Platinum(II) as bifunctional linker in antibody-drug conjugate formation: Coupling of a 4-nitrobenzo-2-oxa-1,3-diazole fluorophore to trastuzumab as a model

Article abstract of DOI:10.1002/cmdc.201402496

The potential of platinum(II) as a bifunctional linker in the coordination of small molecules, such as imaging agents or (cytotoxic) drugs, to monoclonal antibodies (mAbs) was investigated with a 4-nitrobenzo-2-oxa-1,3-diazole (NBD) fluorophore and trastu

Full text of DOI:10.1002/cmdc.201402496

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
DOI: 10.1002/cmdc.201402496
Full Papers
Platinum(II) as Bifunctional Linker in Antibody–Drug
Conjugate Formation: Coupling of a 4-Nitrobenzo-2-oxa-
1,3-diazole Fluorophore to Trastuzumab as a Model
Dennis C. J. Waalboer,*[a] Joey A. Muns,[a] Niels J. Sijbrandi,[b] Richard B. M. Schasfoort,[c]
Rob Haselberg,[d] Govert W. Somsen,[d] Hendrik-Jan Houthoff,[b] and Guus
A. M. S. van Dongen[a]
The potential of platinum(II) as a bifunctional linker in the co-
ordination of small molecules, such as imaging agents or (cyto-
toxic) drugs, to monoclonal antibodies (mAbs) was investigat-
ed with a 4-nitrobenzo-2-oxa-1,3-diazole (NBD) fluorophore
and trastuzumab (Herceptinꢀ) as a model antibody. The effect
of ligand and reaction conditions on conjugation efficiency
was explored for [Pt(en)(L-NBD)Cl](NO3) (en=ethylenediamine),
with L=N-heteroaromatic, N-alkyl amine, or thioether. Conju-
gation proceeded most efficiently at pH 8.0 in the presence of
NaClO4 or Na2SO4 in tricine or HEPES buffer. Reaction of N-co-
ordinated complexes (20 equiv) with trastuzumab at 378C for
2 h, followed by removal of weakly bound complexes with
excess thiourea, afforded conjugates with an NBD/mAb ratio
of 1.5–2.9 that were stable in phosphate-buffered saline at
room temperature for at least 48 h. In contrast, thioether-coor-
dinated complexes afforded unstable conjugates. Finally, sur-
face plasmon resonance analysis showed no loss in binding af-
finity of trastuzumab after conjugation.
Introduction
The coordination of platinum to DNA is an area of intense re-
search that has allowed the development of valuable plati-
num-based chemotherapeutics.[1] Besides DNA, platinum also
readily coordinates to other biomolecules, such as proteins.[2]
Although platinum–protein binding has been studied exten-
sively, this work is largely focused on the study of platinum–
protein binding in cisplatin chemotherapy, where the uncon-
trolled binding of platinum(II) to proteins is believed to cause
dose-limiting toxicity.[3] In this respect, the unique protein
binding properties of platinum(II) still remain largely underex-
ploited. The bivalent nature of cisplatin-type complexes, com-
bined with the protein binding capacity of platinum, opens up
the intriguing option to use these complexes as linkers in bio-
conjugation reactions. The complex [Pt(en)Cl2] has been used
in this context for the conjugation of kinase inhibitors to albu-
min and lysozyme in hepatic and renal targeting as well as tar-
geting to activated endothelium.[4] Extension of this methodol-
ogy to the preparation of antibody–drug conjugates (ADCs),
for selectively targeting diagnostic or therapeutic agents to
tumors or other diseases, would offer potential advantages
over conventional methods (Figure 1).[5] More specifically, this
[a] Dr. D. C. J. Waalboer, J. A. Muns, Prof. Dr. G. A. M. S. van Dongen
Department of Radiology and Nuclear Medicine
VU University Medical Center
Figure 1. Platinum(II)-mediated conjugation of a drug (diagnostic or thera-
peutic compound) to an antibody.
