Exploring Structural Parameters for Pretargeting Radioligand Optimization

Article abstract of DOI:10.1021/acs.jmedchem.7b01108

Pretargeting offers a way to enhance target specificity while reducing off-target radiation dose to healthy tissue during payload delivery. We recently reported the development of an ¹⁸F-based pretargeting strategy predicated on the inverse electron demand Diels-Alder reaction as well as the use of this approach to visualize pancreatic tumor tissue in vivo as early as 1 h postinjection. Herein, we report a comprehensive structure: pharmacokinetic relationship study of a library of 25 novel radioligands that aims to identify radiotracers with optimal pharmacokinetic and dosimetric properties. This investigation revealed key relationships between molecular structure and in vivo behavior and produced two lead candidates exhibiting rapid tumor targeting with high target-to-background activity concentration ratios at early time points. We believe this knowledge to be of high value for the design and clinical translation of next-generation pretargeting agents for the diagnosis and treatment of disease.

Full text of DOI:10.1021/acs.jmedchem.7b01108

Subscriber access provided by LUNDS UNIV
Article
Exploring Structural Parameters for Pretargeting Radioligand Optimization
Jan-Philip Meyer, Paul Kozlowski, James Jackson, Kristen M. Cunanan,
Pierre Adumeau, Thomas R. Dilling, Brian M. Zeglis, and Jason S. Lewis
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01108 • Publication Date (Web): 31 Aug 2017
Downloaded from http://pubs.acs.org on September 3, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 60
Journal of Medicinal Chemistry
Exploring Structural Parameters for Pretargeting
1
2
3
4
5
6
7
8
Radioligand Optimization
1
1
1
2
3
JanꢀPhilip Meyer , Paul Kozlowski , James Jackson , Kristen M. Cunanan , Pierre Adumeau ,
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1,3,4,6*
1,5,6*
Thomas R. Dilling , Brian M. Zeglis
, and Jason S. Lewis
1
Department of Radiology, Memorial Sloan Kettering Cancer Center, New York, NY, USA.
2
Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New
York, NY, USA.
3
Department of Chemistry, Hunter College of the City University of New York, New York, NY,
USA
4
Ph.D. Program in Chemistry, Graduate Center of the City University of New York, New York,
NY, USA
5
Program in Molecular Pharmacology, Memorial Sloan Kettering Cancer Center, New York,
NY USA.
6
Departments of Radiology and Pharmacology, Weill Cornell Medical College, New York, NY,
USA.
Keywords: in vivo pretargeting, bioorthogonal click chemistry, tetrazine radioligand,
pharmacokinetics, PET imaging.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 60
Abstract
Pretargeting offers a way to enhance target specificity while reducing offꢀtarget radiation
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8
dose to healthy tissue during payload delivery. We recently reported the development of an Fꢀ
based pretargeting strategy predicated on the inverse electron demand DielsꢀAlder reaction as
well as the use of this approach to visualize pancreatic tumor tissue in vivo as early as 1 h postꢀ
injection. Herein, we report a comprehensive structureꢀ pharmacokinetic relationship study of a
library of 25 novel radioligands that aims to identify radiotracers with optimal pharmacokinetic
and dosimetric properties. This investigation revealed key relationships between molecular
structure and in vivo behavior and produced 2 lead candidates exhibiting rapid tumor targeting
with high targetꢀtoꢀbackground activity concentration ratios at early time points. We believe this
knowledge to be of high value for the design and clinical translation of nextꢀgeneration
pretargeting agents for the diagnosis and treatment of disease.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Introduction
1
ꢀ6
7
In vivo pretargeting for diagnostic and therapeutic applications has emerged over the last
three decades as a powerful technology to enhance target specificity and reduce offꢀtarget
2,8
effects. Generally speaking, pretargeting strategies strive to combine the inherent advantages
of macromolecular targeting vectors and small molecules, specifically high target specificity
1
,4
and short organ and tissue residence times, respectively. To achieve this, the targeting vector
is administered first and allowed to accumulate at the target site and clear from offꢀtarget organs
prior to the injection of a small effector molecule carrying the payload of interest (e.g.
4,5,9
radionuclide; Fig. 1).
In order to enable their in vivo recombination, both entities are
ACS Paragon Plus Environment

Page 3 of 60
Journal of Medicinal Chemistry
10,11
equipped with complementary functionalities that enable an in vivo ligation reaction.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Appropriate pairs of reaction partners that have been employed in in vivo pretargeting
1
2
13
approaches include streptavidinꢀbiotin , complementary oligonucleotide strands , and click
1
0,14,15
chemistryꢀbased reaction pairs.
While strategies based on streptavidinꢀbiotin have shown
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
16
somewhat deflating outcomes in the clinic , the use of a bispecific, CEAꢀtargeting antibody in
1
7
combination with a radiolabeled hapten peptide has shown very promising clinical results.
One of the newer members of the click chemistry toolbox, the inverse electronꢀdemand
DielsꢀAlder (IEDDA) reaction between transꢀcyclooctene (TCO) and tetrazine (Tz) has proven
2ꢀ5,7,10,11
particularly well suited for in vivo pretargeting.
The IEDDA ligation is selective and
bioorthogonal, but its principal advantage for pretargeting compared to other click reactions —
1
8
14
such as the Staudinger ligation or the strainꢀpromoted alkyneꢀazide cycloaddition (SPAAC)
— lies in its speed. To wit, the firstꢀorder rate constants of the Tz/TCO ligation lie in the realm
4
5
ꢀ1 ꢀ1
of 10 ꢀ10 M s , orders of magnitude faster than the rates of the Staudinger and SPAAC
ꢀ1 ꢀ1
ꢀ1 ꢀ1
11
reactions (0.002 M s and 0.07 M s , respectively). The in vivo feasibility of the IEDDA
reaction between a TCOꢀmodified monoclonal antibody (mAb) and a radiolabeled Tz has been
demonstrated by various groups using a wide range of antigenꢀtargeting immunoconjugates and
1
11
64
99m
18
tetrazines labeled with an array of radionuclides for imaging (including In, Cu, Tc, F,
68
11
1,4,5,19ꢀ21
177
7
Ga, and C)
In terms of pretargeted positron emission tomography (PET) imaging, Zeglis, et al.
presented promising results in 2013 in pretargeting experiments using the gpA33ꢀtargeting mAb
and therapy ( Lu).
6
4
4
huA33ꢀTCO and a Cuꢀlabeled tetrazine radioligand. Shortly thereafter, our laboratories
64
developed a secondꢀgeneration Tz for Cuꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. (a) Schematic illustration of the pretargeting approach: a macromolecular targeting
vector — in our case an antibodyꢀTCO conjugate — is injected first and allowed to reach the
target site while clearing slowly from systemic circulation. After a specific accumulation time,
the small molecule effector probe (in this case a radiolabeled tetrazine probe) is administered
systemically and undergoes bioorthogonal click reaction with the TCOꢀgroups of the
immunoconjugate at the target site. (b) Modular chemistry approach for radioligand design:
radioligands consisted of a Tz moiety for in vivo click chemistry, a linker for altering the
68
3+
biodistribution, and a chelator for the attachment of the positronꢀemitting metal ions Ga and
1
8
2+
[ F]ꢀAlF .
based pretargeted PET imaging applications by integrating the sarcophagine chelator system
3
18
into the radioligand structure. At the same time, our laboratories demonstrated that an Fꢀ
labeled Tzꢀbased radioligand in combination with the carbohydrate antigen 19.9 (CA19.9)ꢀ
22
targeting fully human mAb 5B1ꢀTCO allowed for the successful PET imaging of
1
subcutaneous (s.c.) pancreatic cancer xenografts as early as 1 h postꢀinjection (p.i.) . Critically,
1
8
this new pretargeting approach utilized the shortꢀlived radionuclide F (t = 109 min),
1
/2
ACS Paragon Plus Environment

