A Cyclic-RGD Dinuclear TbIII Macrocyclic Complex as a Tumor Integrin-Selective Luminescent Probe

Article abstract of DOI:10.1002/ejic.201800568

To develop small molecular integrin-selective luminescent imaging probes, we have prepared the binary dipicolinate (DPA) Tb^III dinuclear macrocyclic complex, Tb₂(cRGDfK-ODO2A-dimer) (DPA)₂^2– or complex I, and reference ligands and Tb^III complexes which were purified by HPLC and characterized by NMR, mass spectrometry and luminescence spectroscopy. Luminescence titrations of the structural and bonding model Tb₂(m-ODO2A-dimer)²⁺ complex by DPA^2– ion confirmed the molecular formula of the adduct was Tb₂(m-ODO2A-dimer)(DPA)₂^2–, and the first binary binding constant was determined to be log K₁ = 5.76. At pH 7.4, complex I showed 300 times luminescence enhancement at 544 nm (λ_ex = 278 nm) as compared to that without adding DPA, and was found to bind to α_vβ₃ integrin and human glioblastoma U87MG tumor cells in both specific and non-specific modes, via luminescence spectroscopic and confocal cell imaging competition studies. This makes complex I and its future optimized derivatives potentially feasible for preclinical bioimaging applications, particularly in the time-resolved mode.

Full text of DOI:10.1002/ejic.201800568

A Journal of
Accepted Article
Title: A Cyclic-RGD Dinuclear TbIII Macrocyclic Complex as a Tumor
Integrin-Selective Luminescence Probe
Authors: C. Allen Chang, Ju-Chien Chia, and Syue-Liang Lin
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201800568
Link to VoR: http://dx.doi.org/10.1002/ejic.201800568

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
III
A Cyclic-RGD Dinuclear Tb Macrocyclic Complex as a Tumor
Integrin-Selective Luminescence Probe
C. Allen Chang1,2,3*, Ju-Chien Chia , Syue-Liang Lin
1
2,3
1
Department of Biomedical Imaging and Radiological Sciences, National Yang-Ming University, No.
55, Sec. 2, Li-Nong St., Beitou, Taipei, Taiwan, 112, R.O.C. E-mail: cachang@ym.edu.tw, Tel:
886-2-28267000 Ext. 7943, Fax: +886-2-28238311.
1
+
2
3
Department of Biotechnology and Laboratory Science in Medicine, National Yang-Ming University
Molecular Imaging Research Center, National Yang-Ming University.
*
Corresponding author: C. Allen Chang, E-mail: cachang@ym.edu.tw, Homepage: http://birs-
e.web.ym.edu.tw/files/11-1253-1501-1.php
Abstract
To develop small molecular integrin-selective luminescence imaging probes, we have prepared
III
the binary dipicolinate (DPA) Tb dinuclear macrocyclic complex, Tb
2
(cRGDfK-ODO2A-dimer)
DPA) 2-
or complex I, and reference ligands and Tb complexes which were purified by HPLC and
III
(
2
characterized by NMR, mass spectrometry and luminescence spectroscopy. Luminescence titrations
2
+
2-
of the structural and bonding model Tb
molecular formula of the adduct was Tb
constant was determined to be log K = 5.76. At pH 7.4, complex I showed 300 times luminescence
enhancement at 544 nm (λex = 278 nm) as compared to that without adding DPA, and was found to
bind to α integrin and human glioblastoma U87MG tumor cells in both specific and non-specific
2
(m-ODO2A-dimer) complex by DPA ion confirmed the
(m-ODO2A-dimer)(DPA) 2-, and the first binary binding
2
2
1
β
v 3
modes, via luminescence spectroscopic and confocal cell imaging competition studies. This makes
complex I and its future optimized derivatives potentially feasible for preclinical bioimaging
applications, particularly in the time-resolved mode.
Introduction
Integrins in tumor cells and tumor related vessels contribute to increased cell migration,
invasion, proliferation, metastasis and survival.1-5 Host cellular response to cancer including
angiogenesis is also upregulated by integrins, especially the α
v
β
3
type. Thus, tumor theranostics by
integrin could be an effective
targeting tumor cells and their microenvironment via targeting α
v 3
β
strategy.1,2,6-8 Among various potential theranostic candidates, the Arg-Gly-Asp (RGD) motif which
9
has been found in fibrinogen, fibronectin, plasminogen, metalloprotease-2 and osteopontin is a
specific peptide that targets the over-expressed α
v
β integrin of melanoma, glioblastoma and
3
1
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
pancreatic tumors and their microenvironment. Because linear RGD has a flexible conformation and
1
0
is decomposed relatively easily, cyclic RGD forms such as cRGDfK and cRGDfV with better bio-
stability and stronger α integrin binding ability have been recently used for biomedical
β
v 3
applications.11-13 Although a number of radiolabeled cyclic RGD derivatives are known,14-17 few
small molecular cyclic RGD derivatives have been reported as integrin selective luminescence
imaging probes which could be potentially more convenient to use.18-20
In our translational medicine research program, to develop macrocyclic dinuclear trivalent
III
21
lanthanide (Ln ) complexes as luminescence imaging probes in conjunction with the discovery of
tumor targeting fusion protein drugs,22,23 we have selected the cRGDfK peptide biomarker as a
common delivery vehicle to the tumor cells. As luminescence probes, thermodynamically stable and
III
kinetically inert Ln macrocyclic complexes are preferred as compared to their linear structural
analogues, and their long emission lifetimes are often advantageous for use in the time-resolved
2
4
mode. However, because the electric dipole f-f transitions are Laporte forbidden with very low
extinction coefficients, the employment of the antenna effect via ligand organic chromophores is
III
25,26
often used to enhance the sensitized emission from the Ln center.
In addition, we decided to use
III
the dinuclear Ln systems not only because of the possibly double luminescence signals as
compared to mononuclear ones but also that various mixed metal dinuclear systems such as
IIILn
III
and/or Ln /M systems could be prepared where M could be a transition metal ion to
III
Ln
1
2
examine potential luminescence energy transfer properties and their applications. Mononuclear LnIII
systems that we have studied often did not offer this kind versatility for various applications.27,28 To
understand the fundamental coordination properties, the solution speciation, density functional
theory (DFT) calculations, luminescence properties and promoted nitrophenyl-phosphate hydrolysis
III
rates of the parent dinuclear Ln -m-ODO2A-dimer macrocyclic complexes (Ln
Yb, Y; L = m-ODO2A-dimer) of the current studies have been reported.21
2
L, Ln = La, Eu, Tb,
(cRGDfK-ODO2A-dimer)(DPA) 2
-
Recently, we have found that the newly prepared Tb
2
2
2
-
complex (abbreviated as complex I, DPA is the dipicolinate ion, Scheme 1) was able to selectively
bind α integrin and the human glioblastoma U87MG tumor cells via luminescence spectroscopic
and confocal cell imaging competition studies. This approach has a unique feature in that just
employing the antenna effect using the ligand chromophore in the Tb
(cRGDfK-ODO2A-dimer)2+
β
v 3
2
2
-
was not sufficient for convenient luminescence imaging. However, by adding DPA ions, the
resulting complex I showed 300 times luminescence enhancement29,30 which made it and future
optimized derivatives potentially feasible for practical bioimaging applications. The results are
reported in this paper.
