Harnessing Hypoxia-Dependent Cyanine Photocages for In Vivo Precision Drug Release

Article abstract of DOI:10.1002/anie.202017349

Photocaging holds promise for the precise manipulation of biological events in space and time. However, current near-infrared (NIR) photocages are oxygen-dependent for their photolysis and lack of timely feedback regulation, which has proven to be the major bottleneck for targeted therapy. Herein, we present a hypoxia-dependent photo-activation mechanism of dialkylamine-substituted cyanine (Cy-NH) accompanied by emissive fragments generation, which was validated with retrosynthesis and spectral analysis. For the first time, we have realized the orthogonal manipulation of this hypoxia-dependent photocaging and dual-modal optical signals in living cells and tumor-bearing mice, making a breakthrough in the direct spatiotemporal control and in vivo feedback regulation. This unique photoactivation mechanism overcomes the limitation of hypoxia, which allows site-specific remote control for targeted therapy, and expands the photo-trigger toolbox for on-demand drug release, especially in a physiological context with dual-mode optical imaging under hypoxia.

Full text of DOI:10.1002/anie.202017349

Angewandte
Research Articles
Chemie
How to cite: Angew. Chem. Int. Ed. 2021, 60, 9553–9561
International Edition: doi.org/10.1002/anie.202017349
German Edition: doi.org/10.1002/ange.202017349
Controlled Release
Harnessing Hypoxia-Dependent Cyanine Photocages for In Vivo
Precision Drug Release
Yutao Zhang, Chenxu Yan, Qiaoqiao Zheng, Qian Jia, Zhongliang Wang, Ping Shi, and
Zhiqian Guo*
Abstract: Photocaging holds promise for the precise manip-
ulation of biological events in space and time. However, current
near-infrared (NIR) photocages are oxygen-dependent for
their photolysis and lack of timely feedback regulation, which
has proven to be the major bottleneck for targeted therapy.
Herein, we present a hypoxia-dependent photo-activation
mechanism of dialkylamine-substituted cyanine (Cy-NH)
accompanied by emissive fragments generation, which was
validated with retrosynthesis and spectral analysis. For the first
time, we have realized the orthogonal manipulation of this
hypoxia-dependent photocaging and dual-modal optical sig-
nals in living cells and tumor-bearing mice, making a break-
through in the direct spatiotemporal control and in vivo
feedback regulation. This unique photoactivation mechanism
overcomes the limitation of hypoxia, which allows site-specific
remote control for targeted therapy, and expands the photo-
trigger toolbox for on-demand drug release, especially in
a physiological context with dual-mode optical imaging under
hypoxia.
makes it difficult to directly achieve efficient bond cleavage
in the context of living tissue.[4] Current NIR light-mediated
cages mainly focus on the formation of singlet oxygen (1O2),
via oxidizing of electron-rich olefins, resulting in bond-
cleavage to restore bioactive molecules.[5] Here the major
bottleneck of NIR photocages is that their photolysis reliance
on O2 consumption while hypoxia is ubiquitously aberrant at
tumor.[6] Another challenge in caging design is difficult to
simultaneously acquire in vivo feedback information for
precise control in space and time. Hence, under hypoxia, it is
an urgent conundrum to design such high-performance NIR
photocages for specific targeting therapy.
Cyanine is a well-known NIR chromophore with high
molar extinction coefficient and excellent biocompatibility.[7]
However, its conjugated polymethine skeleton is inherently
prone to light-dependent decomposition.[8] In this regard,
converting the liability of cyanine bleaching for regulating
photo-reactivity might be a novel and feasible strategy to
engineer the NIR photocages. Regrettably, the reported
cyanine-based photocages are all O2-dependent consumption,
1
Introduction
which ascribes to the O2-mediated cleavage of polyenes.[9]
For instance, as a cage scaffold, heptamethine cyanines were
all studied under sufficient O2 supplement (Figure 1), and
meanwhile their on-off fluorescence response mode made it
difficult to achieve timely feedback, thereby limited their
biological applicability. Aiming at these issues, the discovery
of hypoxia-dependent photoactivation that reverses O2 con-
sumption could conquer the limitation of tumor hypoxia, thus
perform precisely spatiotemporal light-mediated therapy.
Herein, we describe a unique cyanine-based photocage
under hypoxia, in which its photolysis simultaneously produ-
ces dual-channel fluorescence and photoacoustic signals to
permit finely manipulation in space and time. The hypoxia-
dependent cage platform is firstly validated through retro-
synthetic dialkylamine-substituted cyanine (Cy-NH), which
undergoes retro-aldol, hydrolysis reaction to form a unique
fragment under NIR light (> 650 nm) irradiation (Figure 1).
Subsequently, the feasibility of cyanine-based photocages for
targeted therapy is demonstrated via an elaborated P(Cy-N-
CPT), which is sequence-activated by tumor acidic pH and
light irradiation under hypoxia. In P(Cy-N-CPT), diblock
copolymer containing ionizable tertiary amine renders ultra-
sensitive pH response for improved tumor-targeting, while
the programmable Cy-NH photoactivation leads to dual-
channel fluorescence and photoacoustic signals (PA).
