PHARMACOKINETICS, PHARMACODYNAMICS AND
DRUG METABOLISM
Elucidation of the Human Serum Albumin (HSA) Binding
Site for the Cu-PTSM and Cu-ATSM Radiopharmaceuticals
NATHAN E. BASKEN, CARLA J. MATHIAS, MARK A. GREEN1
Division of Nuclear Pharmacy, Department of Industrial and Physical Pharmacy, Purdue University, 575 Stadium Mall Drive,
West Lafayette, Indiana
Received 10 May 2008; revised 1 August 2008; accepted 8 August 2008
Published online 20 October 2008 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.21570
ABSTRACT: The Cu-PTSM (pyruvaldehyde bis(N4-methylthiosemicarbazonato)copper-
(II)) and Cu-ATSM (diacetyl bis(N4-methylthiosemicarbazonato)copper(II)) radiophar-
maceuticals exhibit strong, species-dependent binding to human serum albumin (HSA),
while Cu-ETS (ethylglyoxal bis(thiosemicarbazonato)copper(II)) appears to only exhibit
nonspecific binding to human and animal serum albumins. This study examines the
structural basis for HSA binding of Cu-PTSM and Cu-ATSM via competition with drugs
having known albumin binding sites. Warfarin, furosemide, ibuprofen, phenylbutazone,
benzylpenicillin, and cephmandole were added to HSA solutions at drug:HSA mole ratios
from 0 to 8:1, followed by quantification of radiopharmaceutical binding to HSA by
ultrafiltration. Warfarin, a site IIA drug, progressively displaced both [64Cu]Cu-PTSM
and [64Cu]Cu-ATSM from HSA. At 8:1 warfarin:HSA mole ratios, free [64Cu]Cu-PTSM
and [64Cu]Cu-ATSM levels increased 300--500%. This was in contrast to solutions con-
taining ibuprofen, a site IIIA drug; no increase in free [64Cu]Cu-PTSM or [64Cu]Cu-ATSM
was observed except at high ibuprofen:HSA ratios, where secondary ibuprofen binding to
the IIA site may cause modest radiopharmaceutical displacement. By contrast, and
consistent with earlier findings suggesting Cu-ETS exhibits only nonspecific
associations, [64Cu]Cu-ETS binding to HSA was unaffected by the addition of drugs that
bind in either site. We conclude that the species-dependence of Cu-PTSM and Cu-ATSM
albumin binding arises from interaction(s) with the IIA site of HSA. ß 2008 Wiley-Liss, Inc.
and the American Pharmacists Association J Pharm Sci 98:2170 2179, 2009
Keywords: protein binding; albumin; distribution; PET; preclinical pharmacokinetics
INTRODUCTION part of the drug development process, and the FDA
requires that plasma protein binding be report-
Plasma protein binding can have a profound effect ed (http://www.fda.gov/OHRMS/DOCKETS/98fr/
on drug distribution and pharmacokinetics by 00n-1269-nfr0001-03.pdf). While simple documen-
slowing or preventing passive extravasation into tation of the extent of plasma protein binding has
tissues.1,2 Protein binding must be evaluated as utility in assessment of clinical drug delivery phe-
nomenon, knowledge of the structural details of
Correspondence to: Mark A. Green (Telephone: [phone];
small molecule-albumin interactions are some-
Fax: [phone]; E-mail: [e-mail])
what limited for therapeutic drugs, and very scarce
Journal of Pharmaceutical Sciences, Vol. 98, 2170 2179 (2009)
for radiopharmaceuticals.
ß 2008 Wiley-Liss, Inc. and the American Pharmacists Association
2170 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009
Cu-PTSM AND Cu-ATSM DISPLACEMENT FROM HSA 2171
Human serum albumin (HSA) is highly concen- bind in the IIA site of HSA with affinity constants
trated in plasma (46 mg/mL),3 and responsible for ranging from 1.2 103 M 1 (benzylpenicillin) to
the majority of small molecule binding in plasma. 7.0 105 M 1 (phenylbutazone) (Fig. 2).9 Ibupro-
HSA (MW 66,438) has a flexible three domain fen has a similar affinity for the IIIA site
structure with two high-affinity, reversible, drug (Ka ź 2.7 106 M 1) (Fig. 2).9 The other competing
binding sites on subdomains IIA and IIIA.4,5 Hun- agents examined include L-tryptophan, which has
dreds of drugs and endogenous compounds have low affinity (Ka ź 1 104 M 1) for the IIIA site but
been formally classified according to binding in can competitively displace fluorescent probes from
either high affinity site.6--9 High-affinity, limited- that location,28--30 and the beta lactam antibiotic,
capacity, drug binding can occur in one or both of cephmandole, which has been shown to disrupt
these locations, in addition to more widespread bilirubin binding to HSA.31,32 Finally, since one
nonspecific, high-capacity, small molecule-albumin low affinity fatty acid binding site directly overlaps
interactions. the HSA IIA drug binding site,33,34 competitive
Cu-PTSM and Cu-ETS (Fig. 1) have shown binding assays were performed to probe [64Cu]Cu-
promise as perfusion tracers for positron emission PTSM binding to HSA in the presence of four
tomography (PET),10--20 and are suitable for label- saturated and unsaturated fatty acids with chain
ing with any of the positron-emitting copper lengths varying from 10 to 18 carbons.