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
(cRGDfK-ODO2A-dimer)(DPA) 2
-
or complex I. ● =
Scheme 1. The structural formula of the Tb
2
2
2
-
DPA ion.
Results and Discussion
Synthesis of the ligands. The synthesis of various intermediates and macrocyclic ligands
including the cRGDfK-ODO2A-dimer is outlined in Scheme 2. In brief, the coupling agent, 1,3-
bis(bromomethyl)-5-nitrobenzene, was prepared from 5-nitroisophthalic acid via the intermediate 5-
3
1
nitro-1,3-phenylene-dimethanol according to a published method. Two equivalents of the
compound, tricyclo[8,2,1,0]1-oxa-4,7,10-triazacyclotridecane (12N3O-tricyclic), prepared by a
literature published method32,33 were mixed with one equivalent of the coupling agent to result in
1
,3-bis(1,4,10-triaza-7-oxacyclododec-1-ylmethyl)-5-nitrobenzene (nitro-12N3O-tricyclic-dimer).
After alkylation and acidification, the compound, 1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-
oxacyclododec-1-ylmethyl]-5-nitrobenzene hydrochloride salt (nitro-ODO2A-dimer HCl salt) was
obtained. This compound was converted to 1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-
oxacyclododec-1-ylmethyl]-5-aniline hydrochloride salt (amine-ODO2A-dimer HCl salt), and 1,3-
bis[4, 10-bis(carboxy- methyl)-1,4,10-triaza-7-oxacyclododec-1-ylmethyl]-5-isothiocyanatobenzene
hydrochloride salt (SCN-Bn-ODO2A-dimer, HCl salt) using published methods with minor
3
4
modifications. The SCN-Bn-ODO2A-dimer was further derivatized with cRGDfK peptide and
3
5
glycine to obtain the respective cRGDfK-ODO2A-dimer and Gly-ODO2A-dimer. For comparison
and competition experiments, the compound cRGDfK-FITC (FITC = fluorescein isothiocyanate) was
also synthesized. Selected NMR spectra, HPLC chromatograms and mass spectra of various
intermediates and SCN-Bn-ODO2A-dimer derivatives are shown in the supporting information.
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
Scheme 2. The synthesis of the SCN-Bn-ODO2A-dimer and the cRGDfK-ODO2A-dimer.
Luminescence studies. Solutions of Tb (cRGDfK-ODO2A-dimer) and complex I were
2
III
prepared by mixing two equivalents of Tb ions with one equivalent of the appropriate ligand
without or with additional two equivalents of DPA and the final mixtures were slowly adjusted to pH
7
.4 using the 3-(N-morpholino) propanesulfonic acid (MOPS) buffer solution (0.10 M). Figure 1
showed the excitation (λem = 544 nm) and emission (λex = 278 nm) spectra of the Tb (cRGDfK-
2
III
ODO2A-dimer) at pH 7.4 (MOPS buffer). Excitation at 278 nm gave a typical Tb emission
4
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
5
→7F
5
→7F
5
→7F
) and 620
spectrum with typical peaks at 490 nm ( D
4
6
), 544 nm ( D
4
5
), 584 nm ( D
4
4
5
→7F
). Owing to the antenna effect, the luminescence signal at 544 nm is about 35 times
nm ( D
4
3
greater than that excited directly at 355 nm without the antenna effect (data not shown).
Figure 1. The excitation (λem=544 nm, left) and emission (λex=278 nm, right) spectra of
Tb
2
(cRGDfK-ODO2A-dimer), 0.1 mM, pH 7.4, MOPS buffer.
Figure 2 showed the excitation (λem = 544 nm) and emission (λex = 278 nm) spectra of the
complex I (pH 7.4). As shown in Figure 2, the emission peak intensity at 544 nm increased another
00 times as compared to that of the Tb (cRGDfK-ODO2A-dimer) which was mainly due to the
direct DPA ion coordination in addition to displacement of inner-sphere coordinated water
3
2
2
-
2
-
molecules (q) by the DPA ions. The former was similar to the result of the ancillary ligand antenna
2
9
2-
III
effect, i.e. the coordinated DPA ion was closer to the Tb ion with a better energy level overlap
than the substituted benzene ligand backbone. The latter was manifested by the increased lifetime of
III
the Tb ion from 1.32 ms (without DPA, q = 1.6~2.0) to 2.49 ms (with DPA, q = 0.6) in the
2
8
complex. The luminescence decay curves of the Tb (cRGDfK-ODO2A-dimer) without and with
2
DPA, and the estimation of the number of coordinated water molecules, q (λex = 278 nm, λem = 540
nm; pH 7.4, MOPS buffer), were shown in the Supporting information, Figure S1 and Table S1,
respectively. Note that removal of coordinated water molecules (i.e. O-H oscillators) only increased
the luminescence intensity in smaller amount, normally not as high as 300 times enhancement, as
shown in a previously published paper when pH was raised.28
2
+
L2+ complex
It was measured that the absorbance of the model Tb
2
(m-ODO2A-dimer) or Tb
2
with DPA was 25 times greater than that without DPA at the same concentration. Additional quantum
2
+
yield (QY) determinations of the model Tb
values increased from 0.21% without DPA to 2.65% with DPA.36,37 These data confirmed that the
emission brightness (enhancement) ratio to be 25 x 12 = 300 for the Tb L complexes with and
2
L complex without and with DPA showed that the QY
2
3
8
without DPA. This 300 times signal enhancement allowed convenient luminescence imaging at the
later bioimaging stage (vide infra).
5
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
Figure 2. The excitation (λem = 544 nm, left) and emission (λex = 278 nm, right) spectra of the
complex 1, Tb
2
(cRGDfK-ODO2A-dimer)(DPA)
2
, 0.1 mM, pH 7.4, MOPS buffer.
2
+
2-
30,39
Tb
amount of the Tb
Tb (m-ODO2A-dimer) or Tb
model for the DPA ion binding studies. A luminescence titration of the Tb
2
(m-ODO2A-dimer) -DPA binary complex binding studies.
Because of the limited
2
+
2
(cRGDfK-ODO2A-dimer) complex available, the analogous model complex
2
+
L2+ (L = m-ODO2A-dimer) was used as a structural and bonding
L2+ complex (λex = 278
2
2
2
-
2
2
-
nm, λem = 544 nm) by the DPA ion at pH 7.4 showed a saturation curve (Figure 3, left) with a break
point at 2 equivalents of DPA added, indicating that the molecular formula of the adduct was
Tb
2
L(DPA) 2
2
-. This is consistent with that each ODO2A macrocycle in the ditopic ligand occupied 6
III
coordination sites of the 9-coordinated Tb ion, leaving 3 meridional coordination sites for the
2
-
28,30
tridentate DPA ion to bind.