Light-mediated techniques with noninvasive nature and
remote control have become a revolutionary tool in chemical
biology.[1] Photocages that convert light into chemical energy
to manipulate bioactive molecules allows finely spatiotem-
poral control.[2] Due to the unique tissue-transparent charac-
teristic and low phototoxicity, near-infrared (NIR) photons
have attracted immense attention for the design of such
photocages.[3] However, the low energy of NIR-photons
[*] Y. Zhang, Dr. C. Yan, Prof. Dr. Z. Guo
Key Laboratory for Advanced Materials and Institute of Fine
Chemicals, Feringa Nobel Prize Scientist Joint Research Center,
Frontiers Science Center for Materiobiology and Dynamic Chemistry,
School of Chemistry and Molecular Engineering, East China
University of Science & Technology
Shanghai 200237 (China)
E-mail: guozq@ecust.edu.cn
Q. Zheng, Prof. Dr. P. Shi, Prof. Dr. Z. Guo
State Key Laboratory of Bioreactor Engineering, East China University
of Science and Technology, Shanghai 200237 (P. R. China)
Q. Jia, Prof. Dr. Z. Wang
Engineering Research Center of Molecular-imaging and Neuro-
imaging of Ministry of Education, School of Life Science and
Technology, Xidian University, Xi’an, Shaanxi 710126 (China)
This hypoxia-dependent P(Cy-N-CPT) is proven to be
a highly successful one for site-specifically photorelease in
living cells and mice. In fact, this uncaging strategy and dual-
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202017349.
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561
ꢀ 2021 Wiley-VCH GmbH
9553

Angewandte
Research Articles
Chemie
Figure 1. A) The reported photoactivation mechanism of cyanine is via 1O2-mediated oxidative photocleavage of polyenes, and reliance on O2-
dependent consumption.[9] B) Competitive photoactivation pathways dependent on oxygen concentration. With the light irradiation under hypoxia,
free radicals are possibly generated from Cy-NH, then result in the cleavage of Cy-NH, and concurrently produce dual-channel fluorescence
output, thereby allowing precise NIR light-mediated therapy. Note: In Path 2, the photo-induced electron transfer (PeT) process occurs from the
amino group to the methine chain of Cy-NH under hypoxia. The reference Cy-C without a secondary amine group further demonstrated that the
amino group as electron donor plays the key role for the PeT process. This proposed radical-mediated photolysis is consistent with the reported
cyanines.[10f] Finally, with the help of water as a reactant, the photolysis fragments H-Cy are obtained.
modal (dual-channel fluorescence and PA) output are opti-
cally orthogonal: the former allows for directly spatiotempo-
ral control, and the latter permits real-time observation and in
vivo feedback regulation. In detail, PA from the initial
compact micelle of P(Cy-N-CPT) is utilized for in vivo
tracking its biodistribution. When accumulated in acidic
tumor, P(Cy-N-CPT) becomes into dissociated state with
remarkable NIR fluorescence. Subsequently, under hypoxic
tumor, NIR light-initiated photolysis leads to dual-channel
fluorescence for monitoring drug release without blind spot.
This discovery of hypoxia-dependent NIR photocage, com-
bined with dual-modal optical signals, could open up a new
route toward orthogonal remote control through time and
dose of light, allowing site-specifically manipulation for in
vivo targeting therapy.
sumption through the formation of dioxetane (Figure 1).
However, we serendipitously observed and discovered that
even under hypoxia, the photolysis of Cy-NH remains
efficiency by a convenient 670 nm (Æ 10 nm) laser source
(10 mWcmÀ2).
Seeking insights for this photolysis of heptamethine
cyanine under hypoxia, we carried out a set of mechanism
experiments (Figure 2). First of all, high-resolution electro-
spray ionization mass spectrometry (ESI-HRMS, Figure 2A)
was employed to reveal the photolysis fragmentation of Cy-
NH under hypoxic condition by a convenient laser source.
Notably, it was found that the peak at m/z 718.43 (corre-
sponding to [Cy-NHÀC + H+]) and 400.23 (corresponding to
[H-Cy + H+]) could identify the possible radical-based photo-
activation. In detail, the peak at m/z 718.43 could be ascribed
to the reaction of between Cy-NHÀC and H3O+ under NIR-
light irradiation. It seems that the photo-induced electron
transfer (PeT) process from an amino group to the Cy-NH
triplet is very likely occurred,[10] then the exchange of “H”
occurred through a polar process.[11] The elaborated Cy-C as
a reference further confirmed the key role of an amine group
for the PeT process (Figure S1), and thereby leading to the
photolysis of Cy-NH under hypoxia. In the meantime, the
peak at m/z 400.23 could be attributed to H-Cy as the
photolysis fragmentation (Figure 2A), which is clearly differ-
Results and Discussion
Hypoxia-Dependent Photolysis of N-Substituted Heptamethine
Cyanine
The defining feature of cyanines is an odd-carbon polyene
linker with two nitrogen-containing heterocycles, but it is
susceptible to photocleavage of the polyene chain. Indeed,
under normoxia, the 1O2- mediated photooxidation that
derives from heptamethine cyanine photolysis are well
studied. Clearly, this photooxidation of Cy-NH is O2 con-
1
ent from the O2-mediated photolysis fragmentation at the
peak of m/z 440.22 (Figure S2). All these results provided
strong evidences that the hypoxia-dependent photolysis of
9554
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561

Angewandte
Research Articles
Chemie
fragments of Cy-NH. Clearly, the retention
time of the light-initiated fragment of Cy-NH
under hypoxia was observed at 8.30 min,
which is the same as that of H-Cy (Figure 2E
and Figure S10–S13). Meanwhile, the absorp-
tion spectra mapped by the diode array
detector are completely matched (Figure S12
and S13). It was convinced that the light-
initiated fragment of Cy-NH under hypoxia is
H-Cy. Furthermore, the excitation and emis-
sion spectra of the light-initiated fragment of
Cy-NH under hypoxia are nearly identical
with the H-Cy, suggesting that they are the
same compound (Figure 2F, Figure S14, and
Table S1–S2). Taken together, all these results
provided solid evidences that under hypoxic
conditions, the light-initiated photoactivated
pathway of Cy-NH entails cleavage to gener-
ate H-Cy.
