Single-Molecule Force Measurement Guides the Design of Multivalent Ligands with Picomolar Affinity

Article abstract of DOI:10.1002/anie.201814347

Interaction of multiple entities and receptors, or multivalency is widely applied to achieve high affinity ligands for diagnostic and therapeutic purposes. However, lack of knowledge on receptor distribution in living subjects remains a challenge for rati

Full text of DOI:10.1002/anie.201814347

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Single-Molecule Force Measurement Guides the Design of
Multivalent Ligand with Picomolar Affinity
Authors: Zhen Yang, Sheng Jiang, Feng Li, Yatao Qiu, Jianhua Gu,
Roderic I Pettigrew, Mauro Ferrari, Dale J Hamilton, and
Zheng Li
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: Angew. Chem. Int. Ed. 10.1002/anie.201814347
Angew. Chem. 10.1002/ange.201814347
Link to VoR: http://dx.doi.org/10.1002/anie.201814347
http://dx.doi.org/10.1002/ange.201814347

10.1002/anie.201814347
Angewandte Chemie International Edition
COMMUNICATION
Single-Molecule Force Measurement Guides the Design of
Multivalent Ligand with Picomolar Affinity
Zhen Yang,† Sheng Jiang,†*, Feng Li, Yatao Qiu, Jianhua Gu, Roderic I. Pettigrew, Mauro Ferrari, Dale
J. Hamilton, and Zheng Li,*
Abstract: Interaction of multiple entities and receptors, or
multivalency is widely applied to achieve high affinity ligands for
diagnostic and therapeutic purpose. However, lack of knowledge on
receptor distribution in living subjects remains a challenge for rational
structure design. Here, we developed a force measurement platform
to probe the distribution and distance of the cell surface vascular
endothelial growth factor receptors (VEGFR) in live cells, and used
this to assess the geometry of appropriate linkers for distinct
multivalent binding modes. A tetravalent lead ZD-4, which was
developed from an antitumor drug ZD6474 (Vandetanib) with
combined hybrid binding effects, yielded about 2000-fold
improvement in the binding affinity to VEGFR with IC50 value of 25 pM.
We confirmed the improved affinity by the associated increase of
tumor uptake in the VEGFR-targeting positron emission tomography
(PET) imaging using U87 tumor xenograft mouse model. It showed
that radiolabelled ZD-4 resulted in 12 times increase in the tumor
uptake than radiolabelled parent drug ZD6474. Our study suggests
that combining statistical and chelate effect determined by receptor-
receptor distance on live cells for the magnification of multivalent
binding is effective. The force measurement platform we describe
would be also amenable to other cell-surface multivalent ligand design.
active entities.[3] Although proper architectural design to support
multiple bindings plays a key role for the affinity, length of the
linker largely determines the statistical or the chelate effect in the
interaction (Scheme 1). In terms of multivalent ligands targeting
cell-surface receptors however, lack of knowledge on receptor
distribution and distance in live cells often impedes the rational
selection of linkers and scaffold architectures to fit into the
relevant binding mechanisms for construction of multivalent
ligands.
Scheme 1. Two effects that govern bivalent binding to cell-surface receptors:
(a) statistical effect, arisen from increased local concentration of active entities
tethered by short linker in bivalent ligands; (b) chelate effect, arisen from
simultaneously binding multiple receptors by active entities tethered by long
linker in bivalent.
Multivalency governs many biological interactions and has been
widely applied to synthetic compounds in support of targeted
molecular imaging and drug therapy.[1] A number of multivalent
binding mechanisms could be used to acquire high binding affinity,
including chelate effect, statistical effect, steric stabilization,
subsite binding and receptor clustering.[2] The type of multivalent
binding is largely determined by the selection of linkers tethering
Atomic force microscopy (AFM), by the virtue of its spatial
resolution and its ability to function in aqueous systems, has
emerged as a powerful tool to probe biological systems in their
native state at a nanoscale resolution.[4] In this study, we
developed a force measurement platform, based on AFM to tackle
the challenge in spatial information of cell-surface receptors.
Specifically, the interaction of drug compound, ZD6474
(vandetanib; Caprelsa from AstraZeneca), with its targeting to
vascular endothelial growth factor receptors (VEGFR) in live
human umbilical vein endothelial cells (HUVECs) was studied.
For the force measurement, the molecular probe ZD6474 was first
conjugated with thioctic acid (Figure 1a) and then attached to a
gold-coated AFM tip by means of a strong Au-S bond (Figure 1b).