P.O. Box 7057, 1007 MB Amsterdam (The Netherlands)
E-mail: d.waalboer@vumc.nl
[b] Dr. N. J. Sijbrandi, Dr. H.-J. Houthoff
approach would, in principle, allow the coordination of small
molecules via a wide variety of coordinating groups, including
nonconventional functionalities such as N-heteroaryl groups
and S-donors, such as thioethers. Moreover, the introduction
of a charged platinum complex can improve the aqueous solu-
bility of organic compounds, especially at neutral pH, poten-
tially obviating the need for organic co-solvent in subsequent
conjugation reactions. Exposure of antibodies to organic sol-
vents can induce aggregation, and is therefore typically avoid-
LinXis B.V., Heilaarpark 28, 4814 NJ Breda (The Netherlands)
[c] Dr. R. B. M. Schasfoort
Medical Cell Biophysics Group, MIRA institute
University of Twente, P.O. Box 217, 7500 AE Enschede (The Netherlands)
[d] Dr. R. Haselberg, Prof. Dr. G. W. Somsen
Division of BioAnalytical Chemistry
AIMMS Research Group BioMolecular Analysis
VU University Amsterdam, 1081 HV Amsterdam (The Netherlands)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201402496.
ChemMedChem 2015, 10, 797 – 803
797
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Full Papers
ed. Apart from these differences, platinum(II) is also expected
to coordinate to a unique subset of amino acids. Extensive re-
search in this area has shown that cysteine, methionine, and
histidine are the most likely coordination sites for platinum in
proteins.[6] Results from model systems[7] and density functional
theory calculations[8] indicate that binding of platinum(II) to
methionine is kinetically controlled, whereas binding to histi-
dine is thermodynamically controlled. When performed under
thermodynamic conditions, binding of platinum(II) to proteins
also offers the prospect of high selectivity. X-ray diffraction
analysis studies by Calderone et al. on the binding of cisplatin
to bovine erythrocyte copper–zinc superoxide dismutase, for
instance, revealed that cisplatin exclusively binds to His19.[9]
Scheme 1. Synthesis of complex 1. Reagents and conditions: a) 4-piperidine-
carboxylic acid, K2CO3, MeOH, 08C, 15 min, then RT, 2 h, 92%; b) 4-(amino-
methyl)pyridine, BOP, Et3N, MeCN/CH2Cl2, RT, 5 h, 99%; c) [Pt(en)Cl2], AgNO3,
DMF, RT, 40 h, 26%. BOP=(benzotriazol-1-yloxy)tris(dimethylamino)phos-
phonium hexafluorophosphate; en=ethylenediamine.
Results and Discussion
(SEC), and no aggregation was observed. After purification by
PD-10 column, an NBD/mAb ratio of 2.8 was found, as derived
from the absorption of the NBD group (l 472 nm) and the pro-
tein (l 280 nm, corrected for absorption of complex 1 at this
wavelength). Alternatively, the NBD/mAb ratio was determined
by exploiting the hydrolytic instability of the NBD fluorophore
at high pH.[11] Briefly, overnight incubation of the purified con-
jugate at pH 10.5 resulted in complete hydrolysis of the NBD
label to NBD-OH (Figure 3). Separation by SEC then allowed
Synthesis and characterization of complex 1
Screening of reaction conditions was performed with complex
1, in which the NBD group is coordinated to platinum through
a 4-substituted pyridine (Figure 2). Although direct coupling of
4-(aminomethyl)pyridine to the NBD group would be most
Figure 2. Structure of complex 1.
straightforward, we chose to use a short spacer to prevent po-
tential intramolecular coordination of platinum to the NBD
group.[10]
Complex 1 was synthesized in three steps from NBD-Cl (2)
(Scheme 1). After introduction of the pyridine group, coordina-
tion to platinum was effected by activation of [Pt(en)Cl2] with
AgNO3 in DMF followed by reaction with pyridine L1 for 24 h
at room temperature. Purification of the crude product was
performed by preparative HPLC to afford complex 1 in >95%
purity (l 472 nm). In contrast to pyridine L1, complex 1 is read-
ily soluble in water and could be stored as a 5 mm solution in
water (containing 20 mm NaCl) with no decomposition ob-
served by HPLC after storage for at least two months at 48C.
