Engineered recognition of tetravalent zirconium and thorium by chelator-protein systems: Toward flexible radiotherapy and imaging platforms

Article abstract of DOI:10.1021/acs.inorgchem.6b02041

Targeted α therapy holds tremendous potential as a cancer treatment: it offers the possibility of delivering a highly cytotoxic dose to targeted cells while minimizing damage to surrounding healthy tissue. The metallic α-generating radioisotopes ²²⁵Ac and ²²⁷Th are promising radionuclides for therapeutic use, provided adequate chelation and targeting. Here we demonstrate a new chelating platform composed of a multidentate high-affinity oxygen-donating ligand 3,4,3-LI(CAM) bound to the mammalian protein siderocalin. Respective stability constants log β₁₁₀ = 29.65 ± 0.65, 57.26 ± 0.20, and 47.71 ± 0.08, determined for the Eu^III (a lanthanide surrogate for Ac^III), Zr^IV, and Th^IV complexes of 3,4,3-LI(CAM) through spectrophotometric titrations, reveal this ligand to be one of the most powerful chelators for both trivalent and tetravalent metal ions at physiological pH. The resulting metal-ligand complexes are also recognized with extremely high affinity by the siderophore-binding protein siderocalin, with dissociation constants below 40 nM and tight electrostatic interactions, as evidenced by X-ray structures of the protein:ligand:metal adducts with Zr^IV and Th^IV. Finally, differences in biodistribution profiles between free and siderocalin-bound ²³⁸Pu^IV-3,4,3-LI(CAM) complexes confirm in vivo stability of the protein construct. The siderocalin:3,4,3-LI(CAM) assembly can therefore serve as a "lock" to consolidate binding to the therapeutic ²²⁵Ac and ²²⁷Th isotopes or to the positron emission tomography emitter ⁸⁹Zr, independent of metal valence state.

Full text of DOI:10.1021/acs.inorgchem.6b02041

Article
pubs.acs.org/IC
Engineered Recognition of Tetravalent Zirconium and Thorium by
Chelator−Protein Systems: Toward Flexible Radiotherapy and
Imaging Platforms
Ilya Captain,† Gauthier J.-P. Deblonde,† Peter B. Rupert,‡ Dahlia D. An,† Marie-Claire Illy,†
Emeline Rostan,† Corie Y. Ralston,§ Roland K. Strong,*,‡ and Rebecca J. Abergel*,†
†Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States
‡Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109, United States
§Berkeley Center for Structural Biology, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States
S
* Supporting Information
ABSTRACT: Targeted α therapy holds tremendous potential as a cancer treatment: it
offers the possibility of delivering a highly cytotoxic dose to targeted cells while
minimizing damage to surrounding healthy tissue. The metallic α-generating radioisotopes
225Ac and 227Th are promising radionuclides for therapeutic use, provided adequate
chelation and targeting. Here we demonstrate a new chelating platform composed of a
multidentate high-affinity oxygen-donating ligand 3,4,3-LI(CAM) bound to the
mammalian protein siderocalin. Respective stability constants log β110 = 29.65 0.65,
57.26 0.20, and 47.71 0.08, determined for the EuIII (a lanthanide surrogate for AcIII),
ZrIV, and ThIV complexes of 3,4,3-LI(CAM) through spectrophotometric titrations, reveal
this ligand to be one of the most powerful chelators for both trivalent and tetravalent
metal ions at physiological pH. The resulting metal−ligand complexes are also recognized
with extremely high affinity by the siderophore-binding protein siderocalin, with
dissociation constants below 40 nM and tight electrostatic interactions, as evidenced by
X-ray structures of the protein:ligand:metal adducts with ZrIV and ThIV. Finally, differences in biodistribution profiles between
free and siderocalin-bound 238PuIV-3,4,3-LI(CAM) complexes confirm in vivo stability of the protein construct. The
siderocalin:3,4,3-LI(CAM) assembly can therefore serve as a “lock” to consolidate binding to the therapeutic 225Ac and 227Th
isotopes or to the positron emission tomography emitter 89Zr, independent of metal valence state.
227Th-CD33-TTC9 for myeloid leukemia treatment, or 227Th-
INTRODUCTION
■
DOTA-trastuzumab for treating HER-2 positive breast and
Targeted α therapy (TAT), or radioimmunotherapy that uses
α-particle emitting nuclides, is a promising treatment strategy
ovarian cancer.10 Reasons for this slow development are largely
economical and related to costs associated with radionuclide
for small metastatic tumors and other localized diseases. Owing
production, clinical material production, as well as preclinical
to α-particles’ short path length, much of the decay energy may
be deposited into target areas while mitigating damage to
and clinical studies. Nevertheless, a better understanding of
surrounding tissue.1 A number of radionuclides that emit single
radiometal chelation and biodistribution is also direly needed to
α particles, including 211At (t1/2 = 7.2 h), 213Bi (t1/2 = 0.8 h),
advance this field and to provide new chelating platforms that
and 212Bi (t1/2 = 1.0 h), are currently under investigation.2 A
may be tailored on the basis of indication and isotope selection.