Taking the data between 1 and 3 equivalents of DPA added and
fitting them to a two-site saturation ligand binding model gave two identical log K binding constants
2
of 5.38, i.e. K = 1/Kd1 = 1/Kd2 (R =0.996; Equation 1, where Bmax1 and Bmax2 are the maximum
specific bindings to the two sites in the same units as Y, Kd1 and Kd2 are the equilibrium DPA
dissociation constants). This binding constant value was about 0.38 log K unit lower than that of the
more precise value determined by a more rigorous method (vide infra).
Y = Bmax1*[DPA]/(Kd1+[DPA]) + Bmax2*[DPA]/(Kd2+[DPA])
(1)
2
+
2-
Figure 3. Left: Luminescence titration curve of 5 µM Tb (m-ODO2A-dimer) by DPA ion (N =
2
equivalent) at pH 7.4 (MOPS buffer, 25℃). Right: Plot of log(CTb – RCDPA) vs. log (R/(1 – R)) for the
6
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
2
-
2+
binary Tb
2
(m-ODO2A-dimer)(DPA) complex, according to equation 3. [DPA ] = 100 nM, [Tb L ] =
2
5
0 – 5000 nM, pH 7.4 (MOPS/sodium acetate buffer), 25℃. λex = 278 nm and λem = 544 nm were
applied for the determination of Tb L(DPA) emission intensity.
2
A more rigorous method similar to that used by Jones and Vullev was employed to determine
the first DPA ion binding affinity with the Tb
L2+ complex at pH 7.4, 25℃ (Equation 2, where L is
2
-
2
3
9
the m-ODO2A-dimer ligand). The binding affinity of the binary complex (K ) was obtained using
1
the linear relationship in Equation 3:30
2
+
2-
2+
Tb
2
L + DPA  Tb
2
L(DPA) ; K
1
= [Tb
2
L(DPA)]eq /([Tb
2
L ]eq[DPA ]eq
)
(2)
(3)
log(R/(1-R)) = log(CTb – RCDPA) + log K
1
; R = [TbDPA]eq/([TbDPA]eq + [DPA]eq
)
2
+
where CTb is the total concentration of the Tb L complex and R is the normalized integrated
2
emission intensity. Thus, a linear plot of log(CTb – RCDPA) vs. log (R/(1 – R)) gave the respective
2
slope and intercept values of 1.03±0.03 and log K
1
= 5.76±0.19 (R =0.995) for the binary
Tb
2
L(DPA) complex (Figure 3, right). To verify the method, a log K
= 7.41±0.18 for the Tb(DPA)+
1
complex was also determined by us using the same procedure which was in good agreement with the
formation constant obtained by Jones and Vullev at a similar pH. The binding constant for the binary
+
Tb
2
L(DPA) complex was 1.65 log K units lower than that of the Tb(DPA) complex which was
III
probably due to some steric hindrance of the Tb binding sites in the macrocyclic ditopic complex.
2
-
L2+ complex at pH 7.4
Interestingly, the fitting of the titration data of 100 nM DPA ion by the Tb
MOPS buffer, 25℃) to a one-site saturation ligand binding model gave a log K
which was only 0.16 log K unit lower than that obtained by the linear plot according to equation 3
2
(
1
value of 5.60±0.42,
(Figure S2, supporting information).
Note that in the previous study, the structural analogous heterodinuclear TbEu(m-ODO2A-
dimer) or TbEuL complex was predicted to be in the “open-form” conformation in the pH range 6.7–
III
III
1
0.6 assuming that the fluorescence energy transfer (FRET) between the Tb ion and the Eu ion
2
1
could be characterized by the dipole-dipole mechanism. Because the number of inner-sphere
coordinated water molecules (q) of the Tb (cRGDfK-ODO2A-dimer) was determined to be 1.6~2.0,
it was possible that both Tb L and the Tb (cRGDfK-ODO2A-dimer) complexes were in the “closed-
form” conformation at pH 7.4. However, after the binding of the first DPA ion to the Tb
complex, the resulting Tb L(DPA) complex was most likely in the “open-form” conformation,
2
2
2
2
-
L2+
2
2
2
-
making the second DPA ion binding site similar to the first one. Thus, the second stepwise binding
2
-
2-
constant (log K
2
) for the DPA ion to result in the tertiary complex Tb
2
L(DPA)
2
was likely to be
similar to or slightly but not much lower than 5.76. Assuming that the binary and tertiary DPA2-
L(DPA) 2
-
complex, their binding
binding constants of complex I were similar to those of the Tb
2
2
7
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
2
-
strength should be high enough to allow most of the DPA ions remained coordinated at the sub-
micromolar and higher concentrations of complex I for biological studies (vide infra, Figure 5).
α
v
β
3
integrin binding studies. To test whether the cRGDfK moiety in the dinuclear TbIII
complex was able to selectively bind α integrin, receptor binding studies were carried out and
β
v 3
compared with those of c(RGDfK)-FITC (a positive control) and Gly-ODO2A-dimer (a negative
control). It was found that in the absence of added DPA molecules, the low intensity emission spectra
2
+
2+
of the Tb
2
(cRGDfK-ODO2A-dimer) and the Tb (Gly-ODO2A-dimer) were not very much
2
different from noise (Figure S3, supporting information). However, in the presence of added DPA2-
ions, the spectral peak intensities of the complex I (Figure 4, left, cyan) were distinctly greater than
(Gly-ODO2A-dimer)(DPA) 2
-
complex (Figure 4, left, brown).
those of the Tb
2
2
Figure 4. Left: Emission spectra of α
nm, pH 7.4, D-PBS buffer. Green: blank (D-PBS buffer only); cyan: complex I; brown: Tb
ODO2A-dimer)(DPA) ; blue: complex I with 20 fold cRGDfK. Right: Emission spectra of
c(RGDfK)-FITC (green) and FITC (brown), λex = 492 nm, 0.1 mM.
v
β
3
integrin binding with various reagents at 0.1 mM, λex = 278
2
(Gly-
2
-
The luminescence signals for the Tb
to its non-specific binding with α integrin and/or BSA in the blocking buffer solution. If 20 fold of
cRGDfK was added to compete against the complex I for binding with α integrin, the emission
peak intensities were notably decreased to lower than but close to those of the Tb (Gly-ODO2A-
dimer)(DPA) 2-
complex (Figure 4, left, blue). This confirmed the specific binding of complex I to
integrin, and that there was also non-specific binding of complex I with α integrin which was
2
(Gly-ODO2A-dimer)(DPA) 2
2
complex were probably due
β
v 3
β
v 3
2
2
α
v
β
3
β
v 3
-
likely lower than or close to that of the Tb
distinguish if the non-specific binding for the Tb
its binding with the α
Tb (Gly-ODO2A-dimer)(DPA)
was involved. However, given the larger size of the Tb
specific binding results observed above for the complex I, it was possible that the Tb
dimer)(DPA) -part of the complex was most likely involved in the non-specific binding. This non-
2
(Gly-ODO2A-dimer)(DPA) 2
2
complex. It was difficult to
(Gly-ODO2A-dimer)(DPA) 2
-
complex was due to
2
2
v
3
β integrin and/or BSA in the blocking buffer solution, nor which part of the
2
-
2
2
complex, i.e. the Gly-part or the Tb
2
(ODO2A-dimer)(DPA)
-part and the non-
(ODO2A-
2
-part,
2
(ODO2A-dimer)(DPA)
2
2
2
specific binding was likely the reason for the large error bars observed from biological binding
8
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
studies at higher concentrations (vide infra, Figure 5). Figure 4 (right) showed that the cRGDfK-
FITC as the α integrin positive binding control exhibited a characteristic emission band at 517 nm
β
v 3
which was similar to that of the FITC without cRGDfK. Note that the 517 nm emission band of
cRGDfK-FITC was with a similar intensity as compared with that of the complex I at 544 nm, i.e.
