEG transport
forms of camptothecin. For direct condensations the
Mukaiyama reagent was found to produce quite accep-
table yields of diester 3b. Linker groups were also
employed in order to examine the e€ect of modifying
the electron withdrawing a substituent on the rates of
hydrolysis of the promoiety (ester), as well as to permit
high yield conversion to the ®nal PEG transport form.
Anchimeric assistance was observed for those systems
that had NH or NHC=O moieties con®gured in such a
way that a 3, 5, or 6 membered cyclic transition state
could be formed, and increased the rate of hydrolysis of
the drug-ester bond. We have endeavored to correlate
structure and t1/2 hydrolysis with activity in the P388
murine leukemic model. The results of this attempt has
de®ned a range of about 2±6 h for t1/2 circulation of the
transport form in rat plasma which will result in a pre-
dictable per cent ILS. Additionally, HT-29 (human colo-
rectal carcinoma) xenograft studies were used to re®ne
the choice between compounds identi®ed as possessing
similar activities in the P388 model. It was found that a
Cell lines and cytotoxicity assays. Studies using P388/0
cell lines for both IC50 and in vivo screens were main-
tained and conducted as previously reported.9 The solid
tumor HT-29 (human, colon adenocarcinoma) was
obtained from the ATCC (HTB 38) and grown in
DMEM supplemented with 10% FBS. Cells were sub-
cultured once a week and for in vivo experiments
viabilities were >90%. All cell lines were periodically
tested for Mycoplasma and were Mycoplasma free.
In vivo murine leukemia model. Compound 1 and its
prodrug forms were screened for in vivo activity against
the murine leukemia cell line P388/0 (mouse, lymphoid
neoplasm) as previously described.9
In vivo colorectal xenograft. Female nu/nu mice (Harlan
Sprague±Dawley, Madison, WI), 18±24 g and 10±14
weeks old, at onset of treatment, were used to test the
ecacy of camptothecin derivatives in a colorectal

558
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
xenograft model. When tumors reached the average
volume of 250 mm3, the mice were divided into their
experimental groups (n=10) which consisted of control,
topotecan and compounds 1, 3, 11, and 24. This portion
of the study was conducted and results analyzed as pre-
viously reported.21
Camptothecin-20-PEG 40 kDa ester. Method A: DIPC
as the coupling agent (3). PEG 40 kDa diacid 2, (11.5 g,
0.29 mmol,) was dissolved in 200 mL of anhydrous
(anhyd.) methylene chloride at room temperature. The
solution was chilled to 0 ꢀC and DIPC (0.18 mL,
1.15 mmol), DMAP (140 mg, 1.15 mmol) and campto-
thecin (400 mg, 1.15 mmol) were added, in that order,
and stirred for 2 h at 0 ꢀC. The reaction mixture was
allowed to warm to room temperature and left for 16 h.
The solution was concentrated to about 100 mL and ®l-
tered through celite. The ®ltrate was washed with 0.1 N
HCl, dried (anhyd MgSO4), and evaporated under
reduced pressure to yield 3 as a white solid which was
recrystallized from DMF/ether. The solid was ®ltered
and washed with 2-propanol. The product 3 (9.5 g,
82%) was found to be a mixture of monoester 3a and
diester 3b. The structures of the compounds comprising
the mixture were deduced by analysis of the NMR
spectra and determination of per cent camptothecin in
the product using the UV method outlined above: 1H
NMR: d 1.01 (t, J=7.2 Hz, 1H), 1.21 (s, 3H), 1.24 (s,
3H), 1.42 (s, 3H), 1.45 (s, 3H), 2.3 (s, 1H), 3.39 (PEG, t,
J=5.2 Hz), 3.49±3.71 (bs, PEG), 3.93 (PEG, t,
J=5.2 Hz), 4.1±4.32 (m, 6H), 5.44 (s, 1H), 5.5±5.8 (AB
q, J=17.16 Hz, 1H), 7.3 (t, J=7.92 Hz, 1H), 7.6 (s, 1H),
7.7 (t, J=7.26 Hz, 1H), 7.97 (d, J=8.5 Hz, 1H), 8.23 (d,
J=8.5 Hz, 1H), 8.44 (s, 1H); 13C NMR: d 7.25, 19.6,
22.3, 31.4, 42.5, 46.38, 48.68, 53.48, 62.8, 66, 67±71
(PEG), 75, 94.21, 118.8, 127.2, 127.4, 127.5, 127.65,
128.5, 129.5, 130.5, 144.48, 145.67, 151.5, 152.7, 156.53,
166.2, 168.78.