61 62 64
isotopes (60Cu, Cu, Cu, Cu). The structurally
related Cu-ATSM radiopharmaceutical (Fig. 1)
MATERIALS AND METHODS
has shown promise for use in PET imaging to
assess tissue hypoxia.21--25 Cu-PTSM and Cu-
General
ATSM extensively bind HSA (approx. 95% bound
in 40 mg/mL HSA solution), while Cu-ETS binding Copper-64 was obtained as no-carrier-added
64
to HSA appears limited to nonspecific interactions Cu2þ in dilute HCl from the Radionuclide Re-
(approx. 60% bound in 40 mg/mL HSA source for Cancer Applications at Washington Uni-
solution).26,27 The extensive, and species-specific, versity via Isotrace, Inc. (St. Louis, MO). The
binding of Cu-PTSM to HSA is problematic in [64Cu]Cu-PTSM, [64Cu]Cu-ATSM, and [64Cu]Cu-
translation of animal results to humans, as it ETS radiopharmaceuticals were prepared as de-
seemingly limits radiotracer diffusion into tissues scribed previously, and purified to remove traces of
64
exhibiting high rates of perfusion.12,15--17 ionic Cu using C18 SepPak Light (Waters, Mil-
To probe the structural basis for strong HSA ford, MA) solid-phase extraction cartridges.26 Fol-
binding of Cu-PTSM and Cu-ATSM, the present lowing recovery in EtOH, the radiochemical purity
study examined competitive displacement of of each chelate exceeded 99% (silica gel TLC devel-
[64Cu]Cu-PTSM, [64Cu]Cu-ATSM, and [64Cu]Cu- oped with ethanol). Lyophilized human and canine
ETS from HSA by known site-selective albumin- serum albumins were purchased from Sigma
binding drugs. Furosemide, warfarin, phenylbuta- Chemical Company (St. Louis, MO), and reconsti-
zone, and benzylpenicillin (penicillin G) have di- tuted in 0.9% normal saline at 0.66 mM (44 mg/
verse chemical structures (Fig. 2), but commonly mL). The canine serum albumin was a fraction V
Figure 1. Structural formulae of the square-planar copper(II) bis(thiosemicarbazone)
chelates: Cu-PTSM (pyruvaldehyde bis(N4-methylthiosemicarbazonato)copper(II)), Cu-
ATSM (diacetyl bis(N4-methylthiosemicarbazonato)copper(II)), and Cu-ETS (ethylglyoxal
bis(thiosemicarbazonato)copper(II)) radiopharmaceuticals.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009
2172 BASKEN, MATHIAS, AND GREEN
Figure 2. Chemical structures, and affinity constants for HSA binding sites, of ligands
used in competitive binding assays. The binding constants come from references 10, 29,
and 32.
powder, and the HSA was essentially globulin-free, Ultrafiltration Binding Assays
and at least 99.5% fatty acid free. Rat serum was
Stock solutions of each drug were first prepared by
purchased from Sigma and stored frozen until use.
dissolving the drug or drug salts in 0.9% normal
Ibuprofen, cephmandole, and benzylpenicillin
saline solution at 53 mM, with the exception of the
were also purchased from Sigma. Ibuprofen and
cephmandole were sodium salts, with cephman- cephmandole stock solution which was prepared at
26.5 mM due to limited aqueous solubility. Furo-
dole having a vendor reported potency of 921 mg/
semide and phenylbutazone solutions were
mg. The lot of benzylpenicillin (potassium salt)
similarly diluted with saline to 53 mM. The
purchased had a potency of 1596 units per mg
benzylpenicillin and cephmandole stock solutions
reported by the vendor. Phenylbutazone was
were adjusted for potency, so the final drug con-
obtained as PhenylbuteÒ, an injectible 20% solution
centrations as tested were 53 mM. Aliquots of the
for equine analgesia from Phoenix Pharmaceuticals
53 mM stock solutions were serially diluted to
(St. Joseph, MO), containing 10.45 mg/mL benzyl
alcohol as a preservative. Preservative-free furo- produce additional solutions at 26.5, 13.3, 6.7, and
3.3 mM solutions of each drug at pH 7. Then, a 100
semide was obtained as a 5% solution for general
mL aliquot of each drug solution was added to
veterinary administration. Phenylbutazone and
1.0 mL of each serum albumin solution, diluting
furosemide solutions were stored at 2--8 C until
the albumin solutions as tested to 40 mg/mL (at
use. Racemic warfarin sodium was purchased from
corresponding drug:HSA mole ratios of 8:1, 4:1,
Spectrum Chemical Corporation (Gardena, CA) as
2:1, 1:1, and 0.5:1). 100 mL of saline was also added
a crystalline solid.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009 DOI 10.1002/jps
Cu-PTSM AND Cu-ATSM DISPLACEMENT FROM HSA 2173
to a 1.0 mL aliquot of each albumin solution as a of ethanol on HSA binding. Solutions of HSA þ
drug-free control. Then, 2--8 mCi (0.074 to 0.30 [64Cu]Cu-PTSM were incubated at 49 C for
MBq) of [64Cu]Cu-PTSM, [64Cu]Cu-ATSM, or 30 min prior to centrifugation, then assayed by
[64Cu]Cu-ETS was added in a volume of 2--10 mL ultrafiltration to examine any effects of tempera-
to each drug-serum albumin solution and vortex ture on binding. The increased incubation temper-
mixed. In a typical preparation, 300 mL aliquots of ature did not affect the binding of [64Cu]Cu-PTSM.
