Alendronate-functionalized hypoxia-responsive polymeric micelles for targeted therapy of bone metastatic prostate cancer

Article abstract of DOI:10.1016/j.jconrel.2021.04.035

Bone metastasis is one of the leading causes of cancer-related death and remains incurable in spite of great efforts. Bone-targeted nanoparticle-based drug carriers can overcome the difficulties in delivering therapeutic agents to metastatic bone and endo

Full text of DOI:10.1016/j.jconrel.2021.04.035

Journal of Controlled Release 334 (2021) 303–317
Contents lists available at ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel
Alendronate-functionalized hypoxia-responsive polymeric micelles for
targeted therapy of bone metastatic prostate cancer
,
,
Mengmeng Longa 1, Xuemeng Liua 1, Xu Huanga, Min Lua, Xiaomei Wua, Lingyan Wenga,
Qiuping Chenb, Xueting Wanga *, Li Zhua *, Zhongping Chena
,
,
,*
a Institute of Special Environmental Medicine, Nantong University, Nantong, People's Republic of China
b Department of Anesthesiology, Affiliated Hospital of Nantong University, Nantong, People's Republic of China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Bone metastasis is one of the leading causes of cancer-related death and remains incurable in spite of great ef-
forts. Bone-targeted nanoparticle-based drug carriers can overcome the difficulties in delivering therapeutic
agents to metastatic bone and endowing them with a stimuli-responsive feature for controllable drug release can
further maximize their therapeutic outcome. In light of hypoxic microenvironment of bone metastasis, we herein
reported a bone-targeted and hypoxia-responsive polymeric micelle system for effective treatment of bone
metastatic prostate cancer. The micelles were self-assembled from a polyethylene glycol and poly-L-lysine based
copolymer using alendronate as a bone-targeted moiety and azobenzene as a hypoxia-responsive linker, showing
a high affinity to metastatic bone and a high sensitivity in responding to hypoxia in vitro. In vivo studies further
showed that after a selective accumulation in metastatic bone, the micelles could respond to hypoxic bone
metastasis for rapid drug release to an effective therapeutic dosage. As a result, the micelles could suppress tumor
growth in bone and inhibit bone destruction by inhibiting osteoclast activity and promoting osteoblast activity,
achieving an enhanced therapeutic outcome with relieved bone pain and prolonged survival time. Bone-targeted
and hypoxia-responsive nanocarriers therefore represent a promising advancement for treating bone metastasis.
To our best knowledge, it might be the first example of the application of hypoxia-responsive nanocarriers in
treating bone metastasis.
Bone metastasis
Polymeric micelles
Alendronate
Bone-targeted delivery
Hypoxia-responsive release
1. Introduction
pathological and physiological characteristics. As the gold standard
treatments, loco-regional radiotherapy and surgical resection fail to
Cancer metastasis is the leading cause of cancer-related death. Bone
is one of the most common metastatic organs, as its physiological en-
vironments favor homing, survival, colonization, and further growth of
metastatic cancer cells [1]. Bone metastasis has a high incidence in
different cancer types, especially in prostate cancer, the second common
cancer in over 50 years old men with more than 1 million newly diag-
nosed cases annually worldwide. Up to 65–90% of patients with prostate
cancer and nearly all patients who die from prostate cancer display
evidence of bone metastasis [2,3]. Patients with bone metastasis suffer
from a series of skeletal complications such as intolerable pain, fracture,
and hypercalcemia, which lead to increased morbidity, decreased
quality of life, poor prognosis, and high mortality [4].
completely eliminate multiple nodules of bone metastasis and are often
accompanied by a risk of bone-marrow suppression, hypocytosis, and
even less life expectancy with aggravated pain. Conventional chemo-
therapy is the primary systemic treatment to suppress tumor growth;
however, it is difficult for therapeutic agents to achieve an effective
therapeutic dosage in metastatic bone via intravenous administration
[5], due to the marrow-blood barrier and lack of specificity. Moreover,
to acquire effective therapeutic magnitude in bone, therapeutic agents
need to be administered at a high dosage and dosing frequency, which
inevitably leads to dose-limiting side effects. The advancements in un-
derstanding the pathogenesis of bone metastasis shift treatment ap-
proaches from traditional chemotherapy to bone-targeted therapy,
aiming at pain relief and fracture delay for improved quality of life.
Among various bone-targeted agents targeting bone matrix or bone cells
In spite of great progress in medical management by a multidisci-
plinary approach, bone metastasis remains incurable, due to its unique
* Corresponding author.
E-mail addresses: wangxueting@ntu.edu.cn (X. Wang), zhulizhou@ntu.edu.cn (L. Zhu), chenzp@ntu.edu.cn (Z. Chen).
1
These authors contributed equally to this work
https://doi.org/10.1016/j.jconrel.2021.04.035
Received 14 January 2021; Received in revised form 20 April 2021; Accepted 27 April 2021
Available online 29 April 2021
0168-3659/© 2021 Elsevier B.V. All rights reserved.

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
(e.g., osteoblasts, osteoclasts, and osteocytes) [6], nitrogen-containing
bisphosphates (e.g., alendronate, zoledronate, pamidronate, and
olpadronate) are FDA-approved anti-resorptive agents clinically used to
inhibit the activity of osteoclasts, due to their high affinity to bone,
especially to cancer metastatic bone [7]. Unfortunately, they lack anti-
tumor activity and have no improvement in overall survival, making
them at best a palliative [2,8,9]. Moreover, high-dosage and long-term
usage of bisphosphates also contributes to undesirable side effects
such as bone turnover suppression, ocular dysfunction, and renal func-
tion impairment [10]. There is a need for continuous therapeutic
development against bone metastasis.
hypoxia-induced AZO cleavage for micelle disassembly. As a result, the
micelles are capable of suppressing tumor growth in bone and inhibiting
cancer-induced bone destruction and bone pain by balancing bone
turnover, achieving a significant improvement in quality of life and
overall survival (Scheme 1).