The 225Ac and 227Th radioisotopes are members of the
growing subset of the field includes in vivo α-generator
radionuclides 225Ac (t1/2 = 10.0 d), 223Ra (t1/2 = 11.4 d), and
actinide (An) series of elements. They display high
coordination numbers and are best chelated by high-denticity
ligands that contain hard donor atoms, such as the multidentate
227Th (t1/2 = 18.7 d), isotopes that emit multiple α particles in
their decay chains and dramatically increase the potential
delivered dose.3−6 This principle was recently exploited in the
hydroxypyridinone-based (HOPO) compounds, our workhorse
ligands for in vivo actinide decorporation.11 The ligand 3,4,3-
development of Alpharadin, 223RaCl2, a drug for bone
metastases.7 While Alpharadin relies on the natural propensity
LI(1,2-HOPO) is an octadentate, tetraprotic compound
of 223Ra for bone, other specifically targeted α-radiation delivery
strategies use constructs formed with chelating agents to
composed of 4 bidentate 1,2-HOPO metal-binding units
attached onto a spermine (“3,4,3-LI”) scaffold (Figure 1).
complex metallic α-emitters and biological targeting vectors.2
3,4,3-LI(1,2-HOPO) was recently modified to enable mono-
clonal antibody attachment and form a bioconjugate chelator
Though sound in theory, these designs have been slow to
appear in the clinic, with only scarce examples of promising α-
generator immunoconjugates, such as the lintuzumab conjugate
Received: August 23, 2016
225Ac-HuM1958 and the sialic-acid receptor CD33 conjugate
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
ligand complex and Scn. In order to restore complex binding by
Scn, we must use a ligand of high denticity that captures both
3+ and 4+ metals, forms overall negatively charged complexes,
and does not cause steric clashes in Scn’s tight binding pocket.
This work aims to address the inability of the Scn-3,4,3-LI(1,2-
HOPO) system to bind 4+ ions by exploring more suitable
ligands that can complex both 3+ and 4+ metal ions.
METHODS
■
Caution! 232Th, 238Pu, 242Pu, 243Am, and 248Cm are hazardous
radionuclides with high specific activities that should only be
manipulated in specifically designated facilities in accordance with
appropriate safety controls.
General Considerations. Chemicals were obtained from
commercial suppliers and were used as received. The LnCl3·
nH2O lanthanide salts utilized were of the highest purity
available (>99.9%). Stock solutions of 232ThIV and ZrIV were
prepared from 232ThCl4·8H2O (Baker & Adamson, ACS grade)
and ZrCl4 (Sigma-Aldrich, 99.99%), respectively. A stock of
238PuIV was purchased as 238Pu(NO3)4 in 4 M HNO3 from
Eckert & Ziegler (lot 118521). 242Pu was received from Oak
Ridge National Laboratory as PuO2 (lot Pu-242-327 A, 99.93
wt % of metal 242Pu), and a stock solution of 242PuIV was
prepared as described previously.17 The 242Pu isotope was used
for in vitro binding experiments whereas 238Pu was used in
biodistribution studies. Aliquots of acidified stocks of carrier-
free 243Am and 248Cm (95.78% 248Cm, 4.12% 246Cm, 0.06%
245Cm, 0.02% 244Cm/247Cm isotopic distribution by atom
percentage) from Lawrence Berkeley National Laboratory were
used. All solutions were prepared using deionized water purified
by a Millipore Milli-Q reverse osmosis cartridge system, and the
pH was adjusted with concentrated HCl, H2SO4, KOH, or
NaOH when needed. Measurements were conducted at room
temperature using standard instrumentation unless otherwise
noted (see Supporting Information for details).
Figure 1. Complexation of MIII and MIV by the hexadentate
siderophore enterobactin or the octadentate synthetic analogues
3,4,3-LI(CAM) or 3,4,3-LI(1,2-HOPO), when deprotonated. Poly-
amine scaffolds and metal-binding oxygen atoms are highlighted in
blue and red, respectively.
that displayed high in vivo stability when bound to 89Zr, as
imaged by positron emission tomography (PET).12,13 Similarly,
an octadentate Me-3,2-HOPO bioconjugate chelator built on a
slightly different symmetrical polyamine scaffold was recently
prepared for the complexation of 227Th, and it displayed fast
radiolabeling kinetics.14 Although successful, such modifications
necessitate extensive synthetic procedures with low overall
yields,12,14 which prompted us to investigate alternate routes to
link therapeutic and imaging radionuclides to targeting
moieties.
The mammalian 24 kDa glycoprotein siderocalin (Scn) is a
member of the lipocalin protein superfamily. Scn is known to
complement the innate immune system by binding specific
bacterial siderophores and their ferric complexes, a mechanism
that counters the iron acquisition pathway of invading
microorganisms.19 Our recent work revealed that Scn binds
lanthanide (Ln) and An ions precomplexed with a suitable
ligand in solution with very high affinity, through strong
electrostatic interactions.15 Although [FeIII(Ent)]3−, the ferric
complex of the hexadentate catecholate siderophore enter-
obactin (Ent) is Scn’s native ligand,16 the protein’s sterically
hindered binding pocket was shown to bind Ln and An
complexes of Ent (Figure 1).15 More surprisingly, the protein
also accommodates the much stronger Ln and An complexes
formed with the octadentate synthetic analogue 3,4,3-LI(1,2-
HOPO).15 Building on this work, one can envision using the
Scn−ligand−metal system as a radionuclide binder for TAT as
well as a reporter ligand for concurrent diagnostics. The
advantage of using a protein-mediated binding system is 2-fold:
potentially tighter binding to nuclides of interest and easier
attachment to targeting moieties by eliminating the need for
post protein expression bioconjugate chemistry. The afore-
mentioned 225AcIII and 227ThIV show promise in radio-
immunotheraphy, while 89ZrIV is useful as a PET tracer; all of
these metals may be captured by the Scn−ligand system with
some optimization. Upon single deprotonation of each of the
1,2-HOPO units (Figure 1), the overall negative 3,4,3-LI(1,2-
HOPO) complexes of LnIII and AnIII are tightly bound by Scn,
but the formally neutral 3,4,3-LI(1,2-HOPO) complexes of 4+
metals interact weakly.15 The difference in binding originates
from insufficient electrostatic interaction between the metal−
Metal, Ligand, and Protein Working Solutions. The
trivalent lanthanide LnIII working stock solutions were prepared
in standardized 0.1 M HCl. A ZrIV stock solution was prepared
by dissolving ZrCl4 in 3.0 M H2SO4, to prevent hydrolysis. The
metal salt ZrCl4 was handled and stored in a glovebox kept
under inert atmosphere. The ZrIV stock solution was stand-
ardized against EDTA, with xylene orange as the indicator.18
A
ThIV stock solution was prepared in 0.1 M H2SO4. Stock
solutions (4 mM) of Ent and 3,4,3-LI(CAM) were prepared by
direct dissolution of a weighed portion of ligand in DMSO, and
aliquots were removed prior to each set of experiments.