5
5
3
.2x10 vs. 3.5x10 photon counts.
Tumor cell binding studies. The human glioblastoma U87MG cells were used as the α
v 3
β
4
0
integrin-containing cells for the selective receptor binding and competition experiments. The bound
2
-
Tb
2
(cRGDfK-ODO2A-dimer)(DPA)
2
complex ions were determined using a previously constructed
2
calibration curve (dynamic range 0.01-12.0 μM, r =0.9994; Figure S4, supporting information). The
binding curve was shown in Figure 5 (left). A K
d
value 19±10 μM was obtained. Comparing with
integrin
value was one or two
those of cRGDfK-DOTA and cRGDfK-IRDye800 on either cell-based or protein-based α
β
v 3
1
9
binding using NIR or fluorescence polarization competition assays, our K
d
orders of magnitudes greater which might be due to different assay techniques employed and/or the
steric hindrance imposed on the resulting complex I bioconjugate with a short spacer, similar to that
reported for the structural analogous FITC-RGD without galactose or polyethylene glycol (PEG)
2
2
0
spacer. Another possible reason for the low binding affinity of complex I with cell surfaces might
be due to its hydrophilic nature which probably bound less strongly than the more lipophilic
IRDye800 dye molecule, particularly at higher concentrations.
Figure 5. Left: binding curve of various added amounts of complex I (x-axis) towards U87MG cells.
Right: competition of various added amounts of cRGDfK (x-axis) against 10 μM complex I toward
U87MG cell binding.
In the cell competition binding experiments, various amounts of cRGDfK were added to
compete against 10 μM complex I for binding with U87MG cells. Figure 5 (right) showed that the
numbers of bound complex I molecules were decreased as the competing cRGDfK concentrations
were increased. A value of IC50 45 μM was obtained which was 4.5 times of the complex I
1
9
concentration used, this ratio was comparable to that found for the cRGDfK-IRDye800. Thus, the
9
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
specific binding of complex I toward U87MG cells via binding to α
v
β integrin was reasonably
3
confirmed.
Tumor cell imaging studies. The results of the confocal targeting U87MG cell imaging studies
were shown in Figure 6.
2
1
1
0
0
.0
.5
.0
.5
.0
*
A
B
C
D
Figure 6. Left: 60X U87MG confocal cell imaging results: (A) Control (D-PBS and U87MG cells
only), (B) complex I (25 μM), (C) cRGDfK-FITC (5 μM), (D) cRGDfK-FITC (5 μM) + complex I
(
25 μM). Right: relative fluorescence intensities. *The statistical P values were all smaller than 0.05
or even lower between any pair of intensity values in A-D.
In the absence of a suitable excitation laser source at 278 nm for the complex I molecule, we
designed a competition protocol using cRGDfK-FITC (λex = 492 nm) to demonstrate the specific
binding of complex I toward U87MG cells (Figure 6, left). Figure 6A showed the images of the
control (D-PBS and U87MG cells only) with the background fluorescence. Figure 6B showed the
images of the complex I with U87MG cells which showed some residual green fluorescence. Figure
6C showed that after cRGDfK-FITC binding with U87MG cells, clear green FITC fluorescence was
observed. Figure 6D was the result of 25 μM of complex I in competition against 5 μM of cRGDfK-
FITC for U87MG cell binding, and it was seen that the fluorescence intensity was reduced.
Quantitative treatment using the control data as background, the relative fluorescence intensities of
the other three studies were all found to be greater than the background (Figure 6, right). The
competition intensity (Figure 6D) was slightly greater than that of complex I with U87MG cells
without competition (Figure 6B) indicating that there was still residual amount of cRGDfK-FITC left
without being competed away. Thus, the specific binding between complex I with U87MG cells were
1
0
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
further demonstrated.
It is worthwhile to note that most luminescence cell binding and imaging experiments were
performed at a complex I concentration range from sub-micromolar to 0.1 mM, and most of the DPA
III
ligand likely remained coordinated to the Tb ion in the complex I as evidenced by the high
luminescence signals and the linear calibration curve responses in the 0.01-12.0 μM concentration
range. Because the excitation at 278 nm is far shorter than the biological window, complex I at
present could most probably be used for preclinical in vitro bioimaging studies, perhaps in
III
III
22
conjunction with Auger electron emitting Ln or In nuclides for theranostic investigations. For
future studies, in addition to reducing the non-specific binding, It would be helpful to design LnIII
chelates or other substances such as nanocomposites with longer wavelength excitation (e.g. λex
>
80 nm or preferably in the near infrared region) 41-43 or X-ray excitation for more convenient,
44
3
deep-penetrating in vivo bioimaging applications.
Conclusion
In conclusion, we have successfully prepared the dinuclear macrocyclic complex Tb
2
(cRGDfK-
2
-
ODO2A-dimer)(DPA)
enhancement at 544 nm as compared to that without adding DPA (λex = 278 nm). Complex I was
found to bind to α integrin and human glioblastoma U87MG tumor cells in both specific and non-
2
or complex I, in situ. At pH 7.4, complex I showed 300 times luminescence
β
v 3
specific modes, via fluorescence spectroscopic and confocal cell imaging competition studies. This
makes complex I and its future optimized derivatives potentially feasible for preclinical in vitro and
in vivo bioimaging applications, particularly in the time-resolved mode.
Experimental Section
Materials and Standard Solutions. Analytical reagent-grade chemicals and buffers, unless
otherwise stated, were purchased from Sigma (St. Louis, MO, USA), Aldrich (Milwaukee, Wl, USA)
or Merck (Darmstadt, Germany) and were used as received without further purification. Disodium
ethylenediaminetetraacetic acid (Na
2
H EDTA) was purchased from Fisher. The ligand m-ODO2A-
2
dimer (1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-oxa-cyclododec-1-ylmethyl]benzene) was
2
1
prepared and purified according to a published method. The ligand cRGDfK-ODO2A-dimer was
prepared and purified as described below. Carbonate-free deionized water (ddH
2
O) was used for all
solution preparations.
The concentrations of the m-ODO2A-dimer and cRGDfK-ODO2A-dimer stock solutions (ca.
