Rapid and efficient tetrazine-induced drug release from highly stable benzonorbornadiene derivatives

Article abstract of DOI:10.1039/c7cc03477f

A novel class of bioorthogonal release reactions based on benzonorbornadiene derivatives was developed. These carrier molecules are highly stable at physiological conditions, but react rapidly with 1,2,4,5-tetrazines, and near-quantitatively release cargo

Full text of DOI:10.1039/c7cc03477f

ChemComm
View Article Online
View Journal
COMMUNICATION
Rapid and efficient tetrazine-induced drug release
from highly stable benzonorbornadiene
derivatives†
Cite this:DOI: 10.1039/c7cc03477f
Received 4th May 2017,
Accepted 20th May 2017
Minghao Xu,
Julian Tu
and Raphael M. Franzini
*
DOI: 10.1039/c7cc03477f
rsc.li/chemcomm
A novel class of bioorthogonal release reactions based on benzo- Until recently, modified Staudinger reactions were the only
5,14–16
norbornadiene derivatives was developed. These carrier molecules available bioorthogonal release reactions.
The development
are highly stable at physiological conditions, but react rapidly with of the inverse-electron demand Diels–Alder (IEDDA) pyridazine
1
,2,4,5-tetrazines, and near-quantitatively release cargo molecules elimination reaction of carbamate-modified trans-cyclooctenes
17,18
such as drugs and optical reporters.
(TCO) and Tz was a breakthrough with this regard.
is significantly faster than the Staudinger reaction, obviates the
The advent of bioorthogonal reactions opened the unprecedented use of metabolically unstable phosphines, and is widely used in
This reaction
1
7,10,12,13
opportunity to perform chemistry in living organisms. Bio- chemical biology.
Further examples of bioorthogonal
compatible reaction development has focused primarily on cleavage reactions are the strain-promoted 1,3-dipolar cycloaddition
19
transformations that link two molecules, and bioorthogonal of TCO and p-azidobenzylcarbamates, and the Tz-mediated
6,20,21
ligation reactions boast broad applicability in chemical biology, removal of vinyl ethers.
However, reactions need to meet several
materials science, and to localize drugs and imaging agents at strict requirements for in vivo use, including rapid reaction rate,
2
,3
sites of disease. In contrast, the discovery of bioorthogonal near-quantitative payload release, lack of toxicity, and prolonged
cleavage reactions that allow for the controlled release of pay- serum stability. Current reactions meet these conditions only
4
loads has only recently attracted substantial research interest.
partially and further development will be necessary to achieve
Nevertheless, such reactions are valuable in a wide range of the full potential of in vivo drug activation.
applications. Representative examples include biodetection Here we describe benzonorbornadiene (BNBD) derivatives as
and stable carrier molecules that rapidly react with Tz to near-
5
,6
7
8–10
probes, DNA sequencing, protein activity control,
1
1
multiplexed cell imaging. The use of dissociative in vivo quantitatively release a cargo molecule (e.g. cytotoxic agent,
chemistry to enable the spatiotemporally controlled release of optical reporter). Our design harnesses the intrinsic lability of
drugs is particularly appealing because of the potential for isobenzofurans/isoindoles to liberate a molecule of interest.
clinical translation in cancer chemotherapy. For example, These self-immolative heterocycles are readily accessible by the
2
2
implantation of tetrazine (Tz)-modified biomaterials enabled reaction of BNBDs and Tz. Devaraj et al. have used this
the localized activation of doxorubicin prodrugs with superior transformation for nucleic acid sensing but their design lacked
1
2
23
therapeutic index relative to systemic drug administration.
the capability of releasing a molecule. In particular, the
Alternatively, bioorthogonal reactions could activate antibody-drug proposed release molecules consist of 7-aza/oxa-BNBDs with
13
conjugates in vivo. In this pretargeting strategy, an antibody carbamate leaving groups attached via a methylene linker to the
modified with a drug via a chemically-cleavable linker accumulates bridgehead carbon (Scheme 1). The reaction of 7-aza/oxa-BNBDs
at the desired site, and subsequent administration of a trigger with Tz is anticipated to generate an intermediate (I1), which rapidly
1
3
molecule liberates the cytotoxic agent tissue specifically.
2
eliminates N , followed by a retro Diels–Alder cycloreversion,
2
2
The scarcity of bioorthogonal bond-cleavage reactions generating isoindoles/isobenzofurans (I3). These heterocycles
remains a key bottleneck for the advancement of reaction- in turn eliminate the carbamate to liberate the free amine
based applications in chemical biology and smart therapeutics. (Scheme 1a). In contrast to the reaction intermediates, the BNBD
precursors are expected to be highly stable.