The ligand-functionalized tip was brought into contact with the
surface of the HUVECs where the binding of ZD6474/VEGFR
took place. The tip was then withdrawn from the surface, pulling
ligand out of the binding pocket. The force required to separate
the ligand from its specific binding site, the “ligand dissociation
force” was determined. As shown in Figure 1, thiol modified
polyethylene glycol (Thiol PEG) was used as a competing
molecule to reduce the number of ZD6474 probe attached to the
AFM tip, which could reduce the number of binding events even
at a molecular level during each contact by changing the ratio
between the ZD6474 probe and the thiol PEG.
[*]
Dr. Z. Yang, Dr. F. Li, Dr. D. J. Hamilton, Dr. Z. Li
Center for Bioenergetics, Houston Methodist Research Institute
6670 Bertner Avenue, Houston, TX 77030
E-mail: Zli@houstonmethodist.org
Dr. S. Jiang, Dr. Z. Li, Y. Qiu
State Key Laboratory of Natural Medicines, School of Pharmacy and
School of Engineering, China Pharmaceutical University, Nanjing,
21009, China
E-mail: jiangsh9@gmail.com
Dr. J. Gu, Dr. M. Ferrari
Department of NanoMedicine, Houston Methodist Research Institute
Dr. R. I. Pettigrew
Engineering Medicine, Houston Methodist Hospital and Texas A&M
University
Dr. M. Ferrari, Dr. D. J. Hamilton, Dr. Z. Li
Department of Radiology, Department of Medicine, Weil Cornell
Medicine, 1300 York Avenue, New York, NY 10065
[†]
These authors contributed equally
Supporting information for this article is given via a link at the end of
the document.
1
This article is protected by copyright. All rights reserved.

10.1002/anie.201814347
Angewandte Chemie International Edition
COMMUNICATION
revealed a quartet distribution of the dissociation force with peak
values at 66 ± 28 pN, 110 ± 12 pN, 156 ± 24 pN and 201 ± 17 pN
(Figure 2f). As expected, the maximum dissociation force was
reduced to 201 ± 17 pN compared to 728 ± 16 pN in the 100% tip.
More importantly, each peak in the quartet distribution
represented a different number of binding events occurring during
the contact measurements, and the difference between each pair
of adjacent peaks was 45 ± 1 pN, which was in the range of the
smallest peak value of 66 ± 28 pN. These results provided strong
evidence that a single binding of ZD6474 to VEGFR was
associated with a force of 45 ± 1 pN.
Figure 1. (a) Structures of the ZD6474 probe and thiol PEG that were used in
the functionalization of ATM tips. (b) Schematic drawing for AFM force
measurement of ZD6474 binding to VEGFR. Tips functionalized by various
concentrations of ZD6474 probe diluted with thiol PEG, such as the 100%-tip
and 5%-tip depicted in (b) were prepared.
For comparison, AFM tips functionalized with 100% ZD6474
probe and 5% ZD6474 probe diluted with thiol PEG were used for
the force measurement. As shown in Figure 2, both AFM height
images and adhesion force mapping images were recorded.
Interestingly, the protruding area in the force mapping, which
represented the specific binding, seemed to be non-uniformly
distributed across the native cell surface when comparing the
height and force mapping images (Figure 2a-b, c-d). The specific
dissociation events were more concentrated along the cellular
periphery (Figure S1 in SI). It indicated more concentrated
VEGFR distribution at the cell periphery, which is in agreement
with independent immunolabeling observations.[5] In view of this
heterogeneous distribution of VEGFR, the AFM force
measurement was mostly focused on the cell periphery where
there appeared to be abundant expression of the receptors.
To decipher the dissociating force of ZD6474/VEGFR,
histograms were analyzed from statistics of the dissociation force
measured in Figure 2c and 2d. For the group associated with the
100%-tip, the histogram revealed a unimodal distribution of the
dissociating forces with a maxima at 728 ± 16 pN (Figure 2e). This
showed an ensemble result of multiple binding events between
ZD6474 and VEGFR that occurred in the contact experiments.
Since the AFM tip was saturated with the probe molecules, this
force value was solely determined by the contact area of the AFM
tip, which in turn confined the accessible number of receptors to
the given contact area. When the probe coverage on the AFM tip
was diluted, the accessible number of ligand molecules in the
given contact area was dramatically reduced while the number of
receptors remained unchanged. This would prominently decrease
the number of binding events involved during each contact
measurement. With the dilution to 5% of ZD6474, the histogram
Figure 2. The force measurement of ZD6474 to VEGFR in live HUVECs.