2. Materials and methods
2.1. Materials
Methoxy-polyethylene glycol-amine (mPEG-NH2, Mw = 2000),
carboxy-polyethylene glycol-amine (COOH-PEG-NH2, Mw = 2000), and
carboxy-polyethylene glycol-N-hydroxysuccinimide (HOOC-PEG-NHS,
Mw = 2000) were purchased from ToYongBio Tech. Inc. (Shanghai,
China). Sodium dithionite (Na2S2O4), triphosgene, 4, 4′-azodiphenyl-
Over the past decade, great efforts have been made to develop
nanoparticle-based nanocarriers to overcome pharmacokinetic and
pharmacodynamic drawbacks of therapeutic agents for improved drug
accumulation at targeted sites [11]. Nowadays, substantial research has
validated the effectiveness of nanocarriers in treating primary solid tu-
mors; however, the difficulty in crossing the marrow-blood barrier still
makes them not applicable for bone metastasis. Endowing nanocarriers
with a bone-targeted effect is a promising modality to address this issue
[12]. To this end, numerous bisphosphates-functionalized nanocarriers
including liposomes, polymersomes/polymeric micelles, and inorganic
nanoparticles have been designed [13]. Despite substantiated prefer-
ential accumulation in metastatic bone, many of bone-targeted nano-
carriers show no significant survival improvement compared with non-
targeted ones [14–17]. The failure in rapid drug release to an effective
therapeutic dosage to suppress tumor growth in bone is one possible
reason for their unsatisfactory therapeutic outcome. Stimuli-responsive
nanocarriers triggered by bone metastasis microenvironments for
spatiotemporally controllable drug release are thus highly demanded.
It has been reported that bone metastasis is inherently characterized
by acidic pH, redox condition, specific enzymes (e.g., cathepsin K, matrix
metalloproteinases, and vacuolar H+ ATPase), and local hyperthermia.
In light of this, a series of pH-responsive, redox-responsive, enzyme-
responsive, and thermo-responsive nanocarriers have been exploited
to increase therapeutic dosage in metastatic bone [18]. Apart from
aforementioned characteristics, hypoxia is also a hallmark of bone
metastasis, potentiating metastatic tumor cell dormancy, micrometa-
stasis formation, and its development in bone [19]. Although numerous
hypoxia-responsive nanocarriers have been explored [20,21], their
therapeutic outcome in treating bone metastasis has still not been
addressed so far. Our previous findings revealed that hypoxia-responsive
nanocarriers employing nitroaromatics or azobenzene (AZO) as
hypoxia-responsive components could spontaneously respond to
different types of hypoxic cells [22,23], showing the advantages in
response sensitivity and universal applicability over other responsive
systems. For example, acidic microenvironment occurs only after
remarkable tumor growth or in osteolytic bone metastasis [24]; addi-
tional glutathione is needed to trigger the release of a redox-responsive
system at cellular level [25]; the expression of a certain enzyme is var-
iable in different cancer cell types and cancer progression stages, making
enzyme-responsive nanocarriers not universally applicable [18].
Comparatively, hypoxia is a more realistic trigger since it is found in
both bone marrow and bone metastatic nodules and is involved in
lengthy process of bone metastasis [26]. Hypoxia-responsive nano-
carriers are therefore supposed to be particularly valuable in therapeutic
field of bone metastasis.
amine, Nε-benzyloxycarbonyl-L-lysine (L-lysine CBZ), and alendronate
(ALN) were obtained from J&K Scientific LTD (Beijing, China). Doxo-
rubicin hydrochloride (DOX⋅HCl) was purchased from Meryer CO., LTD
(Shanghai, China). N-hydroxysuccinimide (NHS) and 1-ethyl-3(3-dime-
thylpropylamine) carbodiimide (EDC) were acquired from Alfa Aesar
(MA, USA). Near-infrared-emitting (NIR) fluorescent dye Cy5.5 was
purchased from Lumiprobe corporation (FL, USA). Nicotinamide
adenine dinucleotide (NADH) was purchased from Beyotime Biotech-
nology (Shanghai, China). All other chemicals were from Sinopharm
Chemical Reagent CO., LTD. (Shanghai, China). Rat liver microsomes
were obtained from the Research Institute for Liver Diseases (Shanghai,
China). All reagents were of analytical grade and used as received.
Millipore water was used in all experiments unless otherwise indicated.
For cell studies, mouse prostate cancer cell line (RM-1) was acquired
from Cell Resource Center of Institute of Biological Sciences, Chinese
Academy of Sciences (Shanghai, China). Fetal bovine serum (FBS),
0.25% trypsin-EDTA solution, and penicillin-streptomycin solution were
purchased from Gibco Life Technologies (AG, Switzerland). Dulbecco's
Modified Eagle's Medium (DMEM), 4′, 6-diamidino-2-phenylindole
(DAPI), and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
mide (MTT) were obtained from Sigma (MO, USA). Anti-HIF-1α rabbit
polyclonal antibody (No. ab228649) and its corresponding donkey anti-
rabbit secondary antibody conjugated with Alexa Fluor® 555 nm (No.
A31572) were purchased from Abcam (Cambridge, UK) and Thermo
Fisher (MA, USA), respectively.
C57BL/6 mice (6-weeks old, 18–22 g in body weight, male) were
supplied by Animal Center of Nantong University. The protocols for
animal experiments were approved by Animal Care and Use Committee
of Nantong University.
2.2. Synthetic procedures
2.2.1. Copolymer synthesis
Bone-targeted and hypoxia-responsive copolymer consisting of ALN,
PEG, AZO, and PLL (ALN-PEG-AZO-PLL) was synthesized according to
the literature with some modifications [28,29], as shown in Fig. 1.