To test the outlined bioorthogonal release reaction, we
synthesized three BNBD derivatives with oxygen (1), acetamide
Department of Medicinal Chemistry, College of Pharmacy, University of Utah,
3
0 S 2000 E, Salt Lake City, UT-84112, USA. E-mail: raphael.franzini@utah.edu
(2), and Boc-protected nitrogen (3) at position 7 of the BNBD
†
Electronic supplementary information (ESI) available: Detailed experimental
procedures, supporting figures, and tables. See DOI: 10.1039/c7cc03477f
bicycle and a (p-nitrophenylcarbamoyl)methyl substituent at
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
Scheme 1 7-aza/oxa-benzonorbornadiene (BNBD) derivatives as bioorthogonal drug release molecules. (a) Proposed mechanism of tetrazine-inducible
cargo release. (b) Synthesis of 7-aza/oxa-BNBD derivatives 1–3 and doxorubicin prodrug 5. Compound 6 is a control molecule used in kinetics experiments.
the bridgehead carbon (Scheme 1b). The different substituents
in 1–3 were selected to test the influence of the heterocycle on
reaction rate and cargo release. p-Nitroaniline (pNA) was used
as the reporter for bond-cleavage because a bathochromic shift
from lmax = 317 nm to lmax = 378 nm accompanies carbamate to
aniline conversion. The synthesis of 1–3 (Scheme 1b) started from
furfuryl alcohol (1a) and N-acetyl- or N-Boc-(2-hydroxymethyl)pyrroles
(2a, 3a). A [4+2] cycloaddition reaction of these heterocylces with
benzyne afforded the bicyclic structures 1d–3d. The pyrroles required
TBDMS-protection of the hydroxymethyl group (2b, 3b) for the
benzyne reaction. Reaction of hydroxymethyl bicycles 1d–3d
with (4-nitrophenyl)isocyanate provided the desired release
structures 1–3. The straightforward synthesis of these mole- Fig. 1 Release of pNA from 7-aza/oxa-BNBDs. HPLC analysis of reaction 2
cules compares favorably to TCO-prodrugs that require tedious with DPTz (c(1) = c(DPTz) = 6 mM, 24 h, RT). Inset: Color change resulting
from reaction of 1 and DPTz.
multi-step synthesis, including the separation of the axial from
the equatorial stereoisomer.
1
3
We first investigated whether the reaction of BNBD-derivatives kinetics, and the concentration dependence of the rate constants
with Tz released pNA (Fig. 1). Incubation of 1–3 with 3,6-di-(2- agreed with a second-order rate law. In DMSO, the second order rate
À1 À1
À1 À1
pyridyl)-1,2,4,5-tetrazine (DPTz) resulted in a distinct color constants (k
change, which indicated complete DPTz consumption and 0.0044 M
2
) were 0.015 M
s
s
for 1, 0.010 M
for 3 (Table 1 and Fig. S1, ESI†). The differences in
efficient pNA release. 1–3 were completely stable under these the rate constants reflect the increase in steric repulsion from 1
s
for 2, and
À1 À1
2
4
conditions (data not shown). HPLC monitoring of the reaction to 3 although electronic effects may also influence the rate.
(
c(1–3) = 6 mM, c(DPTz) = 18 mM, T = 37 1C, DMSO/H O (9 : 1)) Comparison of the kinetics of the reaction of DPTZ with 1 to that
2
À1 À1
confirmed complete consumption of 1–3, disappearance of with 7-oxo-BNBD (6, Scheme 1b; k = 0.176 M
s ) revealed that
2
DPTz, and formation of two new elution peaks that were the carbamoyl-methyl substituent decreased the reaction rate
identified as pNA and 3,6-di-(2-pyridyl)-1,2-pyridazine (DPPz) B12-fold. Presence of 10% H O accelerated the reaction 1.7 to
based on LC-MS (Fig. 1 and Fig. S2–S4, ESI†). The starting BNBD 1.9-fold (Table 1). This result is consistent with reported rate-
2
2
1,25
2
was the only perceptible peak with an absorbance maximum at enhancing effects of H O on IEDDA reactions.