each saline or protein solution were added to the Unbound [64Cu]Cu-PTSM values in room temper-
ultrafiltration devices in triplicate. The Amicon ature control solutions were 4.2  0.6% (n ź 5) ver-
Centrifree (Millipore, Bedford, MA) ultrafiltration sus 3.8  0.2% (n ź 5) for the test solutions
devices with 30,000-Da NMWL methylcellulose incubated at 49 C. To determine if a 2.5% (v/v)
micropartition membranes were centrifuged ethanol solution would degrade the ultrafiltration
immediately (within 5-min of mixing), as reported membrane, 25 mL ethanol was added to 1.0 mL
64
previously,26 and the quantity of Cu in measured (saline þ [64Cu]Cu-PTSM) and subjected to the ul-
volumes of the albumin and ultrafiltrate solutions trafiltration process. The addition of ethanol did
determined using an automatic gamma counter. not appear to compromise the integrity of the
Unbound radiotracer levels (  % free values) were ultrafiltration device s methyl cellulose mem-
then calculated as described, and corrected for brane. Free tracer in the control (saline) group was
nonspecific binding to the ultrafiltration device 67.2  0.7% free (n ź 5) versus the ethanol test
using the data obtained by analysis of the ultrafil- group which was 68.0  2.3% free (n ź 5). Finally,
trate from protein-free saline solutions.26 To de- to examine any effects of ethanol on HSA
termine if the addition of drug altered nonspecific conformation, 25 mL ethanol was added to 1.0 mL
radiopharmaceutical binding to the ultrafiltration (HSA þ [64Cu]Cu-PTSM). The corrected [64Cu]Cu-
device, additional controls were prepared by add- PTSM % free value was 4.5  0.2% free (n ź 5) for
ing a 100 mL of each 53 mM drug solution to 1.0 mL the HSA þ ethanol solution, compared to 4.0  0.1%
saline with each of the three radiotracers; for each free (n ź 8) for [64Cu]Cu-PTSM binding to ethanol-
drug-tracer combination; no such effect was free HSA. HSA solutions containing 2.5% (v/v)
detected (i.e., measured % free values for the ethanol do not appear to affect the integrity of
[64Cu]Cu-L in protein-free drug solution were iden- the ultrafiltration membrane or the binding of
tical to the % free values measured in saline). In [64Cu]Cu-PTSM to HSA. HSA solutions in
these studies, the presence of [64Cu]Cu-L does not this experiment were maintained at 49 C, and
appreciably alter the number of available HSA [64Cu]Cu-PTSM binding to HSA at this tempera-
binding sites. Calculated ([64Cu]Cu-L):HSA mole ture is comparable to binding to HSA at room
ratios based on the maximum specific activity (171 temperature.
mCi/mg) on date of receipt ranged from (6 10 8):1 To determine whether [64Cu]Cu-PTSM could be
on the date of receipt to (1.8 10 6):1 in the most displaced from HSA by L-tryptophan or other ami-
delayed experiments. no acids, competitive binding assays were also
Additional assays probed [64Cu]Cu-PTSM bind- performed using tryptophan and a mixed amino
ing to HSA in the presence of saturated and unsat- acid solution. L-tryptophan and RPMI-1640 50X
urated fatty acids with chain lengths varying from amino acid medium solution for cell culture (an
10 to 18 carbons. Capric acid, myristic acid, pal- amino acid cocktail containing Arg (10 mg/mL),
mitic acid, and oleic acid (Sigma) were dissolved in Asn (2.5 mg/mL), Asp (1 mg/mL), Cys (3.3 mg/mL),
100% ethanol at 49 C as 145 mM and 24 mM Glu (1 mg/mL), Gly (0.5 mg/mL), His (0.75 mg/mL),
solutions. A 0.60 mM (40 mg/mL) HSA solution Ile (2.5 mg/mL), Leu (2.5 mg/mL), Lys (2 mg/mL),
was prepared using saline, and maintained at Met (0.75 mg/mL), Phe (0.75 mg/mL), Pro (1 mg/
49 C. This temperature was required to keep the mL), Ser (1.5 mg/mL), Thr (1 mg/mL), Trp (0.25 mg/
fatty acids in solution, but was well below the mL), Tyr (1.44 mg/mL), Val (1 mg/mL)) were
denaturation temperature of HSA (60 C).35 A purchasedfromSigmaChemical Co.Multipletrypto-
25 mL aliquot of each fatty acid solution in ethanol, phan solutions (63, 16, 8, and 4 mM) were prepared
corresponding to 1:1 and 6:1 (fatty acid):HSA mole in saline, corresponding to 8:1, 2:1, 1:1, and 0.5:1
ratios, and 2 mLof [64Cu]Cu-PTSM, were added to tryptophan:HSA stoichiometry as tested. Then,
1.0 mL of HSA and vortex mixed prior to being 0.075 mL aliquots of each tryptophan solution and
subjected to the ultrafiltration process. the amino acid solution were added to 0.9 mL of
Additional controls were also required to exam- 0.65 mM (43.3 mg/mL) HSA solution containing
ine the effects of temperature and the addition [64Cu]Cu-PTSM. These solutions were briefly
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009
2174 BASKEN, MATHIAS, AND GREEN
Table 1. Effects of Competing Drugs on the Binding of [64Cu]Cu-PTSM to HSA
Mole Ratio
(Drug:HSA) Warfarina Ibuprofenb Benzylpenicillina Cephmandole Furosemidea Phenylbutazonea
64
% Free Cu-PTSM (corrected)  SD (n 3) in the presence of competing drug
0 4.0  0.1 4.0  0.1 4.0  0.1 4.0  0.1 4.0  0.1 4.0  0.1
0.5 4.6  0.1 3.4  0.1 3.8  0.1 3.5  0.1  
1 5.8  0.7 3.4  0.3 3.5  0.1 3.2  0.2 8.4  0.2 7.1  0.1
2 7.7  0.8 3.2  0.5 3.3  0.2 3.5  0.1  
4 14.8  0.3 4.8  0.2 3.4  0.1 4.2  0.2  
8 20.1  1.5 10.1  0.6 3.3  0.1  32.9  1.1 28.7  0.8
a
Binds IIA site.
b
Binds IIIA site.