Briefly, ALN-PEG-AZO-PLL copolymer was synthesized via the following
steps:1) synthesis of N-carboxy-(Nε-benzyloxycarbonyl)-L-lysine anhy-
dride (L-lysine NCA) via the Fuchs–Farthing method using Nε-carbo-
benzyloxy-L-lysine (L-lysine CBZ) and triphosgene; 2) synthesis of
COOH-PEG-AZO via the amide reaction of COOH-PEG-NHS with 4, 4′-
In this study, we attempt to fabricate bone-targeted and hypoxia-
responsive polymeric micelles, self-assembled from a polyethylene gly-
col (PEG) and poly-L-lysine (PLL) based copolymer using alendronate
(ALN) as a bone-targeted moiety and azobenzene (AZO) as a hypoxia-
responsive linker, for effective treatment of bone metastatic prostate
cancer based on our previous work [27]. The micelles are expected to be
selectively accumulated in metastatic bone, facilitated by bone-targeted
ALN in combination with the enhanced permeation and retention (EPR)
effects; subsequently, they can respond to hypoxic bone metastasis for
rapid payload release to an effective therapeutic dosage, facilitated by
azodiphenylamine; 3) synthesis of COOH-PEG-AZO-poly(Nε-benzylox-
ycarbonyl-L-lysine) (COOH-PEG-AZO-PBLL) via ring-opening polymeri-
zation initiated by amino terminal of COOH-PEG-AZO; 4) synthesis of
ALN-PEG-AZO-PBLL via the amide reaction in the presence of EDC and
NHS; 5) synthesis of ALN-PEG-AZO-PLL after removing amino-
protecting benzyloxycarbonyl group from PBLL. Non-targeted and
non-responsive mPEG-PLL and bone-targeted but non-responsive ALN-
PEG-PLL were also synthesized for comparison. Detailed methods were
described in the Supporting Information.
1H NMR spectra of the copolymers were recorded on a 400 MHz NMR
304

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
Scheme 1. Schematic illustration of ALN-functionalized hypoxia-responsive polymeric micelles for targeted therapy of bone metastatic prostate cancer.
Fig. 1. Synthesis and characterization of the copolymers. Synthetic routes for (A) ALN-PEG-AZO-PLL, (B) mPEG-PLL, and (C) ALN-PEG-PLL. 1H NMR spectra of (D)
ALN-PEG-AZO-PLL, (E) mPEG-PLL, and (F) ALN-PEG-PLL.
spectrometer (Avance III，Bruker) using CDCl3 as the solvent. Gel
permeation chromatography (GPC, CTO-20A, Shimadzu) was used to
determine the molecular weight of the copolymers using
dimethylformamide (DMF) as eluent and polystyrene as standard.
305

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
2.2.2. Preparation of blank and drug-loaded polymeric micelles
PEG-PLL, ALN-PEG-PLL, and ALN-PEG-AZO-PLL were employed to
prepare conventional, bone-targeted, and bone-targeted and hypoxia-
responsive polymeric micelles (denoted as PMs, ALN-PMs, and ALN-
HR-PMs, respectively) via self-assembly technique. In a typical experi-
ment for the synthesis of ALN-HR-PMs, 20 mg of ALN-PEG-AZO-PLL was
dissolved in 4 mL of DMF, and 10 mL of water was added drop by drop
with vigorous stirring. The mixture was then dialyzed against water
(MWCO = 3000) for 24 h to obtain micelle suspension. ALN-HR-PMs
were collected after freeze-drying for further use.
with 100
μ
L of rat liver microsomes and 10 mg of NADH was dispersed in
10 mL of degassed PBS (pH 7.4). The dialysis bag was then dialyzed
against 30 mL of degassed PBS (pH 7.4) with gentle shaking at 37 ◦C.
Degassing was performed throughout the study to maintain hypoxic
condition. At predetermined intervals (0.5, 1, 2, 4, 6, 12 and 24 h), 0.5
mL aliquot of the dialysate was withdrawn for DOX quantification using
HPLC. Equivalent fresh PBS was added immediately when the dialysate
was withdrawn. The micelles alone were studied as a normoxic control.
The cumulative release (%) defined as the total released DOX at pre-
determined intervals relative to the initial loaded DOX was finally
plotted versus time.
To prepare DOX-loaded micelles (PMs/DOX, ALN-PMs/DOX, and
ALN-HR-PMs/DOX), DOX⋅HCl was firstly desalted (Supporting Infor-
mation) and thereafter mixed with the copolymer in DMF for micelle
preparation according to the above-described protocol. Cy5.5-loaded
micelles (PMs/Cy5.5, ALN-PMs/Cy5.5, and ALN-HR-PMs/Cy5.5) were
synthesized using the same protocol for NIR fluorescence imaging. The
ratio of DOX to copolymer was 4:20 (w/w) for DOX loading, while it was
0.1/20 (w/w) for Cy5.5 loading.
2.3.4. In vitro bone-targeted efficiency
The binding efficiency of ALN-functionalized micelles to hydroxy-
apatite (HA) was studied to assess bone-targeted efficiency according to
a previous work [30]. Typically, 3 mg of ALN-HR-PMs/DOX or non-
targeted PMs/DOX were dispersed in 10 mL of water, followed by
addition of 500 mg of HA. The mixture was gently shaken and 200 μL
aliquot of the mixture was withdrawn at predetermined intervals (0.5, 1,
2, 3, and 4 h,). After centrifugation at 3000 rpm for 10 min, the super-
natant containing unbound micelles was separated and treated with 4.8
mL of methanol for DOX quantification based on the procedure
described in Section 2.3.1. The binding efficiency was calculated ac-
cording to the following formula:
2.3. Micelle characterization
2.3.1. Size, surface charge, morphology, and drug loading
The size and morphology of the micelles were observed on a trans-
mission electron microscope (TEM, JEM-1230, Japan). The micelles
were dropped on copper grids and then stained with 1% phospho-
tungstic acid solution (w/v) before observation. Hydrodynamic diam-
eter, polydispersity index (PDI), and zeta potential of the micelles were
investigated by dynamic light scattering technique (DLS, Zetasizer ZS90,
Malvern). To determine drug loading efficiency (DLE, %) and drug
Binding efficiency (%) = (1-unbound DOX/loaded DOX) × 100%.