Figure 3. Method to assess NDB/mAb ratio by UV measurement of NBD-OH
(l=472 nm) and mAb (l=280 nm). a) Hydrolysis of 1–trastuzumab. Re-
agents and conditions: a) glycine buffer (pH 10.5), 378C, 16 h. b) SEC chroma-
tograms of the hydrolysis of 1–trastuzumab (mAb–NBD) at t=0 (left) and
t=16 h (right) at pH 10.5.
Conjugation of complex 1 to trastuzumab
determination of the relative concentration of protein and
NBD-OH. Using this method, an identical NBD/mAb ratio of 2.8
was found. The latter method was preferred and used for sub-
sequent analyses.
The reactivity of complex 1 toward trastuzumab was first as-
sessed in tricine/NaNO3 buffer, typically employed for conjuga-
tion of platinum(II) complexes to albumin and lysozyme.[4a,g–j]
When trastuzumab, used in its commercial formulation, was in-
cubated with complex 1 (20 equiv) for 2 h at 378C, conjuga-
tion was clearly observed by size-exclusion chromatography
Despite the practical advantages of using the NBD group, its
hydrolytically labile nature also imposes some limitations. Con-
trol experiments showed 8 and ~0.2% hydrolysis of the NBD
ChemMedChem 2015, 10, 797 – 803
www.chemmedchem.org
798
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Full Papers
group at pH 8.0 and 7.1, respectively, after incubation at 378C
for 2 h, indicating that NBD/mAb ratios of conjugations per-
formed at pH 8.0 are a slight underestimation (data not
shown).
of salt type on the reaction of monovalent platinum(II) com-
plexes with proteins has never been reported.
To investigate this in greater detail, complex 1 in water
(Milli-Q) was reacted with trastuzumab in tricine buffer in the
presence of a variety of salts (Table 2). Notably, trastuzumab is
Effect of buffer and pH on conjugation efficiency
To explore the effect of pH and buffer on the reaction of com-
plex 1 with trastuzumab, reactions were performed in three
different buffers: MES, HEPES, and tricine, at pH 5.8–8.0 with
overlapping pH ranges to discriminate between buffer and pH
effects. As shown in Table 1, a clear pH effect was observed,
Table 2. Conjugation of complex 1 to trastuzumab: effect of added salt
on conjugation efficiency.[a]
Salt
Conc. [mm]
NBD/mAb[b]
2.7
5.1
3.1
–
–
30
30
Na2SO4
NaOAc
NaCl
6.5
2.8
Table 1. Conjugation of complex 1 to trastuzumab: effect of buffer and
pH on conjugation efficiency.[a]
NaCl
NaNO3
NaClO4
NaCl/NaNO3
NaCl/NaClO4
NaCl/Na2SO4
30
30
30
6.5/30
6.5/30
6.5/30
3.2
3.6 (3.5/3.7)[b]
4.1 (3.6/2.6)[b]
3.1
3.7
4.1
Buffer
pH[c]
NBD/mAb
Buffer
Observed
MES
MES
HEPES
HEPES
HEPES
tricine
tricine+thiourea[b]
5.5
6.4
6.5
7.5
8.5
8.5
8.5
5.8
ND
6.3
7.1
ND
8.0
ND
1.3
1.8
1.7
2.5
2.8
2.8
0.1
[a] Reagents and conditions: trastuzumab 83.1 mm, complex 1 (20 equiv),
tricine (8 mm, pH 8.5), 378C, 2 h. [b] Salt concentration: 60/120 mm.
formulated in a histidine buffer which typically results in a mini-
mum chloride concentration of ~0.9 mm for the conjugation
reactions listed in Table 2. Interestingly, the salt type proved to
have a pronounced effect on the NBD/mAb ratio obtained.
