Phenylalanyl-aminocyclophosphamides as model prodrugs for proteolytic activation: Synthesis, stability, and stereochemical requirements for enzymatic cleavage

Article abstract of DOI:10.1016/j.bmcl.2006.10.017

4-Aminocyclophosphamide (4-NH₂-CPA, 7) was proposed as a prodrug moiety of phosphoramide mustard. Four diastereomers of phenylalanine-conjugates of 4-NH₂-CPA were synthesized and their stereochemistry was assigned based on chromatogr

Full text of DOI:10.1016/j.bmcl.2006.10.017

Bioorganic & Medicinal Chemistry Letters 17 (2007) 517–521
Phenylalanyl-aminocyclophosphamides as model prodrugs for
proteolytic activation: Synthesis, stability, and stereochemical
requirements for enzymatic cleavage
Yongying Jianga and Longqin Hua,b,
*
aDepartment of Pharmaceutical Chemistry, Ernest Mario School of Pharmacy, Rutgers, The State University of New Jersey,
Piscataway, NJ 08854, USA
bThe Cancer Institute of New Jersey, New Brunswick, NJ 08901, USA
Received 31 August 2006; revised 5 October 2006; accepted 5 October 2006
Available online 12 October 2006
Abstract—4-Aminocyclophosphamide (4-NH2-CPA, 7) was proposed as a prodrug moiety of phosphoramide mustard. Four dia-
stereomers of phenylalanine-conjugates of 4-NH2-CPA were synthesized and their stereochemistry was assigned based on chroma-
tographic and spectroscopic data. All diastereomers were stable in phosphate buffer but only the cis-(4R)-isomer of 15 was efficiently
cleaved by a-chymotrypsin with a half-life of 20 min, which is much shorter than the 8.9 h to >12 h half-lives found for the other
diastereomers. LC–MS analysis of the proteolytic products of cis-(4R)-15 indicated that 4-NH2-CPA was released upon proteolysis
and further disintegrated to phosphoramide mustard. These results suggest the feasibility of using peptide-conjugated cis-(4R)-4-
NH2-CPA as potential prodrugs for proteolytic activation in tumor tissues.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Most anticancer agents suffer severe side effects from
their marginal selectivity against cancer cells over
healthy ones. The side effects limit the maximum dose
of these agents to below the amount that can kill all via-
ble tumor cells and the survived tumor cells could be
more drug resistant and malignant.1,2 Among strategies
explored to increase the tumor selectivity of anticancer
agents, tumor-targeted prodrug therapy has attracted
much attention.3–5 In this strategy, an inert prodrug
form of an anticancer agent is selectively activated
through a biochemical mechanism associated with tu-
mor cells, such as hypoxic reduction, enzymatic action,
or receptor recognition. Recently, several proteases have
been identified to be unique for tumor growth and
metastasis, including plasmin,6,7 plasminogen activator,8
matrix metalloproteinases,9,10 and prostate specific anti-
gen (PSA).11,12 The specificity of these enzymes to tumor
cells has provided novel opportunities that are being ex-
plored for tumor-targeted prodrug therapy.
Cyclophosphamide is one of the often used anticancer
drugs in the clinic. Because of its activity against both
cycling and non-cycling cells, it is one of the few antican-
cer agents effective in the treatment of slow-growing sol-
id tumors.13 However, the clinical use of
cyclophosphamide is associated with a life-threatening
side effect, hemorrhagic cystitis, in addition to bone
marrow and GI tract toxicities which are commonly
seen for most other anticancer agents. The existing
knowledge of cyclophosphamide has facilitated the
development of a variety of phosphoramide mustard
prodrugs for tumor-targeted activation.14–17 In the last
few years, we have focused on improving the therapeutic
efficiency of cyclophosphamide through the strategy of
tumor-targeted enzyme prodrug therapy. In one ap-
proach, we developed a series of nitrobenzyl phosphora-
mide mustard prodrugs for Escherichia coli
nitroreductase activation in enzyme prodrug thera-
py.18,19 In this communication, we describe our efforts
toward the conversion of cyclophosphamide into pep-
tide-conjugated 4-aminocyclophosphamide prodrugs in
the form of 6 for tumor-targeted proteolytic activation.