0
.01 M) was determined by pH titration using a standard tetramethylammonium hydroxide
II
2+
(
TMAOH) solution (0. 1 M), with and without an excess of added Zn salt (i.e. [Zn ] > 2.0[L]). The
III
concentration of the Tb nitrate stock solution was ca. 0.01 M and were standardized by EDTA
titration using xylenol orange as indicator. The EDTA solution was standardized by titrating a
1
1
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
calcium carbonate primary standard solution (first dissolved in HCl solution) at pH 10 using
calmagite as the indicator.
2
+
2+
Solutions of Tb
2
(m-ODO2A-dimer) , Tb
2
(cRGDfK-ODO2A-dimer) , Tb (Gly-ODO2A-
2
dimer)(DPA) 2-
and complex I were prepared by mixing two equivalents of Tb ions with one
III
2
equivalent of the appropriate ligand with or without additional two equivalents of DPA and the final
mixtures were slowly adjusted to pH 7.4 using the 3-(N-morpholino) propanesulfonic acid (MOPS)
buffer solution (0.10 M). Mass spectrometry confirmed the presence of Tb (cRGDfK-ODO2A-
2
2
+
+
O
10S+, 1654.36, found 1654.70.
dimer) : [M-H] calc. for Tb
2
C
60
H
88
N
16
The 0.1 M (CH NOH solution was prepared by diluting a 25% (CH
3
)
4
3
4
) NOH-aqueous solution
obtained from Aldrich (carbonate-free) and was standardized by using reagent grade primary
standard potassium hydrogen phthalate. A 0.1 M HCl solution was prepared by diluting a reagent
grade HCl solution and standardized by using the standard (CH
3
)
4
NOH solution. A 1.0 M stock
solution of (CH NCl (Aldrich) was prepared and diluted to 0.1 M for each titration to maintain a
3 4
)
constant ionic strength.
Mass Spectral Measurements. Mass spectrometric analyses were performed by the Center for
Advanced Instrumentation and Department of Applied Chemistry at National Chiao Tung
III
University, Hsinchu, Taiwan, R.O.C. Mass spectra of the Tb -cRGDfK-ODO2A-dimer complex
solutions at pH 7.5 were acquired by direct infusion (2 μL). ESI(+)-MS experiments were carried out
using an Impact HD Q-TOF mass spectrometer (Bruker, Germany) equipped with an electrospray
ionization (ESI) source operating in positive ion mode. The parameters of ESI(+) included 4.5kV for
ion spray voltage, 200°C for capillary temperature, and 6 L/min for sheath gas flow rate. The mass
spectra were collected over the mass range of m/z 400 –3000 at a resolving power of 40000. The
collected data were analyzed using Compass DataAnalysis 4.1 (Bruker, Germany).
Luminescence Measurements. Steady-state luminescence measurements were carried out on an
Edinburgh Instruments FSP920 fluorescence system equipped with a 450W xenon arc lamp as the
illumination source (λex 278 nm, λem 544 nm). Emission light was collected into a TMS300 Czerny-
Turner configuration double grating monochromator and detected by a Hamamatsu R928P
photomultiplier tube in the visible wavelength range. Spectra were recorded by use of F900
Fluorescence spectrometer software. Luminescence decay curves were recorded by use of a μF920H
flashlamp (Lamp Frequency: 100Hz) as the excitation source with the multiple-channel single
photon counting mode (MCS). Data points from 0.1-0.2 half-life up to 4.5-5.5 half-lives were fitted
by a nonlinear least-squares iterative technique (Marquardt-Levenberg algorithm) to obtain
exponential decay half-life (τ) values. The average values of 2-3 measurements from different
samples were reported.
1
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
III
III
For the Tb -m-ODO2A-dimer and Tb -cRGDfK-ODO2A-dimer solutions, all solutions were
prepared from stock solutions of Tb(NO
3
)
3
(~0.01M), m-ODO2A-dimer and cRGDfK-ODO2A-
3
+
dimer (~0.01M). The molar ratio of the Tb ion and the m-ODO2A-dimer or cRGDfK-ODO2A-
dimer ligand was ca. 1.99:1.00 (2Tb-L) to make sure that the ligand was in slight excess. The pH of
each solution was adjusted to desired value by adding 0.1 M (CH
3
4
) NOH solution and appropriate
MOPS buffer solutions. All solutions were allowed to equilibrate overnight prior to luminescence
measurements. The corresponding numbers of inner-sphere coordinated water molecules (q) were
2
8
estimated using the lifetime data in H
absorbance ratio and the quantum yields of the Tb
DPA were determined by using a published method and checked by the determination of that of
Rhodamine B in H
O to be 31% which was exactly the same as that reported.36-38 All experiments
2
O and D
2
O with established empirical equations. The
2
(m-ODO2A-dimer) complexes without and with
2
were repeated and data checked at least twice carefully.
Additional luminescence titrations were performed by adding increasing small amounts (i.e. µL
quantities) of standard dipicolinic acid (DPA, 1.0 mM) into a 2.0 mL, 5 µM Tb (cRGDfK-ODO2A-
2
dimer) solution (pH 7.4, MOPS buffer) and the luminescence intensities were recorded after ~5 min
equilibration time, and plotted against the added quantities of DPA. Sigma plot was used to analyze
the titration data.
Synthesis of 3, 5-bis-hydroxymethylnitrobenzene and 3, 5-bis-bromomethylnitrobenzene.
The two compounds were synthesized and characterized using the published procedures. The overall
yield was ≥ 70%.34
Synthesis of 1-oxa-4,7,10-triazacyclododecane (12N
3
O), and tricyclo[8,2,1,0]1-oxa-4,7,10-
triazacyclotridecane (12N O-tricyclic). The two compounds were synthesized and characterized
3
using the methods similar to the published procedures. The overall yield was ≥ 50%.32,33
Synthesis of 1,3-bis(1,4,10-triaza-7-oxacyclododec-1-ylmethyl)-5-nitro benzene (nitro-
3
4
1
2N
dry acetonitrile was added 0.90 g (0.0029 mol) 1,3-bis(bromomethyl)-5-nitrobenzene in 4 mL dry
acetonitrile dropwise under a N gas atmosphere with stirring. The mixture was heated at 50℃ for 6
3
O-tricyclic-dimer). In a rounded bottle flask containing 0.0073 mol 12N O-tricyclic in 10 mL
3
2
h and cooled to room temperature. The white solid product was washed 3 times with dry acetonitrile.
1
Yield 75.8%. NMR (D
2
O, 400 MHz): H: δ2.91, δ3.21, δ3.36, δ3.45, δ3.66, δ3.96, δ5.03, δ5.72,
1
3
δ8.40, δ8.80; C: δ50.12, δ50.73, δ58.13, δ63.41, δ69.01, δ106.11, δ129.83, δ131.07, δ142.29,
δ149.03.