vortexed and subjected to the ultrafiltration increase in free [64Cu]Cu-PTSM in solution. Fur-
process. ther, this level (approximately 33%) of free tracer
approaches the value of free [64Cu]Cu-PTSM in
canine serum albumin (CSA) solutions
(approximately 42%), believed to indicate only
RESULTS
nonspecific interactions.26
Phenylbutazone has a twofold greater affinity
[64Cu]Cu-L Displacement from HSA by
for HSA than warfarin, (Ka ź 7 105 and 3.4
Competing Drugs
105 M 1, respectively), and more effectively
The binding of [64Cu]Cu-PTSM and [64Cu]Cu- displaced [64Cu]Cu-PTSM. The fractions of un-
ATSM to HSA appear to be selectively disrupted bound tracer in 1:1 and 8:1 phenylbutazone solu-
by the addition of warfarin, phenylbutazone, and tions were 7.1  0.1% and 28.7  0.8%, respectively,
furosemide, drugs known to selectively bind with versus 5.8  0.7% and 20.1  1.5%, in 1:1 and 8:1
high affinity in the IIA site (Tabs. 1 and 2). war farin solutions, respectively. Phenylbutazone
[64Cu]Cu-PTSM and [64Cu]Cu-ATSM are both (log P ź 3.2) is more lipophilic than warfarin
highly bound in drug-free HSA solutions, (log P ź 2.7), and both drugs are considerably more
(4.0  0.1% free, and 5.9  0.4% free, respectively), lipophilic than Cu-PTSM (log P ź 1.9).36,37
and the progressive displacement due to increas- [64Cu]Cu-PTSM displacement from HSA bind-
ing warfarin concentrations is nearly identical ing by site IIA ligands does not appear to be a
for these tracers. At 8:1 warfarin:HSA mole simple function of the individual drug affinities
ratios, % free values for [64Cu]Cu-PTSM and for HSA. Warfarin has more than a 10-fold greater
[64Cu]Cu-ATSM were 20.1  1.5%, and 19.0  affinity for site IIA (Ka ź 3.5 105 M 1) than
1.7%, respectively. The observed displacement of furosemide (Ka ź 2.6 104 M 1), yet even at iden-
[64Cu]Cu-PTSM by furosemide and phenylbuta- tical 1:1 mole ratios, the increase in free tracer due
zone competition at high (8:1) mole ratios is even to furosemide binding (from 4.0% to 8.4% free) is
more dramatic, 32.9  1.1% free, and 28.7  0.8% more dramatic than the increase due to warfarin
free, respectively. Furosemide displacement at an binding (from 4.0% to 5.8% free). The most signifi-
8:1 mole ratio translates into more than a 700% cant drug displacement at high drug concentra-
Table 2. Displacement of [64Cu]Cu-ATSM and [64Cu]Cu-ETS from HSA by Drug Competition
[64Cu]Cu-ATSM [64Cu]Cu-ETS
Mole Ratio (Drug:HSA) Warfarin Ibuprofen Warfarin Ibuprofen
% Free (corrected)  SD (n 3)
0 5.9  0.4 5.9  0.4 40.6  1.1 40.6  1.1
0.5 7.3  0.2 7.1  0.6 42.8  1.6 43.9  2.2
1 9.4  0.8 7.5  0.8 44.2  1.2 42.1  2.7
2 12.4  0.9 8.2  1.1 45.2  1.3 45.8  2.6
4 14.5  0.6 7.3  0.8 42.8  3.4 43.6  3.1
8 19.0  1.7 10.5  1.1 44.7  2.7 49.6  1.2
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009 DOI 10.1002/jps
Cu-PTSM AND Cu-ATSM DISPLACEMENT FROM HSA 2175
tions was observed in solutions containing [64Cu]Cu-radiopharmaceutical binding in the IIA
furosemide (approx. 33% free). Scatchard plot site, rather than reflecting competitive binding at
analysis of furosemide binding to HSA indicates the IIIA site.
there may be at least one additional, lower affinity Cephmandole binding to HSA (Ka ź 1.2 103
(Ka ź 1 10 4 M 1) HSA binding site.38 At high M 1) is disrupted by ibuprofen,31 suggesting this
drug concentrations, furosemide binding to HSA at drug also binds in the IIIA site. However, while
a secondary site may further displace [64Cu]Cu- bilirubin association with HSA is strong (Ka ź
PTSM, if Cu-PTSM shares this common secondary 9.5 107 M 1) and believed to be distinct from the
site. The extremely limited solubility of Cu-PTSM IIA and IIIA drug sites, cephmandole is, surpris-
did not permit direct investigation of a secondary ingly, also known to disrupt the bilirubin--HSA
binding site in this study by Scatchard plot analy- interaction. The data in Table 1 indicates
sis, which typically requires a 100-fold range of [64Cu]Cu-PTSM is not displaced by cephmandole
ligand concentrations. at drug:HSA mole ratios up to 4:1. The apparent
Benzylpenicillin was included in this study as an lack of [64Cu]Cu-PTSM displacement by ceph-
additional site IIA ligand to probe the affinity of mandole competition seems consistent with the
[64Cu]Cu-PTSM for HSA relative to a competing conclusion that Cu-PTSM does not bind the IIIA
drug with a much lower affinity (Ka ź 1.2 site.
103 M 1) for the IIA site.39--41 While [64Cu]Cu-PTSM Tryptophan is the only amino acid known to
was displaced by increasing amounts of IIA drugs strongly bind to HSA, interacting with the IIIA
with high (105 or 107 M 1) affinities for the IIA site, site.28 There was no evidence of [64Cu]Cu-PTSM
no displacement of [64Cu]Cu-PTSM was observed displacement due to tryptophan; even at the high-
even at an 8:1 benzylpenicillin:HSA mole ratio est concentration of tryptophan, 4.8 mM (1.0 mg/
(3.3  0.1% free Cu-PTSM at the 8:1 benzylpenicil- mL), [64Cu]Cu-PTSM remained predominantly
lin:HSA mole ratio vs. 4.0  0.1% free Cu-PTSM in HSA bound (3.2  0.1% free). Similar results were
the drug-free HSA control solution). The relative obtained when trying to disrupt Cu-PTSM binding
affinity of Cu-PTSM for the IIA site of HSA may be to HSA using a mixture of 18 amino acids at (amino
significantly stronger than the binding of benzyl- acid):HSA mole ratios ranging from 0.2:1 (Trp) to
penicillin, or benzylpenicillin binding to HSA does 7.4:1 (Arg). The unbound fraction of [64Cu]Cu-
not disrupt the concurrent binding of Cu-PTSM. PTSM in test solution containing the mixed amino
Ligand binding in the IIIA site does not appear to acid solution was 2.6  0.2% free.