To our knowledge, this is the first time that 4-aminocy-
clophosphamide (4-NH2-CPA, 7) is proposed as a pro-
drug moiety of phosphoramide mustard.
Keywords: Anticancer prodrugs; Cyclophosphamide; 4-Aminocyclo-
phosphamide; Site-specific activation; Proteolytic activation; gem-
Diamines; Phosphoramide mustard.
*
Corresponding author. E-mail: LongHu@rci.rutgers.edu
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.10.017

518
Y. Jiang, L. Hu / Bioorg. Med. Chem. Lett. 17 (2007) 517–521
Cyclophosphamide (1) is a prodrug that requires activa-
tion by cytochrome P450 enzymes in the liver for its
anticancer activity.20,21 The enzymatic oxidation con-
verts 1 to 4-hydroxycyclophosphamide (2) which gener-
ates acrolein (4) and phosphoramide mustard (5) after
ring opening and b-elimination (Scheme 1).20 Acrolein
is responsible for the hemorrhagic cystitis associated
with cyclophosphamide and phosphoramide mustard is
the ultimate alkylating species that cross-links inter-
strand DNA.22
a-chymotrypsin. Because of the presence of two chiral
centers in the oxazaphosphorinane ring, compounds 15
and 16 exist as four configurational diastereomers. They
are referred to as cis-(4R)-, trans-(4R)-, cis-(4S)-, and
trans-(4S)- (Fig. 1) according to the oxazaphosphor-
inane C-4 configuration and the relative orientation
of C-4 substituent to the oxygen atom of P@O bond
in the oxazaphosphorinane ring (cis = SR/RS,
trans = RR/SS).
The two diastereomers of (4R)-16 were synthesized ste-
reospecifically from L-homoserine ((S)-9) while the two
diastereomers of (4S)-16 were synthesized from D-ho-
moserine ((R)-9) as shown in Scheme 2. Z-Phe-OH
was preactivated using HOSu/DCC and the resulting
activated ester reacted with 9 in a mixture of 1 M
KHCO3 and THF yielding the dipeptide 10. Amidation
of 10 was carried out using HOBt/EDC activation fol-
lowed by treatment with saturated ammonium hydrox-
ide,25 and the hydroxyl group in 11 was protected with
a TBDPS group. The bis(trifluoroacetoxy)iodobenzene
(BTI)-mediated Hofmann rearrangement was employed
to convert the amide 12 to the corresponding mono-ac-
ylated gem-diamine derivative 13. This method was cho-
sen over other methods such as the Curtius
rearrangement because of its advantages of mild reac-
tion conditions, high yield, and retention of stereochem-
istry.26,27 The reaction was also conveniently monitored
by the disappearance of the starting material on TLC.
Protection of the hydroxyl group with TBDPS was nec-
essary to avoid an intramolecular cyclization of the iso-
Our design of peptide-conjugates of 4-NH2-CPA (7) as
prodrugs for proteolytic activation takes advantage of
the structural similarity of 7 to 4-hydroxycyclophospha-
mide (2) and the chemical instability of a gem-diamine.
As illustrated in Scheme 1, proteolysis of conjugate 6
generates 7 and a peptide. Under physiological condi-
tions, compound 7 is protonated and forms 3,4-dehy-
drocyclophosphamide (8) after elimination of
ammonia. Compound 7 is known to be in equilibrium
with 2 which is responsible for the generation of phos-
phoramide mustard in the activation mechanism of
cyclophosphamide.20 Chemically, 7 is a gem-diamine,
which is known as a masked aldehyde.23,24 N,N0-Diacy-
lated or dicarbamylated gem-diamines are stable com-
pounds and have been widely used in ‘retro-inverso’
peptide mimetics. Monoacyl or monocarbamyl gem-
diamines are labile to acid- or base-catalyzed elimina-
tion reactions.