Synthesis of 1,3-bis(1,4,10-triaza-7-oxacyclododec-1-ylmethyl)-5-nitro benzene (nitro-
33
1
2N
3
O-dimer). In a rounded bottle flask containing 0.25 g (0.00048 mol) 12N O tricyclic-dimer in
3
7
ml 2N NaOH was added 10 mL ethanol and the mixture was refluxed at 80℃ for 16 h and cooled
1
3
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
to room temperature. The mixture was extracted with chloroform and the chloroform extract was
dehydrated with MgSO , filtered, and the yellow powder product was obtained by evaporation of the
solvent. Yield 83.3%. NMR (CDCl , 400 MHz): 1H: δ2.56, δ2.62, δ2.72, δ3.57, δ3.63, δ7.50, δ7.97;
3C: δ44.49, δ46.57, δ51.87, δ58.82, δ66.47, δ122.90, δ135.82, δ141.14, δ148.38.
4
3
1
Synthesis of 1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-oxacyclododec-1-ylmethyl]-5-
nitrobenzene hydrochloride salt (nitro-ODO2A-dimer, HCl salt).32,33 In a rounded bottle flask
containing 0.50 g (0.001 mol) nitro-12N O-dimer in 10 mL ddH O was added 5 equivalents
3
2
bromoacetic acid and 6N NaOH to adjust and maintain the solution pH ~ 11 at 65℃ for 5 h. After
the reaction, the solution was cooled to room temperature and was added 6N HCl to adjust pH to ~3.
The solution was evaporated to almost dryness and the white NaCl solid was filtered and washed
with 30 mL methanol. The filtrate was evaporated to dryness and re-dissolved in 3 mL ddH O and
2
the pH was adjusted to 1 by adding 6N HCl. The white product was purified by passing through a
+
cation exchange column in the H form and evaporated to dryness. Yield 67.5%. NMR (D
2
O, 400
1
13
MHz): H: δ2.99, δ3.59, δ3.85, δ3.89, δ3.94, δ7.74, δ8.16; C: δ46.8, δ51.9, δ55.1, δ55.3, δ55.8,
+
N O
12+, 725.79, found 726.3.
δ64.0, δ125.1, δ137.8, δ138.7, δ148.2, δ168.1. [M+H] calc. for C32
H
51
7
(
Supporting Information).
Synthesis of 1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-oxacyclododec-1-ylmethyl]-5-
aniline hydrochloride salt (amine-ODO2A-dimer, HCl salt). In a rounded bottle flask containing
.10 g (0.102 mmol) nitro-ODO2A-dimer in 18 ml ddH O was added 22 mg Pd/C and stirred for 2 h
under a H atmosphere. The solution was filtered with celite and washed several times with ddH O.
O and purified by passing
0
2
2
2
The filtrate was evaporated to dryness, dissolved in small amount of ddH
2
+
through a cation exchange column in the H form. The light yellow product was obtained by
1
evaporation the eluate to dryness. Yield 78.9%. NMR (D
2
O, 400 MHz): H: δ2.98, δ3.59, δ3.81,
1
3
δ3.91, δ3.93, δ7.39, δ7.49; C: δ47.1, δ51.8, δ55.0, δ55.8, δ56.0, δ64.0, δ124.9, δ130.6, δ132.8,
+
N O
10+, 696.4, found 696.3. (Supporting
δ138.2, δ168.5. ESI-MS, [M+H] calc. for C32
Information).
H
33
7
Synthesis of 1,3-bis[4, 10-bis(carboxymethyl)-1,4,10-triaza-7-oxacyclododec-1-ylmethyl]-5-
35
isothiocyanatobenzene, hydrochloride salt (SCN-Bn-ODO2A-dimer, HCl salt). In a rounded
bottle flask containing 0.13 g (0.158 mmol) amine-ODO2A-dimer in 5 mL ddH O was added 0.36
mL (0.47 mmol) thiophosgene (CSCl ) in 5 mL CHCl and stirred for 3 h at room temperature. After
the reaction, the aqueous layer was separated from the organic layer, and the organic layer was
2
2
3
extracted with small amounts of ddH
2
O three times and the combined aqueous layer was evaporated
1
to dryness to obtain the yellow product. Yield 70%. NMR (D
2
O, 400 MHz): H: δ3.07, δ3.66, δ3.83,
1
3
δ3.98, δ4.00, δ7.34, δ7.37; C: δ47.1, δ52.0, δ55.1, δ55.6, δ56.1, δ64.3, δ127.5, δ131.2, δ 132.0,
+
N O
10S+, 738.9, found 738.3.( Supporting
δ136.2, δ137.5, δ168.3. ESI-MS, [M+H] calc. for C33
Information).
H
52
7
1
4
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
35
Synthesis of cRGDfK-ODO2A-dimer. The compounds SCN-Bn-ODO2A-dimer (1.1067 mg,
0
.0015 mmol) and cRGDfK (i.e. cyclo[Arg-Gly-Asp-d-Phe-Lys], 1.812 mg, 0.003 mmol) were
added into 1 mL 0.1 M sodium carbonate-bicarbonate buffer solution (pH 9.0) and the mixture was
stirred for 1 day at room temperature. The reaction completion was checked by a HPLC procedure
and the final cRGDfK-ODO2A-dimer product was purified by a similar FPLC method. Yield 84%.
+
N O
17S+, 1342.562, found 1341.637. (Supporting
MALDI-TOF MS, [M+H] calc. for C60
Information).
H
93
16
Synthesis of Gly-ODO2A-dimer. The compounds SCN-Bn-ODO2A-dimer (3.69 mg, 0.005
mmol) and glycine (3.75mg, 0.05 mmol) were added into 1 mL 0.1 M sodium carbonate-bicarbonate
buffer solution (pH 9.0) and the mixture was stirred for 1 day at room temperature. The reaction
completion was checked by a HPLC procedure and the final product was purified by a similar FPLC
method. (Supporting Information).
Synthesis of cRGDfK-FITC. The compounds cRGDfK (0.905 mg, 0.0015 mmol) and FITC
(
(
1.168mg, 0.003 mmol) were added into 1 mL 0.1 M sodium carbonate-bicarbonate buffer solution
pH 9.0) and the mixture was stirred for 1 day at room temperature. The reaction completion was
checked by a HPLC procedure and the final product was purified by a similar FPLC method. ESI-
+
N O
12S+, 994.07, found 993.5. (Supporting Information).
MS, [M+H] calc. for C48
H
53
10
3
+
2+
Tb /Tb (m-ODO2A-dimer) -DPA binary complex binding studies. Association constants
2
3
+
2+
2-
3+
for Tb and Tb
2
(m-ODO2A-dimer) to DPA ion were determined via titration of Tb and Tb (m-
2
2
+
ODO2A-dimer) from 50 nM up to 5 µM against 100 nM DPA in 0.2 M sodium acetate (pH 7.4,
MOPS buffer, 25°C). A linear fit similar to that of the one-step equilibrium model of Jones and
3
3
3+
2+
Vullev was applied as the concentrations of Tb and Tb
2
(m-ODO2A-dimer) were mostly greater
2
-
than [DPA ], and the binding affinity of the binary complex (K ) was obtained using the following
1
2
7
linear relationship :
log(R/(1-R)) = log(CTb – RCDPA) + log K
1
; R = [TbDPA]eq/([TbDPA]eq + [DPA]eq
)
(3)
3
+
2+
where CTb is the total concentration of the Tb ion or the Tb
the normalized integrated emission intensity. Thus, a plot of log(CTb – RCDPA) vs. log (R/(1 – R)) will
produce a linear fit with a slope of unity and a y-intercept equal to the logarithm of K
2
(m-ODO2A-dimer) complex and R is
1
.