2.4. Cell experiments
2.4.1. Cell culture
loading capacity (DLC, %) of the micelles, 200
μL of DOX-loaded micelle
RM-1 cells were cultured in DMEM medium (37 ◦C, 5% CO2, 21% O2)
containing 1% penicillin/streptomycin antibiotic and 10% FBS. The
culture medium was refreshed every 1–2 days and the cells were
passaged when reaching approximately 80% confluence.
suspension (10 mg of the micelles dispersed in 10 mL of water) was
added into 4.8 mL of methanol solution. After sonication for 5 min (10 s
on and 10 s off) with an ultrasonic cell pulverizer (JY92-2D, Zhejiang
Xinzhi Biotech, China) to completely rupture the micelles, the mixture
was centrifuged at 10000 rpm for 10 min and the supernatant with
loaded DOX was collected. A high-performance liquid chromatography
(HPLC, Waters) system equipped with a separation module (Waters
e2695) and a fluorescence detector (2475 FLR Detector) was used to
quantitatively analyze DOX based on a pre-established standard curve.
HPLC conditions were as follows: column, ultimate® AQ-C18 (4.6 ×
2.4.2. Hypoxia-responsive intracellular release
To concretely observe hypoxia-responsive drug release, RM-1 cells
(2 × 104 cells per dish) were seeded in a polystyrene/glass confocal dish
(No. BDD011035, JET BIOFIL) and cultured for 24 h. After treatment
with different drug formulations at 10 μg/mL of DOX concentration for
2 h followed by washing with PBS (pH 7.4) to remove drug-containing
medium, a φ14 mm coverslip was put on the cells cultured on the
glass bottom of confocal dish to create a decreasing oxygen concentra-
tion from its edge to the center [22,31,32]. After another 3 h incubation,
DOX fluorescence of the cells was observed on a confocal laser scanning
microscope (CLSM, TCS SP8, Leica) (488 nm laser and 590 ± 10 nm
emission filter for DOX).
250 mm, 5
μm); mobile phase, methanol/acetonitrile/0.02 mM ammo-
nium phosphate/glacial acetic acid (52:5:43:6, v/v); flow rate, 1 mL/
min; column temperature, 30 ◦C; injection volume, 20
μL; fluorescence
detector, 488 nm in excitation wavelength and 555 nm in emission
wavelength. A standard curve was constructed by plotting peak area as a
function of DOX concentrations ranging from 20 to 100 μg/mL (y =
657.2× – 531.4, R2 = 0.9992). DLE was defined as the weight of loaded
DOX relative to the weight of initially added DOX, while DLC was
defined as the weight of loaded DOX relative to the weight of initially
added DOX and copolymers.
To study hypoxia-responsive drug release in the fixed hypoxic con-
ditions, RM-1 cells were seeded in a glass-bottom culture dish (2 × 105
cells per dish) and cultured for 24 h. Afterwards, the cells were treated
with different drug formulations at 10 μg/mL of DOX concentration for
2 h and washed twice with PBS (pH 7.4). The cells were then divided
into two groups. One group was sequentially cultured under normoxic
condition (21% O2), while the other was cultured under hypoxic con-
dition (1% O2) using a hypoxia incubator (INVIVO2, Ruskinn). Three
hours later, all cells were washed with PBS (pH 7.4) twice and stained
with DAPI. The fluorescence of the cells was observed on the CLSM in
two channels relevant to DAPI and DOX.
2.3.2. In vitro response to hypoxia
A chemical hypoxia (2 mg of Na2S2O4 in 2 mL of water) and a bio-
logical hypoxia (20 μL of rat liver microsomes and 2 mg of NADH in 2 mL
of pH 7.4 PBS) were generated, respectively, to mimic hypoxic bone
metastasis microenvironment. Subsequently, 2 mg of the micelles were
added into above solution and sealed in a quartz cuvette for 12 h. The
changes in characteristic peaks before and after hypoxia treatment were
recorded to study in vitro response to hypoxia on a UV–vis spectropho-
tometer (UV-2401PC, Shimadzu). AZO and the copolymers were also
treated for comparison.
Flow cytometry was further used to quantitatively analyze hypoxia-
responsive intracellular release of DOX. RM-1 cells were cultured in a
petri dish, treated with different drug formulations for 2 h, washed with
PBS (pH 7.4), and further cultured under normoxic or hypoxic condi-
tions for 3 h, as described above. In the following, the cells were
collected through trypsinization and centrifugation and re-suspended in
PBS (pH 7.4) with cell intensity not less than 1 × 106 cells/mL. The red
fluorescence of DOX in FL2 gate (575 nm) was then detected and
quantified on a flow cytometer (Gallios, Beckman Coulter) using 488 nm
2.3.3. In vitro hypoxia-responsive drug release
While Na2S2O4 would quench DOX fluorescence, in vitro hypoxia-
triggered drug release was carried out in the biological hypoxia.
Briefly, in a dialysis bag (MWCO = 3000), 10 mg of DOX-loaded micelles
306

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
laser.
intravenously injected with different Cy5.5-loaded formulations at
equivalent Cy5.5. NIR fluorescence images were acquired and presented
in visible pseudo-color mode at predetermined intervals (0.5, 1, 2, 4, 6,
12, and 24 h) using a small animal dedicated IVIS imaging system
(Lumina II, Caliper Life Sciences), and fluorescence intensity was
analyzed by IVIS imaging software. Following the last imaging at 24 h,
the mice were euthanized, and the tissues of heart, liver, spleen, lung,
kidney, metastatic bone, and contralateral tumor-free bone were har-
vested for ex vivo imaging. Default illumination settings of Cy5.5 filter
were used for imaging throughout the whole study.
2.4.3. Cell migration
Cell migration was assessed by wound healing assay. RM-1 cells (2 ×
104 cells per well) were seeded in 6-well plates, cultured for 24 h, and
treated with different drug formulations at 10
μ
g/mL of DOX concen-
L pipette
tration for 2 h. After removing drug-containing medium, a 10
μ
tip was used to generate a wound area by scraping cell monolayer. The
cells were then incubated under normoxic or hypoxic conditions for 24
h. The wound width was observed and recorded on an inverted micro-
scope. The migration ratio was defined as follows:
Migration ratio (%) = (A0 – Ai) / A0 × 100.
2.5.4. Immunofluorescence staining
Where Ai and A0 were the remaining area of the region without cells
after treatment and the initial area before treatment, respectively. The
area of the selected region was analyzed by Image J software.