Page 5 of 60
Journal of Medicinal Chemistry
resulting in only a fraction of offꢀtarget radiation doses to healthy tissues compared to directly
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
24
89
labeled immunoconjugates with longꢀlived isotopes ( I or Zr, t > 3 days). Despite clear
1
/2
delineation of tumor tissue at early imaging time points demonstrating the general feasibility of
this approach, the relatively low tumorꢀtoꢀbackground activity concentration ratios at even 4 h
p.i., such as tumorꢀtoꢀintestines (1.6 ± 0.1) and tumorꢀtoꢀkidney (1.8 ± 0.4) ratios, inspired us to
undertake a thorough investigation into the fundamental relationships between molecular
structure, pharmacokinetics, and pretargeting performance.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
This firstꢀofꢀitsꢀkind structureꢀpharmacokinetics relationship (SPR) study was further fueled
by the increased popularity of IEDDA pretargeting strategy. However, current approaches lack
fundamental insight into the relationship between physicochemical properties and pretargeting
performance. In order to address those issues, the study at hand was designed with two main
objectives: (1) to identify a radiopharmaceutical lead candidate suitable for clinical
development and (2) to generate experimental evidence for a rational understanding of how
molecular parameters such as overall molecular net charge, distribution coefficient, plasma halfꢀ
life (PHL), and stability influence the in vivo performance of small molecule radioligands in
pretargeting systems.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 60
Linker
PEG (x = 7,11) AA (D, H, K, R)
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Tz
ChelatorPrecursor
Tracer
x
18
1
1
1
1
ꢀ
ꢀ
7
ꢀ
K
ꢀ
NODA
NOTA
NOTA
NOTA
15
16
17
18
[ F]15
1
8
[ F]16
18
[ F]17
1
8
11
ꢀ
[ F]18
18
[
F]19
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
2
ꢀ
ꢀ
ꢀ
K
NOTA
NOTA
NOTA
19
20
21
6
8
[
Ga]19
18
8
[
F]20
KꢀK
6
[
Ga]20
18
[
F]21
KꢀKꢀK
6
8
[ Ga]21
68
2
2
ꢀ
11
KꢀKꢀK
ꢀ
NODA
NOTA
22
23
[ Ga]22
1
8
[ F]23
18
[
F]24
3
11
ꢀ
NOTA
24
6
8
[ Ga]24
18
4
4
7
7
ꢀ
ꢀ
NODA
NOTA
25
26
[ F]25
6
8
[ Ga]26
18
[
F]27
4
11
ꢀ
NODA
27
6
8
[ Ga]27
18
[
F]28
4
4
4
11
7
ꢀ
NOTA
NODA
NODA
28
29
30
6
8
[
Ga]28
18
K
H
[ F]29
1
8
1
8
[
F]30
11
6
[
Ga]30
8
4
4
11
11
R
D
NODA
NOTA
31
32
[ F]31
18
[ F]32
Table 1. All 25 radioligands were employed in the first characterization
process (green). Based on those results, 15 radioligands were selected for the
next step of testing to investigate their performance in pretargeting
experiments (blue). Finally, radioligands [ F]27 and [ Ga]27 were identified
as lead compounds based on their overall tumoral uptake and tumorꢀtoꢀNT
activity concentration ratios (red).
18
68
ACS Paragon Plus Environment

Page 7 of 60
Journal of Medicinal Chemistry
1
2
Tracer library and SPR study design
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The synthesis of the radioligands library as the first step of this study was based on previously
1
,2,4
reported protocols.
New reaction routes and radiolabeling procedures developed within this
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
study are described in the supporting information (Supplementary Sections 2 and 3). Overall,
the radioligands were designed to display structural variation in order to cover a broad spectrum
of physicochemical properties, thereby enabling the study of the relationship between structure
and in vivo behavior (Tab. 1). Each radioligand is composed of 3 different structural building
blocks (Tab. 1; Fig. 2a). First, a Tz component (1–4) was selected for in vivo click chemistry.
Second, a linker moiety consisting of polyethylene glycol [PEG (5) or PEG (6)], amino acids
7
11
(AA) [AA = lysine (K, 7), histidine (H, 8), aspartate (R, 9), and arginine (D, 10)], or a
combination of both is attached. Finally, a bifunctional chelator, either 1,4,7ꢀtriazacyclononaneꢀ
1,4ꢀdiacetic acid (NODA, 11,12) or 1,4,7ꢀtriazacyclononaneꢀ1,4,7ꢀtriacetic acid (NOTA,
18
68
1
3,14), was introduced, allowing for the installation of F and Ga radionuclides. Tz moieties
1
–4 were selected based on their previously reported stability and reaction kinetics with TCO to
4,5,21,23
ensure a wide range of properties.
The use of PEG linkers to modulate in vivo PK of small
molecules has previously been reported, including accelerated nonꢀtarget organ clearance as
1,2,24
well as increased renal clearance.
We included them into our study in order to investigate
their impact on radiotracer PK alone or in combination with AA linkers (which, to the best of
our knowledge, have not yet been systematically reported in any SPR study). The amino acids
lysine, arginine, histidine, and aspartate were regarded as useful structural components to
significantly influence in vivo behavior and PK parameters. It was reasoned that their charged
side chains should have a measureable effect on tracer PK, and would further allow us to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 60
establish a correlation between molecular net charge and PK parameters. Both bifunctional
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
5,26
chelator moieties NODA and NOTA currently find broad application in preclinical
and
27
clinical research for the radiolabeling of biological macromolecules and small molecule
targeting probes. Precursors 15-32 were synthesized in good overall chemical yields (18ꢀ37%)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
via 3–8ꢀstep syntheses, depending on the starting materials (Fig. 2b).
In-situ carboxylic acid activation
and amide bond formation
Boc
H
2
N
NH
O
n=7,11 (5,6)
(1-1.3 eq.)
Non-activated carboxylic acids
H
N
AND/OR
O
OH
O
R2
H
2
N
R
(7-10)
Linker
R2 = NH , COOH
O
OH
2
(
0.8-4 eq.)
N
N
N
N
N
N
N
N
1
. DMF, EDC (2.0 eq.)
HOBt (2.0 eq.), RT, 4-8 h
>75%.
Amine-isothiocyanide addition
1
or
2
. Boc-deprotection (if applicable)
Amide coupling
Simple amide coupling
NODA/NOTA-NCS/NH
2
(11-14)
Boc
Precursors 15-32
H
2
N
NH
O
if NH
DMSO, RT, TEA (1.5 eq.),
h, >80%
2
+ NCSChelator :
n=7,11 (5,6)
NHS-activated ester
(1-1.3 eq.)
1
AND/OR
O
H
H
N
if CO H + NH
DMF, EDC (2.0 eq.)
HOBt (2.0 eq.), RT, 4-8 h
>68%.
:
2, chelator
O
O
H
2
N
R
(7-10)
N
2
N
R2
Linker
O
O
O
OH
R2 = NH , COOH
2
(0.8-4 eq.)
N
N
N
N
1. DMSO, TEA (1.5 eq.), RT,
30-45 min, >90%.
N
N
N
N
4
2. Boc-deprotection (if applicable)
Scheme 1. Reaction scheme exemplifying the synthesis of precursor molecules using Tz 1 and
Tz 4 via two different synthetic routes, both of which furnish final compounds in good overall
yields and high chemical purities. For Tz starting materials such as Tz 1, in situ activation of the
carboxylic acid group under mild conditions was performed using 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminoꢀ
propyl)carbodiimide (EDC) and 1ꢀhydroxybenzotriazole (HOBt), enabling subsequent amide
coupling with the terminal amine of an appropriate linker moiety. NHSꢀactivated, commercially
available starting materials such as Tz 4 could be coupled to the linker moiety under weakly
basic conditions with high conversion rates and without the need for further activation reagents.
After TFAꢀmediated deprotection of the terminal Bocꢀgroup (if applicable), terminal amines (or
carboxylic acids) were reacted with a bifunctional NODA or NOTA chelator construct that
possessed a complementary functional group for chemical ligation.
ACS Paragon Plus Environment