Protein binding experiments. The α
v
3
β integrin (500 ng/mL) was added into the plastic cuvets
for 16 h at 4℃, followed by washing 3 times with Dulbecco's phosphate-buffered saline (D-PBS)
buffer. To the cuvets was then added 3% bovine serum albumin (BSA) blocking buffer and allowed
to stand for 1 h. After washing, to each cuvet was separately added 2.5 mL 100 μM of Tb
2
(cRGDfK-
(without
2
+
2+
2-
ODO2A-dimer) , Tb
2
(Gly-ODO2A-dimer) and Tb
2
(cRGDfK-ODO2A-dimer)(DPA)
2
(Gly-ODO2A-dimer)(DPA) 2
-
and c(RGDfK)-FITC, and each cuvet
and with 2 mM c(RGDfK)), Tb
2
2
1
5
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
solution was allowed to equilibrate for 2 h. Each curvet was then washed 3 times with D-PBS and the
fluorescence emission spectrum was measured. All tests were repeated at least 3 times.
Cell binding experiments. The cells were cultivated in Minimum Essential Medium alpha
(MEM-α) with 10% fetal bovine serum (FBS), 1% penicillin streptomycin (PS) at 5% carbon
6
dioxide/air, 37°C. For the U87MG cell binding experiments, 1×10 U87MG cells were suspended in
the centrifuge tubes and fixed with 4% paraformaldehyde for 10 min at 25℃. The tube solutions
were washed with D-PBS and then added protein-free blocking reagent for 1 h, and washed again
(cRGDfK-ODO2A-dimer)(DPA) 2
-
complex were added and the
with D-PBS. Various amounts of Tb
2
2
tube solutions were allowed to equilibrate for 2 h at 25℃, and then washed with D-PBS to remove
(cRGDfK-ODO2A-dimer)(DPA) 2
-
complex. After counting the number of cells in
the unbound Tb
2
2
3
4
each tube, the cell solutions were transferred into a quartz cell for luminescence spectral
(cRGDfK-ODO2A-dimer)(DPA) 2
-
molecules, using a
measurements to determine the bound Tb
2
2
2
previously constructed calibration curve (dynamic range 0.01-12.0 μM, r =0.9994; Figure S3,
supporting information). In the cell competition binding experiments, various amounts of cRGDfK
were added to compete against 10 μM complex I for binding with U87MG cells. The bound
2
-
Tb
2
(cRGDfK-ODO2A-dimer)(DPA)
2
molecules were determined in the similar ways as that of the
normal cell binding studies. The binding curves and K
d
and IC50 values were obtained using the
GraphPad Prism 5.0 software.
Confocal targeting U87MG cell imaging studies. Cover glass plates were treated with poly-D-
5
lysine and washed with PBS buffer before adding 3×10 U87MG cells for 1 day with the cultivation
medium. The plates were then washed with PBS buffer and the cells were fixed by adding 4%
paraformaldehyde solution (pH 7.4) for 10 min and washed with PBS buffer again. After the addition
of protein-free blocking reagent for another 1 h and washing with D-PBS buffer, the plates were
added four different test reagent solutions to react for 1 h. With D-PBS buffer washing and Hoechst
3
3342 (1:1000) cell nuclei dyeing for 10 min and D-PBS buffer washing, the plates were examined
by the Olmpus FV10i microscope after mounting medium treatment.
Acknowledgements. Financial supports from the Ministry of Science and Technology of Taiwan,
ROC (grant no. NSC-100-2811-M-010-003, NSC102-2627-M-010-001, and MOST103–2627-M-
0
10–001) are acknowledged. Thanks are also due to Dr. J.-J. Li for helpful discussions. Mass
spectrometric analyses were performed by the Center for Advanced Instrumentation and Department
of Applied Chemistry at National Chiao Tung University, Hsinchu, Taiwan, R.O.C. We thank Dr.
Sheng-Cih Huang and Ms. Yun-Ming Li (Mass Laboratory, National Chiao-Tung University,
Hsinchu) for mass spectrometry analysis.
Funding Sources. Ministry of Science and Technology of Taiwan, ROC (grant no. NSC-100-2811-
M-010-003, NSC102-2627-M-010-001, and MOST103–2627-M-010–001)
1
6
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
Supporting Information. Selected NMR spectra with peak assignments, HPLC chromatograms and
mass spectra of various intermediates and SCN-Bn-ODO2A-dimer derivatives, luminescence decay
curves of Tb
2
(cRGDfK-ODO2A-dimer) without and with DPA and estimation of the number of
2
+
coordinated water molecules (q), emission spectra of Tb
2
(cRGDfK-ODO2A-dimer) and Tb (Gly-
2
2
+
2-
ODO2A-dimer) , and calibration curve of Tb
2
(cRGDfK-ODO2A-dimer)(DPA)
2
are shown in the
supporting information.
Abbreviations. DFT: density functional theory; DPA: dipicolinate ion; FITC: fluorescein
isothiocyanate; FRET: fluorescence energy transfer; MOPS: 3-(N-morpholino) propanesulfonic acid;
m-ODO2A-dimer: 1,3-bis[4,10-bis(carboxymethyl)-1,4,10-triaza-7-oxa-cyclododec-1-ylmethyl]
benzene; SCN-Bn-ODO2A-dimer: 1,3-bis[4, 10-bis(carboxymethyl)-1,4,10-triaza-7-oxacyclododec-
1
-ylmethyl]-5-isothiocyanatobenzene.
References
1
2
3
4
5
6
7
8
9
1
1
.
.
.
.
.
.
.
.
.
E. Ruoslahti, M.D. Pierschbacher, Science, 1987, 238, 491-497.
J.S. Desgrosellier, D.A. Cheresh, Nature Rev. Cancer, 2010, 10, 9-22.
J. Folkman, Nature Med., 1995, 1, 27-31.
D. Hanahan, J. Folkman, Cell, 1996, 86, 353-364.
G. Bergers, E. Laura, L.E. Benjamin, Nature Rev. Cancer, 2003, 3, 401-410.
N. Ferrara, R.S. Kerbel, Nature, 2005, 438, 967-974.
E. Pasquier, M. Kavallaris, N. André, Nature Rev. Clin. Oncol., 2010, 7, 455-465.
P. Carmeliet, R.K. Jain, Nature, 2011, 473, 298-307.
H.M. Sheldrake, L.H. Patterson, Curr. Cancer Drug Targets, 2009, 9, 519-540.