The phenylalanine-conjugated 4-NH2-CPA (H-Phe-
NH-CPA, 16) and their benzyloxycarbonyl (Z)-protect-
ed analogs 15 were chosen as model prodrugs to exam-
ine the synthetic feasibility, chemical stability, and
proteolytic activation of this class of compounds. Selec-
tion of the phenylalanine residue was based on the need
of a chromophore to monitor reactions by UV and the
known cleavage of amide bond after phenylalanine by
cyanate
intermediate
during
the
Hofmann
rearrangement to form 2-oxazolidinone.28
The TBDPS group in 13 was removed with TBAF at
room temperature, which was accompanied by the for-
mation of Z-Phe-NH2 as a side product. Cyclization of
1,3-aminoalcohol 14 with bis(dichloroethyl)phospho-
O
P
O
P
CYP450
(liver)
O
P
HO
O
O
+
O
O
CHO
P
H2N
N(CH2CH2Cl)2
N
H
N(CH2CH2Cl)2
HO
N
H
N(CH2CH2Cl)2
2
H
O
N(CH2CH2Cl)2
4
H2N
5
1
3
-H2O
+H2O
O
O
P
O
O
P
O
P
H+
-NH3
Proteolysis
(tumor)
O
N
H
N
H
H2N
N
H
N(CH2CH2Cl)2
N
N(CH2CH2Cl)2
Peptide
N(CH2CH2Cl)2
6
7
8
Scheme 1. Proposed degradation mechanism of peptide-conjugated 4-aminocyclophosphamide as compared to the activation of cyclophosphamide.
O
O
H
N
H
N
H
H
N(CH2CH2Cl)2
N(CH2CH2Cl)2
R-Phe
R-Phe
O
N
N
P
P
H
N(CH2CH2Cl)2
H
O
cis-(4R)-
trans- (4R)-
N(CH2CH2Cl)2
O
P
H
O
H
O
P
O
N
H
N
H
R-Phe
N
R-Phe NH
H
cis-(4S)-
trans -(4S)-
Figure 1. The four diastereomers of Z-protected and unprotected phenylalanine-4-aminocyclophosphamides 15 (R = Z) and 16 (R = H).

Y. Jiang, L. Hu / Bioorg. Med. Chem. Lett. 17 (2007) 517–521
519
Scheme 2. Synthesis of diastereomers of H-Phe-NH-CPA (16) from D- or L-homoserine. Reagents and conditions: (i) Z-Phe-OSu, 1 M KHCO3,
THF, rt; (ii) HOBt, EDC, THF, rt, then satd NH3 (aq); (iii) TBDPS-Cl, imidazole, DMF, rt; (iv) BTI, CH3CN/H2O = 1:1, rt; (v) TBAF, THF, rt; (vi)
Cl2PON(CH2CH2Cl)2, TEA, THF, 0 ꢁC, rt; (vii) H2, 10% Pd/C, MeOH.
ramidic dichloride gave the cis and trans diastereomers
of 15 that were easily separated by silica gel flash column
chromatography, yielding a faster eluting cis diastereo-
mer in 16–28% yield and a slower eluting trans diaste-
reomer in 12–16% yield. The overall cyclization yield
was apparently affected by significant degradation of
14 to Z-Phe-NH2 during the lengthy reaction time
(48 h) in addition to the low nucleophilicity of gem-
diamines and the competing intermolecular cyclization,
which have been observed in the synthesis of other sim-
ilar cyclophosphamide analogs.29,30 The cis- and trans-
configurations of 15 were conveniently assigned accord-
ing to their chromatographic behavior, 1H and 31P
chemical shifts as shown in Table 1.31,32 It should be
noted that TLC on silica gel was used to help assign
the cis/trans-diasteromers as the normal phase TLC is
more reliable than reversed-phase HPLC in predicting
relative polarity. Compared to the trans-diastereomer,
the cis-isomer elutes more quickly on silica gel, has a
more upfield C-4 proton signal, a more downfield C-5
proton signal, and a more upfield phosphorus signal.