In vivo circulatory retention. Circulatory retention stu-
dies were performed in 25 g, nontumor bearing CD1
female mice (Charles River Laboratories, Stone Ridge,
NY). Mice received an i.v. bolus of 12 mg of test article
(8 mg/kg CPT equiv) via the tail vein and were exsan-
guinated over a 48 h period (3 mice/time point). Exsan-
guination was conducted in unconscious (100% CO2)
animals via orbital bleeding into a sterile tube to remove
a minimum of 1.0 mL whole blood/mouse. Blood samples
were processed and assayed as previously described.21
Half-life values of the t1/2a (distribution) and t1/2b
(elimination) phases for PEG-camptothecins were
calculated using
a two compartment, intravenous
bolus, ®rst order elimination model (WinNonlin,
Scienti®c Consulting Inc., Apex, NC). The correla-
tion between observed and predicted model time
point values was greater than 97% for each test
article examined.
Chemical methods. Unless stated otherwise, all reagents
and solvents were used without further puri®cation.
NMR spectra were obtained using a 270 MHz spectro-
meter and deuterated chloroform as the solvent unless
speci®ed. PEG diols (40 kDa) were obtained from Serva
(Crescent Chemical Company, NY). PBS bu€er was
purchased from Sigma Chemical Company. All PEG
compounds were dried under vacuum or by azeotropic
distillation from toluene prior to use. Elemental analysis
were performed by Galbraith Laboratories, Knoxville,
TN, and FAB-MS analyses were done at the mass
spectrometry facility of Yale Medical School, New
Haven, CT.
Method B: 2-Chloro-1-methylpyridinium iodide (Muk-
aiyama reagent) (3b). A mixture of PEG 40 kDa diacid
(2, 5 g, 0.125 mmol), and camptothecin (0.2 g,
0.57 mmol) in toluene (150 mL) was azeotroped for 2 h.
The reaction mixture was cooled to room temperature
and the solvent was completely removed by distillation
in vacuo. Anhyd. dichloromethane (200 mL) was added
to the reaction mixture and the solution was chilled to
0 ꢀC for 15 min followed by the addition of 2-chloro-1-
methylpyridinium chloride (0.25 g, 0.98 mmol) and
DMAP (0.25 g, 2.05 mmol). The reaction mixture was
cooled to room temperature and the solvent was com-
pletely removed by distillation in vacuo. Anhyd.
dichloromethane (200 mL) was added to the reaction
mixture and the solution was chilled to 0 ꢀC for 15 min
followed by the addition of 2-chloro-1-methylpyr-
idinium chloride (0.25 g, 0.98 mmol) and DMAP (0.25 g,
2.05 mmol). The reaction mixture was allowed to warm
to room temperature slowly and stirred for 48 h. The
solution was washed with 0.5 N HCl (2Â50 mL), dried
(anhyd. MgSO4), and evaporated under reduced pres-
sure to yield a light-yellow solid that was recrystallized
from 2-propanol (500 mL) to give a white solid, fol-
lowed by a second recrystallization from DMF (60 mL).
Abbreviations. DIPC (diisopropylcarbodiimide), DMAP
(dimethylaminopyridine), TFA (tri¯uoroacetic acid),
DMF (dimethylformamide), EEDQ (2-ethoxy-1-(ethoxy-
carbonyl)-1,2-dihydroquinoline).
Analysis of 20-camptothecin PEG 40 kDa esters. The
percentage of camptothecin in the PEG-camptothecin
esters was determined using an identical UV assay as
previously published9 for PEG-taxol.
Determination of rates of hydrolysis of camptothecin-
PEG esters. The rates of hydrolysis were obtained by
employing a C-8 reverse-phase column (Zorbax1 SB
300 A, 4.6 mmÂ50 mm), using a gradient mobile phase
consisting of (a) 0.05% TFA in water and (b) 0.05%
TFA in acetonitrile.

R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
559
The solid was ®ltered and washed with 25 mL of ice-cold
DMF:ethyl ether (1:1, v/v) followed by ethyl ether
(3Â50 mL) to give 3b (4.7 g, 92%) as a pale-yellow solid.
HPLC analysis indicated that the product contained
PEG 40 kDa diamine dihydrochloride (8). Polyethylene
glycol 40 kDa (500 g, 12.5 mmol) was placed in an 1 L
round-bottomed ¯ask and warmed to 70 ꢀC under
vacuum (0.1 Torr) for 16 h to remove traces of water.
To the dried PEG-OH was added thionyl chloride
(100 mL) and the reaction mixture was kept at 70 ꢀC for
16 h. 2-Propanol (5 L) was added to the reaction mix-
ture. The solution was cooled overnight at room tem-
perature. The precipitated product was ®ltered and
washed with 2-propanol followed by recrystallization
from 2-propanol to yield 500 g (98%) of colorless crys-
talline PEG 40 kDa dichloride: 13C NMR: d 42.18,
69.94±71.76 (PEG). A solution of PEG 40 kDa dichlor-
ide (50 g, 1.25 mmol) in 30% ammonia solution
(400 mL) was placed in a sealed polypropylene bottle
and heated at 60 ꢀC for 3 days. Subsequent removal of
the solvent from the reaction mixture followed by
recrystallization from 2-propanol (1.5 L) yielded 8 (44 g,
87%): 13C NMR: d 39.20, 66.04±71.49 (PEG).