Hypoxia-Dependent Dual-Mode NIR Photocages
After reviewing this hypoxia-dependent
photoactivation, we reasoned that Cy-NH
could be a promising NIR light-initiated cage
scaffold for optically orthogonal manipulation
of fluorescence and photoacoustic (PA) sig-
nals. The sequence-activated photocage P(Cy-
N-CPT) was elaborately designed (Figure 3),
which is composed of an ionizable tertiary
amine-containing diblock copolymer as ultra
Figure 2. A) ESI-HRMS spectra of Cy-NH after NIR-light irradiation; spectral profiles of
Cy-NH (10 mM) were carried out with a 670 nm (Æ10 nm) laser source (10 mWcmÀ2
)
under B,C) hypoxia or D) normoxia in aqueous solution. E) HPLC traces of Cy-NH and H-
Cy in air and N2. F) Excitation and emission spectra of H-Cy and the NIR light-initiated
fragments of Cy-NH under hypoxia.
Cy-NH occurs (Figure 1B and Figure S3–S4), that is, free
radicals are generated from Cy-NH under NIR light irradi-
ation, and thereby with the help of water as a reactant the
photolysis fragments of H-Cy are obtained under hypoxia.
To further verify this proposed radical-mediated photo-
activation under hypoxia, the retrosynthetic analysis of
photolysis was carried out (Figure S5) and the synthetic route
of H-Cy was elaborately depicted in Scheme S2. Luckily, we
successfully obtained the synthesized H-Cy (Figure S6–S8).
Subsequently, the spectra of H-Cy and the photolysis frag-
ments of Cy-NH were analyzed. Specifically, the spectra of
Cy-NH were recorded under normoxia or hypoxia with NIR-
light irradiation, respectively. As shown in Figure 2B, under
hypoxia, the NIR absorbance of Cy-NH at ca. 740 nm
decreased in aqueous solution, while a concomitant new
broad absorption peak appeared. Notably, the photolysis of
Cy-NH under hypoxia showed a new strong emission with
a tailed long-wavelength band (Figure 2C). In contrast, under
normoxic with the same NIR light irradiation, there was
observed similar absorption spectral change but non-fluores-
cence (Figure 2D and Figure S9). It was indicated that the
photolysis fragmentation of Cy-NH under hypoxia is obvi-
ously distinct from that of under normoxia via well-defined
1O2-mediated photoactivation.
pH-sensitive unit to small pH differences between acidic
tumor cells and blood, and is covalently bonded with
a fluorescent Cy-NH cage for dual-modal optical signals.
We, therefore, performed light-initiated activation in
aqueous solution to examine the feasibility of dual-modal
optical signals. Firstly, the low critical micelle concentration of
P(Cy-N-CPT) at 1.06 mgmLÀ1 indicated its initial formation of
compact nanoparticles in aqueous solution (Figure 4A). The
results of transmission electron microscopy (TEM) and
dynamic light scattering (DLS) confirmed that P(Cy-N-
CPT) forms compact and uniform nanoparticles with a hy-
drated particle size of about 60 nm (Figure 4B,C). Interest-
ingly, when the hydrophobic block is protonated, the original
compact micelle becomes disassemble (Figure 4D). Concur-
rently, the PA properties of P(Cy-N-CPT) towards concen-
tration and assembled or disassembled micelles were inves-
tigated. Notably, the PA signals were plotted in a normalized
phantom mold at various concentrations of P(Cy-N-CPT)
(Figure 4E,F). The PA intensities at 720 nm indicated the
superior PA properties of P(Cy-N-CPT) in the prosthesis
(Figure 4G). Notably, there was a good linear relationship
between the concentration and the mean PA signal (R2 =
0.986 and R2 = 0.987) (Figure S15). Thus, this intrinsic photo-
acoustic signal of P(Cy-N-CPT) could be utilized to real-time
monitor its bio-distribution.
To validate that H-Cy are completely the same as that of
the hypoxic photolysis of Cy-NH, we employed high-perfor-
mance liquid chromatography (HPLC) to trace the photolysis
Next, we focused on whether P(Cy-N-CPT) could sense
a subtle pH difference between the acidic endocytic organ-
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9555

Angewandte
Research Articles
Chemie
Figure 3. A) Schematic illustration of photocage with dual-modal PA and dual-channel fluorescence output. B) Schematic diagram of NIR light-
initiated tumor-targeted precision drug release. At normal tissue with neutral pH, the photocage self-assembles as a compact micelle, with
significant PA signal and non-fluorescence at 810 nm. At acidic tumor under hypoxia, the initial compact micelle dissociates with a dramatically
increased fluorescence signal at 810 nm (FL channel 1). Subsequently, NIR light-initiated dual-channel fluorescence (a remarkable shift from 810
to 530 nm, FL channel 2) combined with PA signals occurs for tracking site-specifically activation of camptothecin (CPT) release.
Figure 4. A) Critical micelle concentration (CMC) of P(Cy-N-CPT) at 378C. B) Size distribution of P(Cy-N-CPT). TEM images of P(Cy-N-CPT) in
C) neutral or D) acid PBS solution (Scale bar=100 nm). Photoacoustic spectra of P(Cy-N-CPT) in the normalized phantom mold at E) pH 5.8 and
F) 7.4 PBS solution. G) Photoacoustic images at 720 nm of P(Cy-N-CPT) as a function of the concentration in the normalized phantom mold.