Recombinant human Scn was prepared as previously
described.19
Incremental Spectrophotometric Titrations. This
method was used to determine the protonation constants of
3,4,3-LI(CAM) as well as the stability constants of its
complexes formed with EuIII, ZrIV, and 232ThIV. The
experimental titration setup is similar to previously described
systems, and details are provided as Supporting Information.20
For the 3,4,3-LI(CAM) protonation (and EuIII-3,4,3-LI(CAM)
complexes), titrations were performed with an initial
concentration of 50 μM of 3,4,3-LI(CAM) (and 50 μM of
EuIII) resulting in absorbance values between 0 and 1.0
throughout the titration. For the ZrIV and ThIV complexes,
titrations were performed similarly but in the presence of
DTPA to avoid the formation of metal hydroxides at low pH,
before the uptake by 3,4,3-LI(CAM).
B
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Protein-Binding Assay and Crystallography. Scn bind-
ing of metal−ligand complexes was measured by fluorescence
quenching, as described previously15,21 and detailed in the
Supporting Information. Diffraction-quality crystals were grown
by vapor diffusion from drops containing the ternary metal−
chelator−protein complex according to standard methods
(Supporting Information). Diffraction data were collected on
beamline 5.0.2 at the Advanced Light Source (ALS, Berkeley,
CA) and processed as detailed in Supporting Information.
Crystallographic statistics are reported in Supporting Informa-
tion Table S1. Final models have been deposited in the PDB.22
In Vivo Biodistribution Assay. All procedures and
protocols used in the presented in vivo studies were reviewed
and approved by the Institutional Animal Care and Use
Committee at Lawrence Berkeley National Laboratory and
performed in AAALAC accredited facilities. The animals used
were adult female CD-1 mice (180 7 days old, 40.8 5.8 g).
Solutions of 238Pu complexed by 3,4,3-LI(CAM) and Scn:3,4,3-
LI(CAM) were prepared in situ at molar ratios protein:li-
gand:238Pu of 0:100:1 and 100:100:1, respectively, by mixing
and incubating the appropriate quantities of 238Pu(NO3)4,
ligand, and protein in phosphate-buffered saline (PBS) to reach
HOPO) due to CAM units requiring further deprotonation for
metal binding (Figure 1). 3,4,3-LI(CAM) was synthesized here
from readily available building blocks using a process developed
in-house (Scheme 1). The new preparation moves away from
a
Scheme 1. Synthesis of 3,4,3-LI(CAM).
a
(a) H2SO4, MeOH, 65°C, 16 h (88%). (b) Dichlorodiphenyl-
methane, 160°C, 1 h. (c) 50/50 THF/H2O, reflux 5 h (81% over 2
steps). (d) (COCl)2, toluene, cat. DMF; then spermine, Et3N, THF,
50°C, O/N (78%). (e) AcOH/H2O + conc HCl, 16 h (90%).
a
238Pu concentration of 12 ng L−1. Protein solutions were
washed thrice with PBS using 10 kDa molecular weight cutoff
membrane-based centrifugal filters, and all solutions were filter-
sterilized (0.22 μm) prior to injection. Under isoflurane
anesthesia, groups of three normally fed mice were injected
intravenously with 0.2 mL of a complex solution (370 Bq per
mouse). After injection of the 238Pu tracer, mice were weighed,
identified, and housed in groups of three in plastic stock cages
lined with a 0.5 cm layer of highly absorbent low-ash pelleted
cellulose bedding (Alpha-dri) for separation of urine and feces.
Mice were given water and food ad libitum and euthanized at 4,
24, or 48 h after tracer injection. All experiments using 238Pu
tracers were managed as metabolic balance studies, in which
tissues and excreta were analyzed for 238Pu by liquid
scintillation counting on a Perkin-Elmer Packard Tri-Carb
model B4430. The methods of sample collection, preparation,
radioactivity measurements, and data reduction have been
published previously.23−25 Those methods provide quantitative
measurements of radioactivity in biological samples; material
recoveries averaged about 99% of the amount injected in these
experiments.
using harsh reaction conditions by using the protected
diphenylmethylene acetal derivative (5), which greatly
simplifies purification of the final product.
Affinity of 3,4,3-LI(CAM) toward 3+ and 4+ Metals. A
comprehensive solution thermodynamic analysis was per-
formed to characterize the affinity of 3,4,3-LI(CAM) for
trivalent and tetravalent metals and to understand the effect of
substituting 1,2-HOPO for CAM binding units on the
octadentate spermine scaffold. The protonation constants of
3,4,3-LI(CAM) were determined by spectrophotometric
titrations, and eight protonation equilibria were assigned to
sequential removal of two protons from each of the four CAM
units (Table 1). Previous studies of Ent and other CAM-
containing synthetic analogues established that the protonation
constants (pKa1−pKa4) of the meta-hydroxyl oxygen atoms are
well-separated from the ortho-hydroxyl oxygen atoms (pKa5−
pKa8).28 The last four pKa values are most relevant to metal
binding as moieties corresponding to these values have to be
deprotonated at physiological pH in order to bind the metal
ions. The overall acidity of 3,4,3-LI(CAM) can be defined as
∑pKa5−8 = 27.8 versus 3,4,3-LI(1,2-HOPO)’s 21.2,29 with
lower values representing higher acidity. 3,4,3-LI(CAM) is
therefore less prone to bind hard Lewis acids at low pH than its
1,2-HOPO analogue, due to competition between metal uptake
and protonation of the CAM moieties.