1
ꢁ90% of diester: H NMR: d 1.01 (t, J=7.2 Hz, 1H),
2.3 (m, 1H), 2.80 (s, PEG), 3.49±3.71 (bs, PEG), 3.93 (t,
J=5.2 Hz, PEG), 4.4 (s, 4H), 5.44 (s, 1H), 5.5±5.8 (AB
q, J=17.16 Hz, 1H), 7.3 (s, 1H), 7.6 (m, 1H), 7.7 (t,
J=7.26 Hz, 1H), 7.97 (d, J=8.5 Hz, 1H), 8.23 (d,
J=8.5 Hz, 1H), 8.44 (s, 1H); 13C NMR: d 7.15, 31.28,
49.58, 66.70, 67.58±71.83 (PEG), 95.39, 119.78, 127.63,
127.79, 127.87, 128.10, 129.12, 130.26, 130.97, 144.95,
146.01, 148.37, 151.80, 156.84, 166.73, 169.21.
Mono t-butylester of diglycolic acid (5). A solution of
diglycolic anhydride (10 g, 0.09 mol) and DMAP (10.5 g,
0.09 mol) in dry t-butanol (75 mL) was stirred at re¯ux
temperature for 18 h. The solvent was removed under
reduced pressure and the residue was dissolved in water
(100 mL). The aq solution was acidi®ed to pH 2.5±3.0
with 1 N HCl and extracted with dichloromethane.
Removal of solvent from the dried extracts yielded
12.3 g (75%) of the mono t-butylester of diglycolic acid
5: 13C NMR: d 27.84, 68.45, 68.78, 82.42, 169.54,
172.85.
PEG 40 kDa di-N-methylamine dihydrochloride (9). A
solution of PEG 40 kDa dichloride (50 g, 1.25 mmol) in
40% methylamine (400 mL) was placed in a sealed
polypropylene bottle and heated at 60 ꢀC for 3 days.
Subsequent removal of the solvent from the reaction
mixture followed by recrystallization from 2-propanol
(1.5 L) yielded 9 (44 g, 87%): 13C NMR: d 33.10, 48.38,
66.18±71.60 (PEG).
Camptothecin-20-monoester of diglycolic acid (7). A
mixture of 5 (4.2 g, 0.02 mol), camptothecin (4.0 g,
0.01 mol), DMAP (2.7 g, 0.02 mol), and DIPC (2.8 g,
0.02 mol) in anhyd. dichloromethane (40 mL) was stir-
red for 18 h at room temperature. The reaction mixture
was washed with water, then saturated aq sodium
bicarbonate, 0.1 N HCl and again with water. The
organic layer was dried (anhyd MgSO4) and the solvent
was removed in vacuo. Recrystallization of the resultant
solid from dichloromethane/ether gave 6 (3.1 g, 54%):
1H NMR: d 8.35(s), 8.15±8.18 (d), 7.89±7.92 (d), 7.80
(m), 7.63±7.66 (m), 7.20 (s), 5.35±5.72 (AB q,) 5.21 (s),
4.45±4.48 (d), 4.11±4.13 (d), 2.2±2.3 (m), 1.45 (s), 1.00
(t); 13C NMR: d 7.38, 27.86,31.53, 49.72, 66.96, 67.45,
68.36, 76.38, 81.75, 95.54, 119.94, 127.76, 127.99,
128.21, 129.32, 130.39, 130.97, 145.13, 146.25, 148.55,
151.94, 157, 166.98, 168.52, 168.97. MS (FAB)
(M+H)+ 521.4. A solution of compound 6 (0.8 g,
1.5 mmol) in dichloromethane (8 mL) and TFA (4 mL)
was stirred at room temperature for 30 min. The solvent
was removed under reduced pressure, and the resulting
solid was recrystallized from dichloromethane/ether to
yield 7 (0.6 g, 82%): 1H NMR: d 8.44 (s), 8.27±8.34 (m),
7.95 (d), 7.86 (m), 7.70 (m), 7.34 (s), 5.37±5.75 (AB q),
5.31 (s), 4.47±4.5 (d), 4.39±4.41 (d), 2.18±2.26 (m), 1.03
(t); 13C NMR (DMSO-d6): d 171.15, 169.27, 167.41,
156.84, 152.63, 148.19, 146.41, 145.45, 131.88, 130.71,
130.11, 129.22, 128.83, 128.29, 128.02, 119.08, 95.24,
76.72, 67.57, 67.33, 66.62, 50.54, 30.4, 7.84. MS (FAB)
(M+H)+ 465.3.
PEG 40 kDa diamide of acid 7 (10). A mixture of 7
(0.14 g, 0.3 mmol), PEG 40 kDa diamine hydrochloride
8 (3.0 g, 0.075 mmol), DMAP (55 mg, 0.45 mmol), and
DIPC (38 mg, 0.3 mmol) in anhyd. dichloromethane
(30 mL) was stirred for 18 h at room temperature.