Whereas addition of NaCl and NaOAc only showed a modest
increase at 30 mm, NaNO3, NaClO4, and especially Na2SO4
showed a substantially higher NBD/mAb ratio. Increasing the
salt concentration to 60 and 120 mm was ineffective for
NaNO3, whereas a strong decrease was observed for NaClO4 at
[a] Reagents and conditions: trastuzumab 83.1 mm, complex 1 (20 equiv),
buffer (8 mm), NaCl (6.5 mm), 378C, 2 h. [b] Conjugation in the presence
of thiourea (10 mm). [c] pH measured at the start of reaction at 208C.
with higher NBD/mAb ratios observed at higher pH. Because
coordination of platinum(II) to sulfur donors such as methio-
nine and disulfides is believed to be relatively insensitive to
pH, this trend implies that histidine groups, with pKa values
typically in the range of 5.5–7.5,[12] are likely involved in bind-
ing to complex 1.
2À
120 mm. The effect of NO3À, ClO4À, and SO4 was somewhat
diminished when the ClÀ concentration was increased to
6.5 mm.
MES and HEPES buffers reportedly show negligible metal
binding, whereas tricine does show some affinity for metals,
The significant differences between the various salts are not
easily explained if only the interaction of the salt with the plati-
num complex is considered. Apparently, at the concentrations
used, none of the salts tested interfere with the binding of
complex 1 to trastuzumab, as the reaction is least efficient
when no salt is added. One could speculate that certain salts
could protect the platinum complex from aquation, which at
pH 8.0 is expected to result in deprotonation of the water
ligand ([Pt(dien)(H2O)]; pKa =6.0)[19c] to give an unreactive hy-
droxide complex. It is, however, well known that in contrast to
such as Cu2+ [13]
Comparison of overlapping pH ranges for
.
these three buffers, however, showed no significant difference
in the number of NBD groups introduced. To exclude binding
of complex 1 to trastuzumab via pathways not mediated by
platinum(II), the reaction was also performed in the presence
of thiourea, an excellent ligand for platinum(II). These condi-
tions effectively inhibited the conjugation reaction for complex
1, indicating that binding is indeed platinum mediated. Reac-
tion conditions with pH>8.0 were not explored due to the un-
stable nature of the NBD group under basic conditions.
À
À
NO3 and SO42À, ClO4 does not coordinate to platinum(II).[14]
An intriguing and more likely alternative to explain the dif-
ference in conjugation efficiency is the interaction of the salt
with the protein. Recently, Weis and co-workers[15] reported on
the effect of Hofmeister series salts on the local flexibility of an
IgG1 monoclonal antibody, as measured by hydrogen/deuteri-
um exchange coupled to mass spectrometry. Compared with
Effect of salt on conjugation efficiency
Both organic and inorganic salts are routinely used in buffers
and as stabilizers in antibody formulation. Depending on salt
type and concentration, their presence could significantly
impact the efficiency of the conjugation reaction. In previous
studies,[4a,g–j] NaNO3 was often employed in conjugation reac-
tions of platinum-based linkers with lysozyme or albumin.
However, to the best of our knowledge, a study on the effect
2À
ClÀ (0.1m), both SO4 and SCNÀ were found to increase the
flexibility of a model IgG1 at 0.5m, albeit not necessarily in the
same areas. At 0.5m, ClÀ was found to induce an overall de-
crease in flexibility of IgG1. Zhang-van Enk et al. recently
ChemMedChem 2015, 10, 797 – 803
www.chemmedchem.org
799
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Full Papers
2À
showed that, depending on pH, SO4 and SCNÀ strongly de-
To this end, 1–trastuzumab was prepared under identical
conditions, and the crude reaction mixture was treated with
methionine or thiourea for 30–60 min before being purified by
PD-10 column. Whereas methionine (50 mm) only partially de-
creased the release of NBD label for 1–trastuzumab upon stor-
age in PBS, thiourea (10 mm) completely abrogated release of
the NBD label. These data show that, relative to methionine,
thiourea more efficiently removes weakly bound complex 1. As
is evident from the lower NBD/mAb ratios listed in Table 3, this
stabilize the human IgG1 Fc fragment at concentrations below
50 mm.[16] When the results from Table 2 are aligned with the
Hofmeister series of anions, one could speculate that the
2À
À
higher conjugation efficiencies for SO4 and ClO4 might, at
least in part, be explained by an increase in the flexibility of
IgG1, thereby facilitating access to coordination sites in the
protein for platinum. (Figure 4).