directly affect the interaction of [64Cu]Cu-PTSM or The interaction of [64Cu]Cu-ETS with HSA was
[64Cu]Cu-ATSM with HSA. The number of site IIIA unaffected by the addition of representative drugs
drugs with solubilities acceptable for use in ultra- with known affinities for either the IIA or IIIA sites
filtration assays is extremely limited. Ibuprofen (Tab. 2). [64Cu]Cu-ETS binding in the drug-free
binding in the IIIA site has been confirmed by a HSA solution was 40.6  1.1% free. At 8:1 drug:H-
crystal structure.42 Ibuprofen has a high affinity SA mole ratios with warfarin and ibuprofen, Cu-
for the IIIA site (Ka ź 2.7 106 M 1), but there was ETS was 44.7  2.7% free, and 49.6  1.2% free,
no displacement of [64Cu]Cu-PTSM or [64Cu]Cu- respectively. Thus, binding of Cu-ETS to HSA
ATSM from HSA observed for ibuprofen:HSA mole appears to be limited to nonspecific interactions.
ratios up to 4:1 (Tabs. 1 and 2). At an 8:1 ibupro- Unlike Cu-PTSM and Cu-ATSM, there is no evi-
fen:HSA mole ratio, however, there were increases dence that Cu-ETS binds with high affinity to the
in unbound tracer: 10.1  0.6% free and 10.5  1.1% IIA or IIIA drug sites.
free for [64Cu]Cu-PTSM and [64Cu]Cu-ATSM, re-
spectively. The reported ibuprofen:HSA adduct
[64Cu]Cu-PTSM Displacement from HSA by
crystal structure revealed a secondary ibuprofen
Fatty Acids
binding site, overlapping a fatty acid binding site,
which bridges the IIA and IIB subdomains.42 Cur- While we have previously seen no influence of fatty
ry further noted anomalous electron density with- acid depletion on the binding of [67Cu]Cu-PTSM in
in the IIA site, suggesting ibuprofen may actually commercially available HSA preparations,27 ESR
bind with very low affinity in the IIA site, but it studies show Cu-PTSM can be displaced from HSA
was not sufficient to allow inclusion in the final by the addition of supraphysiologic levels of stearic
refined model. In our test solutions with the high- acid and the spin-labeled analogue, 5-doxyl stearic
est high ibuprofen concentrations, we believe this acid.43 The results in Table 3 indicate varying levels
secondary ibuprofen binding may be affecting of [64Cu]Cu-PTSM displacement by capric acid
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009
2176 BASKEN, MATHIAS, AND GREEN
Table 3. [64Cu]Cu-PTSM Displacement from HSA Binding by Fatty Acids
Fatty Acid:HSA Ratio Capric Myristic Palmitic Oleic
% Free [64Cu]Cu-PTSM  SD (n)a
1:1 7.5  0.3 (4) 6.8  0.3 (4) 5.6  0.2 (3) 6.0  0.3 (4)
6:1 12.6  3.5 (5) 14.6  0.8 (5) 9.6  0.3 (4) 12.7  0.7 (4)
a
The value of unbound [64Cu]Cu-PTSM in fatty acid-free (HSA þ ethanol) solutions was
4.5  0.2%.
(CH3(CH2)8COOH), myristic acid (CH3(CH2)12 Further competitive binding assays evaluated
COOH), palmitic acid (CH3(CH2)14COOH), and [64Cu]Cu-PTSM, [64Cu]Cu-ATSM, and [64Cu]Cu-
oleic acid (CH3(CH2)7CHźCH(CH2)7COOH). At ETS binding to canine serum albumin (CSA) in the
1:1 fatty acid:HSA ratios, there is little effect of presence of warfarin and ibuprofen. There is no
fatty acid addition on the HSA binding of [64Cu]Cu- evidence of high affinity CSA binding for any
PTSM, while at 6:1 fatty acid:HSA mole ratios the of these bis(thiosemicarbazone)copper(II) com-
level of free [64Cu]Cu-PTSM was approximately plexes, with approximately 40% free reported for
doubled. each tracer in CSA.26 Thus, these studies were
undertaken with the expectation that the added
drugs should not alter the levels of free [64Cu]Cu-
[64Cu]Cu-L Displacement from Rat and Dog Serum
PTSM, [64Cu]Cu-ATSM, or [64Cu]Cu-ETS in ca-
Albumin by Drug Competition
nine serum albumin solutions. Indeed, the free
[64Cu]Cu-PTSM binding to rat serum in the pres- fraction of [64Cu]Cu-PTSM in 40 mg/mL CSA sol-
ence of representative site IIA drugs warfarin, utions containing 8:1 mole ratios of warfarin were
phenylbutazone, and furosemide was also exam- 36.1  2.8% (n ź 4) and 32.8  0.7% (n ź 4), respec-
ined (Tab. 4). Protein binding of [64Cu]Cu-PTSM is tively. Unbound levels of [64Cu]Cu-ATSM and
comparable between solutions of rat serum albu- [64Cu]Cu-ETS assayed in CSA solutions contain-
min (RSA) and rat serum itself.26 The affinity of ing warfarin were 37.0% and 40.4%, respectively.
[64Cu]Cu-PTSM for RSA is relatively high, but Free [64Cu]Cu-ATSM and [64Cu]Cu-ETS in CSA
significantly less than the affinity for HSA, solutions with added ibuprofen were 43.8% and
14.1  0.3% free vs. 4.0  0.1% free for RSA and 39.5%, respectively. These results further suggest
HSA, respectively.26 The affinities of warfarin and that [64Cu]Cu-PTSM has a unique affinity for HSA,
phenylbutazone for RSA (Ka ź 8.3 105 and and that any binding of these three radiopharma-
4.6 105 M 1), respectively, and HSA (Ka ź ceuticals to CSA is limited to nonspecific interac-
3.4 105 and 7.0 105 M 1), respectively, are tions. The affinity of racemic warfarin for CSA has
nearly identical.44 An affinity constant for furose- been reported as 7.4 104 M 1; while this is ap-
mide binding to RSA has not been reported, how- proximately 5 times less than the warfarin affinity
ever, one study reported furosemide binding to for HSA (Ka ź 3.4 105 M 1), warfarin clearly has
RSA as high as 99.1%.45 As observed with HSA, a significant affinity for CSA.46
warfarin, phenylbutazone, and furosemide were
able to competitively displace [64Cu]Cu-PTSM
from protein binding sites in rat serum (Tab. 4). DISCUSSION
The observed competitive displacement of
[64Cu]Cu-PTSM from RSA suggests that RSA con- These competitive binding assays provide some
tains a site analogous to the IIA site of HSA capable structural details of copper chelate binding to HSA.