Page 9 of 60
Journal of Medicinal Chemistry
18
68
68
All Fꢀ and Gaꢀlabeled [t ( Ga) = 68 min] Tzꢀderived radioligands were furnished in high
1
/2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8
68
radiochemical yields [RCYs, >55% ( F); >83% ( Ga)], with specific activities (SAs) of >19
MBq/nmol and high radiochemical and radionuclidic purity (Supplementary Section 3). For in
vitro analysis, purified tracers were incubated in human serum at 37 °C to analyze their stability
under physiological conditions. Further, the distribution coefficient (logD, n = 3) of all
radioligands in a 1:1 mixture of PBS:1ꢀoctanol was determined using the shakeꢀflask method.
Subsequently, the tracers (4–8 MBq, 0.5–1 nmol) were injected into healthy athymic nude mice
via the lateral tail vein. At various time points (between 2–120 min p.i.), blood was drawn via
the lateral tail vein or saphenous vein (n = 4) to calculate PHLs (all radioligands, n = 1) and
plasma stabilities (n = 3). Tracers (11–14 MBq, 0.8–1.3 nmol) were injected into healthy
athymic nude mice, and general biodistribution experiments for all tracers were performed
using serial PET imaging hourly between 1–4 h p.i., unless stated otherwise (n = 4). Decayꢀ
corrected PET imaging data and reconstructed 3D images of the tracer distribution were then
used to determine radioactivity concentrations (given as percent injected dose per gram, %ID/g)
in the kidney and large intestine (quantitative ROI analysis). Additional ex vivo organ uptake
values were determined for selected compounds and were found to be in line with imageꢀ
derived uptake data, justifying the use of imageꢀderived organ uptake data for the majority of
compounds in order to reduce the number of animals euthanized. Data analysis led to the
selection of 15 radioligands (Tab. 1, highlighted in blue) that were further tested in in vivo
pretargeting experiments using athymic nude mice bearing subcutaneous pancreatic ductal
adenocarcinoma (PDAC) xenografts (n = 4). A TCOꢀmodified immunoconjugate of the
CA19.9ꢀtargeting antibody 5B1 (5B1ꢀTCO; 1.3 nmol, 200 µg per mouse) was injected 72 h
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 60
prior to the injection of the small molecule radioligands (1.3–1.6 nmol, 0.6ꢀ1.4 µg, 1–1.2 eq.).
PET images were acquired hourly between 1–4 h p.i., unless stated otherwise. In addition, ex
vivo biodistribution experiments were conducted for selected tracers in order to obtain
quantitative information of tracer distribution in up to 14 organs. Ultimately, this investigation
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
18
68
identified two radioligands — [ F]27 and [ Ga]27 — as the most promising lead compounds
Tab. 1, highlighted in red).
(
Physicochemical and pharmacokinetic properties of Tz-derived PET tracer
Structural, in vitro, and pharmacokinetic (PK) data obtained for all 25 tracers were used to
investigate potential correlations between physicochemical parameters. Tracer stabilities (given
as %ꢀintact, n = 3) in human serum (incubation for 4 h at 37 °C, Fig. 3c) ranged from >90%
1
8
68
([ F]23) to 54 ± 7% ([ Ga]24) (Fig. 3a, Supplementary Section 4). Considering all of the
structural elements of the radioligands, we reasoned that both the tetrazine moiety and the metal
complex would have an impact on in vitro (and presumably in vivo) stability. However, we
found that instability was due primarily to the decomposition of the tetrazine moiety, and that
the radioactive metal complexes were stable over the course of our experiments. The majority
of the radioligands did not show any elevated protein binding (<2% of total radioactivity,
18
18
68
68
Supplementary Section 4), with the exception of [ F]20, [ F]21, [ Ga]21, [ Ga]22, and
1
8
[
F]29, each of which exhibited up to 16% proteinꢀbound radioactivity. This result may be
explained by the high lysine content residues and positive charge of these radioligands,
2
4
prompting electrostatic interactions between the tracers and plasma proteins.
ACS Paragon Plus Environment