Removal of the solvent in vacuo and recrystallization
from 2-propanol gave 10 (2.8 g, 90%). 13C NMR: d
168.39, 168.09, 166.51, 156.81, 151.80, 148.44, 146.22,
144.76, 131.04, 130.32, 129.17, 128.17, 127.91, 127.71,
119.77, 95.05, 76.72, 66.73±71.89 (PEG), 49.64, 38.24,
31.38, 7.0.
PEG 40 kDa di-N-methylamide of acid 7 (11). A mixture
of 7 (0.14 g, 0.3 mmol), PEG 40 kDa di-N-methylamine
hydrochloride 11, (3.0 g, 0.075 mmol), DMAP (55 mg,
0.45 mmol), and DIPC (38 mg, 0.3 mmol) in anhyd
dichloromethane (30 mL) was stirred for 18 h at room
temperature. Removal of the solvent in vacuo and
recrystallization from 2-propanol gave 11 (2.8 g, 90%):
13C NMR: d 168.73, 168.09, 166.51, 156.58, 151.80,
148.18, 145.91, 144.77, 130.84, 130.03, 128.95, 127.99,
127.71, 127.61, 127.40, 119.77, 95.05, 76.72, 66.46±71.65
(PEG), 49.41, 48.3, 47.5, 35.24, 33.3, 31.06, 6.97.
Camptothecin-20-ester of hydroxyacetic acid (14). To a
suspension of camptothecin (1.4 g, 4.02 mmol) in
CH2Cl2 (500 mL) was added benzyloxyacetic acid 12

560
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
(2 g, 12.06 mmol, prepared by the aqueous hydrolysis of
benzyloxyacetyl chloride), DIPC (1.9 mL, 12.06 mmol)
and DMAP (982 mg, 8.04 mmol) at 0 ꢀC and stirring
was continued for 3 h. The resulting yellow solution was
concentrated to about 100 mL and washed with 1 N
HCl (2Â10 mL), followed by 1% aq sodium bicarbonate
solution (2Â10 mL). The organic layer was dried
(anhyd MgSO4) and evaporated in vacuo to give a yel-
low solid which was recrystallized from ethyl acetate.
The product was triturated with methanol (10 mL), and
the slurry ®ltered to yield camptothecin-20-ester of
similar manner as compound 16, using PEG 40 kDa
di-N-methyldiamine 9 as starting material. 13C NMR d
166.63, 166.08, 156.17, 151.39, 147.84, 145.41, 144.35,
130.84, 130.03, 128.95, 127.99, 127.71, 127.61, 127.40,
118.68, 94.30, 76.72, 65.66±71.65 (PEG), 61.3, 49.37,
48.52, 48.15, 47.82, 35.14, 34.54, 33.15, 30.97, 21.62,
6.97.
Camptothecin-20-glycinate TFA salt (22). t-Boc-glycine
18 (1.8 g, 9.39 mmol) was dissolved in 700 mL of anhyd
methylene chloride at room temperature and to this
solution were added DIPC (1.5 mL, 9.39 mmol), DMAP
(765 mg, 6.26 mmol) and camptothecin (1.09 g,
3.13 mmol) at 0 ꢀC. The reaction mixture was allowed to
warm to room temperature and left for 16 h. The solu-
tion was washed with 0.1 N HCl, dried and evaporated
under reduced pressure to yield a white solid, which was
recrystallized from methanol to give camptothecin-20-
ester of t-Boc-glycine 20: 1H NMR(DMSO-d6) d 0.9
(t), 1.3 (d), 1.6 (s), 2.1 (m), 4 (m), 5.3 (s), 5.5 (s), 7.3 (s),
7.5±8.8 (m). Anal. (C27H27N3O7. 0.5 H2O) C, H, N.
Compound 20 (1.19 g, 2.12 mmol) was dissolved in a
mixture of methylene chloride (15 mL) and TFA
(15 mL) and stirred at room temperature for 1 h. Solvent
was removed and the solid was recrystallized from
methylene chloride and ether to give 1 g of product 22.
1H NMR (DMSO-d6) d 1.0 (t), 1.6 (d), 2.2 (m), 4.4 (m),
5.4 (s), 5.6 (s), 7.2 (s), 7.7±8.8 (m); 13C NMR (DMSO-
d6) d 7.5, 15.77, 30.09, 47.8, 50.27, 66.44, 77.5, 94.92,
119.10, 127.82, 128.03, 128.62, 128.84, 129.75, 130.55,
131.75, 144.27, 146.18, 147.90, 152.24, 156.45, 166.68,
168.69.