Measurements that used AFM tips functionalized with 100% (a, c, e) and 5% (b,
d, f) concentration of ZD6474 probe are shown and compared: (a, b), 3D height
images of the cells obtained during the AFM force measurement; (c, d),
Adhesion force mapping presented as 3D images corresponding to the height
images in each group; (e, f), Histograms of ZD6474/VEGFR specific dissociative
forces obtained from statistics of the area presented as the specific binding.
1024 histogram bins were used.
As mentioned above, the contact area of the AFM tip restricted
the number of receptors accessed in the force measurements,
and the number of receptors within the detecting area (NVEGFR
could be readily calculated from the AFM force measurement as
NVEGFR=Fensemble/Fsingle
)
,
where Fensemble was the ensemble of multiple dissociation forces
obtained by the 100%-tip, and Fsingle was the dissociation force of
ZD6474/VEGFR at a single-molecule level that was obtained by
the 5%-tip. As determined from the data in Figure 2, the value of
Fensemble was 728 ± 16 pN and Fsingle was 45 ± 1 pN, and
consequently NVEGFR was calculated to be 16 within the effective
contact area of the AFM tip measuring about 250 nm2. Using a
maximizing minimum algorithm, the maximal distance between
the two adjacent VEGFR on cell periphery in live HUVECs was
determined as about 48 Å (details in the SI). To our best
knowledge, this is the first-time report of receptor distance in live
cells measured by AFM.
2
This article is protected by copyright. All rights reserved.

10.1002/anie.201814347
Angewandte Chemie International Edition
COMMUNICATION
3, and the measured IC50 values were summarized in Table 1. As
designed, a linker with length of ~14 Å that was greatly shorter
than the proximal VEGFR distance in HUVECs, was used to
tether ZD6474 entities and construct a bivalent ZD-2, in which the
statistical effect was believed to be dominated in the binding. A
second bivalent ZD-3 with linker of ~47 Å that corresponded to
the estimated distance of neighboring VEGFR was also
constructed to assess the chelate effect. The IC50 of the
monovalent ZD-1 was determined to be 47 ± 3.0 nM, whereas
IC50 of the bivalent designs ZD-2 and ZD-3 were found to be 1.8
± 0.2 nM and 2.2 ± 0.2 nM, respectively. The deceased IC50 of
ZD-2 and ZD-3 thus represented an approximate 20-fold
improvement in the affinity compared to that of ZD-1. It suggests
the distinction of statistical and chelate effects in the bivalent
binding by using the right lengths of linkers, and both statistical
and chelate effects lead to increased receptor binding.
Table 1. Linker length and IC50 of compounds ZD1-5.
Compounds
Valency
Linker Length[a] [Å]
IC50 [b] [nM]
ZD-1
ZD-2
ZD-3
ZD-4
ZD-5
1
2
2
4
4
n/a
47 ± 3.0
~14
1.8 ± 0.2
~47
2.2 ± 0.2
~47 and ~14
~14
0.025 ± 0.002
0.89 ± 0.08
[a] Linker length is calculated by distance measurement using Chem3D Ultra
8.0. [b] IC50 is the concentration of an inhibitor required to reduce the specific
binding by 50%.
Scheme 2. Chemical structure of compounds ZD1-5. The parent ligand entities
were highlighted in red.
With availability of distinguished linkers to arise different
multivalent bindings, the hybrid tetravalent ZD-4 was constructed
to effectively combine the statistical and chelate effect in one
structure for further optimization of binding affinity. As evaluated
by the displacement/competition assay, ZD-4 showed
unprecedentedly increased binding affinity with measured IC50 as
low as 0.025 ± 0.002 nM, which presents almost 2000-fold
improvement over ZD-1, and nearly 100-fold increase compared
to bivalent ZD-2 and ZD-3. As a comparison, another tetravalent
ZD-5 that only used short linkers of ~14 Å tethering the active
entities, in which only a statistical effect would be expected to be
visited in the multivalent binding, was also study. Its IC50 was
revealed 0.89 ± 0.08 nM, which showed only a 2-fold increase
compared to the bivalent ligands and a 35-fold less potent than
the hybrid ZD-4. The comparison of ZD-4 with ZD-5 strongly
supports our proposition that optimized multivalent binding can be
achieved by hybrid multivalent design.
Furthermore, the improved binding affinity of ZD-4 to VEGFR
was investigated using in vivo VEGFR-targeting positron
emission tomography (PET) imaging of U87 glioblastoma-bearing
mice. Increased tumor uptake of the imaging tracer will be
associated with improved VEGFR targeting.[7] Upon radiolabelled
with 64Cu, the PET radiotracer 64Cu-ZD4 (Scheme S1 in SI) with
specific activity of 0.3-0.5 MBq/nmol was compared with its parent
drug ZD6474 derived radiotracer 64Cu-ZD6474 for their tumor
uptake. Specifically, microPET/CT scans were performed on a
U87 xenograft mouse model at 2 h, 6 h and 24 h after ~3.7 MBq
(100uCi) injection of 64Cu-ZD4 or 64Cu-ZD6474 (n=6).