Page 11 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Correlation diagrams showing how structural and physicochemical parameters (e.g.
stability, linker, molecular charge, logD, PHL) influence each other. (a) Tracer stabilities after a
3
h incubation time in human serum are shown after tracers were divided into groups with the
68
same linker and radionuclide. Serum stabilities were generally >60%, except for Gaꢀlabeled
compound 46. (b) LogD values (median shown, n = 3) summarized as a group diagram. (c)
distribution coefficients in relation to molecular charge (under physiological conditions)
showed a significant correlation. (dꢀf) in vivo PHLs plotted as a group diagram (d), in
dependence of molecular net charge (e), as well as plotted against logD (f). Lysineꢀcontaining
compounds showed overall fast clearance from circulation, with PEGylated compounds having
3ꢀtimes longer PHLs. Interestingly, a positive correlation between a tracer’s PHL and its logD
value was found.
Not surprisingly, a significant negative correlation between the logD value of a tracer and its
overall molecular charge under physiological conditions was found (pꢀvalue <0.0001,
Supplementary Sections 4 and 9). Generally speaking, the closer the net charge of a tracer is to
0
, the higher its partition coefficient. However, it is not quite that simple. For instance,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 60
radioligands possessing the same overall molecular charge but a different number of formally
charged functional groups could still show significant differences in their distribution
coefficients. One explanation could be enhanced solvation by water molecules through chargeꢀ
dipole interactions: if the molecule exhibits a greater number of formal charges, it would lead to
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
28
a higher degree of solvation. For example, radioligands with a net charge of 0 containing a
PEG linker (or no linker) exhibited logD values between ꢀ1.0 and ꢀ1.5, while radioligands with
the same overall charge but with lysine residues in the linker displayed logD values between ꢀ
1
.8 and ꢀ3.1. PHLs for the radioligands correlated (pꢀvalue <0.0001, Supplementary Sections 5
and 9) with their net charge. Tracers with lower (more negative) logD values and higher net
charges possessed shorter PHLs (Fig. 3d,e). The replacement of a PEG linker by AAs or the
incorporation of AA(s) to an already existing PEG linker resulted in reduced PHLs. PEGꢀ
18
68
containing tracers [ F]21–[ Ga]28 — even though ranging from ꢀ1 to +1 in terms of net charge
18
68
—
possessed calculated PHLs of >10 min. For instance, lead compounds [ F]27 and [ Ga]27
exhibited PHL of 17.1 and 15.1 min, respectively. In contrast, AAꢀcontaining radioligands
exhibited PHLs <10 min. Tracers that solely contained lysine moieties as linkers displayed the
shortest PHLs, values that decreased further as the number of lysine residues increased. For
6
8
example, [ Ga]22 — with a charge of +4 and 3 lysine residues — possessed a PHL of 1.9 min,
the shortest of all tracers. Longitudinal PET imaging experiments conducted in healthy animals
revealed a significant difference in biodistribution patterns and clearance pathways between
tracers (Supplementary Section 7). In general, lysineꢀcontaining radioligands — regardless of
the presence of a PEG group — showed fast clearance from circulation and relatively high
kidney radioactivity concentration of >5%ID/g as early as 1 h p.i. Radioactivity uptake and
ACS Paragon Plus Environment

Page 13 of 60
Journal of Medicinal Chemistry
retention increased in a nearly linear fashion with the number of lysine residues per radioligand
Fig. 4a-c). High retention of the tracers containing lysine residues was likely due to
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
reabsorption of those tracers in the proximal tubules of the kidney. In fact, recent studies have
shown that peptides high in lysine residues are powerful kidney targeting agents, facilitating the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
29,30
uptake and retention of those constructs in the renal clearing organs.
All other radioligands
that did not contain lysine residues exhibited significantly lower kidney activity concentrations
6
8
(<3.5 %ID/g, Fig. 4). Compound [ Ga]30, for instance, exhibited the lowest uptake values of
1
.2 ± 0.3 %ID/g at 2 h p.i., though it also displayed higher liver (>1.8 %ID/g) and intestinal
uptake (>3 %ID/g). Generally, tracers with overall net charges were cleared faster from
circulation through globular filtration, whereas compounds with low or no net charge exhibited
elongated circulation times and were predominantly cleared hepatically and excreted via the
intestines.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. (a) Chemical structures and maximum intensity projections (MIPs) acquired 2 h and 4
1
8
18
18
18
18
h p.i. for 5 selected radioligands ([ F]19, [ F]20, [ F]21, [ F]27, and [ F]30) in healthy
athymic nude mice (B = bladder, K = kidney, I = intestine). Tracers (11ꢀ14 MBq, 0.8ꢀ1.3 nmol)
were administered via the lateral tail vein (t = 0). Notably, lysineꢀcontaining radioligands
1
8
18
18
18
[
F]19, [ F]20, and [ F]21 showed fast renal clearance, whereas [ F]30 showed more
18
predominantly uptake in the intestines. Lead compound [ F]27 showed clearance via both
excretion routes. The increasing uptake and retention in the kidneys facilitated by an increasing
number of lysine residues in the radioligand structure is well visible for [ F]19ꢀ21. (b) Kidney
uptake of radioligands at 2 h p.i. is shown as a function of molecular net charge. (c) Kidney
uptake values at 2 h p.i. dependent on the linker and radionuclide. Solely lysine containing
radioligands exhibited elevated kidney uptake values of up to 16.0 ± 5.8 (2 h p.i.) and 21.1 ± 5.6
%ID/g (4 h p.i.) for [ Ga]22. Even one lysine residue had a dramatic effect on kidney uptake as
shown for [ F]29, where a lysine residue was integrated between the PEG linker and the
chelator moiety, increasing kidney uptake from 2.6 ± 0.3 ([ F]25, TzꢀPEG ꢀNODA) to 8.3 ±
1.5 %ID/g ([ F]29, TzꢀPEG ꢀLysineꢀNODA). Substitution of lysine by other amino acids with
18
68
1
8
7
18
7
1
8
7
either positively or negatively charged side chains reduced kidney uptake to below 4 %ID/g.
ACS Paragon Plus Environment

Page 15 of 60
Journal of Medicinal Chemistry
Evaluation of pretargeting performance in a PDAC xenograft model
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Based on this in vitro and PK data, 15 radioligands were selected for in vivo pretargeting
experiments to probe correlations between PK parameters and pretargeting performance.
Compounds were selected in order to cover a broad range of structural and physicochemical
diversity. Generally, we reasoned that PHL and the primary clearance pathway of a tracer
should have significant impact on its pretargeting performance. PHL would determine how
much tracer molecules reach the target site before clearance, and the clearance pathway should
influence the targetꢀtoꢀbackground ratio, as renal clearance should reduce background noise
more quickly than hepatic clearance.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
All initial pretargeting experiments were carried out in mice bearing s.c. BxPC3 xenografts.
6
Approximately 3ꢀ4 weeks after inoculation (2.5 x 10 cells), TCOꢀmodified antiꢀCA19.9 mAb
5B1ꢀTCO (1.33 nmol, 200 µg in 150 µL 0.9% saline) was injected via the tail vein. 72 hours
later, Tzꢀderived radioligands (1.3–1.6 nmol, 1–1.2 eq., in 150 µL 0.9% saline, containing <5%
1
,2
v/v ethanol) were injected into the opposite lateral tail vein (t = 0).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. (a) MIPs and transverse image slices acquired in pretargeted PET imaging
experiments 2 h after radioligand administration. White arrows indicate tumor tissue. To this
end, 5B1ꢀTCO (200 µg, 1.33 nmol, in 150 µL 0.9% saline) was injected into the lateral tail vein
7
2 h prior to the radioligand (1.3–1.6 nmol in 150 µL 0.9% saline, containing <5% v/v ethanol).
1
8
18
[ F]19 exhibited slightly higher tumor uptake (2.2 ± 0.2 %ID/g, 1 lysine residue) than [ F]20
1.7 ± 0.2 %ID/g, 2 lysine residues), where the tumor was not visible due to the high
background of kidney. However, those uptake values were significantly lower compared to the
(
18
68
PEGylated lead compounds [ F]27 (7.6 ± 1.8 %ID/g) and [ Ga]27 (6.6 ± 1.4 %ID/g).
1
8
Incorporation of a lysine residue as for [ F]29 (PEG + Lysine) increased kidney uptake, while
decreasing PHL and tumoral uptake, indicating a positive correlation. (bꢀd) Plots showing
tumoral uptake in dependence of in vivo stability (2 h p.i.) and PHL (2, 4 h p.i.). In vivo stability
(
shown as %ꢀintact tracer and determined at 2 h p.i.) of all radioligands employed in
18
pretargeting experiments was >50%, except for [ F]24 with a measured stability of 31.6 ±
.3%. In vivo stability had no measurable effect on the pretargeting performance, most likely
6
because the majority of intact tracer had undergone click reaction at the target site within the
first 2 h of the experiment. Tumor uptake of radioligands did however correlate with their
PHLs: the longer the PHL the higher the overall tumoral uptake.
PET imaging and ex vivo biodistribution experiments were conducted at 2 h and 4 h p.i. Tumor
uptake values of all the radioligands were determined via PET image analysis (ROI analysis) as
ACS Paragon Plus Environment