1
benzyloxyacetic acid (13) (1.3 g, 65%). H NMR: d 1.0
(t), 1.84 (s), 2.1±2.3 (m), 4.31 (s), 4.59±4.69 (q), 5.28 (s),
5.4±5.8 (dd), 7.22 (s), 7.27 (s), 7.3±7.38 (m), 7.6±7.7 (m),
7.81±7.87 (m), 7.92±7.95 (d), 8.19±8.22 (d), 8.39 (s); 13C
NMR: d 7.52, 31.74, 49.90, 66.59, 67.16, 73.27, 76.38,
95.80, 120.27, 127.97, 128.10, 128.43, 129.54, 130.63,
131.15, 136.81, 145.39, 146.36, 148.81, 152.22, 157.26,
167.19, 169.52. MS (FAB) (M+H)+ 497.6. A suspen-
sion of 13 (1 g, 2.01 mmol) and 10% Pd/C (500 mg) in
ethanol (100 mL) was degassed by sparging with nitrogen.
Cyclohexene (5 mL) was added, and the mixture was
re¯uxed for 20 h. The catalyst was ®ltered and the sol-
vent was removed in vacuo followed by recrystallization
of the solid product from acetonitrile to give 14 (500 mg,
1
50%): H NMR (DMSO-d6): d 1.0 (t), 1.84 (s), 2.1±2.3
(m), 3.1 (s), 4.31 (s), 4.0±4.69 (m), 5.28 (s), 5.6 (m) 5.8 (m),
7.22 (s), 7.6±7.7 (m), 7.6±7.95 (m), 8.0±8.2 (m), 8.3 (s),
8.7 (s); 13C NMR (DMSO-d6) d 7.5, 30.14, 50.20, 59.37,
66.21, 75.83, 79.14, 94.95, 118.84, 127.71, 127.97, 128.52,
128.87, 129.77, 130.42, 131.58, 145.35, 145.95, 147.85,
152.29, 156.53, 167.21,171.69. MS (FAB) (M+H)+ 407.4.
PEG dicarbamate derivative of camptothecin-di-20-
Camptothecin-20-sarcosinate TFA salt (23). A mixture
of sarcosine (5 g, 56.1 mmol), t-Boc anhydride (14.7 g,
67.4 mmol) and sodium hydroxide (4.5 g, 112.3 mmol)
in water (25 mL) was stirred at room temperature for
18 h. The reaction mixture was cooled to 0 ꢀC and was
acidi®ed to pH 3 with 6 N HCl and extracted with
ethyl acetate. Evaporation of the solvent gave t-Boc
hydroxyacetate (16).
A solution of 14 (240 mg,
0.59 mmol) and N,N0-carbonyldiimidazole (288 mg,
1.77 mmol) in chloroform (80 mL) was stirred at 50 ꢀC
for 18 h. Removal of solvent in vacuo followed by tri-
turation of the residue with ethyl acetate gave the car-
bonylimidazole derivative of camptothecin 15 as a pale-
1
1
yellow solid (170 mg, 58%). H NMR d 1.0 (t), 2.1±2.5
sarcosine 19 as a clear oil. H NMR d 4.58 (m), 3.0 (s),
(m), 5.1 (d), 5.28 (s), 5.4±5.8 (dd), 7.0 (s), 7.5 (s), 7.7 (m),
7.9 (m), 8 (d), 8.1 (s), 8.3 (d), 8.4 (s). MS (FAB)
(M+H)+ 501.6. A solution of 15 (74 mg, 0.15 mmol)
and PEG 40 kDa diamine 8 (2 g, 0.05 mmol) in 2-pro-
panol (20 mL) was re¯uxed for 12 h. The solvent was
removed under reduced pressure to yield 16 as a solid,
which was further puri®ed by recrystallization from 2-
propanol: 13C NMR d 167.24, 166.44, 156.77, 154.83,
152.0, 148.0, 145.96, 144.35, 130.81, 130.10, 129.20,
128.09, 127.92, 127.89, 127.78, 127.60, 127.50, 119.72,
95.70, 76.72, 65.66±71.65 (PEG), 61.3, 49.51, 40.58,
31.30, 6.97.
4.0 (m). Compound 19 (1.63 g, 8.61 mmol) was dissolved
in 100 mL of anhyd methylene chloride at room tem-
perature and to this solution were added DIPC (1.3 mL,
8.61 mmol), DMAP (725 mg, 5.74 mmol) and camp-
tothecin (1 g, 2.87 mmol) at 0 ꢀC. The reaction mixture
was allowed to warm to room temperature and left for
2 h. The solution was washed with 0.1 N HCl, dried and
evaporated under reduced pressure to yield a white solid
which was recrystallized from 2-propanol to give camp-
tothecin-20-t-Boc-sarcosinate (21, 750 mg, 50.3%).