Incremental spectrophotometric titrations were then carried
out to determine the formation of EuIII, ZrIV, or ThIV complexes
with 3,4,3-LI(CAM). Because of the relatively short half-life of
225Ac and the scarce availability of the longer-lived 227Ac, EuIII
was used here as a nonradioactive Ln surrogate for AcIII. On the
basis of previous solution thermodynamic studies of LnIII
complexes of 3,4,3-LI(1,2-HOPO) and other common
polyaminocarboxylate ligands,24 it is reasonable to expect
similar stability constants for EuIII and AcIII complexes of 3,4,3-
LI(CAM). The 3,4,3-LI(CAM) octadentate ligand showed a
very high affinity for both 3+ and 4+ ions (Table 1). The
stability constants of [EuIII-3,4,3-LI(CAM)]5−, [ThIV-3,4,3-
LI(CAM)]4−, and [ZrIV-3,4,3-LI(CAM)]4− are several orders
of magnitude higher than those of their 1,2-HOPO counter-
RESULTS AND DISCUSSION
■
Synthesis of Octadentate Ligand 3,4,3-LI(CAM). Since
electrostatic interactions between Scn and Ln/An complexes
play a key role in binding, we chose to explore ligands that
would form overall negative complexes with both 3+ and 4+
metals. Although Scn exhibits a broad, degenerate recognition
mechanism for native siderophores, previous studies probing
the extent of Scn binding to synthetic siderophore analogues
showed that the Scn binding site allows only limited changes to
its ligands.21,26 Thus, the simplest way to correct the binding
would be to employ ligands with similar structural features. We
chose to follow Occam’s razor by using 3,4,3-LI(CAM) (Figure
1), a compound first prepared by Raymond and co-workers for
plutonium decorporation purposes.27 This octadentate ligand
leverages grafting of catecholamide (CAM) moieties found in
microbial siderophores on the spermine scaffold to form a
hybrid version of Ent and 3,4,3-LI(1,2-HOPO) that should (i)
display increased complex stability over Ent due to its higher
denticity and (ii) bear more negative charges than 3,4,3-LI(1,2-
C
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Protonation and EuIII, ZrIV, and ThIV Complex
Formation Constants for 3,4,3-LI(CAM)
interactions with the Scn protein. This represents a large
advantage over 3,4,3-LI(1,2-HOPO), for which the complexes
with MIII at pH 7.4 have only one negative charge and are
neutral in the case of MIV ions.
a
species
mlh
log βmlh
pKa
3,4,3-LI(CAM)
12.50 0.30
24.50 0.35
35.81 0.55
45.41 0.47
54.11 0.23
61.93 0.56
68.67 1.45
73.24 1.38
29.65 0.65
41.75 0.06
46.79 0.14
57.26 0.20
64.25 0.32
47.71 0.08
55.36 0.09
3,4,3-LI(1,2-HOPO)
6.64
Scn Recognition of 3,4,3-LI(CAM)-Metal Complexes.
As described in several previous reports,15,21,33 the affinity of
Scn for ligands or metal−ligand complexes is quantified by
monitoring protein fluorescence quenching upon ligand or
complex binding. The equilibrium dissociation constant of Scn
for the apo form of the ligand 3,4,3-LI(CAM), Kd = 1.2 0.4
nM, is nearly identical to that determined for the native
siderophore apo-Ent,21 indicating that the addition of a fourth
CAM unit does not affect ligand recognition by the protein.
Subsequent determination of Kd values for various metal
complexes of 3,4,3-LI(CAM) (MIII = Sm, Eu, Gd, 243Am, or
248Cm, and MIV = Zr, 232Th, or 242Pu) confirms tight binding to
the protein, independent of the metal valence, with values well
below 40 nM (Figure 2). As hypothesized, these data
LH7−
011
012
013
014
015
016
017
018
110
111
112
110
111
110
111
12.50 0.30
12.00 0.30
11.31 0.55
9.60 0.12
8.70 0.27
7.82 0.57
6.73 0.61
4.57 0.29
6−
5−
4−
3−
2−
−
LH2
LH3
LH4
LH5
LH6
LH7
LH8
[EuL]5−
[EuLH]4−
[EuLH2]3−
[ZrL]4−
[ZrLH]3−
[ThL]4−
[ThLH]3−
LH3−
011
012
013
014
110
111
110
110
6.64
5.68
5.01
3.87
2−
LH2
12.32
−
LH3
17.33
LH4
21.20
[EuL]−
[EuLH]
[ZrL]
[ThL]
20.2
24.8
43.1
40.1
a
I = 0.1 M (KCl), T = 25 °C. Errors correspond to standard deviations
from at least three independent titrations. Protonation and EuIII,6
ZrIV,7 and ThIV complex formation constants previously reported for
8
3,4,3-LI(1,2-HOPO) are also given for comparison.
parts, with log β110 values of 29.7, 47.7, and 57.3, respectively.