of binding [64Cu]Cu-PTSM. The site IIA drug ligands warfarin, phenylbuta-
Table 4. [64Cu]Cu-PTSM Displacement from Rat Serum by Competitive Drug Binding
Mole Ratio (Drug:RSA) Warfarin Phenylbutazone Furosemide
% Free [64Cu]Cu-PTSM  SD (n ź 3)
0 14.7  1.2 14.7  1.2 14.7  1.2
1 19.8  0.5 18.6  0.4 20.2  0.7
8 24.1  0.5 20.6  0.4 28.6  0.6
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009 DOI 10.1002/jps
Cu-PTSM AND Cu-ATSM DISPLACEMENT FROM HSA 2177
zone, furosemide, and to a lesser extent, endoge- copper(II) are subtle (Fig. 1). However, it appears
nous fatty acids, were shown to displace [64Cu]Cu- that additional steric bulk introduced by the ethyl
PTSM from binding to HSA. Similar displacement vs. methyl substituent on the diimine chelate back-
due to warfarin binding was observed in HSA bone is sufficient to preclude strong binding of
solutions containing [64Cu]Cu-ATSM. The addi- Cu-ETS in the HSA site that binds Cu-PTSM and
tion of warfarin, furosemide, and phenylbutazone Cu-ATSM.
also elevated unbound [64Cu]Cu-PTSM in rat se- The poor aqueous solubility of Cu-PTSM pre-
rum. Radiotracer liberation was only detected in vents the direct experimental calculation of an
HSA solutions containing ligands with high affini- affinity constant for binding to HSA. Attempts to
ties for the IIA site, suggesting the high affinity competitively displace [64Cu]Cu-PTSM from the
HSA binding of [64Cu]Cu-PTSM and [64Cu]Cu- HSA binding site using excess Cu-PTSM were
ATSM arises from interactions in the IIA binding plagued by precipitation of the poorly soluble Cu-
site of HSA. The crystal structures of warfarin- PTSM chelate.
--HSA and phenylbutazone--HSA adducts show The competitive displacement of Cu-PTSM
drug binding causes only subtle, localized changes and Cu-ATSM from HSA by common therapeutic
in protein configuration, which further suggests drugs is unlikely to have clinical significance with
the observed [64Cu]Cu-PTSM and [64Cu]Cu-ATSM regard to radiopharmaceutical distribution and
displacement may be due to direct competition for kinetics, since these competition studies employed
the IIA site, and not allosteric release from a dis- the competing drugs at levels that could not be
tant binding site. sustained in vivo. Warfarin is a powerful anticoag-
[64Cu]Cu-PTSM or [64Cu]Cu-ATSM displace- ulant with a narrow therapeutic range (2--5 mg/
ment was not observed in HSA solutions contain- mL), and a relatively long pharmacological half
ing even high concentrations of the site IIIA ligand life, averaging 40 h.47 Though warfarin binding
L-tryptophan. The IIIA ligand ibuprofen also did was shown to displace [64Cu]Cu-PTSM, it could not
not disrupt HSA binding of [64Cu]Cu-PTSM or be administered at the doses needed to preclude
[64Cu]Cu-ATSM, except at an 8:1 ibuprofen:HSA HSA interaction with the [60,61,62,64Cu]Cu-PTSM
mole ratio, where we believe ibuprofen s secondary or [60,61,62,64Cu]Cu-ATSM radiopharmaceuticals
binding to the IIA site likely explains the observed in vivo. Furosemide is >98% protein-bound in
modest radiopharmaceutical displacement. Ceph- humans and can be acutely administered as
mandole is also likely a site IIIA ligand (based on its an intravenous bolus, but similarly would not
displacement due to ibuprofen binding), that has saturate even 50% of available HSA molecules at
also been shown to displace bilirubin, which may a typical adult single dose (40 mg).47
bind HSA at a site distinct from the IIA and IIIA Fatty acid binding to HSA is physiologically
sites. Like ibuprofen and L-tryptophan, cephman- important, as HSA is the primary transport vehicle
dole exhibited no tendency to displace Cu-PTSM of insoluble fatty acids in plasma. Typical fatty acid
from HSA. Benzylpenicillin, with a relatively low loading of HSA in circulation is 0.1--2 mol of fatty
affinity (Ka ź 1.2 103 M 1) for site IIA, similarly acids of varying chain lengths heterogeneously
failed to displace radiotracer at any concentration. distributed among as many as seven binding
The extent of benzylpenicillin protein binding in sites.33 Protein X-ray crystallography has revealed
human plasma is only 50%, so a relatively low the relationship of fatty acid binding sites and
affinity for the IIA site may be insufficient to chain length, and shown that a single low affinity
disrupt [64Cu]Cu-PTSM binding in that location. fatty acid binding site overlaps the IIA drug bind-
[64Cu]Cu-ETS has a drastically different protein ing site.34
binding profile than its structurally similar coun- The combined masses of palmitic and oleic acids
terparts, Cu-PTSM and Cu-ATSM. The addition of represent 50% of the total fatty acid present in
representative drugs known to bind in the IIA and adult human serum,48 while myristic acid repre-
IIIA sites did not affect the level of unbound sents 2% of the mass fraction, and capric acid is
[64Cu]Cu-ETS in either HSA or CSA solutions, present only in traceamounts.Thus, while oleic and
suggesting that any measured binding was due to palmitic acids might be the most likely to displace
nonspecific interactions. Similarly, the nonspecific Cu-PTSM in vivo, we believe this is unlikely under
binding of [64Cu]Cu-PTSM or [64Cu]Cu-ATSM to natural circumstances; palmitic acid (the most
CSA was not affected by the introduction of any abundant fatty acid in human serum at 0.074 g/
drug ligand. The structural differences between L) would typically be present at a 0.4:1 palmita-
these three bis(thiosemicarbazone) chelates of te:HSA stoichiometry, well below the 6:1 stoichi-
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009
2178 BASKEN, MATHIAS, AND GREEN
ometry required for even moderate Cu-PTSM dis- on human serum albumin. Mol Pharmacol 12:
1052 1061.
placement from HSA in our study (Tab. 3).