These differences between the cis- and trans-isomers of
15 are similar to those observed among the isomers of
4-nitrophenyl-cyclophosphoramide analogs.18 The Z
group in each isomer of 15 was removed by hydrogenol-
ysis to give the corresponding unprotected isomer of 16.
All four diastereomers of 16 were incubated in, pH 7.4,
phosphate buffer (100 mM) at 37 ꢁC. HPLC analysis of
the incubation mixtures showed no significant changes
over a period of 3 days (<10%), suggesting that the ami-
no acid- or peptide-conjugated 4-NH2-CPA is stable un-
der these conditions.
The diastereomers of 15 and 16 were evaluated as sub-
strates of a-chymotrypsin by incubating each compound
(0.2 mM) in 100 mM phosphate buffer, pH 7.4, at 37 ꢁC.
Na-Benzoyl tyrosine ethyl ester (BTEE) was used as a
positive control of the activity of a-chymotrypsin. After
the reaction was initiated by the addition of a-chymo-
trypsin, aliquots were withdrawn at various time inter-
vals, quenched with acetonitrile, and analyzed by
HPLC. The N-terminal capping group was found to
be required for the proteolysis of these conjugates as
none of compound 16 isomers were cleaved in the pres-
ence of a-chymotrypsin over an incubation time of 12 h.
Table 1. Analytical data of diastereomers of Z-Phe-NH-CPA (15) and H-Phe-NH-CPA (16)
Compound
Rfa/trb
MS
NMR d (ppm)
1H (C-5, 2H)
1H (C-4, 1H)
31Pe
cis-(4R)-15
trans-(4R)-15
cis-(4S)-15
0.35a
0.25a
0.45a
0.25a
15.6b
16.2b
16.8b
16.1b
557.3c
5.37–5.24
5.40–5.35
5.34–5.12
5.35–5.27
5.37–5.22
5.43–5.35
5.33–5.25
5.54–5.50
2.20–1.85
2.05–1.86
2.05–1.52
1.84–1.60
2.27–1.85
2.09–1.83
2.31–1.44
2.16–1.77
11.4
13.5
11.4
13.4
11.0
13.7
11.6
13.5
557.3c
557.3c
557.3c
423.1123d
423.1135d
423.1140d
423.1135d
trans-(4S)-15
cis-(4R)-16
trans-(4R)-16
cis-(4S)-16
trans-(4S)-16
a TLC (on silica gel) developing solvents: CH2Cl2/MeOH = 20:1.
b HPLC retention time (min) on C18 column (5 lm, 4.6 · 250 mm) with a gradient elution of 4–76% methanol in 15 min.
c MH+ recorded on LC–MS (ESI, isotopic peaks were not listed).
d HRMS (FAB), m/z calcd for C16H26Cl2N4O3P [MH]+ 423.1120.
e The 31P chemical shift was determined by using 5% H3PO4 in D2O as an internal standard.

520
Y. Jiang, L. Hu / Bioorg. Med. Chem. Lett. 17 (2007) 517–521
The absence of an N-terminal capping group not only
decreases the affinity of the phenylalanine conjugate to
a-chymotrypsin, but generates a positive charge close
to the cleavage site, which may have resulted in the loss
of substrate activity.33 This was confirmed by the incu-
bation of phenylalanine-conjugated homoserine amide
with a-chymotrypsin, which was not cleaved without
an N-terminal capping group.
both showing half-lives greater than 12 h. The proteoly-
sis of cis-(4R)-15 was also compared to that of the
dipeptide Z-Phe-L-Hse-NH2 ((S)-11) in Figure 3. cis-
(4R)-15 was hydrolyzed by the enzyme 1.6· more quick-
ly than the dipeptide (S)-11, which has a half-life of
32 min, indicating that the C-terminal 4-NH2-CPA ren-
dered the conjugate an even better substrate to the en-
zyme. When the reaction mixtures of cis-(4R)-15 with
a-chymotrypsin were analyzed by LC–MS, Z-Phe-OH
was clearly identified, but the released 4-NH2-CPA (7)
was not observed. Instead, we have observed com-
pounds corresponding to 4-hydroxycyclophosphamide,
phosphoramide mustard, and derivatives with their dis-
tinct isotopic patterns in MS (data not shown). This sug-
gests that compound 7 was not stable under the assay
conditions and may have disintegrated according to
our proposed mechanism shown in Scheme 1.