elles of cancer cells (5.0–6.0) and blood (7.4). As shown in
Figure 5A, P(Cy-N-CPT) displayed a sharp pH transition
(pKa = 6.0, DpHON/OFF = 0.3). At pH 7.4, P(Cy-N-CPT)
formed compact self-assembled micelles, which completely
9556
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561

Angewandte
Research Articles
Chemie
Figure 5. A) Fluorescence intensity as a function of pH alteration. P(Cy-N-CPT) keeps silent at high pH (7.4, 100 mM PBS buffer). At pH values
below its pKa transition (pH 6.0), P(Cy-N-CPT) is activated as a result of micelle dissociation. The pH response is extremely sharp
(DpHON/OFF =0.3). B) Emission spectra (lex =720 nm) with various pH values (from 6.2 to 5.9). C) Reversibility study of P(Cy-N-CPT) toward pH
at 378C. D) Stability study of P(Cy-N-CPT) in PBS and fetal bovine serum (FBS) over 48 h at 378C. Spectral characterization of P(Cy-N-CPT) at
pH 5.9 in PBS buffer under hypoxia upon light irradiation (670Æ10 nm, 20 mWcmÀ2) E) absorption and emission spectra at F) lex =720 nm and
G) lex =480 nm. H) Time-dependence of fluorescence changes at 810 and 530 nm. Fluorescence responses P(Cy-N-CPT) toward various amino
acids, enzymes and serum potential markers in PBS buffer (pH 7.4) at 378C: I) lem =810 nm and J) lem =535 nm, including blank, glycine, L-
glutamine, L-arginine, isoleucine, threonine, histidine, lysozyme, pepsin, SNA, tyrosinase, GSH, cysteine, Ab, glucose, BSA, MgSO4, NaCl, GGT,
CaCl2, H+ (pH 5.9), H+ +light (pH 5.9); each concentration is 50 mM.
silences the emission at 810 nm of Cy-N-CPT unit (Fig-
ure 5B). The phenomenon of the fluorescence “OFF” state
resulting from the compact micelles could be further con-
firmed by the typical ACQ (aggregation-caused quenching)
effect (Figure S16). In contrast, when the pH was decreased to
5.9, the original compact micelle was quickly disassembled,
thereby leading to a dramatic light-up emission at 810 nm
(Figure 5A,B). P(Cy-N-CPT) exhibited reversible pH-in-
duced micelle transition with excellent sensitivity (Fig-
ure 5C), which is consistent with the TEM imaging results.
Notably, the initial formed compact nanoparticles of P(Cy-N-
CPT) kept stable for a long time in serum and neutral medium
(Figure 5D), thereby protecting this photocage before photo-
activation in heterogenous biological environment.
We subsequently performed light-initiated photolysis to
validate dual-channel fluorescence output (a remarkable shift
from 810 to 530 nm). As expected, upon NIR light irradiation
with a convenient laser source, P(Cy-N-CPT) (at pH 5.9,
below the transition) exhibited a remarkable shift change in
absorption spectra, along with color change from blue to red
in aqueous solution. As shown in Figure 5E, the absorption
peak of P(Cy-N-CPT) at 720 nm was decreased and concom-
itantly a new broad absorption band was observed. Mean-
while, the emission spectra of P(Cy-N-CPT) underwent
remarkable changes: the original NIR fluorescence at
810 nm (lex = 720 nm) was gradually disappeared, and a new
broad emission appeared and sharply increased (Fig-
ure 5F,G). It took about 180 min to reach the equilibrium of
the NIR light-triggered photolysis (Figure 5H). Moreover,
the interference analysis revealed that P(Cy-N-CPT) shows
excellent selectivity over other biological analytes including
amino acids, and proteins (Figure 5I,J). This dual-channel
fluorescence output of P(Cy-N-CPT) allows for real-time
monitoring of the NIR light-initiated release through time
and dose of light without blind spot.
Clearly, the aforementioned light-initiated dual-mode
optical signals (PA and dual-channel fluorescence output)
could achieve precisely orthogonal manipulation in space and
time for targeted therapy (Figure 3). In detail, the significant
PA signal shows only slight changes during the micellar
dissociation, thus tracking the delivery and biodistribution of
P(Cy-N-CPT). Concurrently, the dramatic light-up fluores-
cence at 810 nm accurately guides for in vivo uncaging, and
subsequently a remarkable shift from 810 to 600 nm (tailed
emission) timely monitors this photolysis and camptothecin
(CPT) release. Meanwhile, as feedback information, the NIR
light-triggered fluorescence signal at 600 nm can regulate the
external light dose (Figure 3B). The hypoxia-dependent
photolysis and dual-modal optical output are orthogonal,
enabling NIR light-initiated irradiation to precisely manipu-
late where, when and how the intact and active drugs are
delivered.
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9557

Angewandte
Research Articles
Chemie
Tracking NIR Light-Initiated In Vivo Behavior in Living Cells with
Dual-Channel Fluorescence
cellular uptake. In contrast, upon 670 nm light irradiation,
a sharply decreased NIR fluorescence at 810 nm was found,
and concomitantly a new enhanced emission at 535 nm was
observed. Clearly, this dual-channel fluorescence output
monitored the behaviors of the quick uptake and specific
activation in a light-initiated manner (Figure 6B,D). These
cell imaging results suggested that P(Cy-N-CPT) could be
visualized to effectively sense the acidic endocytosis of tumor
cells and then specific light-triggered release with dual-
channel fluorescence changes.
To evaluate the light-dependent cytotoxicity, cell viability
experiments of P(Cy-N-CPT) and its monomer Cy-N-CPT
towards cancer cells (A549 and HepG2) were performed by
MTT assay. In the absence of NIR light irradiation, P(Cy-N-
CPT) and its monomer Cy-N-CPT exhibited negligible
cytotoxic effects on cancer cells at the studied concentrations
(0–10 mM), indicative of their good biocompatibility. After
the external NIR light irradiation, P(Cy-N-CPT) showed
significant cytotoxicity even at low concentrations (Fig-
ure 6E,F). Obviously, external light initiation is a key factor
for the cytotoxicity of P(Cy-N-CPT) with the release of active
CPT but no preference for the cell type. All these results
clearly indicate that P(Cy-N-CPT) is solely light-dependent
cytotoxicity in living cells. As a consequence, through the
orthogonal time and dosage of light irradiation, the cellular
behavior of P(Cy-N-CPT) becomes observable with the aid of
dual-channel fluorescence, from effectively targeting tumor
cells to specifically light-initiated drug release.