6. Kragh-Hansen U, Chuang VTG, Otagiri M. 2002.
Practical aspects of the ligand-binding and enzymat-
ic properties of human serum albumin. Biol Pharm
CONCLUSIONS
Bull 25:695 704.
7. Tillement JP, Houin G, Zini R, Urien S, Albengres E,
The binding of Cu-PTSM and Cu-ATSM to HSA is
Barre J, Lecomte M, D Athis P, Sebille B. 1984. The
selectively disrupted by drugs known to possess
binding of drugs to blood plasma macromolecules:
high affinities for the IIA binding site of albumin.
Recent advances and therapeutic significance. In:
Drugs that bind the IIIA site did not generally
Testa B, editor. Advances in drug research 13.
appear to disrupt Cu-PTSM or Cu-ATSM binding
London: Academic Press Inc. pp 59 94.
Ź
to HSA; however, the IIIA drug ibuprofen, which 8. Sjoholm I, Ekman B, Kober A, Ljungstedt-Pahlman
Ź
I, Seiving B, Sjodin T. 1979. Binding of drugs to
appears to have some secondary affinity for the IIA
serum albumin XI: The specificity of three binding
site,42 was found to modestly elevate free Cu-
sites as studied with albumin immobilized in micro-
PTSM and Cu-ATSM levels at a high ibuprofen:H-
particles. Mol Pharmacol 16:767 777.
SA mole ratio. The interaction of the structurally
9. Peters T. 1996. All about albumin. San Diego, CA:
related Cu-ETS radiopharmaceutical with HSA is
Academic Press Inc.
unaffected by drugs that bind at either the IIA or
10. Green MA, Klippenstein DL, Tennison JR. 1988.
IIIA site, consistent with previous conclusions that
Copper(II) bis(thiosemicarbazone) complexes as
any observed serum protein binding of this tracer is
potential tracers for evaluation of cerebral and
limited to nonspecific interactions.26,27 Displace-
myocardial blood flow with PET. J Nucl Med
ment of Cu-PTSM due to site IIA drugs was also
29:1549 1557.
detected in rat serum solutions, while there was no 11. Shelton ME, Green MA, Mathias CJ, Welch MJ,
Bergmann SR. 1989. Kinetics of copper-PTSM in
evidence of increased Cu-PTSM, Cu-ATSM, or Cu-
isolated hearts: A novel tracer for measuring blood
ETS in canine serum albumin solutions using
flow with positron emission tomography. J Nucl Med
drugs which bind in either site.
30:1843 1847.
12. Shelton ME, Green MA, Mathias CJ, Welch MJ,
Bergmann SR. 1990. Assessment of regional
ACKNOWLEDGMENTS
myocardial and renal blood flow using copper-PTSM
and positron emission tomography. Circulation 82:
This work was supported by a research grant from
990 997.
the Purdue Research Foundation, and R01-
13. Herrero P, Markham J, Weinheimer CJ, Anderson
CA092403. The development of Cu-64 production
CJ, Welch MJ, Green MA. 1993. Quantification of
at Washington University School of Medicine was
regional myocardial perfusion with generator-pro-
supported by NCI grant R24 CA86307. 62
duced Cu-PTSM and positron emission tomogra-
phy. Circulation 87:173 183.
14. Herrero P, Hartman JJ, Green MA, Anderson CJ,
REFERENCES
Welch MJ, Markham J. 1996. Assessment of region-
al myocardial perfusion with generator-produced
62
1. Winters ME. 1994. Basic Clinical Pharmacokinetics. Cu-PTSM and PET in human subjects. J Nucl Med
37:1294 1300.
Philadelphia, PA:Lippincott Williams & Willkins.
9 17. 15. Beanlands RS, Muzik O, Mintun MA, Mangner T,
2. Bertucci C, Domenici E. 2002. Reversible and cova- Lee K, Petry N. 1992. The kinetics of copper-62-
lent binding of drugs to human serum albumin: PTSM in the normal human heart. J Nucl Med
Methodological approaches and physiological rele- 33:684 690.
vance. Curr Med Chem 9:1463 1481. 16. Wallhaus TR, Lacy J, Whang J, Green MA,
3. Lenter C, editor. 1984. Geigy scientific tables, Vol. 3. Nickles RJ, Stone CK. 1998. Human biodistri-
West Caldwell, NJ: Ciba Geigy Corp. pp 136 bution and dosimetry of the PET perfusion
139. agent copper-62-PTSM. J Nucl Med 39:1958
4. Sudlow G, Birkett DJ, Wade DN. 1975. The 1964.
characterization of two specific drug binding sites 17. Wallhaus TR, Lacy J, Stewart R, Bianco J, Green
on human serum albumin. Mol Pharmacol 11: 824 MA, Nayak N. 2001. Copper-62-pyruvaldehyde bis-
832. (N-4-methylthiosemicarbazone) PET imaging in the
5. Sudlow G, Birkett DJ, Wade DN. 1976. Further detection of coronary artery disease in humans.
characterization of specific drug binding sites J Nuc Cardiol 8:67 74.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009 DOI 10.1002/jps
Cu-PTSM AND Cu-ATSM DISPLACEMENT FROM HSA 2179
18. Green MA, Mathias CJ, Willis L, Handa R, Miller M, 32. Brodersen R. 1982. Physical chemistry of bilirubin:
Hutchins G. 2005. Evaluation of Cu-ETS as a PET Bindingtomacromolecules andmembranes.In: Heir-
tracer of regional renal perfusion. J Labelled Compd wegh KPM, Brown SB, editors. Bilirubin. Vol. I.
Radiopharm 48:S164. Chemistry. Boca Raton, FL: CRC Press. pp 75 123.