Following ex vivo NIR imaging at 24 h post-injection, tumor tissue
was stripped from the harvested metastatic bone, followed by dehy-
dration and transparentization for paraffin sectioning. The obtained
slices were rehydrated and sequentially incubated with anti-HIF-1
α
2.4.4. Cytotoxicity
rabbit polyclonal antibody (1:250) overnight at 4 ◦C, its corresponding
donkey anti-rabbit fluorescence-conjugated secondary antibody conju-
gated with Alexa Fluor® 555 nm (1:800) at room temperature for 1 h,
and DAPI for 8 min. After rinsing with PBS (pH 7.4) and covering with a
In vitro cytotoxicity was assessed by MTT assay. RM-1 cells seeded in
96-well plates (3000 cells per well) were cultured and treated with
different drug formulation for 2 h. After removing drug-containing
medium, the cells were cultured under normoxic or hypoxic condi-
tions for 48 h. The cells were then washed with PBS (pH 7.4) and treated
for MTT assay using an automatic microplate reader (SN209941, Bio-
TEK) at 570 nm.
coverslip, the fluorescence of DAPI, HIF-1α, and Cy5.5 was observed
using their corresponding fluorescence filters on the CLSM.
2.5.5. Pain-related behavior evaluation
On the 7th day after modeling, tumor-bearing mice were intrave-
nously injected with different drug formulations at equivalent DOX
dosage (5 mg/kg body weight) every four days. The mice treated with
PBS (pH 7.4) were studied as a control group. Pain-related behaviors
such as number of flinches, spontaneous lifting time, and movement
scores of tumor-bearing limb were measured during a 4-min period
according to our previous work [27]. The mice were acclimated in a
glass box for 10 min before measurement and pain-related behavior
evaluation started from the 1st day after modeling. Meanwhile, the
survival of the mice was checked every day.
2.5. Animal experiments
2.5.1. Murine model of bone metastatic prostate cancer
RM-1 cell suspension of PBS (pH 7.4) (5 × 107 cells/mL) was placed
on ice for use. After anesthetization with 5% chloral hydrate by intra-
peritoneal injection, the femur of the right hind limb of the mice was
exposed carefully by surgery. Then, 20 μL of above-prepared cell sus-
pension was carefully injected into the femur with a 29-gauge needle
and subsequently, the pinhole was sealed with bone wax. The mice after
modeling were given food and water freely.
2.5.6. Micro-computed tomography (micro-CT) imaging
On the 21st day after modeling, the mice were euthanized and
metastatic bone was harvested for Micro-CT two-dimensional (2D)
scanning on a Micro-CT (SKYSCAN 1176, Bruker, Belgium). Following
three-dimensional (3D) reconstruction, bone analysis software was used
to evaluate bone mineral density (BMD) and trabecular bone number
(Tb.N). Healthy bone from healthy mice was also scanned as a control.
2.5.2. Pharmacokinetics and tissue distribution
Pharmacokinetic study was performed in healthy C57BL/6 mice. The
mice were intravenously injected with different drug formulations at
equivalent DOX dosage (5 mg/kg body weight). At predetermined in-
tervals (0.05, 0.25, 0.5, 1, 2, 4, 12 and 24 h), the mice were euthanized,
and 0.5 mL of blood was taken from mouse orbital. The blood was
centrifuged at 8000 rpm for 10 min and the collected plasma was mixed
with 3 mL of chloroform and methanol (4:1, v/v). The mixed plasma was
vortex for 5 min and centrifuged at 8000 rpm for 10 min to collect
extracted DOX in the upper chloroform layer. After evaporating chlo-
roform, 3 mL of methanol was added into the samples to completely
precipitate plasma protein, followed by centrifugation at 8000 rpm for
another 10 min. The methanol layer containing DOX was collected and
2.5.7. Bone staining
Hematoxylin and eosin (H&E), tartrate-resistant acid phosphatase
(TRAP), alkaline phosphatase (ALP) and TRAP/ALP staining of meta-
static, tumor-free, and healthy bone were performed to further assess the
therapeutic outcome of different drug formulations. Bone decalcifica-
tion was necessary before prepared into slice for staining. Typically, the
harvested metastatic bone from the mice euthanized on the 21st day
after modeling was put into periodate-L-lysine-paraformaldehyde (PLP)
solution for 24 h for fixing. After terminating fixing using PBS (pH 7.4),
the samples were placed into 5% glycerol-PBS, 10% glycerol-PBS, and
filtered with syringe filter (0.45 μm in pore diameter). Finally, 20 μL of
obtained methanol solution was injected into HPLC for DOX
quantification.
Tissue distribution was assessed in tumor-bearing C57BL/6 mice. On
the 7th day after modeling, tumor-bearing mice were intravenously
injected with different drug formulations at equivalent DOX dosage (5
mg/kg body weight). At predetermined intervals (1 and 12 h), the mice
were euthanized for prefusion-fixation and the tissues of heart, liver,
spleen, lung, kidney, metastatic bone, and contralateral tumor-free bone
were then harvested. Following tissue homogenization using a homog-
enizer (FastPrep-24TM 5G, Shanghai Boyi Biotech, China), the samples
were treated with 5 mL of chloroform and methanol (4:1, v/v) for DOX
qualification according to the protocol used for blood sample
preparation.
◦
15% glycerol-PBS solutions, respectively, for 12 h at 4 C for rinsing.
Subsequently, the samples were transferred to ethylenediaminetetra-
◦
acetic acid (EDTA)-glycerol solution for 10–15 days at 4 C for decal-
cification. The decalcification process was observed under the
microscope. The decalcified samples were then immersed in 15% su-
crose-7.5% glycerol-PBS, 15% sucrose-PBS, 7.5% sucrose-PBS, and PBS
(pH 7.4) solutions, respectively, for 12 h at 4 ◦C for rinsing. Finally, the
samples were dehydrated with graded ethanol, transparentized with
xylene, embedded in paraffin, and sectioned into 6 μm slices for H&E
staining or TRAP, ALP, and TRAP/ALP staining using TRAP/ALP stain-
ing kit® (Wako, Osaka, Japan), based on the supplier's recommended
protocol.