Representative whole-body 3D images at 24 h post injection (p.i.)
Figure 3. Competition binding curves of compounds ZD1-5 to HUVECs. In the
binding assay, 64Cu-ZD6474 was used as binding probe and the ZD compound
from the design was individually supplemented as competition agent. The IC50
was calculated from the competitive binding curve. Data shown are represented
as mean ± SD (n = 6 per group).
With the actual distribution information of the receptors in hand,
the lengths of the linkers can be rationally selected to design
statistical and chelate bivalent binding. Scheme 2 presented a
series of multivalent ligands that were used in this study of hybrid
multivalent binding. All these compounds were prepared and
characterized accordingly (see SI for details), and their binding
affinity to VEGFR was measured by the IC50 values required to
reduce the binding of the parent entity, ZD6474 (ZD-1), in the
binding assay of ligand to cell surface receptors as previously
described.[6] The inhibitory binding curve was presented in Figure
3
This article is protected by copyright. All rights reserved.

10.1002/anie.201814347
Angewandte Chemie International Edition
COMMUNICATION
are shown in Figure 4a. High tumor uptake of 64Cu-ZD4 was
observed as indicated by the arrow, and revealed 64Cu-ZD4 to be
a suitable PET probe for in vivo tumor VEGFR imaging. In
comparison, the tumor uptake of 64Cu-ZD6474 was revealed by
the PET imaging to be much less. For each PET scan, regions of
interest (ROIs) quantification analysis of tumor uptake was
performed to obtain an imaging ROI–derived percentage of the
injected radioactive dose per gram of tissue (%ID/g). From ROI
analysis of decay-corrected PET images, the tumor uptake of
64Cu-ZD4 probe was 6.16 ± 0.30 %ID/g as early as 2 h p.i., and
increased to 7.24 ± 0.23 %ID/g and 7.90 ± 0.39 %ID/g at 6 h and
24 h p.i., respectively. For the parent monomer radiotracer, the
tumor uptake of 64Cu-ZD6474 was only 0.30 ± 0.11 %ID/g at 2 h,
0.66 ± 0.04 %ID/g at 6 h, and 0.46 ± 0.14 %ID/g at 24 h p.i.. The
following histological study confirmed the abundant expression of
VEGFR in the U87 tumor, which accounted for the high uptake of
VEGFR-specific radiotracers. The in vivo PET imaging clearly
showed that the improved binding of hybrid ZD-4 led to a
significantly increased tumor uptake of 64Cu-ZD4, which was 12-
times higher than that of radiolabeled parent monomer 64Cu-
ZD6474 (P < 0.0001). The high tumor uptake and favorable
pharmacological properties, suggests 64Cu-ZD4 a suitable PET
radiopharmaceutical. Promising radionuclide-based therapeutic
products could be produced by incorporation into developed
agents with appropriate radionuclides.
platform that was capable of probing the actual distribution of cell-
surface receptors could greatly facilitate the rational design of
multivalent drugs.
Acknowledgements
This work was supported by the Houston Methodist Research
Institute. Additional support was acquired from the following
sources: DoD BCRP Breakthrough Award (BC141561P1),
George and Angelina Kostas Research Center for Cardiovascular
Nanomedicine award, the Ernest Cockrell Jr. Presidential
Distinguished Endowed Chair, the National Institutes of Health
(U54CA143837, U54CA151668, U54CA210181), Elaine and
Marvy A. Finger Distinguished Endowed Chair, China National
Natural Science Foundation (81773559, 21472191), the Double
First-Class University Project (CPU2018GY03, CPU2018GY24),
and the Project of State Key Laboratory of Natural Medicines,
China Pharmaceutical University (SKLNMZZRC201810). We
thank the electron microscopy and small animal imaging core
facility of Houston Methodist Research Institute. Dr. Zhong Xue is
acknowledged for providing algorithm calculation support.
Keywords: atomic force microscopy • ligand design •
multivalency • PET imaging • VEGFR
[1]
a) M. Mammen, S.-K. Choi, M. Whitesides George, Angew. Chem. Int.
Ed. 1998, 37, 2754-2794; b) C. Fasting, C. A. Schalley, M. Weber, O.
Seitz, S. Hecht, B. Koksch, J. Dernedde, C. Graf, E.-W. Knapp, R. Haag,
Angew. Chem. Int. Ed. 2012, 51, 10472-10498; c) L. Xu, J. S. Josan, J.