Ź
19. Green MA, Mathias CJ, Willis LR, Handa RK, Miller 33. Bhattacharya AA, Grune T, Curry S. 2000. Crystal-
62
MA, Hutchins GD. 2006. Evaluation of Cu-ETS as lographic analysis reveals common modes of binding
a PET tracer of renal perfusion in a model of acute of medium and long-chain fatty acids to human
focal renal injury. J Nucl Med 47:146P. serum albumin. J Mol Biol 303:721 732.
20. Green MA, Mathias CJ, Willis LR, Handa RK, 34. SimardJR,Zunszain PA, Hamilton JA,Curry S.2006.
Lacy JL, Miller MA. 2007. Assessment of Cu-ETS Location of high and low affinity fatty acid binding
as a PET radiopharmaceutical for evaluation of sites on human serum albumin revealed by NMR
regional renal perfusion. Nucl Med Biol 34:247 drug-competition analysis. J Mol Biol 361:336 351.
255. 35. Watanabe H, Kragh-Hansen U, Tanase S, Nakajou
21. Lewis JS, Welch MJ. 2001. PET imaging of hypoxia. K, Mitari M, Iwao Y, Maruyama T, Otagiri M. 2001.
Q J Nucl Med 45:183 188. Conformational stability and warfarin-binding
22. Lewis JS, Sharp TL, Laforest R, Fujibayashi Y, properties of human serum albumin studied by re-
Welch MJ. 2001. Tumor uptake of copper-diacetyl- combinant mutants. Biochem J 357:269 274.
bis(N-4-methylthiosemicarbazone): Effect of changes 36. Koizumi K, Ikeda C, Ito M, Suzuki J, Kinoshita T,
in tissue oxygenation. J Nucl Med 42:655 661. Yasukawa K, Hannai T. 1998. Influence of glycosyl-
23. Yuan H, Schroeder T, Bowsher JE, Hedlund LW, ation on the drug binding of human serum albumin.
Wong T, Dewhirst MW. 2006. Intertumoral differen- Biomed Chromatogr 12:203 210.
ces in hypoxia selectivity of the PET imaging agent 37. John EK, Green MA. 1990. Structure-activity
Cu-64(II)-diacetyl-bis(N-4-methylthiosemicarbazone). relationships for metal-labeled blood flow tracers:
J Nucl Med 47:989 998. Comparison of keto aldehyde bis(thiosemicarbazone)
24. Lewis JS, Herrero P, Sharp TL, Engelbach JA, Fu- copper(II) derivatives. J Med Chem 33:1764 1770.
jibayashi Y, Laforest R. 2002. Delineation of hypoxia 38. Viani A, Cappiello M, Silvestri D, Pacifici GM. 1991.
in canine myocardium using PET and copper(II)- Binding of furosemide to albumin isolated from hu-
diacetyl-bis(N-4-methylthiosemicarbazone). J Nucl man fetal and adult serum. Dev Pharmacol Ther
Med 43:1557 1569. 16:33 40.
25. Blower PJ, Castle TC, Cowley AR, Dilworth JR, 39. Joos RW, Hall WH. 1969. Determination of binding
Donnelly PS, Labisbal E. 2003. Structural trends in constants of serum albumin for penicillin. J Phar-
copper(II) bis(thiosemicarbazone) radiopharma- macol Exp Ther 166:113 118.
ceuticals. Dalton Trans 23:4416 4425. 40. Kunin CM. 1966. Clinical pharmacology of the new
26. Basken NE, Mathias CJ, Lipka AE, Green penicillins. Clin Pharmacol Ther 7:180 188.
MA. 2008. Species dependence of [64Cu]Cu-bis 41. Kunin CM. 1966. Inhibitors of penicillin binding to
(thiosemicarbazone) radiopharmaceutical binding serum proteins. J Lab Clin Med 65:416 431.
to serum albumins. Nucl Med Biol 35:281 286. 42. Ghuman J, Zunszain PA, Petitpas I, Bhattacharya
27. Mathias CJ, Bergmann SR, Green MA. 1995. AA, Otagiri M, Curry S. 2005. Structural basis for
Species-dependent binding of copper(II) bis- the drug-binding specificity of human serum albu-
(thiosemicarbazone) radiopharmaceuticals to se- min. J Mol Biol 353:38 52.
rum albumin. J Nucl Med 36:1451 1455. 43. Yuan H, Antholine WE, Subczynski WK, Green MA.
28. McMenamy RH, Oncley JL. 1958. The specific bind- 1996. Release of Cu-PTSM from human serum albu-
ing of l-tryptophan to serum albumin. J Biol Chem min after addition of fatty acids. J Inorg Biochem
255:1436 1447. 61:251 259.
29. Cunningham VJ, Hay L, Stoner HB. 1975. The 44. Herve F, Rajkowski K, Martin MT, Dessen P, Cit-
binding of l-tryptophan to serum albumins in the tanova N. 1984. Drug binding properties of rat a1-
presence of non-esterified fatty acids. Biochem J fetoprotein. Biochem J 221:401 406.
146:653 658. 45. Hammarlund MM, Paalzow LK. 1982. Dose-depen-
30. Yamasaki K, Kuga N, Takamura N, Furuya Y, dent pharmacokinetics of furosemide in the rat.
Hidaka M, Iwakiri T, Nishii R, Okamura M, Kodama Biopharm Drug Dispos 3:345 359.
H, Kawai K, Arimori K. 2005. Inhibitory effects of 46. O Reilly RA. 1971. Interaction of several coumarin
amino-acid fluids on drug binding to site II of human compounds with human and canine plasma albu-
serum albumin in vitro. Biol Pharm Bull 28: 549 min. Mol Pharmacol 7:209 218.
552. 47. Lacy CF, Armstrong LL, Goldman MP, Lance LL.
31. Nerli B, Romanini D, Pico G. 1997. Structural spec- 2003. Drug information handbook. Hudson, OH:
ificity requirements in the binding of beta lactam Lexi-Comp Inc.
antibiotics to human serum albumin. Chem Biol 48. Lenter C, editor. 1984. Geigy scientific tables, Vol.
Interact 104:179 202. 3.West Caldwell, NJ:Ciba Geigy Corp. p 122.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 98, NO. 6, JUNE 2009