2.5.3. In vivo NIR imaging
On the 7th day after modeling, tumor-bearing mice were
307

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
2.5.8. Safety assessment
micelles and played a considerable role in their size. The size determined
by DLS technique is calculated from the fluctuations of scattered light
intensity due to the Brownian movement of nanoparticles and the large
and heavy nanoparticles move slowly in comparison to the small and
light ones [35]. After loading substantial DOX, the Brownian movement
of the micelles decreases, leading to the increase of calculated size.
Moreover, DOX adsorbed onto the surface of the micelles also contrib-
uted to the size increase.
On the 21st day after modeling, the tissues of heart, liver, spleen,
lung, and kidney were harvested from the mice euthanized for bone
staining. After paraffin sectioning, histopathological examination of the
tissues was carried out by H&E staining.
2.6. Data analysis
We next studied the response of the micelles to hypoxic conditions
generated by Na2S2O4 and liver microsomes/NADH, respectively.
Na2S2O4 is a classic deoxidizer [36], which can rapidly reduce oxygen
concentration to zero (Supplementary Fig. S4), while liver microsomes
can induce some redox reactions to mimic in vivo hypoxic conditions
when cooperating with NADH [37]. When exposed to either Na2S2O4 or
liver microsomes/NADH, AZO in ALN-HR-PMs will undergo rapid
reductive cleavage, triggering the disassembly of ALN-HR-PMs for drug
release (Fig. 2A). Hypoxia-induced AZO cleavage in ALN-HR-PMs was
validated by UV–vis spectra. ALN-HR-PMs showed two characteristic
peaks at 392 and 430 nm (Fig. 2B), originating from AZO (Supplemen-
tary Fig. S5); after hypoxic treatment, these two peaks disappeared,
indicating AZO cleavage. In contrast, PMs and ALN-PMs showed no
difference before and after hypoxic treatment (Fig. 2C and D). The
disassembly of ALN-HR-PMs, triggered by hypoxia-induced AZO cleav-
age, was further confirmed by TEM observation. As revealed in Fig. 2E,
ALN-HR-PMs as well as PMs and ALN-PMs were of regular spherical
structure; once exposed to hypoxia, intact ALN-HR-PMs were cleaved
and no regular structure was observed. In contrast, no change was found
in PMs and ALN-PMs. It was worth noting that TEM size of three micelles
was approximate (less than 100 nm), which was significantly smaller
than DLS size. The sample prepared for TEM needs dehydration and
immobilization, leading to its size always smaller than that in solvent-
swollen state for DLS [38]. The disassembly of ALN-HR-PMs was sup-
posed to accelerate the release of their payload, as validated by the
cumulative release profiles of the micelles. As depicted in Fig. 2F, PMs/
DOX and ALN-PMs/DOX exhibited similar drug release kinetics with less
than 30% of loaded DOX released in 24 h, regardless of normoxia or
hypoxia. This slight release should be attributed to DOX adsorbed onto
the surface of the micelles. Although DOX released from ALN-HR-PMs/
DOX under normoxic condition was equivalent to that from PMs/DOX
and ALN-PMs/DOX, more than 40% of loaded DOX was released under
hypoxic condition in 1 h and more than 70% was released in 24 h. The
release of ALN-HR-PMs/DOX was dramatically accelerated under hyp-
oxic condition, indicating their high sensitivity to hypoxia.
All data were analyzed by GraphPad Prim 7 software and statistically
compared using unpaired student's t-test for two groups or one-way
analysis of variance (ANOVA) test for multiple groups. A p-value less
than 0.05 (p < 0.05) was defined as a statistically significant difference.
3. Results and discussion
3.1. Synthesis and characterization of the micelles, in vitro response to
hypoxia, and bone-targeted efficiency
To fabricate ALN-HR-PMs, we firstly synthesized amphiphilic ALN-
PEG-AZO-PLL copolymer composed of ALN as bone-targeted moiety,
PEG and PLL as hydrophilic and hydrophobic blocks, respectively, and
AZO as hypoxia-responsive linker of PEG and PLL. Heterobifunctional
HOOC-PEG-NHS was utilized as starting material, NHS terminal of
which was conjugated with AZO. The residual amino terminal of AZO
initiated ring-opening polymerization of L-lysine NCA for the formation
of amino-protected PLL. After conjugation with ALN using carboxyl
terminal of PEG followed by amino-protecting group removal, ALN-
PEG-AZO-PLL was obtained (Fig. 1A). PEG-PLL and ALN-PEG-PLL
were also synthesized based on the similar procedures (Fig. 1B and C).
M
w values of obtained PEG-PLL, ALN-PEG-PLL, and ALN-PEG-AZO-PLL
were 6278, 6770, and 7074, respectively, as determined by GPC (sup-
plementary Fig. S1). Three copolymers had a low PDI(Mw/Mn < 1.2)
with the degree of polymerization around 29 for PEG-PLL and 31 for
ALN-PEG-PLL and ALN-PEG-AZO-PLL. 1H NMR results confirmed the
successful synthesis of three copolymers. ALN-PEG-AZO-PLL showed
characteristic peaks at 6.96 and 7.53 ppm from AZO and at 5.32 ppm
from ALN, differentiating it from PEG-PLL and ALN-PEG-PLL (Fig. 1D-1F
and supplementary Fig. S2).