0. S. Liu, Mol. Pharmaceutics, 2006, 3, 472-487.
1. F.K. Gaertner, H. Wester, M. Schwaiger, A. Beer, Eur. J. Nucl. Med. Mol. Imaging, 2012, 39,
S126-S138.
1
2. P. Burke, S. DeNardo, L. Miers, K. Lamborn, S. Matzku, G. DeNardo, Cancer Res., 2002, 62,
4
263–4272.
1
1
1
1
1
1
3. R. Stupp, M.E. Hegi, T. Gorlia, et al. Lancet Oncol., 2014, 15, 1100-1108.
4. S. Liu, Bioconjugate Chem., 2015, 26, 1413–1438.
5. S. Liu. Molecular Pharmaceutics, 2006, 3, 472-487.
6. S. Liu. Advanced Drug Delivery Reviews, 2008, 60, 1347-1370.
7. S. Liu. Bioconjugate Chemistry, 2009, 20, 2199-2213.
8. T. Ito, M. Inoue, K. Akamatsu, E. Kusaka, K. Tanabe, S.-i. Nishimoto, Bioorg. Med. Chem.
Lett., 21, 3515-3518.
1
9. H.M. Shallal, S.R. Banerjee, A. Lisok, R.C. Mease, M.G. Pomper, Bioconjugate Chem., 2014,
2
5, 393-405.
1
7
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
2
0. Y. Zheng, S. Ji, A. Czerwinski, F. Valenzuela, M. Pennington, S. Liu, Bioconjugate Chem.,
014, 25, 1925-1941.
2
2
1. C.A. Chang, H.-Y. Lee, S.-L. Lin, C.-N. Meng, T.-T. Wu, Chem. Eur. J. 2018, 24, 6442-6457.
2. J.-J. Li; K.-L. Lan; S.-F. Chang, Y.-F. Chen, W.-C. Tsai, P.-H. Chiang, M.-H. Lin, W.B. Fischer,
Y.-S. Shih, S.-H. Yen, R.-S. Liu, Y.-G. Tsay, H.-E. Wang, C.A. Chang, Bioconjugate Chem.,
2
2
015, 26, 2481–2496.
3. J.-J. Li, S.-F. Chang, I.-I. Liau, P.-C. Chan, R.-S. Liu, H.-E. Wang, C.A. Chang, J. Biomed. Sci.
016, 23, 15.
2
2
2
2
2
4. N. Gahlaut, L.M. Miller, Cytometry A, 2010, 77, 1113-1125.
5. J.-C.G. Bünzli, Chem. Rev. 2010, 110, 2729–2755.
6. N. Sabbatini, S. Perathoner, V. Balzani, B. Alpha, J.-M. Lehn, In Supramolecular Chemistry; V.
Balzani, Ed.; D. Reidel Publishing Co.: Dordrecht, 1987.
2
2
7. C.A. Chang, in Handbook on the Physics and Chemistry of Rare Earths, Vol. 51 (Eds.: J.-C. G.
Bünzli, V. K. Pecharsky), Elsevier, 2017, Ch. 297, pp. 169-299.
8. C.A. Chang, I.-F. Wang, H.-Y. Lee, C.-N. Meng, K.-Y. Liu, Y.-F. Chen, T.-H. Yang, Y.-M. Wang,
Y.-G. Tsay, Dalton Trans. 2012, 41, 14697-14706
2
9. M.L. Cable, J. P. Kirby, H.B. Gray, A. Ponce, Acc. Chem. Res., 2013, 46, 2576-2584.
0. M.L. Cable, J.P. Kirby, D.J. Levine, M.J. Manary, H.B. Gray, A. Ponce, J. Am. Chem. Soc.,
3
2
009, 131, 9562-9570.
1. K.P. Nandre, S.V. Bhosale, K.V.S.R. Krishna, A. Guptac, S.V. Bhosale, Chem. Commun., 2013,
9, 5444-5446.
3
3
3
3
3
4
2. Dischino, E. Delaney, J. Emswiler, G. Gaughan, J. Prasad, S. Srivastava, M. Tweedle, Inorg.
Chem. 1991, 30, 1265-1269.
3. C. Gløgård, R. Hovland, S.L. Fossheim, A.J. Aasen, J.J. Klaveness, Chem. Soc., Perkin Trans.
2
, 2000, 1047–1052.
4. M. Le Baccon, F. Chuburu, L. Toupet, H. Handel, M. Soibinet, I. Déchamps-Olivier, J.-P.
Barbier, M. Aplincourt, New J. Chem. 2001, 25, 1168-1174.
5. D.N. Pandya, N. Bhatt, A.V. Dale, J.Y.; Kim, H. Lee, Y.S. Ha, J.-E. Lee, G.I. An, J. Yoo,
Bioconjugate Chem., 2013, 24, 1356-1366.
3
3
3
3
4
4
6. J. C. Boyer, F. C. Van Veggel, Nanoscale, 2010, 2, 1417-1419.
7. D. Magde, G.E. Rojas, and P. Seybold, Photochem. Photobiol, 1999, 70, 737-744.
8. N. C. Shaner, P. A. Steinbach, R. Y. Tsien, Nat. Methods, 2005, 2, 905909.
9. G. Jones, V.I. Vullev, J. Phys. Chem. A 2002, 106, 8213–8222.
0. S.A. Altman, L. Randers, G. Rao, Biotechnol. Prog., 1993, 9, 671–674.
1. Y.O. Fung, W. Wu, C.-T. Yeung, H-K. Kong, K.K.-C. Wong, W.-S. Lo, G.-L. Law, K.-L. Wong,
C.-K. Lau, C.-S. Lee, W.-T. Wong, Inorg. Chem., 2011, 50, 5517-5525.
4
4
2. D. Arosio, L. Manzoni, E. M. V. Araldi, C. Scolastico, Bioconjugate Chem., 2011, 22, 664–672.
3. S. Cui, D. Yin, Y. Chen, Y. Di, H. Chen, Y. Ma, S. Achilefu, Y. Gu, ACS Nano, 2013, 7, 676-
1
8
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800568
6
88.
4
4. C.-C. Hsu, S.-L. Lin, and C.A. Chang, ACS Appl. Mater. Interfaces, 2018, 10, 7859-7870.
Table of Contents Graphic
III
2-
The cyclic-RGD dinuclear Tb macrocyclic complex Tb
pH 7.4 showed 300 times luminescence (λex = 278 nm) enhancement at 544 nm as compared to that
without adding DPA and was found to selectively bind to α integrin and human glioblastoma
2
(cRGDfK-ODO2A-dimer)(DPA)
2
at
β
v 3
U87MG tumor cells via fluorescence spectroscopic and confocal cell imaging competition studies,
making it potentially feasible for practical bioimaging applications, particularly in the time-resolved
mode.
III
Keywords: Cyclic RGD; Dinuclear Tb macrocyclic complex; Luminescence probe; Integrin-
selective binding; Intensity enhancement
1
9
This article is protected by copyright. All rights reserved.