Compound 21 (750 mg) was dissolved in methylene
chloride (4 mL) and TFA (4 mL) and the solution
was stirred at room temperature for 1 h. Ether
(10 mL) was added and the precipitated solid was ®l-
tered and dried to give 23 (550 mg, 85%) as a yellow
PEG di-N-methylcarbamate derivative of camptothecin-
di-20-hydroxyacetate (17). This was prepared in
a

R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
561
1
1
powder. H NMR(DMSO-d6) d 1.0 (t), 2.2 (m), 2.7 (s),
67%). H NMR d 1.0 (t), 1.84 (s), 2.1±2.3 (m), 3.9±4.4
(q), 5.28 (s), 5.4±5.8 (dd), 7.2 (s), 7.27 (s), 7.6±7.7 (m),
7.81±7.87 (m), 7.92±7.95 (d), 8.19±8.22 (d), 8.39 (s); 13C
NMR d 7.52, 24.97, 31.77, 49.97, 67.16, 76.53, 95.73,
120.29, 128.05, 128.17, 128.39, 129.64, 130.65, 131.17,
144.94, 146.48, 148.84, 152.18, 157.24, 165.97, 166.83.
Anal. (C22H17BrN2O5.0.5 H2O) C, H, N.
2.84 (s), 4.4±4.6 (dd), 5.11 (brs), 5.34 (s), 5.65 (s), 7.36
(s), 7.6±8.3 (m), 8.74 (s), 9.46 (s); 13C NMR (DMSO-d6)
d 7.55, 30.21, 32.49, 47.87, 50.22, 66.40, 77.72, 95.34,
118.84, 127.74, 127.95, 128.76, 129.67, 130.51, 131.65,
144.64, 147.88, 152.24, 156.48, 158.23, 166.78. Anal.
(C25H22F3N3O7.0.5 H2O) C, H, N.
Camptothecin-di-20-ester of PEG 40 kDa glycine (24). A
solution of 1.0 g (0.025 mmol) of PEG 40 kDa dicar-
boxylic acid in toluene (35 mL) was azeotroped by dis-
tillation of 15 mL of toluene for 2 h. The mixture was
cooled to room temperature and 22 (70 mg, 0.13 mmol),
EEDQ (30 mg, 0.12 mmol), and DMAP (50 mg,
0.4 mmol) were added. The reaction mixture was stir-
red at 55±60 ꢀC overnight and the solvent was removed
by distillation in vacuo. The residue was crystallized
from 200 mL of 2-propanol and washed with anhy-
drous ethyl ether to give the product as a white solid
(0.8554 g, 86% yield). HPLC assay indicated that the
product contained >90% of diester 24. 1H NMR d 1.00
(t, J=7.2 Hz, 1H), 2.20 (m, 1H), 2.80 (s, PEG), 3.38±
4.15 (m, PEG), 4.43 (s,4H), 5.30 (s, 1H), 5.43±5.64 (AB
q, J=17.16 Hz, 1H), 7.32 (s, 1H), 7.70 (t, J=7.26 Hz,1
H), 7.84 (t, J=7.25 Hz, 1H), 7.95 (d, J=8.5 Hz, 1H),
8.26 (d, J=8.5 Hz, 1H), 8.46 (s, 1H); 13C NMR d 7.12,
31.30, 39.93, 49.53, 66.57, 68.32±71.62 (PEG), 95.63,
119.54, 127.57, 127.,75, 128.05, 129.20, 130.16, 130.89,
144.97, 145.97, 148.32, 151.78, 156.77, 166.57, 168.45,
170.27.
Reaction of 27 with PEG diamine hydrochloride 8 (28).
A solution of PEG 40 kDa diamine hydrochloride 8
(5 g, 0.125 mmol), 20-bromoacetyl camptothecin 26
(322 mg, 0.63 mmol) and triethylamine (208 mL,
1.5 mmol) in anhyd methylene chloride (75 mL) was
stirred at room temperature for 3 days. The solvent was
evaporated under reduced pressure and the residual
solid obtained was ®rst recrystallized from DMF and
then using 2-propanol to give 28 as a white solid (4.4 g,
87.4%). 13C NMR d 6,99, 32.16, 38.50, 47.97, 48.39,
66.17±70.84 (PEG), 75.96, 95.02, 119.26, 127.18, 128.10,
128.25, 129.22, 130.19, 131.20, 145.57, 146.64, 147.82,
152.27, 156.71, 167.25.
Reaction of 27 with PEG di-N-methyldiamine hydro-
chloride 9 (29). A solution of PEG 40 kDa di-N-
methyldiamine hydrochloride 9 (5 g, 0.125 mmol), 20-
bromoacetyl camptothecin 26 (322 mg, 0.63 mmol) and
triethylamine (208 mL, 1.5 mmol) in anhyd. methylene
chloride (75 mL) was stirred at room temperature for
three days. The solvent was evaporated under reduced
pressure and the solid obtained was ®rst recrystallized
from DMF and then using 2-propanol to give 29 as a
white solid (4.5 g, 90%): 13C NMR d 6.84, 30.62, 41.53,
49.33, 53.96, 64.97, 66.12, 67±70.5 (PEG), 94.82,118.53,
127.21, 127.40, 127.55, 127.91, 128.59, 129.85, 130.81,
144.34, 145.84, 147.84, 151.23, 156.19, 165.77.
Camptothecin-20-ester of PEG 40 kDa sarcosinate (25).