We next investigated the feasibility of this light-initiated
photocage for real-time tracking of cellular uptake and
precise drug release for further biological application. A549
(Human adenocarcinoma cell) and HepG2 cells (Human
hepatoma cell) were selected as cell models. Owing to the
ultra-sensitive pH response, it is expected that P(Cy-N-CPT)
keeps non-fluorescence in normal tissue, while specifically
senses acidic tumor with light-up fluorescence. To gain insight
into the light-mediated behavior, the intracellular photolysis
of P(Cy-N-CPT) was intuitively observed by confocal laser
microscopy. As shown in Figure 6, upon incubation P(Cy-N-
CPT) with A549 and HepG2 cells, a ca. 20-fold enhancement
of NIR emission at 810 nm was observed (Figure 6A,C and
Figure S17). Notably, without washing process, the sharp
enhancement fluorescence demonstrated that P(Cy-N-CPT)
could accurately identify the acid endocytic organelles of
tumor cells. Meanwhile, there was observed only one-channel
NIR fluorescence without light irradiation, indicating that this
photocage remains intact and non-specific cleavage during
Mapping the Biodistribution via Photoacoustic Signals
Photoacoustic imaging is a powerful imaging modality
capable of scalable spatial resolution and centimeter pene-
tration.[12] When designing PA probe, it is essential for the
chromophore featuring three key characteristics of NIR
absorbance, large extinction coefficient and low fluorescence
quantum.[13] In this cyanine-based photocage, P(Cy-N-CPT)
initially forms compact micelles under natural physiological
condition, thus leading to completely fluorescence quenching,
while it still has a NIR absorbance with large extinction
coefficient. Thus, these characteristics of P(Cy-N-CPT) make
it as a perfect candidate to achieve significant photoacoustic
signals, then for tracking in vivo biodistribution.
The high-resolution photoacoustic images of P(Cy-N-
CPT) were recorded in a time-dependent method after
intravenous injection with A549 xenograft tumor-bearing
mice. As shown in Figure 7 (top), before intravenous injec-
tion, the tumor site exhibited initial weak PA signals at
720 nm on account of the intrinsic background signal. Where-
as, at 24 h post-injection of P(Cy-N-CPT), the PA signal
within the tumor showed a dramatic increment and reached
up to a maximum, suggestive of the effective tumor-targeted
accumulation (Figure 7 top and Figure S18). Furthermore, the
3D images of the tumor region provided solid evidence that
P(Cy-N-CPT) passively permeated into the deep tumor
(Figure 7, bottom). In contrast, other organs (such as spleen
and intestine) displayed very weak photoacoustic signal.
Obviously, the PA signals of P(Cy-N-CPT) could track it in
Figure 6. In vitro analysis of P(Cy-N-CPT) in A549 and HepG2 (cancer
cells) with and without light irradiation (670Æ10 nm, 20 mWcmÀ2).
Confocal microscopy images of cells incubated with P(Cy-N-CPT)
(10 mM) for 2 h: A,C) without light irradiation (À); B,D) under light
irradiation (+). The images were collected (A1, B1, C1 and D1) at the
red channel (780Æ20 nm, lex =633 nm) and (A2, B2, C2 and D2) at
the green channel (535Æ20 nm, lex =488 nm). E,F) Light-dependent
cytotoxicity of P(Cy-N-CPT) in cell lines.
9558
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561

Angewandte
Research Articles
Chemie
venous injection of A549 xenograft tumor-bearing mice,
a specific tumor-targeted accumulation of P(Cy-N-CPT) was
visualized (Figure 8A,B). Consistently, the ex vivo imaging of
the excised tumors further confirmed a high accumulation
and then acidic activation with NIR light irradiation at tumor
site (Figure 8C,D, and Figure S19, S20). Accordingly, there
was much weaker fluorescence in the liver and almost no
fluorescence in the other organs (heart, liver, spleen, lung,
and kidney). The excellent tumor-targeting ability could be
attributed to the synergistic effects of P(Cy-N-CPT): passive
targeting from the EPR effect and activatable targeting
(ultra-sensitive and reversible pH response). Notably, as
a control of H+ inhibition, NaHCO3 solution was intra-
tumorally injected before P(Cy-N-CPT) administration. In-
deed, there was almost no fluorescence signal in the NaHCO3
pre-treated group (Figure S21). All these results suggested
that P(Cy-N-CPT) possesses striking characteristics of tumor-
targeting, and then its self-assembly dissociation generates
significant NIR signals, thereby allowing the guidance of light
irradiation at the right site.
According to the in vivo imaging results without NIR light
irradiation, the accumulation of fluorescence signals at the
tumor site is time-dependent increased up to a maximum at
24 h (Figure 8A), which is consistent with the PAI results.
Then, at the optimal time for in vivo photoactivation, we
evaluated the light-initiated efficiency of P(Cy-N-CPT) at
24 h and the right position (timely fluorescence feedback on
810 nm). As shown in Figure 8E,F, under in vivo NIR-light
(670 nm) irradiation, the emission signal at 810 nm decreased
(yellow-hot channel), and concurrently new emission signal at
600 nm occurred (rainbow channel), confirming the desirable
real-time tracking properties for site-specifically drug deliv-
ery at the tumor. Furthermore, the ex vivo studies also
illustrated that the activated CPT was abundantly accumu-
lated at the tumor and scarcely any appeared in other organs
(Figure 8G,H). Both the in vivo and ex vivo imaging indicated
that P(Cy-N-CPT) could be in vivo tracked via the activatable
emission at 810 nm and thereby precisely guide the NIR light
irradiation; then the concurrent emission at 600 nm is utilized
for timely monitoring the drug release.