Consequently, 3,4,3-LI(CAM) is one of the strongest ligands
ever reported for the chelation of both trivalent and tetravalent
f-elements. For comparison, a cyclic octadentate terephthala-
mide derivative was recently designed to bind Th4+ in vivo and
showed an unprecedented affinity for Th4+ with a log β110
(ThL4−) value of 53.7.30 It is worth mentioning, however, that
terephthalamide ligands would not be recognized by Scn, due
to steric hindrance generated by the bulky substituents on the
para position of the catecholate-like unit. To inspect the pH
dependency of metal complex formation, speciation diagrams
were calculated for 3,4,3-LI(CAM) in the presence of 1
equivalent of EuIII, ZrIV, or ThIV (Figures S1−S3). Both ZrIV
and ThIV complexes start forming at around pH 3, with the
mono and fully deprotonated species, [MIVLH]3− and
[MIVL]4−, being predominant at physiological pH (7.4). This
behavior departs from that of 3,4,3-LI(1,2-HOPO), with which
4+ metal complexes are formed even under very acidic
conditions (pH < 0).31,32 For EuIII, complexation by 3,4,3-
LI(CAM) starts at pH 5, and the monoprotonated complex,
[EuIIILH]4−, is the only species present at pH 7.4. Similar to
what is observed with 4+ metals, the pH at which EuIII-3,4,3-
LI(CAM) complexes start forming is higher than in the case of
EuIII-3,4,3-LI(1,2-HOPO) species that already appear at pH ∼
1 under those same conditions.29 However, it is important to
note the multiple negative charges of the 3,4,3-LI(CAM)
complexes with 3+ and 4+ metal complexes under physiolog-
ically relevant conditions ([MIIILH]4−, [MIVLH]3−, and
[MIVL]4−), which are now capable of forming electrostatic
Figure 2. Scn dissociation constants determined from fluorescence
quenching analyses for MIII and MIV complexes formed with Ent15 (a,
top left), 3,4,3-LI(1,2-HOPO)15 (b, top right), or 3,4,3-LI(CAM) (c,
bottom left), and crystal pictures for the Scn adducts formed with the
ZrIV (d, bottom middle) and ThIV (e, bottom right) complexes of
3,4,3-LI(CAM). The charges of the discussed metal complexes vary
from 0 to 4− at pH 7.4 and are indicated above each bar; asterisks
indicate that no binding was observed. Ent and 3,4,3-LI(1,2-HOPO)
data are plotted on the basis of previously reported values,15 except for
the Zr complex Kd values that were determined in this work; the
affinity observed with the ferric complex of Ent is shown for reference,
as it is the native Scn ligand.
demonstrate a substantial difference in protein recognition
between the 3,4,3-LI(CAM) and 3,4,3-LI(1,2-HOPO) com-
plexes of tetravalent metals. Use of the diprotic CAM units in
lieu of the monoprotic 1,2-HOPO moieties led to the
formation of negatively charged complexes, enabling electro-
static interactions with the protein trilobal calyx. Moreover,
while the addition of a fourth CAM metal-binding group in the
octadentate 3,4,3-LI(CAM) was important for increased
stability of the metal−ligand complexes at pH 7.4, it did not
prevent the high Scn affinities initially observed with
hexacoordinated Ent complexes. Subtle differences were
observed with the recognition patterns: for similarly charged
D
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
complexes, weaker binding was observed with actinide
complexes (AmIII, CmIII, ThIV, and PuIV) as compared to
corresponding lanthanide or d-block metal complexes (SmIII,
EuIII, GdIII, ZrIV) in the case of 3,4,3-LI(CAM). The opposite
trend had been noted with Ent complexes of MIII metals15 and
was confirmed here with ZrIV, while no significant differences
are discernible with 3,4,3-LI(1,2-HOPO).
Structural Characterization of Scn-3,4,3-LI(CAM) Ad-
ducts. X-ray crystallography was used as previously described15
to determine the structures of the Scn adducts formed with the
232Th-3,4,3-LI(CAM) and Zr-3,4,3-LI(CAM) complexes
(Table S1). As expected, and as observed in previous Scn
complex structures (e.g., with 243Am-3,4,3-LI(1,2-HOPO)),15
the compounds are bound in the deeply recessed trilobal
binding site, or calyx, of Scn (Figure 3a−c). As in previous
structures of Scn bound to CAM-bearing ligands with An ions,
only one CAM substituent is ordered in the crystal structures
along with the bound metal, Zr or Th. This was likely a result of
the remainder of the chelator sampling multiple conformations
between molecules in the crystal, but clearly confirmed binding
of chelator and metal in both adducts. The CAM substituent
bound in the key binding pocket in the Scn calyx, between the
side-chains of two bracketing lysine residues (K125 and K134;
Figure 3d). The structure of the Scn calyx is highly conserved
with prior structures, reflecting its rigidity, with the side-chains
of two residues, W79 and R81, as the only elements flexing to
accommodate different ligands (Figure 3e). In prior Ent or Ent
analogue chelator/actinide structures,15 where two of the three
CAM groups are also disordered, the side-chain of W79
adopted an unusual rotamer, flipping inward toward the metal
to contribute a cation−π interaction. However, in these
structures, this side-chain adopts more conventional orienta-
tions, either sampling multiple rotamers, rendering it
disordered in the crystallographic analysis (in the Zr complex),
or lying against the calyx wall in the Th complex. Apart from
these two side-chains, the overall impression conveyed by these
and previous results is that the chelator flexes and distorts to fit
in an essentially rigid and unyielding calyx.
Biodistribution Evaluation of Scn-3,4,3-LI(CAM) Ad-
ducts. To evaluate the in vivo retention and excretion patterns
of MIV-3,4,3-LI(CAM) complexes and their Scn adducts,
238PuIV was used as a radiolabel. 238Pu was selected as a
surrogate radiometal to mimic 227Th or 89Zr for the following
reasons: (i) It forms tetravalent coordination species under
physiological conditions. (ii) Its ionic radius is between those of
ZrIV and ThIV.32 (iii) It is easier to handle and allows for more
accurate metabolic balance experiments due to its relatively
long radioactive half-life (87.8 years) and low specific activity
(0.63 TBq/g), compared to the therapeutic 227Th (18.7 d; 1139
TBq/g) and imaging 89Zr (78.4 h; 16,630 TBq/g) isotopes.