Diacid 2 (2 g, 0.05 mmol) was dissolved in 30 mL of
anhyd methylene chloride at room temperature and to
this solution were added DIPC (30 mL, 0.20 mmol),
DMAP (24 mg, 0.20 mmol) and 20-camptothecin sar-
cosinate TFA salt 23 (112 mg, 0.21 mmol) at 0 ꢀC. The
reaction mixture was allowed to warm to room tem-
perature and left for 16 h. The solution was evaporated
under reduced pressure to yield a white solid which was
recrystallized from 2-propanol to give 25 (1.4 g, 69%).
NMR data is in agreement with the structure.
Acknowledgements
We gratefully acknowledge the contributions of the
pharm±tox group of Rita Linberg, Scott Shen, and
Dr Kwok Shum; and the assistance of Doug MacDo-
nald and Hong Hsu of the analytical group. We also
greatly appreciate the secretarial e€orts of Ms Vivian
Laurel.
Camptothecin-20-ester of bromoacetic acid (27). To a
suspension of camptothecin (1 g, 2.87 mmol) in anhyd
CH2Cl2 (700 mL) were added bromoacetic acid 26
(1.2 g, 8.61 mmol), DIPC (1.3 mL, 8.61 mmol) and
DMAP (700 mg, 5.74 mmol) at 0 ꢀC and stirring was
continued for 4 h. The resulting yellow solution was
concentrated to about 100 mL and washed with 1 N
HCl (2Â10 mL), followed by 1% aq sodium bicarbonate
solution (2Â10 mL). The organic layer was dried
(anhyd MgSO4) and evaporated in vacuo to give a yel-
low solid which was recrystallized from ethyl acetate.
This crude product was triturated with methanol
(10 mL), and the slurry was ®ltered to yield 27 (0.9 g,
References and Notes
1. Gros, L.; Ringsdor€, H.; Schlupp, H. Angew Chem. Int. Ed.
Engl. 1981, 20, 305±325.
2. (a) Kopecek, J.; Biomaterials 1984, 5, 19±25. (b) Gardner,
C. R. Biomaterials 1985, 6, 153±160. (c) Kopecek, J.; Duncan,
R. J. Control Release. 1987, 6, 315±327. (d) Friend, D. R.;
Pangburn, S. Med. Res. Revs. 1987, 1, 53±106. (e) Drobnik,
J. Adv. Drug Delivery Rev. 1989, 3, 229±245. (f) Sezaki, H.;
Takakura, Y.; Hashida, M. Med. Res. Revs. 1989, 3, 247±266.
(g) Maeda, H. Med. Res. Revs. 1991, 6, 181±202. (h) Maeda,

562
R. B. Greenwald et al./Bioorg. Med. Chem. 6 (1998) 551±562
H.; Seymour, L. W.; Miyamoto, Y. Bioconjugate Chem. 1992,
3, 351±362. (I) Duncan, R. Anti-Cancer Drugs 1992, 175±210.
(j) Seymour, L. W. Crit. Rev. Ther. Drug Carrier Syst. 1992, 9,
135±137. (k) Duncan, R. Clin. Pharmacokinet. 1994, 27, 290±
306. (l) Takakura, Y.; Hashida, M. Crit. Rev. Oncology/Hema-
tology 1995, 18, 207±231. (m) Matthews, S. E.; Pouton, C. W.;
Threadgill, M. D. Adv. Drug Delivery Rev. 1996, 18, 219±267.
(n) Duncan, R.; Dimitrijevic, S.; Evagorou, E. G. STP Pharma.
Sciences 1996, 6, 237±263.
16. (a) Deutsch, H. M.; Glinski, J. A.; Hernandez, M.; Haug-
witz, R. D.; Narayanan, V. L.; Su€ness, M.; Zalkow, L. H. J.
Med. Chem. 1989, 32, 788±792. (b) Mathew, A. E.; Mejillano,
M. R.; Nath, J. P.; Himes, R. H.; Stella, V. J. J. Med. Chem.
1992, 35, 145±151.
17. Rao Vishnuvajjala, B.; Garzon-Aburbeh, A. U.S Patent
4,943,579, 1990.
18. Bundgaard, H.; Falch, E.; Jensen, E. J. Med. Chem. 1989,
32, 2503±2507.
3. Hurwitz, E.; Wilchek, M.; Pitha, J. J. Appl. Biochem. 1980,
2, 25±35.
19. Greenstein, J. P.; Winitz, M. Chemistry of the amino acids,
Vol 1. New York: John Wiley, 1961, 498±499.
4. Ueda, Y.; Munechika, K.; Kikukawa, A.; Kanoh, Y.;
Yamanouchi, K.; Yokoyama, K. Chem. Pharm. Bull. 1989, 37,
1639±1641.