We prepared a
lmax = 317 nm and intermediates I1–I3 or side-products with water-soluble Tz (PEG–Tz; Scheme 1a) to test the solvent effect
trapped pNA were undetectable. Isoindole/isobenzofuran on the kinetics. Increasing the H
2
O content further (DMSO/PBS,
À1 À1
decomposition products were visible in the HPLC traces at later 3 : 2) accelerated the reaction rate of 2 (k = 0.135 M
s ) and 1
2
À1 À1
measurements. These results demonstrated the rapid and high- (k = 0.190 M
s ) with PEG–Tz. The reaction of 1 and 2 with Tz
2
yielding release of pNA.
We further measured the kinetics of the reaction of 1–3 and Tz. aromatic azides (k
Photospectrometric analysis of DPTz disappearance (labs = 525 nm) of vinyl ethers (k
was significantly faster than the release reaction of TCO with
À1 À1 19
2
= 0.027 M
= 0.00021 M
s ), Tz-induced uncaging
À1 À1 20,21
2
s
),
and dissociative
À1 À1
1
in the presence of excess BNBDs revealed pseudo-first order Staudinger reactions (k
2
= B0.001 M
s ). Only the IDEEA
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017

View Article Online
ChemComm
Communication
Table 1 Second-order rate constants (k
tetrazines
2
) for reactions of 1–3 and (Fig. S5, ESI†). A singlet peak at 5.59 ppm was present in early
a
1
H NMR spectra but disappeared with prolonged incubation.
Integration of this peak accounted for the difference between
DMSO
90%DMSO/H
2
O
60%DMSO/PBS
consumed 1 and released pNA, and we postulate that it corresponds
to the heterocyclic intermediate I3 (Fig. S10, ESI†). In contrast,
consumption of 2 and DPPz formation occurred in parallel with
À1 À1
Pb Tz
[M
s ]
1
2
3
6
1
2
DPTz
DPTz
DPTz
DPTz
PEG–Tz
PEG–Tz
0.015 Æ 0.008
0.010 Æ 0.0004
0.028 Æ 0.0003
0.017 Æ 0.002
—
—
—
—
1
pNA generation, and H NMR peaks consistent with the structure of
0.0044 Æ 0.0009 0.0084 Æ 0.0016
0.176 Æ 0.004
—
I3 were absent (Fig. 2a and c). Peaks corresponding to the 1,4-isomer
of intermediate I2 were not visible but may be formed in quantities
below the experimental detection limit or as transient species. These
results demonstrate that isoindoles I3 release amines rapidly and
near-quantitatively.
—
—
0.058 Æ 0.0003 0.190 Æ 0.029
0.020 Æ 0.0007 0.135 Æ 0.010
a
The reactions were monitored by time-dependent absorbance mea-
surements at labs = 525 nm and T = 37 1C.
Given the Tz-induced liberation of pNA, we evaluated the
For several potential of BNBDs in a prodrug activation strategy. A 7-acetamide–
1
7,18
pyridazine release reaction surpasses this rate.
bioorthogonal cycloaddition reactions, simple structural changes BNBD doxorubicin prodrug (5) was synthesized via 2d (Scheme 1b).
were sufficient to enhance reaction rates by orders of magnitude HPLC analysis of the reaction between 5 and PEG–Tz (DMSO,
and efforts are ongoing to further accelerate the reaction of BNBDs T = 37 1C, c(5) = 0.2 mM, c(PEG–Tz) = 1.6 mM) showed rapid
with Tz for in vivo applications.