Figure 7. In vivo photoacoustic images of tumor tissue (white ellipse)
at different times (top) and the 3D-images (bottom) at 24 h after the
intravenous injection of P(Cy-N-CPT) (CPT-equivalent dose of
0.1 mgkgÀ1, lex =720 nm). The white arrow indicates the tumor site.
Note: In vivo mouse imaging experiments were repeated three times
(n=3).
vivo distribution and then visualize the location, shape, and
size of solid tumors for accurately guiding NIR-light irradi-
ation.
In this way, harnessing hypoxia-dependent photocaged
cyanine, we take the dual-modality of fluorescence and
photoacoustic imaging of P(Cy-N-CPT), tracking where,
when and how the intact and active prodrugs are delivered
in vivo. Specifically, photoacoustic signal has been serviced
for real-time in vivo tracing where the intact prodrugs are
located. At the same time, fluorescence signal (810 nm) from
tumor tissue can be employed as the spatial guidance for the
light irradiation, while according to timely feedback on
another channel fluorescence signal (600 nm), the site-specif-
ically initiation and dosage of light irradiation can finely
regulate the activation and drug release.
Harnessing Dual-Mode Optical Signals for Precision Drug
Release in Space and Time
We next validate the orthogonal application of this
hypoxia-dependent photocaging and dual-modal optical out-
put, precisely in vivo manipulating where, when and how the
intact and active drugs are delivered in living mice. In P(Cy-
N-CPT), the ultra-sensitive pH-induced activation under
acidic tumor and remarkable shift from dual-channel emis-
sion response made it perfect to real-time guide NIR-light Conclusion
stimulus and track the CPT release. In detail, the yellow-hot
channel (emission at 810 nm) represents the disassemble of
P(Cy-N-CPT) and the rainbow channel (emission at 600 nm)
monitors light-triggered active drug release. After the intra-
In summary, we revealed a hypoxia-dependent cyanine-
based photolysis mechanism, where P(Cy-N-CPT) leads to
dual-mode optical signals under hypoxia, enabling precisely
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9559

Angewandte
Research Articles
Chemie
Figure 8. In vivo dual-channel NIR-fluorescence imaging of A549 xenograft tumor-bearing mice at various time (0, 3, 6, 9, 12, 24 h or under
irradiation 10 min, 0.5 h, 1 h, 1.5 h and 2 h) after intravenous injection of P(Cy-N-CPT) E–H) with or A–D) without light irradiation (670Æ10 nm,
200 mWcmÀ2) administered at a CPT-equivalent dose of 0.1 mgkgÀ1. The white arrow indicates the tumor site. Ex vivo NIR-fluorescence imaging
of excised organs (heart, liver, spleen, lung, kidney and tumor) at 24 h or 36 h after the intravenous injection of P(Cy-N-CPT) (C, D, G and H).
Note: In vivo mouse imaging experiments were repeated three times (n=3).
remote control in space and time. We conducted a detailed Acknowledgements
retrosynthetic analysis and synthesized the corresponding
This work was supported by NSFC/China (21788102,
21878087 and 21908060), Shanghai Municipal Science and
Technology Major Project (Grant 2018SHZDZX03), the
Innovation Program of Shanghai Municipal Education Com-
mission, the Shuguang Program (18SG27), the China Post-
doctoral Science Foundation (2019M651417).
fragments of Cy-NH through Vilsmeier–Haack, Aldol, and
hydrolysis reaction, which confirmed our discovery of hypo-
xia-dependent photoactivation mechanism. This photocage
P(Cy-N-CPT) was sequentially triggered by acidic pH and
exogenous NIR light, thereby simultaneously producing dual-
channel fluorescence and PA signals. Notably, unlike 1O2-
mediated photooxidation reliance on O2-dependent con-
sumption, our discovery of this cyanine photolysis conquers
the limitation of tumor hypoxia, allowing remote control for Conflict of interest
targeted therapy. We have for the first time established finely
The authors declare no conflict of interest.
orthogonal manipulation of hypoxia-dependent photocaging
and dual-mode optical signals in living cells and tumor-
bearing mice, thereby realizing spatiotemporal control and in
vivo feedback regulation on light-mediated therapy. This
hypoxia-dependent photolysis design strategy for dual-mo-
dality imaging would inspire the creation of a new generation
of NIR photocages in a physiological context, thus greatly
expanding the photo-triggered toolboxes for deep tissue
penetration or phototherapy.
Keywords: cyanine ·fluorescence imaging ·
hypoxia-dependent photolysis ·near-infrared photocages ·
photoacoustic imaging
[1] a) H. B. Cheng, Y. Li, B. Z. Tang, J. Yoon, Chem. Soc. Rev. 2020,
49, 21 –31; b) X. Li, J. F. Lovell, J. Yoon, X. Chen, Nat. Rev. Clin.
Oncol. 2020, 17, 657 –674; c) J. B. Grimm, B. P. English, H. Choi,
A. K. Muthusamy, B. P. Mehl, P. Dong, T. A. Brown, J. Lippin-
cott-Schwartz, Z. Liu, T. Lionnet, L. D. Lavis, Nat. Methods
2016, 13, 985 –988; d) A. L. Vahrmeijer, M. Hutteman, J. R.
van der Vorst, C. J. van de Velde, J. V. Frangioni, Nat. Rev. Clin.
Oncol. 2013, 10, 507 –518; e) L. Wu, A. C. Sedgwick, X. Sun,
9560 www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561

Angewandte
Research Articles
Chemie
S. D. Bull, X. P. He, T. D. James, Acc. Chem. Res. 2019, 52, 2582 –
2020, 132, 12252 –12259; d) R. R. Nani, J. A. Kelley, J. Ivanic,
2597; f) S. Long, Q. Qiao, L. Miao, Z. Xu, Chin. Chem. Lett.