20. Sheehan, J. C.; Lengyel, I. (J. Am. Chem. Soc. 1964, 86,
1356±1359) reported the preparation of a-lactams (from N-
alkyl amides), and their rapid hydrolysis in the presence of
water. Under similar conditions N,N-dialkylamides do not
form aziridinones: Scrimin, P.; D'Angeli, F.; Cavicchioni, G.
Synthesis 1982, 1092±1094.
5. Danhauser-Reidl, S.; Hausmann, E.; Schick, H.-D.; Bender,
R.; Dietzfelbinger, H.; Rastetter, J.; Hanauske, A.-R. Investi-
gational New Drugs 1993, 11, 187±195.
21. Conover, C. D.; Pendri, A.; Lee, C.; Gilbert, C. W.; Shum,
K. L.; Greenwald, R. B. AntiCancer Res. 1997, 17, 3361±3368.
6. See also: Duncan, R.; Cable, H. C.; Lloyd, L. B.; Rejma-
nova, P.; Kopecek, J. Makromol Chem. 1984, 184, 1997±2008.
Rajmanova, P.; Kopecek, J.; Duncan, R.; Lloyd, J. B. Bioma-
terials 1985, 6, 45±48.
22. Levy, R. K.; Hall, I. H.; Lee, K.-H. J. Pharm. Sci. 1983,
1158±1161.
7. Vasey, P. A.; Duncan, R.; Kaye, S. B.; Cassidy, J. Eur. J.
Cancer 1995, 31A, S193.
23. (a) Bernhard, S. A.; Berger, A.; Carter, J. H.; Katchalski,
E.; Sela, M.; Shalitin, Y. J. Am. Chem. Soc. 1962, 84, 2421±
2434. (b) Mitra Tadayoni, B. M.; Friden, P. M.; Walus, L. R.;
Musso, G. F. Bioconjugate Chem. 1993, 4, 139±145.
8. Greenwald, R. B.; Pendri, A.; Conover, C.; Gilbert, C.;
Yang, R.; Xia, J. J. Med. Chem. 1996, 39, 1938±1940.
24. Gogate, U. S.; Repta, A. J.; Alexander, J. Int. J. Pharma-
ceutics 1987, 40, 235±248.
9. Greenwald, R. B.; Gilbert, C. W.; Pendri, A.; Conover,
C. D.; Xia, J.; Martinez, A. J. Med. Chem. 1996, 39, 424±431.
25. (a) Fife, T. H.; DeMark, B. R. J. Am. Chem. Soc. 1976, 98,
6978±6982. (b) Saari, W. S.; Schwering, J. E.; Lyle, P. A.;
Smith, S. J.; Englehardt, E. L. J. Med. Chem. 1990, 33, 97±101.
10. (a) Maeda, H. J. Control Release. 1992, 19, 315±324. (b)
Maeda, H.; Matsumura, Y. Crit. Rev. Ther. Drug Carrier Syst.
1989, 6, 193±210.
26. It should be noted that our results for Topotecan in the
HT-29 model are virtually identical to those previously repor-
ted.27
11. Carl, P. L.; Chakravarty, P. K.; Katzenellenbogen, J. A. J.
Med. Chem. 1981, 24, 479±480.
12. Mukaiyama, T.; Usai, M.; Shimada, E.; Saigo, K. Chem.
Lett. 1975, 1045.
27. Luzzio, M. J.; Besterman, J. M.; Evans, M. G.; Lackey, K.;
Leitner, P. L.; McIntyre, G.; Morton, B.; Myers, P. L.; Peel,
M.; Siscom, J. M.; Sternbach, D. D.; Tong, W. Q.; Truesdale,
A.; Uehling, D. E.; Vuong, A.; Yates, J. J. Med. Chem. 1995,
38, 395±401.
13. Wall, M. E.; Wani, M. C.; Cook, C. E.; Palmer, K. H.;
McPhail, A. T.; Sim, G. A. J. Am. Chem. Soc. 1966, 88, 3888±
3890.
14. Employing diglycolic anhydride with paclitaxel has been
shown to give a slightly water soluble ester prodrug: Nicolaou,
K. C.; Riemer, C.; Kerr, M. A.; Rideout, D.; Wraisidlo, W.
Nature 1993, 364, 464±466.
28. A dose response study employing 24 in the HT-29 model
resulted in substantial tumor reduction with as little as 1 mg/kg/
mouse. A similar response from 3 required 2.5 mg/kg/mouse.21
29. Kingsbury, W. D.; Boehm, J. C.; Jakas, D. R.; Holden, K.
G.; Hecht, S. M.; Gallagher, G.; Caranfa, M. J.; McCabe, F.
L.; Faucette, L. F.; Johnson, R. K.; Hertzberg, R. P. J. Med.
Chem. 1991, 34, 98±107.
15. (a) Colla, L.; DeClerq, E.; Busson, R.; Vanderhaeghe, H. J.
Med. Chem. 1983, 26, 602±604. (b) Bundgaard, H.; Larsen, C.;
Thorbek, P. Int. J. Pharm. 1984, 18, 67±77.