Page 17 of 60
Journal of Medicinal Chemistry
well as via ex vivo biodistribution analysis for selected compounds (Supplementary Sections 6
and 7). The median tumoral activity concentration values at 2 h p.i. ranged from 7.6 ± 1.8
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
18
18
%ID/g for Fꢀlabeled lead compound [ F]27, to 1.7 ± 0.8 %ID/g for the fastꢀclearing tracer
1
8
[
F]20 (Figs. 5 and 6). Lysineꢀcontaining radioligands generally possessed tumor uptake values
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of <3%ID/g at all time points, with tumor uptake values of <2 %ID/g for compounds solely
containing a lysine linker. The highest tumor uptake values were observed for radioligands
containing PEG linkers only. The introduction of AA linker led to accelerated plasma clearance,
and thus reduced the tumoral uptake due to fewer tracer molecules capable of reaching the
tumor site. As expected, the PHL of a radiotracer correlated (pꢀvalue <0.0001, Supplementary
Section 9) with the tracer’s overall tumoral uptake, as illustrated for the 2 h and 4 h time points
(Fig. 5c,d). Additionally, tracer stabilities in vivo were determined to be >50% at 3 h p.i. and
did not correlate with tumoral uptake (Fig. 5b).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. (a) 3D bar diagram summarizing tumoral uptake values (Zꢀaxis) for all evaluated
radioligands at 2 h p.i. (Compound numbers are given for selected compounds only). Uptake
values are presented in relation to the employed linker moiety (Xꢀaxis) and molecular net
charge (Yꢀaxis). In opposition to kidney uptake data, tumoral uptake values of lysine containing
and other, fastꢀclearing radioligands were generally low (<3.5 %ID/g for all time points). Tracer
with PHLs of >10 min generally exhibited higher tumoral uptake, and the highest values were
18
measured for PEGylated compounds based on Tz 4, including lead compounds [ F]27 and
68
[
Ga]27. (b) Summary of tumor uptake values of all 15 radioligands tested in pretargeting
experiments using the PDAC model.
68
In case of tracer [ Ga]24, possessing low stabilities in both human serum (66.9 ± 7.9%) and in
vivo (31 ± 6.3%), stability may indeed explain the relatively poor tumoral uptake of 3.1 ±
0
.7%ID/g at 2 h p.i., despite exhibiting a PHL (15.5 min) that would allow for higher
accumulation at the target site over time.
ACS Paragon Plus Environment

Page 19 of 60
Journal of Medicinal Chemistry
Identification of lead candidates
Taking all of the data into account, our study identified two lead candidates: [ F]27 and
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
18
6
8
[ Ga]27. Both tracers possessed good stabilities in vitro as well as in vivo and exhibited PHLs
(17.1 and 15.1 min, respectively) that allowed for a fast enough clearance from circulation
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
18
without impairing tumor accumulation. [ F]27 showed tumor uptake values of 7.6 ± 1.8 %ID/g
at 2 h p.i. and 8.8 ± 1.7 %ID/g after 4 h, as well as promising tumorꢀtoꢀNT (nonꢀtarget) activity
6
8
concentration ratios, although it did display some uptake in the intestines (Fig. 8a,b). [ Ga]27
exhibited tumoral uptake values of 6.8 ± 1.4 %ID/g at 2 h p.i. and 7.1 ± 1.8 %ID/g at 4 h p.i.
18
and boasted superior tumorꢀtoꢀNT activity concentration ratios compared to its Fꢀlabeled
counterpart, especially with regard to the large intestines (>3 at 2 h p.i. compared to ~1.2 for
1
8
[
F]27; Supplementary Section 6).
Figure 6. (a) Transverse (top) and coronal (middle) slices as well as MIPs (bottom) of
18
68
pretargeted PET imaging experiments using lead compounds [ F]27 (left) and [ Ga]27 (center)
in the BxPC3ꢀbased PDAC model. PET images show high tumor uptake of >6%ID/g 2 h p.i. In
68
68
addition, the Gaꢀlabeled tracer [ Ga]27 was tested in a pretargeting study using a
subcutaneous CRC model (SW1222, right). The gpA33ꢀtargeting immunoconjugate huA33ꢀ
6
8
TCO (0.95 nmol) was injected 48 h prior to tracer [ Ga]27. Good delineation of tumor tissue
was achieved as early as 1 h p.i., with the tumor showing highest uptake at 3 h p.i. These data
confirming results obtained from the PDAC model, and supported the overall promising
performance of both radioligands. (b) Quantitative comparison of uptake values calculated for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 60
1
8
tumor, blood, large intestines, and kidneys. [ F]27 showed overall the highest tumor uptake
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
68
(8.8 ± 1.7 %ID/g), but also the highest residual radioactivity in the intestines. [ Ga]27 showed
highest tumorꢀtoꢀclearance organ ratios in both models, but exhibited relatively high residual
blood radioactivity concentrations, most likely due to residual antibody in the bloodstream at
the time of tracer injection.
6
8
However, [ Ga]27 showed elevated activity concentrations in the blood pool in both tumor
models with tumorꢀtoꢀblood ratios of 0.8 ± 0.2 and 1.7 ± 0.4 at 2 h and 4 h p.i., respectively, in
the PDAC model, suggesting that a significant amount of 5B1ꢀTCO was still present in
circulation when the tracer was administered. High tumor uptake as well as similar residual
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
68
blood radioactivity concentrations were observed when [ Ga]27 was tested in a s.c. colorectal
cancer (CRC, SW1222 cell line, Supplementary Section 7) model, performed to demonstrate the
versatility and modularity of this approach. To this end, the antiꢀA33 mAb huA33ꢀTCO was
31,32
injected 48 h prior to the radioligand.
Despite a relatively low tumorꢀtoꢀblood activity
68
concentration ratio due to residual circulating mAb, [ Ga]27 showed excellent tumor uptake as
early as 1 h p.i., with maximum uptake values of 7.7 ± 1.0 %ID/g at 4 h p.i. (Fig. 8a,b). In
68
accordance to our data showing the influence of molecular net charge, [ Ga]27 (+1) showed
1
8
significantly reduced uptake in the intestines in both in vivo models compared to [ F]27 (0).
Instead, enhanced renal clearance was observed that led to an increased tumorꢀtoꢀbackground
signal. Furthermore, dosimetric calculations based on ex vivo biodistribution data gave wholeꢀ
1
8
68
body effective doses of 0.009 mSv/MBq ([ F]27) and 0.013 mSv/MBq ([ Ga]27) for both lead
8
9
compounds. Compared to approaches using longerꢀlived radionuclides ( Zr), this means a >70ꢀ
fold reduction in radiation doses to healthy tissue (Supplementary Section 8). Ultimately,
1
8
68
[
F]27 and [ Ga]27 showed the most favorable pretargeting PK with high tumor accumulation
ACS Paragon Plus Environment