2019, 30, 573 –576.
M. J. Schnermann, Chem. Sci. 2015, 6, 6556 –6563; e) V.
Rapozzi, L. Beverina, P. Salice, G. A. Pagani, M. Camerin,
L. E. Xodo, J. Med. Chem. 2010, 53, 2188 –2196; f) Z. Guo, Y.
Ma, Y. Liu, C. Yan, P. Shi, H. Tian, W.-H. Zhu, Sci. China Chem.
2018, 61, 1293 –1300.
[2] a) T. Wei, S. Lu, J. Sun, Z. Xu, X. Yang, F. Wang, Y. Ma, Y. S. Shi,
X. Chen, J. Am. Chem. Soc. 2020, 142, 3806 –3813; b) W. Zhang,
F. Huo, Y. Yue, Y. Zhang, J. Chao, F. Cheng, C. Yin, J. Am. Chem.
Soc. 2020, 142, 3262 –3268; c) P. Klµn, T. Solomek, C. G. Bochet,
A. Blanc, R. Givens, M. Rubina, V. Popik, A. Kostikov, J. Wirz,
Chem. Rev. 2013, 113, 119 –191; d) M. Li, A. P. Dove, V. X.
Truong, Angew. Chem. Int. Ed. 2020, 59, 2284 –2288; Angew.
Chem. 2020, 132, 2304 –2308; e) Q. Lin, L. Yang, Z. Wang, Y.
Hua, D. Zhang, B. Bao, C. Bao, X. Gong, L. Zhu, Angew. Chem.
Int. Ed. 2018, 57, 3722 –3726; Angew. Chem. 2018, 130, 3784 –
3788; f) N. Umeda, H. Takahashi, M. Kamiya, T. Ueno, T.
Komatsu, T. Terai, K. Hanaoka, T. Nagano, Y. Urano, ACS
Chem. Biol. 2014, 9, 2242 –2246; g) Y. Chen, L. Li, W. Chen, H.
Chen, J. Yin, Chin. Chem. Lett. 2019, 30, 1353 –1360.
[9] a) J. Peng, A. Samanta, X. Zeng, S. Han, L. Wang, D. Su, D.
Loong, N. Kang, S. Park, A. H. All, W. Jiang, L. Yuan, X. Liu,
Y. T. Chang, Angew. Chem. Int. Ed. 2017, 56, 4165 –4169;
Angew. Chem. 2017, 129, 4229 –4233; b) A. P. Gorka, R. R.
Nani, J. Zhu, S. Mackem, M. J. Schnermann, J. Am. Chem. Soc.
2014, 136, 14153 –14159; c) A. P. Gorka, M. J. Schnermann,
Curr. Opin. Chem. Biol. 2016, 33, 117 –125; d) R. R. Nani, A. P.
Gorka, T. Nagaya, T. Yamamoto, J. Ivanic, H. Kobayashi, M. J.
Schnermann, ACS Cent. Sci. 2017, 3, 329 –337; e) M. S. Baptista,
J. Cadet, P. D. Mascio, A. A. Ghogare, A. Greer, M. R. Hamblin,
C. Lorente, S. C. Nunez, M. S. Ribeiro, A. H. Thomas, M.
Vignoni, T. M. Yoshimura, Photochem. Photobiol. 2017, 93,
912 –919.
[3] a) Y. Zhu, H. Jia, G. Pan, N. W. Ulrich, Z. Chen, F. Wu, J. Am.
Chem. Soc. 2018, 140, 4062 –4070; b) K. Mitra, C. E. Lyons,
M. C. T. Hartman, Angew. Chem. Int. Ed. 2018, 57, 10263 – [10] a) M. Heilemann, E. Margeat, R. Kasper, M. Sauer, P. Tinnefeld,
10267; Angew. Chem. 2018, 130, 10420 –10424; c) E. D. Ander-
son, A. P. Gorka, M. J. Schnermann, Nat. Commun. 2016, 7,
13378; d) E. Palao, T. Slanina, L. Muchova, T. Solomek, L. Vitek,
P. Klan, J. Am. Chem. Soc. 2016, 138, 126 –133; e) T. A. Shell,
J. R. Shell, Z. L. Rodgers, D. S. Lawrence, Angew. Chem. Int. Ed.
2014, 53, 875 –878; Angew. Chem. 2014, 126, 894 –897.
J. Am. Chem. Soc. 2005, 127, 3801 –3806; b) G. T. Dempsey, M.
Bates, W. E. Kowtoniuk, D. R. Liu, R. Y. Tsien, X. Zhuang, J.
Am. Chem. Soc. 2009, 131, 18192 –18193; c) A. P. Demchenko,
Methods Appl. Fluoresc. 2020, 8, 022001; d) V. Glembockyte, G.
Cosa, J. Am. Chem. Soc. 2017, 139, 13227 –13233; e) P.
Holzmeister, A. Gietl, P. Tinnefeld, Angew. Chem. Int. Ed.
2014, 53, 5685 –5688; Angew. Chem. 2014, 126, 5792 –5796; f) Y.
Gidi, L. Payne, V. Glembockyte, M. S. Michie, M. J. Schner-
mann, G. Cosa, J. Am. Chem. Soc. 2020, 142, 12681 –12689.
[11] a) H. Kasai, Z. Yamaizumi, M. Berger, J. Cadet, J. Am. Chem.
Soc. 1992, 114, 9692 –9694; b) J. Cadet, T. Douki, J.-L. Ravanat,
Acc. Chem. Res. 2008, 41, 1075 –1083.
[12] a) J. Ouyang, L. Sun, Z. Zeng, C. Zeng, F. Zeng, S. Wu, Angew.