and complete doxorubicin release (Fig. S12, ESI†). We further
To quantify pNA release and to analyze the mechanism of the tested Tz-triggered drug release in cell viability assays. Addition
reaction of 1–3 and DPTz, we monitored the transformation by of doxorubicin-prodrug 5 followed by PEG–Tz caused dose-
1
H NMR (Fig. 2 and Fig. S5–S8, ESI†). Solutions of 1–3 and DPTz in dependent cytotoxicity in A549 pulmonary adenocarcinoma
6 6 2
DMSO-d or DMSO-d /D O (9 : 1) were incubated at 37 1C and cells (EC50(5 + 200 mM PEG–Tz) = 96 nM), rivaling that of
1
H NMR spectra were recorded periodically over 24 h. At these doxorubicin (EC50(Dox) = 88 nM, Table 2). Conversely, the
conditions, DPTz near-completely consumed BNBD derivatives prodrug 5 alone was essentially non-toxic in the tested concen-
within 2 h in agreement with results from HPLC analysis (Fig. 1) tration range (EC50(5) 4 10 mM; Table 2) and only at the highest
and kinetics measurements (Table 1). NMR peaks corresponding to concentrations was mild cytotoxicity observed. The combination
pNA emerged concomitant with disappearance of 1–3, indicating of high BNBD stability and facile drug release will be essential
rapid and efficient cargo liberation (Fig. 2). The measured pNA for achieving a high therapeutic index in targeted drug delivery
release from 1 and 2 in DMSO-d
range of 80–90%; release of pNA from 3 was less efficient (Fig. 2b). 100 mM PEG–Tz (EC50(5 + 100 mM PEG–Tz) = 99 nM) but
Experiments with PEG–Tz provided similar results (Fig. S9, ESI†). gradually decreased when lowering the PEG–Tz concentration
6 2
/D O at 6 h and 24 h was in the approaches. The cytotoxic effect of 5 was nearly preserved at
1
H NMR integration analysis revealed that in case of 1, further (Table 2). At the lowest tested concentration c(PEG–Tz) =
pNA formation was delayed relative to BNBD consumption 25 mM, the EC50 was 0.521 mM, which corresponds to 10–20%
1
Fig. 2 H NMR analysis of bioorthogonal bond cleavage reaction of BNBDs and Tz. (a) Time-dependent pNA release from 2. (b) Quantification of the
release of p-nitroaniline (pNA) from 7-aza/oxa-BNBD 1–3. (c) Concentrations of starting material (2) and reaction product (DPPz, pNA) as a function of
time. Fig. 1. Release of pNA from 7-aza/oxa-BNBDs. HPLC analysis of reaction 2 with DPTz (c(2) = c(DPTz) = 6 mM, 24 h, RT). Inset: Color change
resulting from reaction of 1 and DPTz.
This journal is ©The Royal Society of Chemistry 2017
Chem. Commun.

View Article Online
Communication
ChemComm
a
Table 2 Cytotoxicity of activated prodrug 5, 72 h incubation at 37 1C
BNBDs may find widespread applications in chemical biology for
example in detection probes and for enzyme activation.
We gratefully acknowledge financial support from the Uni-
versity of Utah, the Huntsman Cancer Institute, and the USTAR
initiative. We thank the group of Dr S. Sharma (HCI) for help
with cell experiments.
Compounds
Doxorubicin
EC50 (mM) in A549 cells
0.088 Æ 0.031
0.096 Æ 0.022
0.099 Æ 0.029
0.128 Æ 0.017
0.521 Æ 0.192
410
5
5
5
5
5
2
+ PEG–Tz (200 mM)
+ PEG–Tz (100 mM)
+ PEG–Tz (50 mM)
+ PEG–Tz (25 mM)
+ PEG–Tz (100 mM)
410
Notes and references
PEG–Tz
4200
1
2
3
E. M. Sletten and C. R. Bertozzi, Angew. Chem., Int. Ed. Engl., 2009,
8, 6974–6998.
J. M. Mejia Oneto, M. Gupta, J. K. Leach, M. Lee and J. L. Sutcliffe,
Acta Biomater., 2014, 10, 5099–5105.
R. Rossin, P. R. Verkerk, S. M. van den Bosch, R. C. Vulders, I. Verel,
J. Lub and M. S. Robillard, Angew. Chem., Int. Ed. Engl., 2010, 49,
a
The proliferation assay was performed in at least triplicate and EC50
values were derived from the normalized cell growth.
4
released doxorubicin and the combination with 5 was 420-fold
3375–3378.
more toxic than the prodrug alone, emphasizing the high
stability of 5 (Table 2). Tz molecules are rather non-toxic, and
4 J. Li and P. R. Chen, Nat. Chem. Biol., 2016, 12, 129–137.
5 R. M. Franzini and E. T. Kool, J. Am. Chem. Soc., 2009, 131,
13
1
6021–16023.
mice in a previous animal study showed no adverse reactions to
6
H. Wu, S. C. Alexander, S. Jin and N. K. Devaraj, J. Am. Chem. Soc.,
2016, 138, 11429–11432.
À1 8
repeated intravenous injection of doses as high as 1250 mmol kg
.
7
8
9
D. R. Bentley, et al., Nature, 2008, 456, 53–59.
J. Li, S. Jia and P. R. Chen, Nat. Chem. Biol., 2014, 10, 1003–1005.
J. Luo, Q. Liu, K. Morihiro and A. Deiters, Nat. Chem., 2016, 8,
1027–1034.
Indeed, control samples with PEG–Tz showed no cytotoxicity even
at the highest concentration tested (200 mM). Also, combination of
PEG–Tz and 2 resulted in minimal toxicity in the tested concen-
tration range (Fig. S15, ESI†) demonstrating that cells tolerate
isoindole decomposition products well.