All four isomers of 15 with a Na-Z group were cleaved
by a-chymotrypsin under near physiological conditions
with the concurrent release of Z-Phe-OH. As shown in
Figure 2, a significant difference of substrate activity ex-
ists between the diastereomers. The cis-(4R)-15 was the
best substrate of a-chymotrypsin with a half-life of
20 min. It was hydrolyzed by the enzyme about 27·
more quickly than trans-(4R)-15, which has a half-life
of 8.9 h. The two (4S)-15 isomers were hydrolyzed much
more slowly than their corresponding (4R)-isomers,
In summary, 4-aminocyclophosphamide (4-NH2-CPA,
7) was shown for the first time as a prodrug moiety
of phosphoramide mustard. The four diastereomers
of phenylalanine-conjugated 4-NH2-CPA were synthe-
sized stereospecifically from the D- and L-homoserine
and were found to be stable under physiological con-
ditions. Proteolysis of the Z-Phe conjugates by a-chy-
motrypsin was dependent upon their stereochemical
configurations. The cis-(4R) isomer was most efficient-
ly cleaved by a-chymotrypsin with a half-life of
20 min, significantly shorter than the 8.9 to >12 h
half-lives of the other three isomers. LC–MS analysis
of the proteolytic products of cis-(4R)-15 indicated
that 4-NH2-CPA was released, but disintegrated into
phosphoramide mustard through elimination of
ammonia and conversion to 4-hydroxycyclophospha-
mide. These results suggest that peptide-conjugated
4-aminocyclophosphamides could potentially be used
as prodrugs for proteolytic activation to improve the
therapeutic effectiveness of cyclophosphamide in the
treatment of cancer.
1.0
0.8
0.6
0.4
0.2
0.0
0
2
4
6
8
10
12
Time (hour)
Figure 2. a-Chymotrypsin digestion of compound cis-(4R)-15 (d,
t1/2 = 20 min), trans-(4R)-15 (s, t1/2 = 8.9 h), cis-(4S)-15 (.,
t1/2 P 12 h), and trans-(4S)-15 (,, t1/2 P 12 h). Shown is the
disappearance of the substrates.
Acknowledgments
We gratefully acknowledge the financial support of
Grant SNJ-CCR 700-009 from the State of New Jersey
Commission on Cancer Research, a pilot grant from the
Gallo Prostate Cancer Center of the Cancer Institute of
New Jersey, and Grant RSG-03-004-01-CDD from the
American Cancer Society.
1.0
0.8
0.6
0.4
0.2
0.0
References and notes
1. Eltahir, A.; Heys, S. D.; Hutcheon, A. W.; Sarkar, T. K.;
Smith, I.; Walker, L. G.; Ah-See, A. K.; Eremin, O. Am. J.
Surg. 1998, 175, 127.
2. Young, S. D.; Hill, R. P. J. Natl. Cancer Inst. 1990, 82,
371.
3. Sinhababu, A. K.; Thakker, D. R. Adv. Drug Delivery
Rev. 1996, 19, 241.
4. Dubowchik, G. M.; Walker, M. A. Pharmacol. Ther. 1999,
83, 67.
5. Hu, L. IDrugs 2004, 7, 736.
0
20
40
60
80
100
120
140
160
Time (min)
Figure 3. a-Chymotrypsin digestion of compounds cis-(4R)-15 (d and
s, t1/2 = 20 min) and Z-Phe-L-Hse-NH2 ((S)-11, and ,,
t1/2 = 32 min). Shown are the disappearance of the substrates (closed
.
symbols) and appearance of Z-Phe-OH (open symbols).

Y. Jiang, L. Hu / Bioorg. Med. Chem. Lett. 17 (2007) 517–521
521
6. Reuning, U.; Magdolen, V.; Wilhelm, O.; Fischer, K.;
Lutz, V.; Graeff, H.; Schmitt, M. Int. J. Oncol. 1998, 13,
893.