Chem. Int. Ed. 2020, 59, 10111 –10121; Angew. Chem. 2020, 132,
10197 –10207; b) C. Liu, S. Zhang, J. Li, J. Wei, K. Mullen, M.
Yin, Angew. Chem. Int. Ed. 2019, 58, 1638 –1642; Angew. Chem.
2019, 131, 1652 –1656; c) J. Qi, Y. Fang, R. T. K. Kwok, X.
Zhang, X. Hu, J. W. Y. Lam, D. Ding, B. Z. Tang, ACS Nano
2017, 11, 7177 –7188; d) Z. Sheng, B. Guo, D. Hu, S. Xu, W. Wu,
W. H. Liew, K. Yao, J. Jiang, C. Liu, H. Zheng, B. Liu, Adv.
Mater. 2018, 30, 1800766; e) X. Zhen, J. Zhang, J. Huang, C. Xie,
Q. Miao, K. Pu, Angew. Chem. Int. Ed. 2018, 57, 7804 –7808;
Angew. Chem. 2018, 130, 7930 –7934.
[13] a) Y. Ren, A. C. Sedgwick, J. Chen, G. Thiabaud, C. V. Chau, J.
An, J. F. Arambula, X. P. He, J. S. Kim, J. L. Sessler, C. Liu, J.
Am. Chem. Soc. 2020, 142, 16156 –16160; b) D. Xi, M. Xiao, J.
Cao, L. Zhao, N. Xu, S. Long, J. Fan, K. Shao, W. Sun, X. Yan, X.
Peng, Adv. Mater. 2020, 32, 1907855; c) L. Teng, G. Song, Y. Liu,
X. Han, Z. Li, Y. Wang, S. Huan, X. B. Zhang, W. Tan, J. Am.
Chem. Soc. 2019, 141, 13572 –13581; d) Z. Hosseinaee, M. Le, K.
Bell, P. H. Reza, Photoacoustics 2020, 20, 100207; e) I. T. Ho,
J. L. Sessler, S. S. Gambhir, J. V. Jokerst, Analyst 2015, 140,
3731 –3737.
[4] a) J. M. Silva, E. Silva, R. L. Reis, J. Controlled Release 2019,
298, 154 –176; b) T. Solomek, J. Wirz, P. Klµn, Acc. Chem. Res.
2015, 48, 3064 –3072.
[5] a) C. G. Dariva, J. F. J. Coelho, A. C. Serra, J. Controlled Release
2019, 294, 337 –354; b) P. Rajaputra, M. Bio, G. Nkepang, P.
Thapa, S. Woo, Y. You, Bioorg. Med. Chem. 2016, 24, 1540 –
1549; c) P. Thapa, M. Li, M. Bio, P. Rajaputra, G. Nkepang, Y.
Sun, S. Woo, Y. You, J. Med. Chem. 2016, 59, 3204 –3214; d) F.
Zhang, Q. Wu, H. Liu, Wiley Interdiscip. Rev. Nanomed.
Nanobiotechnol. 2020, 12, 1643.
[6] a) X. Li, S. Yu, Y. Lee, T. Guo, N. Kwon, D. Lee, S. C. Yeom, Y.
Cho, G. Kim, J. Huang, S. Choi, K. T. Nam, J. Yoon, J. Am.
Chem. Soc. 2019, 141, 1366 –1372; b) X. Zhao, S. Long, M. Li, J.
Cao, Y. Li, L. Guo, W. Sun, J. Du, J. Fan, X. Peng, J. Am. Chem.
Soc. 2020, 142, 1510 –1517; c) H. Li, X. Li, W. Shi, Y. Xu, H. Ma,
Angew. Chem. Int. Ed. 2018, 57, 12830 –12834; Angew. Chem.
2018, 130, 13012 –13016; d) S. Y. Lim, K. H. Hong, D. I. Kim, H.
Kwon, H. J. Kim, J. Am. Chem. Soc. 2014, 136, 7018 –7025; e) N.
Narayanaswamy, S. Narra, R. R. Nair, D. K. Saini, P. Kondaiah,
T. Govindaraju, Chem. Sci. 2016, 7, 2832 –2841; f) J. V. Frangio-
ni, Curr. Opin. Chem. Biol. 2003, 7, 626 –634.
[7] a) X. Li, N. Kwon, T. Guo, Z. Liu, J. Yoon, Angew. Chem. Int.
Ed. 2018, 57, 11522 –11531; Angew. Chem. 2018, 130, 11694 –
11704; b) D. M. Gilkes, G. L. Semenza, D. Wirtz, Nat. Rev.
Cancer 2014, 14, 430 –439; c) M. S. Nakazawa, B. Keith, M. C.
Simon, Nat. Rev. Cancer 2016, 16, 663 –673; d) A. Sharma, J. F.
Arambula, S. Koo, R. Kumar, H. Singh, J. L. Sessler, J. S. Kim,
Chem. Soc. Rev. 2019, 48, 771 –813.
[8] a) L. Jiao, Y. Liu, X. Zhang, G. Hong, J. Zheng, J. Cui, X. Peng, F.
Song, ACS Cent. Sci. 2020, 6, 747 –759; b) X. Chen, X. Peng, A.
Cui, B. Wang, L. Wang, R. Zhang, J. Photochem. Photobiol. A
2006, 181, 79 –85; c) D. H. Li, C. L. Schreiber, B. D. Smith,
Angew. Chem. Int. Ed. 2020, 59, 12154 –12161; Angew. Chem.
Manuscript received: December 31, 2020
Revised manuscript received: February 4, 2021
Accepted manuscript online: February 10, 2021
Version of record online: March 17, 2021
Angew. Chem. Int. Ed. 2021, 60, 9553 –9561
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
9561