We directly tested the stability of 5 in DMSO/PBS and human
serum. 5 was inert for 48 h, and no free doxorubicin or
doxorubicin-containing side products were observed by HPLC
1
1
1
1
0 G. Zhang, J. Li, R. Xie, X. Fan, Y. Liu, S. Zheng, Y. Ge and P. R. Chen,
ACS Cent. Sci., 2016, 2, 325–331.
1 M. Mondal, R. Liao, L. Xiao, T. Eno and J. Guo, Angew. Chem., Int. Ed.
Engl., 2017, 56, 2636–2639.
2 J. M. Mejia Oneto, I. Khan, L. Seebald and M. Royzen, ACS Cent. Sci.,
2016, 2, 476–482.
3 R. Rossin, S. M. van Duijnhoven, W. Ten Hoeve, H. M. Janssen,
L. H. Kleijn, F. J. Hoeben, R. M. Versteegen and M. S. Robillard,
Bioconjugate Chem., 2016, 27, 1697–1706.
(
Fig. S11 and S13, ESI†). The quantity of 5 decreased with
longer incubation times but no free doxorubicin was detectable
Fig. S13, ESI†). In light of the instability of doxorubicin in
14 M. J. Hangauer and C. R. Bertozzi, Angew. Chem., Int. Ed. Engl., 2008,
(
47, 2394–2397.
serum, we reasoned that decomposition of doxorubicin rather 15 R. van Brakel, R. C. Vulders, R. J. Bokdam, H. Grull and M. S. Robillard,
Bioconjugate Chem., 2008, 19, 714–718.
6 K. Gorska, A. Manicardi, S. Barluenga and N. Winssinger, Chem.
Commun., 2011, 47, 4364–4366.
than the BNBD linker caused the observed effect. To test this
hypothesis, we measured the serum-stability of 2 and this
1
BNBD was completely stable until the end of the analysis at 17 R. M. Versteegen, R. Rossin, W. ten Hoeve, H. M. Janssen and
M. S. Robillard, Angew. Chem., Int. Ed. Engl., 2013, 52, 14112–14116.
one week and no traces of pNA were formed (Fig. S14, ESI†).
1
8 X. Fan, Y. Ge, F. Lin, Y. Yang, G. Zhang, W. S. Ngai, Z. Lin, S. Zheng,
J. Wang, J. Zhao, J. Li and P. R. Chen, Angew. Chem., Int. Ed. Engl.,
2016, 55, 14046–14050.
Additionally, BNBD derivatives are expected to retain their
reactivity in contrast to TCO-derived molecules that gradually
1
3,17
19 S. S. Matikonda, D. L. Orsi, V. Staudacher, I. A. Jenkins, F. Fiedler,
deactivate by spontaneous trans/cis isomerization.
J. Chen and A. B. Gamble, Chem. Sci., 2015, 6, 1212–1218.
In conclusion, we have developed a novel and promising
bioorthogonal reaction for the traceless release of a molecular
cargo. In our design, derivatives of BNBD react with Tz and liberate
a drug or optical reporter via unprecedented hydrolysis-susceptible
isoindole/isobenzofuran intermediates. The reaction exhibits
favorable characteristics including quantitative as well as
prompt payload release, rapid bimolecular reaction, low reagent
toxicity, and exceptional probe stability at physiological conditions.
The straightforward synthesis of the BNBD precursors is another
advantage of the release design. This reaction boasts great promise
for application in reaction-based prodrug strategies. Additionally,
2
0 K. Neumann, S. Jain, A. Gambardella, S. E. Walker, E. Valero,
A. Lilienkampf and M. Bradley, ChemBioChem, 2017, 18, 91–95.
1 E. Jimenez-Moreno, Z. Guo, B. L. Oliveira, I. S. Albuquerque,
A. Kitowski, A. Guerreiro, O. Boutureira, T. Rodrigues, G. Jimenez-
Oses and G. J. Bernardes, Angew. Chem., Int. Ed. Engl., 2016, 56,
243–247.
2
2
2
2
2
2 G. M. Priestley and R. N. Warrener, Tetrahedron Lett., 1972, 13,
4295–4298.
3 H. Wu, B. T. Cisneros, C. M. Cole and N. K. Devaraj, J. Am. Chem.
Soc., 2014, 136, 17942–17945.
4 A. C. Knall, M. Hollauf and C. Slugovc, Tetrahedron Lett., 2014, 55,
4763–4766.
5 A. C. Knall, M. Hollauf, R. Saf and C. Slugovc, Org. Biomol. Chem.,
2016, 14, 10576–10580.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2017