Page 21 of 60
Journal of Medicinal Chemistry
with minimal offꢀtarget radiation. As a result, both compounds are currently being evaluated to
enter the clinical development phase.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discussion
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Since the advent of in vivo pretargeting as a diagnostic and therapeutic tool, clinical success
so far has been limited. Despite the ability to successfully target tumor tissue, peripheral issues
(such as the immunogenicity of streptavidin) have impaired clinical development, though recent
results obtained with bispecific antibodies are extremely promising. The question whether click
chemistryꢀbased approaches — most notably those based on the IEDDA reaction — will have
more success in the clinic has yet to be answered.
In recent years, the IEDDA reaction utilizing fullꢀsized, TCOꢀmodified mAbs in
combination with small, Tzꢀderived imaging probes has attracted by far the most attention in
the scientific community. Due to its nearly ideal properties such as speed, selectivity, and
bioorthogonality, the IEDDA reaction is currently regarded as an approach with high clinical
potential. Despite its popularity, limited efforts to investigate the influence of small molecule
PK have been reported. Thus, a comprehensive investigation of how small molecule in vivo
behavior would significantly improve our current understanding of how molecular parameters
influence pretargeting performance, and increase the potential for designing and identifying
clinical candidates. The feasibility of combining the pretargeting approach with shortꢀlived (t1/2
1
, 21
< 2 h) radionuclides was only recently demonstrated.
Improving the small molecule’s PK in
order to achieve high targetꢀtoꢀNT activity concentration ratios as early as possible after tracer
administration would be highly valuable from a clinical perspective. In order to achieve this
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 60
goal, effector molecules with the ability to accumulate quickly at the target site while
simultaneously exhibiting short offꢀtarget residence times are needed. Taken together, these
issues motivated us to pursue a comprehensive SPR study. The study at hand includes the in
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8
68
vitro and in vivo evaluation of 25 novel Fꢀ and Gaꢀlabeled Tzꢀderived radioligands capable
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of in vivo click reaction with clinically validated mAbs as pretargeting constructs.
Molecular and structural parameters were correlated with in vitro and in vivo results in order
to investigate relationships between structure and in vivo behavior. Overall, four major trends
were observed. First, the overall molecular charge of the radioligand had a significant impact on
in vivo behavior, as high net charge reduced lipophilicity and shortened the PHL. Second, the
introduction of amino acid residues into the radioligand structure did not only lead to reduced
circulation times but also influenced whether the tracer was eliminated renally or hepatically.
Charged molecules were predominantly cleared through glomerular filtration in the kidneys,
whereas compounds without an overall net charge displayed higher uptake in the liver and
intestines. Importantly, the introduction of lysine residues into the radioligand structure
dramatically decreased the tracer’s PHL (<6 min), while it was quickly and significantly taken
up and retained in the proximal tubules of the kidneys. Kidney uptake values of these
68
radioligands were up to 20 %ID/g at 3 h p.i. ([ Ga]22), while their tumoral uptake values were
dramatically reduced to <2%ID/g. Those results appeared to be in line with recently published
data demonstrating that peptideꢀsequences high in lysine residues are powerful delivery agents
2
9,30
to specifically target the proximal tubules of the kidneys.
Third, the substitution of lysine by
other amino acids (such as arginine) resulted in a significant reduction of kidney targeting
ability. Those results are consistent with data presented in this study, showing that kidney
ACS Paragon Plus Environment

Page 23 of 60
Journal of Medicinal Chemistry
uptake was decreased >4ꢀfold when lysine was substituted by an arginine or histidine. Most
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
8
notably, the incorporation of a lysine residue into the structure of [ F]25 to from radioligand
1
8
18
18
[ F]29 dramatically changed in vivo behavior. Compared to [ F]25, [ F]29 exhibited a
significantly reduced PHL and a lower tumor uptake value of 2.9 ± 0.3%ID/g at 2 h p.i., with
the majority of radioactive tracer being accumulated in the kidneys (8.3 ± 2.1%ID/g, Fig. 6a,
Supplementary Section 7). And finally, tumor uptake was mainly determined by the PHL of the
radioligand. We found a significant and positive correlation between a radioligand’s tumor
uptake and PHL (pꢀvalue <0.0001), explaining the observations made in our in vivo studies. The
PHL of a radioligand had to be sufficiently long — approximately >10 min — to enable
effective tumor accumulation.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Overall, we observed that the incorporation of amino acid residues — and with that the
alteration of molecular net charge — had a more profound effect on tracer PK then the
introduction of PEGꢀlinkers alone. Although PEGꢀlinker increased the kidneyꢀtoꢀintestines
uptake ratios as expected, no change in PHL could be observed. Instead, our results suggest that
the molecular net charge had the most significant impact on radioligand PK.
18
68
Lead compounds [ F]27 and [ Ga]27 contained PEG ꢀlinkers, exhibited PHLs of
11
approximately 15 min, and possessed net molecular charges of 0 and +1, respectively. Even
18
18
though the Fꢀlabeled tracer [ F]27 was predominantly excreted via the intestines, it showed
the highest overall tumoral uptake of all radioligands investigated, with 8.8 ± 1.7 %ID/g at 4 h
6
8
68
p.i. The Gaꢀlabeled lead compound [ Ga]27 was tested in 2 tumor models with different
targeting vectors and showed high tumor uptake values in both models at 4 h p.i.: 6.9 ±
1
.8%ID/g (PDAC, BxPC3) and 7.7 ± 1 %ID/g (CRC, SW1222). However, as for all other
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 60
radioligands that were employed in pretargeting experiments, both lead compounds showed
significant residual blood radioactivity concentrations that cannot be explained by the clearance
of the tracer alone (Supplementary Section 7). Instead, we believe that residual mAb circulating
in the blood at the time of tracer injection and subsequent click chemistry in circulation led to
elevated blood radioactivity concentrations. We applied the established accumulation intervals
between the mAbꢀTCO and tracer injections of several days (2 and 3 days for huA33 and 5B1,
respectively) to allow mAb accumulation at the target site and clear from circulation. Due to the
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
,31
inherent instability of the TCO group (formation of the more stable cisꢀisomer over time) and
internalization of some antibodies (e.g. 5B1) the accumulation interval cannot be extended
infinitely. Instead, a mAbꢀdependent time interval between the injections of mAbꢀTCO and the
3
Tz probe has to be determined. Indeed, pretargeting studies conducted by our groups that used
longer accumulation intervals for 5B1ꢀTCO did not lead to dramatically improved tumorꢀtoꢀ
2
blood activity concentration ratios. Although a decreased antibody concentration in the blood
pool can be achieved, the increased decomposition of TCO at the tumor site led to overall
2
reduced tumoral uptake and TTB values. In case of nonꢀinternalizing huA33ꢀTCO, however,
3
longer intervals of up to 120 h yielded slightly higher TTB ratios. New technologies such as
33
innovative clearing agents or the use of smaller antibody fragments have the potential to
further improve this strategy by reducing residual activity concentrations in the blood.
We initially speculated that the pretargeting performance of a radioligand could be improved
by solely increasing the renalꢀtoꢀgut excretion ratio. However, our data suggest that it is not that
simple. In order to facilitate sufficient delivery of radioligand to the tumor site a compromise,
or “sweet spot”, between the excretion pathways, PHL, and tumor signal needs to be identified,
ACS Paragon Plus Environment