Due to its amphiphilic nature, ALN-PEG-AZO-PLL as well as PEG-PLL
and ALN-PEG-PLL could self-assemble in aqueous condition to form
ALN-HR-PMs as well as PMs and ALN-PMs. PMs, ALN-PMs, and ALN-HR-
PMs showed 169.1 nm, 175.7 nm, and 186.2 nm in hydrodynamic
diameter, while their zeta potential was 43.5 mV, 36.7 mV, 35.3 mV,
respectively, as determined by DLS (Supplementary Table S1 and
Fig. S3). ALN conjugation led to the relatively large size of ALN-PMs
compared with PMs, and AZO insertion further contributed to the size
increase of ALN-HR-PMs. Positive surface charge of the micelles was
owing to PLL with abundant amino groups, which was slightly reduced
after ALN conjugation. The size of nanoparticles plays a vital role in drug
delivery to tumor. It has been reported that nanoparticles of 100–200
nm in diameter can effectively escape renal filtration to remain in blood
circulation system [33], further favoring their extravasation from cir-
culation system into metastatic bone [34]. Thus, ALN-HR-PMs as well as
PMs and ALN-PMs show a suitable size for effective drug delivery to
bone metastasis. In the following, DOX as a model anticancer drug was
loaded into the micelles to form PMs/DOX, ALN-PMs/DOX, and ALN-
HR-PMs/DOX. At the fixed ratio of DOX to copolymer (4:20, w/w),
obtained micelles exhibited high DLE and DLC. As demonstrated in
supplementary Table S1, DLE of PMs/DOX, ALN-PMs/DOX, and ALN-
HR-PMs/DOX was 69.3%, 75.1%, and 87.1%, while their DLC was
13.9%, 15.0%, and 17.4%, respectively. The higher DLE of ALN-PMs/
DOX than PMs/DOX indicated that ALN conjugation facilitated the
adsorption of DOX onto the surface of the micelles. Meanwhile, rigid
AZO provided more inner space for DOX loading and ALN-HR-PMs/DOX
thus showed higher DLE and DLC than ALN-PMs/DOX. It could also be
found that DOX loading had a slight effect on the surface charge of the
Hydroxyapatite (HA) is the main inorganic component in bone.
Hydroxyl and phosphonate groups of ALN have a high affinity to HA
[15] and in vitro bone-targeted efficiency of ALN was thus evaluated by
studying its binding with HA. After mixing ALN-HR-PMs/DOX dispersed
in water with water-insoluble HA, the solution became clear and red
ALN-HR-PMs/DOX almost completely precipitated, indicating their
specific binding with HA. Comparatively, the solution with mixed PMs/
DOX remained red, although some red precipitates were observed, due
to non-specific adsorption (Fig. 2H). The binding rate was further
quantified and as shown in Fig. 2G, up to 96% of ALN-HR-PMs/DOX was
bound to HA in 4 h, while only 25% of ALN-free PMs/DOX was bound.
This result indicated that ALN-HR-PMs/DOX had a strong affinity to
bone for bone-targeted drug delivery.
3.2. Hypoxia-responsive intracellular drug release for improved anti-
cancer efficacy in vitro
The therapeutic outcome of nanoparticles as drug carriers is highly
associated with their uptake by cells and subsequent drug release. We
firstly studied hypoxia-responsive drug release by covering a glass
coverslip on the cultured cells, where a non-quantitative gradient hyp-
oxia was generated with the oxygen concentration decreasing from the
edge to center of the coverslip. CLSM images showed that the
308

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
Fig. 2. In vitro response to hypoxia and bone-targeted efficiency evaluation. (A) schematic illustration of hypoxia-induced AZO cleavage for triggered drug release of
ALN-HR-PMs. UV–vis spectra of (B) ALN-HR-PMs, (C) PMs, and (D) ALN-PMs before and after hypoxic treatment. (E) TEM images of PMs, ALN-PMs, and ALN-HR-
PMs before and after hypoxic treatment. (F) Cumulative drug release of PMs, ALN-PMs, and ALN-HR-PMs before and after hypoxic treatment. (G) Optical photo-
graphs of binding of PMs and ALN-HR-PMs with water-insoluble HA and (H) corresponding binding rate. Na2S2O4 and rat liver microsomes/NADH were used to
generate the chemical and biological hypoxic conditions, respectively. Data were presented as mean ± SD (n = 3).
fluorescence intensity of DOX gradually increased from the normoxic
edge to the hypoxic center in ALN-HR-PMs/DOX treated cells, as a result
of hypoxia-induced disassembly of ALN-HR-PMs/DOX. No fluorescence
difference between the edge and center was observed in AZO-free PMs/
DOX or ALN-PMs/DOX treated cells (Fig. 3A). CLSM observation was
further confirmed by line scanning. As revealed in Fig. 3B, the fluores-
cence intensity of DOX in AZO-free groups, acquired by Image J soft-
ware, remained low (< 50) throughout the whole coverslip;
comparatively, DOX intensity in ALN-HR-PMs/DOX treated cells was
hypoxia-dependent, reaching its peak (> 200) in the mostly severe
hypoxic center.
hypoxic condition than under normoxic condition (Fig. 3C). Microscopic
observation also revealed that the released DOX was mainly located in
the cytoplasm without permeating into the nucleus. Hypoxia-responsive
drug release was further quantitatively analyzed by flow cytometry
(Fig. 3D). As expected, DOX release in AZO-free groups was similar
under normoxic and hypoxic conditions (3.16/3.20 for PMs/DOX and
3.81/3.78 for ALN-PMs/DOX, hypoxia/normoxia); DOX release in ALN-
HR-PMs/DOX treated cells was considerably accelerated under hypoxic
condition, showing 2.02-fold higher than under normoxic condition
(8.33/4.12). DOX intensity under normoxic condition could be also used
to compare the difference in cell uptake of the micelles. It could be seen
that cell uptake of PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX
was 3.20, 3.78, and 4.12, respectively. The relatively high uptake of
ALN-PMs/DOX in comparison to PMs/DOX should be due to ALN
conjugation. This inference did not mean the existence of HA on the
surface of RM-1 cells. The abundant hydroxyl groups should account for
increased cell uptake through the formation of covalent bonds with
In the following, a hypoxia incubator was employed to study
hypoxia-responsive drug release at the fixed oxygen conditions. In line
with the results from the coverslip experiments, AZO-free PMs/DOX or
ALN-PMs/DOX treated cells showed no difference in DOX intensity
under normoxic (21% O2) and hypoxic (1% O2) conditions, while there
was a stronger fluorescence for ALN-HR-PMs/DOX treated cells under
309

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
Fig. 3. Hypoxia-responsive intracellular drug release for improved anti-cancer in vitro. (A) CLSM observation of DOX release in RM-1 cells treated with PMs/DOX,
ALN-PMs/DOX, and ALN-HR-PMs/DOX in a non-quantitative gradient hypoxia and (B) corresponding quantitative analysis by the line scanning. The cells were
covered by a coverslip after removing drug-containing medium and the oxygen concentration decreased from the edge to center of the coverslip. (C) CLSM
observation of DOX release in RM-1 cells treated with PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX under the fixed normoxic (21% O2) and hypoxic (1% O2)
conditions generated by the hypoxia incubator. The nuclei were stained with DAPI. (D) Quantitative analysis of DOX fluorescence in RM-1 cells treated with PMs/
DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX under normoxic and hypoxic conditions by flow cytometry. (E) Wound healing assay of RM-1 cells treated with PMs/
DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX under normoxic and hypoxic conditions and (F) corresponding quantitative analysis. Data were acquired by Image J
and expressed as mean ± SD (n = 5). **p < 0.01 and ***p < 0.001. (G) MTT assay of PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX against RM-1 cells under
normoxic and hypoxic conditions after 24 h incubation. Data were expressed as mean ± SD (n = 6). ***p < 0.001.