(iv) Natural Zr or other commonly available isotopes such as
232Th (14 Gyr; 4.07 kBq/g) would not exhibit enough activity
to allow for radioanalysis. In this in vivo stability experiment,
238Pu−ligand complex solutions were formed in situ (ligand:Pu
and Scn:ligand:Pu ratios of 100:1 and 100:100:1, respectively)
and administered intravenously. Mice were euthanized 4, 24, or
48 h after the metal injection, and tissues and excreta were
radioanalyzed for 238Pu content (Figure 4). Independent of the
presence of Scn in the administered solution, about 30% of the
injected 238Pu was excreted by 48 h, and 238Pu excreta content
steadily increased, suggesting delayed clearance of the
complexes. The rate of 238Pu elimination observed for 238Pu-
Figure 3. Crystallographic analyses of the binding of 232Th-3,4,3-
LI(CAM) and Zr-3,4,3-LI(CAM) by Scn. (a) A molecular surface
representation of the calyx of Scn is colored by atom type (C, gray; O,
red; N, blue), with the 243Am-3,4,3-LI(1,2-HOPO) complex15 shown
in a licorice-stick representation, also colored by atom type. The metal
ion is shown as a small sphere, colored magenta. Surfaces contributed
by the side-chains of W79 and R81 are colored yellow and green,
respectively. Two prominent blue-tipped protuberances sticking into
the calyx correspond to the side-chains of K125 and K134, which
bracket the crucial binding pocket in the calyx. (b) A view into the Scn
calyx of the 232Th-3,4,3-LI(CAM) complex structure, shown as in part
a. In this structure, the side-chain of W79 is disordered, indicated by
the yellow dashed circle, due to sampling of multiple rotamers in the
complex, allowing much more of the surface of R81 to be seen. Only
one substituent CAM group and the actinide (aqua sphere) are
ordered and visualized in the complex crystal structure, as had been
seen in a previous hexadentate CAM/actinide/Scn structure.15 The
one ordered CAM moiety sits in the key binding pocket defined by the
side-chains of K125 and K134. (c) A view into the Scn calyx of the Zr-
3,4,3-LI(CAM) complex structure, shown as in parts a and b. In this
structure, the side-chains of W79 and R81 have repositioned to
accommodate this ligand. Like the 232Th structure, only one
substituent CAM group and the Zr atom (orange sphere) are ordered.
(d) A view perpendicular to that in parts a−c highlights the differential
packing of the CAM or HOPO substituents in the key binding pocket.
Other atoms in the chelator have been removed for clarity, and the
molecular surface has been rendered partially transparent. Metal and
carbon atoms have been recolored to indicate the complex: 243Am-
3,4,3-LI(1,2-HOPO) in magenta, 232Th-3,4,3-LI(CAM) in aqua, and
Zr-3,4,3-LI(CAM) in green. (e) The side-chains of W79 and R81 and
the connecting protein backbone have been isolated and superimposed
from the three complexes; they are shown in a licorice-stick
representation, colored as in part d. This view highlights the different
rotamers selected in the different complexes, the only element of
conformational flexibility in the extremely rigid Scn calyx.15
3,4,3-LI(CAM) is strikingly different from that observed for the
238Pu-3,4,3-LI(1,2-HOPO) complex in previous studies (albeit
performed in a different strain of mice and with younger
animals), in which quantitative excretion was observed by 24 h.
In both Scn-bound and free 238Pu-3,4,3-LI(CAM) cases, and at
all time points, more 238Pu was found in the urine than in the
feces. However, the kidney and liver contents reveal a key
difference in excretion pattern: when free, the 3,4,3-LI(CAM)
complex is predominantly found in the liver at early time points
after administration and follows a biliary pathway, similar to
what is known for HOPO complexes. However, insertion
within the protein favors elimination through the renal system,
with up to ∼52% of 238Pu found in the kidneys 4 h after
administration of the Scn adducts, a burden that subsequently
E
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Detailed experimental procedures; NMR and mass
spectra of protected and deprotected 3,4,3-LI(CAM);
examples of spectrophotometric titrations and speciation
diagrams for the scavenging of EuIII, ZrIV, and ThIV by
3,4,3-LI(CAM); and crystallographic statistics for
Scn:3,4,3-LI(CAM) bound to ZrIV and ThIV (PDF)
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: rstrong@fhcrc.org.
*E-mail: rjabergel@lbl.gov.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This material is based upon work supported by the U.S.
Department of Energy, Office of Science, Office of Basic Energy
Sciences, Chemical Sciences, Geosciences, and Biosciences
Division at the Lawrence Berkeley National Laboratory under
Contract DE-AC02-05CH11231 (R.J.A.), and by the National
Institutes of Health under award number R01DK073462
(subcontract to R.K.S.). R.J.A. is the recipient of a U.S.
Department of Energy, Office of Science Early Career Award.
M.-C.I. is the recipient of a mobility grant from the ParisTech
chair of nuclear engineering supported by Areva. The Advanced
Light Source (ALS) is supported by the Director, Office of
Science, Office of Basic Energy Sciences, of the U.S.
Department of Energy under Contract DE-AC02-
05CH11231. We thank M. Allaire, S. Morton, J. Bramble, K.
Engle, I. Tadesse, and M. Dupray for their assistance in
planning and implementing diffraction data collection on
radioactive crystals at the ALS 5.0.2 beamline. We also thank
K. Raymond for providing us with purified Ent. The content is
solely the responsibility of the authors and does not necessarily
represent the official views of the National Institutes of Health.
Figure 4. Biodistribution of PuIV when administered as a free or Scn-
bound 3,4,3-LI(CAM) complex. Healthy mice injected intravenously
with 238Pu−ligand and 238Pu−ligand−protein solutions (370 Bq/
mouse); mice euthanized at 4, 24, or 48 h. “ART” and “Soft” stand for
“abdominal remaining tissues” and other “soft tissues”, respectively.
slowly decreases. In combination with significantly faster rates
of excretion and considerably lower skeleton and soft tissue
burden when compared with free 238Pu, this major difference
between kidney versus liver 238Pu retention of the Scn-bound
versus free complex provide evidence for the high in vivo
stability of the Scn:[PuIV(3,43-LI(CAM))] adduct.