Page 25 of 60
Journal of Medicinal Chemistry
and that achieving high on one of those parameters will most likely have a negative impact on
the others. As an example, short PHL and fast organ clearance had a negative effect on tumor
accumulation. On the other hand, molecules with no net charge or low overall net charge
exhibited longer PHLs and higher tumoral uptake values but also higher background noise
levels due to a slower excretion predominantly via the intestines.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
At this point of the discussion, it is important to highlight the focus of this study to develop
radioligands specifically for pretargeting approaches utilizing shortꢀlived radionuclides with
physical halfꢀlives <2 h. Short physical halfꢀlives, however, limit the time window for highꢀ
quality PET image acquisition to a maximum of 6–8 h after tracer injection (preclinically).
Thus, radioligands have to possess PK properties that allow for successful tumor visualization
with high tumorꢀtoꢀbackground ratios at early time points, preferably between 1–4 h p.i.
64
Conversely, approaches that apply longerꢀlived radionuclides such as Cu (t = 12.7 h) allow
1
/2
image acquisition at later time points (up to 48 h p.i.), and are thus less sensitive to early
radioligand PK. At the same time it is important to note that time restraints are easily justified
and outꢀweighed by the dosimetric advantages that the use of shortꢀlived radionuclides entails.
6
8
In addition, the ability of virtually any major hospital and cancer center to produce Gaꢀ and
1
8
Fꢀlabeled clinicalꢀgrade radiopharmaceuticals on site and on a daily basis is a tremendous
logistical advantage and thus justifies the development of these tracers.
Ultimately, this SPR study revealed fundamental relationships between the structure and in
vivo behavior of tetrazineꢀbased radioligands, allowing for a more rational approach to design
nextꢀgeneration radiopharmaceuticals for in vivo pretargeting. As discussed earlier, the second
goal of this study was to generate candidates worthy of clinical translation. We believe that with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 60
18
68
lead compounds [ F]27 and [ Ga]27 we have identified 2 promising candidates that show high
tumor targeting abilities and favorable dosimetries, justifying an IND application in the near
future.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
General Information: All reactions involving tetrazine moieties were carried out using
aluminum foil covered reaction vials due to light sensitivity of the tetrazine structure. Further,
all reaction mixtures were purified using preparative HPLC with the following gradient: 5ꢀ95%
of MeCN in H O over 20 min, 8 mL/min. The final purity of all molecules was determined
2
using analytical HPLC (see Table 2 for precursors 15-32, gradient: 5ꢀ95% of MeCN in H O
2
over 20 min, 1 mL/min) to achieve purities of >95%. Retention times (R ) were assigned to all
t
compounds. Upon purification and freezeꢀdrying, all final precursor compounds 15-32 were
aliquoted into 25 µL (50 nmol) fractions using metalꢀfree, anhydrous DMSO. It is worth noting
that reaction conditions, in particular for the formation of amide bonds, had to be designed with
respect to the wellꢀknown sensitivity of tetrazines to harsh synthetic conditions, such as high pH
23,34
and elevated temperatures . For instance, the presence of bases such as pyridine,
triethylamine and reagents such as N’,N’ꢀdicylcohexylcarbodiimide and Nꢀhydroxysuccinimide
used for in situ carboxylic acid activation, led to decomposition of the tetrazine moiety, which
was easily visible by eye due to loss of purple color and the absence of absorption at 525 nm.
To overcome this limitation, we found that amide bonds could be formed in high yields and
without impairing the tetrazine moiety by using 2 equivalents of both 1ꢀethylꢀ3ꢀ(3ꢀ
demethylamunopropyl)carbodiimide and hydroxybenzotriazole in dry DMF or DMSO. The
ACS Paragon Plus Environment