Table 3. Incubation of 1–trastuzumab with S-donor ligands to remove
weakly bound platinum complexes.[a]
Competitor
Conc. [mm]
Incubation time [min]
NBD/mAb
Figure 4. The Hofmeister series of anions and observed NBD/mAb ratio at
pH 8.0 as shown in Table 2.
–
–
50
10
50
–
60
30
60
3.2
2.7
2.0 (2.5)[b]
2.0
methionine
thiourea
thiourea
[a] Reagents and conditions: trastuzumab 83.1 mm, complex 1 (20 equiv),
tricine (8 mm, pH 8.5), 378C, 2 h, then addition of competitor S-donor
compound and incubation for time specified. [b] Conjugation performed
with NaClO4 (30 mm).
Stability and mass spectrometry analysis of 1–trastuzumab
Among the various potential binding partners in proteins, histi-
dine, methionine, and cysteine are expected to be preferential
binding sites for platinum(II) at near neutral pH.[6] The stability
of the respective platinum–amino acid complexes does not
necessarily have to be the same. For instance, it is believed
that in cisplatin therapy, the rescue agent sodium diethyldi-
thiocarbamate (DDTC) is capable of reversing the binding of
platinum(II) to methionine residues, but not for platinum(II)–
cysteine bonds.[4,6e,17] This raises the question if the conjugates
prepared display a heterogeneous stability, with both weakly
and strongly bound Pt complexes present in the conjugate.
Indeed, when 1–trastuzumab was stored in phosphate-buf-
fered saline (PBS) for 48 h at room temperature, analysis by
SEC showed a partial release of the NBD linker, as shown in
Figure 5. It was hypothesized that a post-treatment step with
an S-donor ligand might be able to remove weakly bound
platinum complexes causative to this release.
procedure partially reverses the binding of complex 1 to tras-
tuzumab. Because extending the thiourea treatment from 30
to 60 min at a fivefold higher concentration did not result in
a further decrease in the NBD/mAb ratio, the thiourea appa-
rently only selectively removes a weakly bound subpopulation
of complex 1 (Table 3).
Besides weakly bound complex 1, nonspecific binding of un-
reacted complex 1 to trastuzumab could be an alternative ex-
planation for these findings. In this scenario, thiourea and me-
thionine would rapidly scavenge the remaining complex 1, re-
sulting in a more hydrophilic dicationic complex with possibly
decreased affinity for the lipophilic protein surface. Although
difficult to disprove, complex 5 was synthesized to assess the
propensity of NBD–platinum complexes for nonspecific bind-
ing to trastuzumab (Figure 6). Reaction of the more lipophilic
complex 5 with trastuzumab afforded an NBD/mAb ratio of
<0.1. Because only an insignificant amount of complex 5 was
still present after PD-10 purification, it is unlikely that the thio-
Figure 5. SEC chromatograms of 1–trastuzumab (NBD/mAb: 3.2) after PD-10
purification and storage in PBS at RT for 48 h. Reference retention times 1–
trastuzumab: 4.3 min, NBD-OH: 9.4 min, complex 1: 10.6 min, pyridine L1:
16.2 min (not detected). a) No post-treatment step before PD-10 purification.
b) Post-treatment step with methionine (50 mm) for 60 min at 378C before
PD-10 purification. c) Post-treatment step with thiourea (10 mm) for 30 min
at 378C before PD-10 purification.
Figure 6. Lipophilic complex 5 used in control experiment to exclude non-
specific (not platinum-mediated) binding to mAb. Reagents and conditions:
trastuzumab (83.1 mm), complex 5 (20 equiv), tricine (8 mm, pH 8.5), 378C,
2 h.
ChemMedChem 2015, 10, 797 – 803
www.chemmedchem.org
800
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim