Phosphate ester serum albumin affinity tags greatly improve peptide half-life in vivo

Article abstract of DOI:10.1016/S0960-894X(03)00209-9

A series of phosphate ester based small molecules designed to bind tightly to serum albumin were applied to the amino terminus of an anticoagulant peptide in an effort to increase its protein binding in vivo. The tagged peptides exhibited high affinity for both rabbit and human serum albumin when passed through liquid chromatographic columns with serum albumins incorporated into the stationary phase. The peptides were then administered intravenously to rabbits and found to have a greater than 50-fold increase in plasma half life. The highest affinity peptides showed a reduction in bioactivity consistent with their sequestration away from their protein target in the presence of 0.1% rabbit serum albumin.

Full text of DOI:10.1016/S0960-894X(03)00209-9

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1513–1515
Phosphate Ester Serum Albumin Affinity Tags Greatly
Improve Peptide Half-Life In Vivo
Kerry Zobel,a Michael F. T. Koehler,a,b,* Maureen H. Beresini,c Lisa D. Carisc and
Daniel Combsd
aDepartment of Protein Engineering, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA
bDepartment of Bioorganic Chemistry, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA
cDepartment of BioAnalytical Research and Development, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA
dDepartment of Clinical and Experimental Pharmacology, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA
Received 10 January 2003; revised 22 February 2003; accepted 27 February 2003
Abstract—A series of phosphate ester based small molecules designed to bind tightly to serum albumin were applied to the amino
terminus of an anticoagulant peptide in an effort to increase its protein binding in vivo. The tagged peptides exhibited high affinity
for both rabbit and human serum albumin when passed through liquid chromatographic columns with serum albumins incorpo-
rated into the stationary phase. The peptides were then administered intravenously to rabbits and found to have a greater than
50-fold increase in plasma half life. The highest affinity peptides showed a reduction in bioactivity consistent with their sequestra-
tion away from their protein target in the presence of 0.1% rabbit serum albumin.
# 2003 Elsevier Science Ltd. All rights reserved.
Phage display offers a rapid and convenient technology
for identifying and optimizing peptides which bind to a
target protein.1 A number of these peptides have been
shown to have potent biological activity,2 and could
potentially serve as therapeutics. There are still sig-
nificant obstacles to oral delivery of peptides, including
their rapid proteolysis and poor permeability in the
gut.3 These issues have been addressed through novel
delivery technologies, and thus close analogues of peptide
hormones such as somatostatin and gonadotropin releas-
ing hormone are used clinically by delivering them
through an osmotic pump or other slow release formula-
tion.4 Even if the problems of peptide delivery can be
solved, their rapid clearance once in circulation remains
an unaddressed obstacle to their general therapeutic use.
that this approach had merit, but it was unclear whether
the maximum effect had been achieved. We therefore set
out to design tags with higher affinity for serum albumin
to determine whether we could further reduce their clear-
ance and increase their circulating half life.
In the course of our initial work, we noted that serum
albumin binding has also been investigated in the
development of magnetic resonance imaging (MRI)
contrast agents. Contrast agents are employed during
MRI procedures as a tool to enhance the magnetic sus-
ceptibility of water molecules in the body, and their
effectiveness is increased through association with a
macromolecule.6 A rationally designed contrast agent,
MS-325 (Fig. 1), has been developed which binds to
human serum albumin (HSA).7 This molecule is
bifunctional, with the diethylenetriaminepentaacetate
(DTPA) portion of the molecule serving to bind the
gadolinium ion while the diphenylcyclohexanol phos-
phate ester moiety binds to HSA. We envisioned incor-
porating this and similar phosphate esters into albumin
affinity tags to determine if they could further extend
the in vivo half-life of peptide 1c.
Recently, we reported an increase to peptide half-life in
vivo by attaching serum albumin binding tags to the
amino terminus of an anticoagulant peptide 1a.5
Appending a napthalene acylsulfonamide tag to give
peptide 1b increased its half-life nearly four-fold relative
to peptide 1a (Fig. 1). From this initial success, it was clear
As we had done previously, we designed our tags to be
coupled to the amino terminus of peptide 1 through an
*Corresponding author. Tel.: +1-650-225-8135; fax: +1-640-225-
3734; e-mail: mkoehler@gene.com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00209-9

1514
K. Zobel et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1513–1515
Following their synthesis, peptides 5a–c were examined
in a coagulation factor VIIa (fVIIa) ELISA binding
assay5 to determine how the albumin affinity tags affec-
ted their ability to bind their protein target (Table 1).
All the peptides examined bound with similar affinity to
fVIIa in the presence of 1% ovalbumin or 0.1% HSA as
a bulking agent. However, when the assay was done in
the presence of 0.1% rabbit serum albumin (RSA),
activity was reduced for peptides 1b, 5a and 5c, indicat-
ing that RSA binding might be reducing the ability of
the tagged peptides to bind fVIIa by sequestering the
bound peptide from fVIIa. This effect is commonly
observed in small molecules which bind to serum pro-
teins, and indicates that compound 5a may bind more
tightly to RSA than our previously synthesized com-
pounds.
Figure 1. Previously described anticoagulant peptides 1a–c and serum
albumin binding MRI contrast reagent MS-325.
As had been done previously for compounds 1a and 1b,
compounds 5a–c were evaluated for their ability to bind
to both HSA and RSA using an affinity chromato-
graphy assay (Table 1). HSA and RSA were immobi-
lized on sepharose gel columns,10 and 20 mL of a 0.1
mM solution of the compounds injected while the UV
absorbance at 295 nm was monitored to determine the
retention time of each. The mobile phase consisted of 20
mM phosphate buffer at pH 7.2, 150mM NaCl, 2 mM
b-cyclodextrin11 and 0.0015% ProClin 300 as a pre-
servative. Compounds 5a–c were all strongly retained
on the columns, as expressed in their k0/(k0+1) values
near the theoretical maximum of one. Conjugate 5a
binds less well to HSA than 5b and 5c which were simi-
lar to 1b. Compounds 5a–c were indistinguishable from
1b on the RSA column, indicating that the sensitivity
limit of this approach may have been reached.
amide linkage. In this case, a six carbon linker was used
to couple the peptide and the phosphate ester portions
of the molecule (Scheme 1). Ethyl 6-hydroxyhexanoate
was coupled with 2-cyanoethyl N,N-diisopropyl-chloro-
phosphoramidite,8 giving phosphoramidite 2. Three
different alcohols, selected for their demonstrated affi-
nity for albumin as well as similarity to the previously
prepared 1b were then coupled and oxidized to give
phosphodiesters 3a–c. Subsequently, the cyanoethyl
protecting group was removed using ammonia in
methanol and the esters cleaved to the free acids 4a–c.
Each acid was then coupled to the amino terminus of
peptide 1 on resin with PyBOP and DIPEA.9 The peptides
were cleaved from the resin using standard TFA condi-
tions, then oxidized with I2 in acetic acid to form the
intramolecular disulfide bond. Purification using reverse
phase high performance liquid chromatography (HPLC)
gave compounds 5a–c which each eluted as a single peak
on an analytical HPLC column and exhibited the desired
Peptides 5a–c were subsequently evaluated along with
the previously described peptides 1a and 1b in vivo.
According to the previously described procedure,5 the
peptides were administered at 2 mg/kg intravenously to
male New Zealand White rabbits. Citrated blood sam-
ples were drawn prior to dosing and then at regular
intervals thereafter. Following sample workup, the
plasma concentration of each peptide was evaluated by
comparing the sample activity in the fVIIa binding
assay to that of known standards, which were handled
identically. In assaying both the samples and the stan-
dard, 1% bovine serum albumin was used as the bulk-
ing agent, which did not significantly affect the IC50
values of the peptides (data not shown). The con-
centrations obtained were used to compute pharmaco-
kinetic parameters using a non-compartmental model
(Table 2). The newly designed peptides had sharply
reduced clearance rates and significantly increased half-
lives in vivo (Fig. 2). The longest acting peptide, 5a, had
a half-life of 222 min, 7.4-fold longer than 1b, the long-
est acting peptide in our previous work and 53.1-fold
longer than the unmodified peptide 1a. Peptides 5b and
5c showed improvements of 8.3-fold and 18.9-fold
respectively, over 1a.
M
+ ion by electrospray mass spectrometry.
Scheme 1. (a) 2-cyanoethyl N,N-diisopropylchlorophosphoramidite,
DIPEA, CH2Cl2; (b) ROH, 1-H tetrazole, t-butylhydroperoxide,
CH3CN; (c) 2M NH3 in MeOH; (d) LiOH, THF, H2O; (e) peptide 1c
unoxidized and with protected sidechains on resin, DIPEA, PyBOP,
DMA, CH2Cl2; (f) 2.5% TIPSH, 2.5% H2O in TFA; (g) I2/HOAc.
From these results, it is clear that optimizing the albu-
min binding properties of our affinity tags has resulted
in a dramatically reduced clearance rate and a com-

K. Zobel et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1513–1515
1515
Table 1. Affinities of tagged peptides for fVIIa in the presence of different bulking proteins, and albumin column binding data
Compd
IC50, (nM) 1% ovalbumin
IC50, (nM) 0.1% HSA
HSA k0/(k0+1)
IC50, (nM) 0.1% RSA
RSA k0/(k0+1)
1a
1b
5a
5b
5c
9.2
3.6
4.0
1.9
2.0
11
0.21
0.97
0.78
0.92
0.92
9.7
24
90
9.2
22
0.39
0.98
0.97
0.97
0.98
2.5
3.4
2.0
2.4
k0=(trꢀt0)/t0, with tr being the retention time of the compound and t0 the retention time of DMSO.
fold increase in half life over the unmodified peptide 1a,
and its 222 min half life allows for animal studies with a
managable dosing schedule as well as a reduction in
both the amount of peptide required and the variation
in serum peptide levels between doses. Additional
experimentation with both tether composition and
length is underway and may allow us to access these
improvements in reducing clearance while maintaining
the potent bioactivity of the peptides.
Table 2. PK parameters calculated from the concentrations of pep-
tide over time in rabbit blood samples
Compd
Clearance (mL/min/kg)
t1/2 (min)
Dose (mg/kg)
1.11
n
1a
1b
5a
5b
5c
26.8
4.18
302.0
222
1
1.64
0.185
0.707
0.505
2.0
2.0
2.0
2
2
2
34.8
79.1
Acknowledgements
The authors would like to thank Robert Lazarus and
Simon Williams for helpful discussions.
References and Notes
1. Sidhu, S. A. Curr. Opin. Biotechnol. 2000, 11, 610.
2. (a) Wrighton, N. C.; Farrell, F. X.; Chang, R.; Kashyap,
A. K.; Barbone, F. P.; Mulcahy, L. S.; Johnson, D. L.; Bar-
rett, R. W.; Jolliffe, L. K.; Dower, W. J. Science 1996, 273,
458. (b) Dennis, M. S.; Eigenbrot, C.; Skelton, N. J.; Ultsch,
M. H.; Santell, L.; Dwyer, M. A.; O’Connell, M. P.; Lazarus,
R. A. Nature 2000, 404, 465.
3. Sood, A.; Panchagnula, R. Chem. Rev. 2001, 101, 3275.
4. (a) Kutscher, B.; Bernd, M.; Beckers, T.; Polymeropoulos,
E. E.; Engel, J. Angew. Chem., Int. Ed. Engl. 1997, 36, 2148.
(b) Heron, I.; Thomas, F.; Dero, M.; Gancel, A.; Ruiz, J. M.;
Schatz, B.; Kuhn, J. M. J. Clin. Endocrinol. Metab. 1993, 76,
721. (c) Lancranjan, I.; Bruns, C.; Grass, P.; Jaquet, P.; Jer-
vell, J.; Kendall-Taylor, P.; Lamberts, S. W. J.; Marbach, P.;
Ørskov, H.; Pagani, G.; Sheppart, M.; Simionescu, L. Meta-
bolism 1996, 45, 67.
Figure 2. Average concentration of peptides 1a, 1b and 5a–c in rabbit
plasma samples.
5. Koehler, M. F. T.; Zobel, K.; Beresini, M. H.; Caris, L. D.;
Combs, D.; Paasch, B. D.; Lazarus, R. A. Bioorg. Med. Chem.
Lett. 2002, 12, 2883.
6. Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B.
Chem. Rev. 1999, 99, 2293.
mensurately longer half life. Conjugate 5a is especially
effective in this regard with a half life in circulation of
nearly 4 h. However, as the affinity of the tags for serum
albumin has increased, a reduction in their ability to
bind fVIIa in the presence of RSA has also been
observed. This indicates that the peptide is less active or
inactive when bound to RSA, despite the inclusion of a
flexible six carbon linker between the albumin binding
phosphate ester and the bioactive peptide.
7. (a) Lauffer, R. B.; Parmelee, D. J.; Dunham, S. U.; Ouellet,
H. S.; Dolan, R. P.; Witte, S.; McMurry, T. J.; Walovitch,
R. C. Radiology 1998, 207, 529. (b) Lauffer, R. B.; Parmelee,
D. J.; Ouellet, H. S.; Dolan, R. P.; Sajiki, H.; Scott, D. M.;
Bernard, P. J.; Buchanan, E. M.; Ong, K. Y.; Tyeklar, Z.;
Midelfort, K. S.; McMurry, T. J.; Walovitch, R. C. Acad.
Radiol. 1996, 3, S356.
8. Sinha, N. D.; Biernat, J.; Koster, H. Tetrahedron Lett.
1983, 24, 5843.
9. Coste, J.; Le-Nguyen, D.; Castro, B. Tetrahedron Lett.
1990, 31, 205.
10. Massolini, G.; Aubry, A.-F.; McGann, A.; Wainer, I. W.
Biochem. Pharm 1993, 46, 1285.
11. Hage, D. S.; Sengupta, A. J. Chromatogr. B, Biomed. Sci.
Appl 1999, 724, 91.
Despite this potential drawback, the general approach
of reducing the clearance rate of peptides in vivo by
appending tags to them which confer affinity for serum
albumin has been further validated through these
experiments. Conjugate 5a exhibits a greater than 50-