CONCLUSIONS
■
Results obtained with the Scn:3,4,3-LI(CAM) system suggest a
novel concept that enables the engineering of new radiophar-
maceuticals comprising protein targeting vectors, such as
antibodies, coupled with radionuclide capture moieties based
on Scn. The protein’s highly specific binding to 3,4,3-
LI(CAM)-MIV and 3,4,3-LI(CAM)-MIII complexes eliminates
the need for costly synthetic bioconjugation of ligands to
targeting molecules. Interestingly, one can envision a system
where both imaging (89ZrIV) and therapeutic (227ThIV or
225AcIII) metallic radioisotopes may be used in conjunction for
dual diagnostics/treatment applications. Our results illustrate
the promise of this system. It is also likely that, with further
tuning, a more effective protein−ligand arrangement may be
developed: current work is underway to explore other ligands
based on the 3,4,3-LI scaffold in order to reconcile the metal/
protein binding compromise by incorporating both monoprotic
HOPO and diprotic CAM units. Finally, future efforts will focus
on evaluating the kinetics of release of different radionuclides
from the Scn:ligand systems, as well as the distribution of
daughter products. We will perform in vitro and in vivo studies
with the more suited medical isotopes 225Ac, 227Th, and 89Zr, in
lieu of the low-activity surrogates used in the presented solution
studies.
REFERENCES
■
(1) Mulford, D. A.; Scheinberg, D. A.; Jurcic, J. G. The promise of
targeted α-particle therapy. J. Nucl. Med. 2005, 46, 199S−204S.
(2) Kim, Y. S.; Brechbiel, M. W. Tumor Biol. 2012, 33, 573−590.
(3) McDevitt, M. R.; Ma, D.; Lai, L. T.; Simon, J.; Borchardt, P.;
Frank, R. K.; Wu, K.; Pellegrini, V.; Curcio, M. J.; Miederer, M.;
Bander, N. H.; Scheinberg, D. A. Tumor Therapy with Targeted
Atomic Nanogenerators. Science 2001, 294, 1537−1540.
(4) Aghevlian, S.; Boyle, A. J.; Reilly, R. M. Radioimmunotherapy of
cancer with high linear energy transfer (LET) radiation delivered by
radionuclides emitting α-particles or Auger electrons. Adv. Drug
Delivery Rev. 2015, DOI: 10.1016/j.addr.2015.12.003.
(5) Knapp, F. F.; Dash, A. Alpha Radionuclide Therapy. In
Radiopharmaceuticals for Therapy; Springer India: New Delhi, 2016;
pp 37−55.
(6) Read, E. D.; Eu, P.; Little, P. J.; Piva, T. J. The status of
radioimmunotherapy in CD20+ non-Hodgkin’s lymphoma. Targeted
Oncology 2015, 10, 15−26.
(7) Liepe, K. Alpharadin, a 223Ra-based alpha-particle-emitting
pharmaceutical for the treatment of bone metastases in patients with
cancer. Current opinion in investigational drugs (London, England: 2000)
2009, 10, 1346−1358.
(8) Ravandi, F.; Pagel, J. M.; Park, J. H.; Douer, D.; Estey, E. H.;
Kantarjian, H. M.; Cicic, D.; Scheinberg, D. A. Phase I trial of the
targeted alpha-particle nano-generator actinium-225 (225Ac)-lintuzu-
mab (anti-CD33) in combination with low-dose cytarabine (LDAC)
for older patients with untreated acute myeloid leukemia (AML).
Blood 2013, 122, 1460−1460.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.6b02041.
■
S
F
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(9) Hagemann, U. B.; Wickstroem, K.; Wang, E.; Shea, A. O.;
Sponheim, K.; Karlsson, J.; Bjerke, R. M.; Ryan, O. B.; Cuthbertson, A.
S. In vitro and in vivo efficacy of a novel CD33 targeted thorium-227
conjugate for the treatment of acute myeloid leukemia. Mol. Cancer
Ther. 2016, 15, 2422.
(10) Heyerdahl, H.; Abbas, N.; Brevik, E. M.; Mollatt, C.; Dahle, J.
Fractionated therapy of HER2-expressing breast and ovarian cancer
xenografts in mice with targeted alpha emitting 227 Th-DOTA-p-
benzyl-trastuzumab. PLoS One 2012, 7, e42345.
(11) Bunin, D. I.; Chang, P. Y.; Doppalapudi, R. S.; Riccio, E. S.; An,
D. D.; Jarvis, E. E.; Kullgren, B.; Abergel, R. J. Efficacy and Safety
Toxicology of Hydroxypyridinonate Actinide Decorporation Agents in
Rodents: Towards a Safe and Effective Human Dosing Regimen.
Radiat. Res. 2013, 179, 171−182.
(12) Deri, M. A.; Ponnala, S.; Kozlowski, P.; Burton-Pye, B. P.; Cicek,
H. T.; Hu, C.; Lewis, J. S.; Francesconi, L. C. p-SCN-Bn-HOPO: A
Superior Bifunctional Chelator for 89Zr ImmunoPET. Bioconjugate
Chem. 2015, 26, 2579−2591.
(13) Deri, M. A.; Ponnala, S.; Zeglis, B. M.; Pohl, G.; Dannenberg, J.
J.; Lewis, J. S.; Francesconi, L. C. Alternative Chelator for Zr-89
Radiopharmaceuticals: Radiolabeling and Evaluation of 3,4,3-(LI-1,2-
HOPO). J. Med. Chem. 2014, 57, 4849−4860.