Page 27 of 60
Journal of Medicinal Chemistry
procedures and analytical data for the synthesis of all precursor molecules can be found in the
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Supplementary Information (Sections 2 and 3). Summarized below (Table 2) are the yield and
18
68
purity data obtained for all precursor and tracer molecules. Further, the general F and Gaꢀ
labeling procedures that were developed for this study are presented.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Synthesis of precursor molecules 15-32:
4-(1,2,4,5-tetrazin-3-yl)benzoic acid. The title compound was synthesized according to the
2
3
previously published procedure from Karver, et al. Obtained analytical data were consistent
1
with the reported data. H NMR (500 MHz, DMSOꢀd ) δ (ppm): 10.66 (s, 1H), 8.62 (d, J = 8.3
6
ꢀ
Hz, 2H), 8.22 (d, J = 8.3 Hz, 2H). MS (ESI) m/z [M + Cl] : 239.2.
2,2'-(7-(4-(4-(1,2,4,5-tetrazin-3-yl)benzamido)benzyl)-1,4,7-triazonane-1,4-diyl)diacetic
acid (Tz-1-NODA, 15). 4ꢀ(1,2,4,5ꢀtetrazinꢀ3ꢀyl)benzoic acid (26.7 µmol) was dissolved in
DMSO (0.5 mL) before pꢀBnꢀNODAꢀNCS (26.7 µmol) and triethylamine (5 µL) were added.
The pink reaction mixture was stirred at room temperature for 1 h (general procedure for the
isothiocyanateꢀamine addition reaction). After completion of the reaction (monitored by LCꢀ
MS), precursor 15 was purified (R = 12.4 min). The collected fraction was concentrated under
t
reduced pressure and dried over night under high vacuum. Precursor 15 (purity >97%) was
1
obtained as pink oil (8.9 mg, 58%). H NMR (500 MHz, DMSOꢀd ) δ (ppm): 10.32 (s, 1H),
6
8
.41 (d, J = 8.2 Hz, 2H), 8.29 (d, J = 8.3 Hz, 2H), 7.62 (d, J = 7.8Hz, 2H), 7.37 (d, J = 7.5 Hz,
+
2H), 4.16 (s, 2H), 3.12–3.04 (m, 12H), 2.42–2.11 (m, 5H). MS (ESI) m/z [M + H] : 535.4.
+
HRMS (ESI) m/z calcd. for C H N NaO [M + Na] : 557.3256 found: 557.3348.
26 30
8
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 60
6
2
N -(4-(1,2,4,5-tetrazin-3-yl)benzoyl)-N -(tert-butoxycarbonyl)lysine (Tz-1-Lysine-Boc). Tz 1
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
5 mg, 26.7 µmol) was dissolved in dry DMF (0.5 mL) under nitrogen atmosphere before Bocꢀ
Lysine (5.8 mg, 29 µmol), 1ꢀethylꢀ3ꢀ(3ꢀdimethylaminopropyl)carbodiimide (10.7 mg, 60 µmol),
and hydroxybenzotriazole (6.8 mg, 60 µmol) were added to the pink solution. The reaction
mixture was stirred at room temperature for 4ꢀ6 h until all starting material was converted
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
according to LCꢀMS (general amide coupling conditions). The mixture was purified (R = 14.1
t
min) and the product solution was subsequently concentrated and dried to yield the above
1
shown compound as pink oil (6.4 mg, 58%). H NMR (500 MHz, DMSOꢀd ) δ (ppm): 10.41 (s,
6
1
6
H), 8.52 (t, J = 7.9 Hz, 1H), 8.40 (d, J = 7.9 Hz, 2H), 7.76 (s, 1H), 7.54 (d, J = 8.0 Hz, 2H),
.99 (d, J = 7.7 Hz, 1H), 3.00 (s, 2H), 2.84–2.73 (m, 3H), 1.70–1.49 (m, 4H), 1.42 (s, 9H).
+
(ESI) m/z [M + H] : 431.5.
2,2',2''-(2-(4-(6-(4-(1,2,4,5-tetrazin-3-yl)benzamido)-2-((tert-butoxycarbonyl)amino)-
hexanamido)-benzyl)-1,4,7-triazonane-1,4,7-triyl)triacetic acid (Tz-1-Lysine-Boc-NOTA). The
title compound was obtained using Tzꢀ1ꢀLysineꢀBoc (5 mg, 12.1 µmol) and an equimolar
amount of pꢀBnꢀNOTAꢀNCS, as well as the general conditions for the isothiocyanateꢀamine
addition reaction. The product was purified (R = 16.6 min) and dried under reduced pressure
t
1
and furnished as pink oil (8.2 mg, 78%). H NMR (500 MHz, Chloroformꢀd) δ (ppm): 10.20 (s,
1
H), 8.53 (d, J = 8.0 Hz, 2H), 7.73 (d, J = 7.9 Hz, 2H), 7.56 (d, J = 8.2 Hz, 2H), 7.47 (d, J = 7.9
Hz, 2H), 7.37 (t, J = 7.9 Hz, 3H), 7.31 – 7.25 (m, 7H), 6.72 (s, 1H), 5.52 (s, 1H), 4.49 (dd, J =
9.8, 6.8 Hz, 7H), 4.16 (d, J = 21.1 Hz, 2H), 3.12–3.09 (m, 4H), 2.02–1.71 (m, 9H), 1.41 (s, 9H).
+
(ESI) m/z [M + H] : 821.7.
ACS Paragon Plus Environment

Page 29 of 60
Journal of Medicinal Chemistry
2
,2',2''-(2-(4-(6-(4-(1,2,4,5-tetrazin-3-yl)benzamido)-2-aminohexanamido)benzyl)-1,4,7-
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
triazonane-1,4,7-triyl)triacetic acid (16). Tzꢀ1ꢀLysineꢀBocꢀNOTA (7 µmol) was dissolved in
DCM (0.5 mL) before TFA (0.3 mL) was added dropwise to the solution under vigorous
stirring. The mixture was stirred at room temperature for 45 min before DCM was removed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
under reduced pressure (General TFAꢀdeprotection procedure). The remaining crude product
1
was purified (R = 12.3 min, >96%), yielding precursor 14 as pink oil (4.8 mg, 90%). H NMR
t
(500 MHz, DMSOꢀd ) δ (ppm): 10.10 (s, 1 H), 8.42 (d, J = 8.1 Hz, 2 H), 7.78 (d, J = 8.1 Hz, 2
6
H), 7.43 (d, J = 9.1 Hz, 2 H), 7.34 (d, J = 8.9 Hz, 2 H), 6.72 (s, 1 H), 5.52 (s, 1 H), 4.49 (m, 1
H), 4.16 (m, 2 H), 3.82–3.69 (m, 6 H), 3.12–3.09 (m, 9 H), 2.02–1.71 (m, 15 H); MS (ESI) m/z
+
+
[
M + H] : 721.4. HRMS (ESI) calcd. for C H N NaO [M + Na] m/z 743.3381 found:
34
44 14
8
743.3373.
tert-butyl-(1-(4-(1,2,4,5-tetrazin-3-yl)phenyl)-1-oxo-5,8,11,14,17,20,23-heptaoxa-2-
azapentacosan-25-yl)carbamate (Tz-1-PEG -NHBoc). Tzꢀ1 (10 mg, 0.05 mmol) was dissolved
7
in anhydrous dimethylsulfoxide (DMSO, 0.5 mL) before Oꢀ(2ꢀaminoethyl)ꢀOꢀ[2ꢀ(bocꢀ
amino)ethyl]hexaethylene glycol (0.0375 mmol) and TEA (0.0057 mL, 0.0375 mmol) were
added. The reaction mixture was stirred at room temperature for 1 h (general NHS ester
coupling procedure). After completion of the reaction (monitored by LCꢀMS) the product was
1
isolated and purified using preparative HPLC (R = 15.3 min). H NMR (500 MHz,
t
Chloroformꢀd) δ (ppm): 10.21 (s, 1H), 8.58 (d, J = 8.2 Hz, 2H), 7.54 (d, J = 8.2 Hz, 2H), 6.89
(m, 2H), 6.45 (m, 3H), 5.08–4.88 (m, 4H), 4.56 (d, J = 8.0 Hz, 2H), 3.63 (d, J = 8.0 Hz, 2H),
3
.54 (dt, J = 10.4, 4.9 Hz, 6H), 3.42 (q, J = 5.2 Hz, 3H), 3.33–3.27 (m, 5H), 2.37 (t, J = 10.2
ACS Paragon Plus Environment