Vagner, M. R. Caplan, V. J. Hruby, E. A. Mash, R. M. Lynch, D. L. Morse,
R. J. Gillies, Proc. Natl. Acad. Sci. 2012, 109, 21295-21300; d) S. Bhatia,
L. C. Camacho, R. Haag, J. Am. Chem. Soc. 2016, 138, 8654-8666; e)
S. Cecioni, A. Imberty, S. Vidal, Chem. Rev. 2015, 115, 525-561; f) G. V.
Dubacheva, T. Curk, R. Auzély-Velty, D. Frenkel, R. P. Richter, Proc.
Natl. Acad. Sci. 2015, 112, 5579; g) M. Honer, L. Gobbi, L. Martarello, R.
A. Comley, Drug Discov. Today 2014, 19, 1936-1944; h) M. Srinivasarao,
C. V. Galliford, P. S. Low, Nat. Rev. Drug Discov. 2015, 14, 203-219; i)
L. Lang, W. Li, N. Guo, Y. Ma, L. Zhu, D. O. Kiesewetter, B. Shen, G.
Niu, X. Chen, Bioconjugate Chem. 2011, 22, 2415-2422.
[2]
[3]
a) J. E. Gestwicki, C. W. Cairo, L. E. Strong, K. A. Oetjen, L. L. Kiessling,
J. Am. Chem. Soc. 2002, 124, 14922-14933; b) L. L. Kiessling, J. E.
Gestwicki, L. E. Strong, Curr. Opin. Chem. Biol. 2000, 4, 696-703.
a) V. M. Krishnamurthy, V. Semetey, P. J. Bracher, N. Shen, G. M.
Whitesides, J. Am. Chem. Soc. 2007, 129, 1312-1320; b) M.-H. Li, H.
Zong, P. R. Leroueil, S. K. Choi, J. R. Baker, Bioconjugate Chem. 2017,
28, 1649-1657.
Figure 4. Whole-body microPET/CT image of U87 glioblastoma-bearing mice.
(a) Representative PET/CT images of the tumor-bearing mice were displayed
24 h p.i. of 64Cu-ZD6474 and 64Cu-ZD4 radiotracers. Tumor is indicated by a
white arrowhead. (b) Comparison of decay-corrected ROI analysis of 64Cu-
ZD6474 and 64Cu-ZD4 in tumor. Data shown are represented as mean ± SD (n
= 6 per group, P < 0.0001). (c) The imaged U87 glioblastoma xenograft was
histologically studied using haemotoxylin and eosin (H&E) and VEGFR2-
antibody immunohistochemistry (VEGFR2 IHC) staining.
[4]
a) C.-K. Lee, Y.-M. Wang, L.-S. Huang, S. Lin, Micron 2007, 38, 446-
461; b) X. Huang, X. Li, Q. Wang, J. Dai, J. Hou, L. Chen, Biomaterials
2013, 34, 6139-6146; c) D. J. Müller, M. Krieg, D. Alsteens, Y. F. Dufrêne,
Curr. Opin. Biotechnol. 2009, 20, 4-13.
[5]
[6]
N. Almqvist, R. Bhatia, G. Primbs, N. Desai, S. Banerjee, R. Lal, Biophys.
J. 2004, 86, 1753-1762.
a) Y.-s. Kim, F. Li, B. E. O’Neill, Z. Li, Bioconjugate Chem. 2013, 24,
1937-1944; b) F. Li, S. Jiang, Y. Zu, D. Y. Lee, Z. Li, J. Nucl. Med. 2014,
55, 1525-1531.
In summary, we successfully demonstrated an AFM force
measurement platform that was applied to guide the efficient
design of VEGFR-targeting hybrid ZD-4 from drug ZD6474.
Compound ZD-4 exhibited picomolar receptor affinity, increased
tumor uptake and favorable pharmacological properties, which
showed high potential in diagnostic imaging and therapeutic
treatment of VEGFR-related diseases. Our study suggested a
useful strategy to design ligands with dramatic increased binding
affinity from known drug by effectively combining various
multivalent binding mechanisms. And the force measurement
[7]
a) J. Folkman, New England J. Med. 1971, 285, 1182-1186; b) N. Ferrara,
H.-P. Gerber, J. LeCouter, Nat. Med. 2003, 9, 669-676.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.201814347
Angewandte Chemie International Edition
COMMUNICATION
Entry for the Table of Contents
Layout 2:
COMMUNICATION
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
Efficient hybrid multivalent design leads to highly increased tumor uptake.
5
This article is protected by copyright. All rights reserved.