(14) Ramdahl, T.; Bonge-Hansen, H. T.; Ryan, O. B.; Larsen, Å.;
Herstad, G.; Sandberg, M.; Bjerke, R. M.; Grant, D.; Brevik, E. M.;
Cuthbertson, A. S. An efficient chelator for complexation of thorium-
227. Bioorg. Med. Chem. Lett. 2016, 26, 4318−4321.
(15) Allred, B. E.; Rupert, P. B.; Gauny, S. S.; An, D. D.; Ralston, C.
Y.; Sturzbecher-Hoehne, M.; Strong, R. K.; Abergel, R. J. Siderocalin-
mediated recognition, sensitization, and cellular uptake of actinides.
Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 10342−10347.
(16) Clifton, M. C.; Corrent, C.; Strong, R. K. Siderocalins:
siderophore-binding proteins of the innate immune system. BioMetals
2009, 22, 557−564.
defending against mycobacterial infections through iron sequestration.
Structure 2005, 13, 29−41.
(27) Weitl, F. L.; Raymond, K. N. Specific sequestering agents for the
actinides. 3. Polycatecholate ligands derived from 2, 3-dihydroxy-5-
sulfobenzoyl conjugates of diaza-and tetraazaalkanes. J. Am. Chem. Soc.
1980, 102, 2289−2293.
(28) Loomis, L. D.; Raymond, K. N. Solution equilibria of
enterobactin and metal-enterobactin complexes. Inorg. Chem. 1991,
30, 906−911.
́
(29) Abergel, R. J.; D’Aleo, A.; Ng Pak Leung, C.; Shuh, D. K.;
Raymond, K. N. Using the Antenna Effect as a Spectroscopic Tool:
Photophysics and Solution Thermodynamics of the Model Lumines-
cent Hydroxypyridonate Complex [EuIII(3, 4, 3-LI (1, 2-HOPO))]−.
Inorg. Chem. 2009, 48, 10868−10870.
(30) Pham, T. A.; Xu, J.; Raymond, K. N. A Macrocyclic Chelator
with Unprecedented Th4+ Affinity. J. Am. Chem. Soc. 2014, 136, 9106−
9115.
(31) Deblonde, G. J.; Sturzbecher-Hoehne, M.; Abergel, R. J.
Solution thermodynamic stability of complexes formed with the
octadentate hydroxypyridinonate ligand 3, 4, 3-LI (1, 2-HOPO): a
critical feature for efficient chelation of lanthanide (IV) and actinide
(IV) ions. Inorg. Chem. 2013, 52, 8805−8811.
(32) Sturzbecher-Hoehne, M.; Choi, T. A.; Abergel, R. J.
Hydroxypyridinonate complex stability of group (IV) metals and
tetravalent f-block elements: The key to the next generation of
chelating agents for radiopharmaceuticals. Inorg. Chem. 2015, 54,
3462−3468.
(33) Abergel, R. J.; Wilson, M. K.; Arceneaux, J. E.; Hoette, T. M.;
Strong, R. K.; Byers, B. R.; Raymond, K. N. Anthrax pathogen evades
the mammalian immune system through stealth siderophore
production. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 18499−18503.
(17) Gorden, A. E.; Shuh, D. K.; Tiedemann, B. E.; Wilson, R. E.; Xu,
J.; Raymond, K. N. Sequestered Plutonium:[PuIV{5LIO(Me-3, 2-
HOPO)}2]The First Structurally Characterized Plutonium Hydrox-
ypyridonate Complex. Chem. - Eur. J. 2005, 11, 2842−2848.
(18) Welcher, F. J. Analytical Uses of Ethylenediamine Tetraacetic Acid;
D. Van Nostrand Co., 1958
(19) Goetz, D. H.; Holmes, M. A.; Borregaard, N.; Bluhm, M. E.;
Raymond, K. N.; Strong, R. K. The neutrophil lipocalin NGAL is a
bacteriostatic agent that interferes with siderophore-mediated iron
acquisition. Mol. Cell 2002, 10, 1033−1043.
(20) Sturzbecher-Hoehne, M.; Deblonde, G.-P.; Abergel, R. Solution
thermodynamic evaluation of hydroxypyridinonate chelators 3, 4, 3-LI
(1, 2-HOPO) and 5-LIO (Me-3, 2-HOPO) for UO2 (VI) and Th
(IV) decorporation. Radiochim. Acta 2013, 101, 359−366.
(21) Abergel, R. J.; Moore, E. G.; Strong, R. K.; Raymond, K. N.
Microbial evasion of the immune system: structural modifications of
enterobactin impair siderocalin recognition 1. J. Am. Chem. Soc. 2006,
128, 10998−10999.
(22) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.
N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data
Bank. Nucleic Acids Res. 2000, 28, 235−242.
(23) Kullgren, B.; Jarvis, E. E.; An, D. D.; Abergel, R. J. Actinide
chelation: biodistribution and in vivo complex stability of the targeted
metal ions. Toxicol. Mech. Methods 2013, 23, 18−26.
́
(24) Sturzbecher-Hoehne, M.; Leung, C. N. P.; D’Aleo, A.; Kullgren,
B.; Prigent, A.-L.; Shuh, D. K.; Raymond, K. N.; Abergel, R. J. 3, 4, 3-
LI (1, 2-HOPO): In vitro formation of highly stable lanthanide
complexes translates into efficacious in vivo europium decorporation.
Dalton Trans. 2011, 40, 8340−8346.
(25) Durbin, P. W.; Kullgren, B.; Ebbe, S. N.; Xu, J.; Raymond, K. N.
Chelating agents for uranium (VI): 2. Efficacy and toxicity of
tetradentate catecholate and hydroxypyridinonate ligands in mice.
Health Phys. 2000, 78, 511−521.
(26) Holmes, M. A.; Paulsene, W.; Jide, X.; Ratledge, C.; Strong, R.
K. Siderocalin (Lcn 2) also binds carboxymycobactins, potentially
G
DOI: 10.1021/acs.inorgchem.6b02041
Inorg. Chem. XXXX, XXX, XXX−XXX