19. Hu, L.; Yu, C.; Jiang, Y.; Han, J.; Li, Z.; Browne, P.;
Race, P. R.; Knox, R. J.; Searle, P. F.; Hyde, E. I. J. Med.
Chem. 2003, 46, 4818.
7. de Groot, F. M.; van Berkom, L. W.; Scheeren, H. W. J.
Med. Chem. 2000, 43, 3093.
8. Eisenbrand, G.; Lauck-Birkel, S.; Tang, W. C. Synthesis
1996, 1246.
9. Birkedal-Hansen, H.; Moore, W. G.; Bodden, M.
K.; Windsor, L. J.; Birkedal-Hansen, B.; DeCarlo,
A.; Engler, J. A. Crit. Rev. Oral. Biol. Med. 1993,
4, 197.
10. Ray, J. M.; Stetler-Stevenson, W. G. Eur. Respir. J. 1994,
7, 2062.
20. Borch, R. F.; Millard, J. A. J. Med. Chem. 1987, 30, 427.
21. Kwon, C. H. Arch. Pharm. Res. 1999, 22, 533.
22. Cox, P. J. Biochem. Pharmacol. 1979, 28, 2045.
23. Loudon, G. M.; Almond, M. R.; Jacob, J. N. J. Am.
Chem. Soc. 1981, 103, 4508.
24. Fletcher, M. D.; Campbell, M. M. Chem. Rev. 1998, 98,
763.
25. Manolopoulou, A. P. C.; Tsegenidis, T. Eur. J. Med.
Chem. 1992, 27, 949.
26. Loudon, G. M.; Radhakrishna, A. S.; Almond, M. R.;
Blodgett, J. K.; Boutin, R. H. J. Org. Chem. 1984, 49,
4272.
11. Denmeade, S. R.; Nagy, A.; Schally, A.; Issacs, J. T. Proc.
Am. Assoc. Cancer Res. 1998, 39, 3765.
12. Coombs, G. S.; Bergstrom, R. C.; Pellequer, J. L.; Baker,
S. I.; Navre, M.; Smith, M. M.; Tainer, J. A.; Madison, E.
L.; Corey, D. R. Chem. Biol. 1998, 5, 475.
13. Colvin, O. M. Curr. Pharm. Des. 1999, 5, 555.
14. Borch, R. F.; Canute, G. W. J. Med. Chem. 1991, 34,
3044.
27. Pallai, P.; Goodman, M. J. Chem. Soc., Chem. Commun.
1982, 280.
28. Yu, C.; Jiang, Y.; Liu, B.; Hu, L. Tetrahedron Lett. 2001,
42, 1449.
29. Walsh, S. P.; Shulman-Roskes, E. M. J. Labelled Compd.
Radiopharm. 1995, 36, 1193.
15. Li, Z.; Han, J.; Jiang, Y.; Browne, P.; Knox, R. J.; Hu, L.
Bioorg. Med. Chem. 2003, 11, 4171.
30. Martinek, T.; Forro, E.; Gunther, G.; Sillanpaa, R.;
Fulop, F. J. Org. Chem. 2000, 65, 316.
16. Foster, E. L. J. Pharm. Sci. 1978, 67, 709.
17. Hernick, M.; Borch, R. F. J. Med. Chem. 2003, 46, 148.
18. Jiang, Y.; Han, J.; Yu, C.; Vass, S. O.; Searle, P. F.;
Browne, P.; Knox, R. J.; Hu, L. J. Med. Chem. 2006, 49,
4333.
31. Shih, Y. E.; Wang, J. S.; Chen, C. T. Heterocycles 1978, 9,
1277.
32. Boyd, V. L.; Zon, G.; Himes, V. L.; Stalick, J. K.; Mighell,
A. D.; Secor, H. V. J. Med. Chem. 1980, 23, 372.
33. Neurath, H.; Schwert, G. W. Chem. Rev. 1950, 46, 69.