proteins on cell surface [25,39,40]. On the other hand, the higher uptake
of ALN-HR-PMs/DOX might result from DOX release triggered by local
hypoxia generated by clustered cell growth [41]. This result was in good
consistence with the line scanning profiles, where ALN-HR-PMs/DOX
generally showed high fluorescence intensity throughout the whole
coverslip, even in the normoxic edge (Fig. 3B).
PMs/DOX and ALN-PMs/DOX under normoxic and hypoxic condi-
tions. MTT assay further confirmed that ALN-HR-PMs/DOX were of high
cytotoxicity under hypoxic condition. As shown in Fig. 3F, three drug
formulations exhibited DOX concentration-dependent cytotoxicity and
AZO-free PMs/DOX and ALN-PMs/DOX had no difference under nor-
moxic and hypoxic conditions; comparatively, the cytotoxicity of ALN-
HR-PMs/DOX was considerably augmented under hypoxic condition,
showing a significant difference compared with the other groups (p <
0.001, hypoxia versus normoxia; p < 0.001, ALN-HR-PMs/DOX versus
PMs/DOX and ALN-PMs/DOX under hypoxic condition). MTT assay
performed in MDA-MB-231 cells displayed a similar tread (supplemen-
tary Fig. S6). The results from wound healing assay and MTT assay were
well consistent. Improved in vitro anti-cancer efficacy of ALN-HR-PMs/
DOX under hypoxic condition indicates that internalized micelles
remain a whole under normoxic condition for reduced side effect and
however, intracellularly release DOX to a high therapeutic dosage under
In vitro anti-cancer efficacy of the micelles under normoxic and
hypoxic conditions was evaluated by wound healing assay and MTT
assay against RM-1 cells. Wound healing assay showed that cell migra-
tion was effectively inhibited after treated with ALN-HR-PMs/DOX
under hypoxic condition (p < 0.01, hypoxia versus normoxia), showing
a significantly high inhibition ratio compared with the other groups (p
< 0.001, ALN-HR-PMs/DOX versus PMs/DOX and ALN-PMs/DOX under
hypoxic condition) (Fig. 3E and F). There was no difference observed in
the cells treated with PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX
under normoxic condition as well as in the cells treated with AZO-free
310

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
hypoxic condition for improved bioavailability. ALN-HR-PMs/DOX can
deliver DOX to cells via fusing with cell membrane, micropinocytosis
after destabilization by cell membrane components when adsorbed on
cell surface, or nanoparticle-mediated endocytosis [42]. Only
nanoparticle-mediated endocytosis can ensure that the micelles are
internalized as a whole. Thus, nanoparticle-mediated endocytosis is the
key entry pathway for the micelles [43].
Additionally, the disassembly of ALN-HR-PMs composed of ALN-
PEG-AZO-PLL copolymer under hypoxic condition will produce ani-
line, which is also toxic to cells and organs. Thereby, clarifying that
Fig. 4. Improved pharmacokinetic performance for bone-targeted drug delivery and subsequent hypoxia-responsive drug release. (A) The plasma DOX concentration
over time. Healthy mice were intravenously injected with DOX, PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at 5 mg DOX/kg body weight. (B and C) DOX
content in spleen, heart, liver, kidney, lung, metastatic bone, and tumor-free bone at 1 h and 12 h post-injection. Tumor-bearing mice were intravenously injected
with DOX, PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at 5 mg DOX/kg body weight. Data were presented as mean ± SD (n = 3). *p < 0.05 and **p < 0.01.
(D) In vivo NIR fluorescence images acquired at 0, 0.5, 1, 2, 4, 6, 12, and 24 h post-injection. Yellow circle indicated the tumor-free bone site and blue circle indicated
the metastatic bone site. The color bar indicated fluorescence intensity from (red) low to high (yellow). (E) Ex-vivo fluorescence images of the tissues harvested from
the mice after the last whole-body imaging. H: heart; Li: liver; Sp: spleen; Lu: lung; Ki: kidney; T: tumor-bearing metastatic bone; N: normal contralateral tumor-free
bone. (F) Fluorescence ratio of the metastatic bone site to the tumor-free bone site. The fluorescence intensity of the selected sites indicated by yellow or blue circle
was acquired by IVIS imaging software. Data were presented as mean ± SD (n = 3). **p < 0.01. (G) In vivo immunofluorescence staining of the tumor tissue stripped
from the metastatic bone after ex-vivo fluorescence imaging. For NIR imaging and immunofluorescence staining, tumor-bearing mice were intravenously injected
with 100
μL of PMs/Cy5.5, ALN-PMs/Cy5.5, and ALN-HR-PMs/Cy5.5. (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)
311

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
hypoxia-responsive drug release of ALN-HR-PMs but not the production
of aniline contributes to increased cytotoxicity is necessary. As revealed
in supplementary Fig. S7A, blank ALN-HR-PMs with the concentrations
unilateral femur of mice is the most common for preclinical study, which
is also a classic model to evaluate cancer-induced bone pain [46].
Herein, bone metastasis model was established by injecting RM-1 cells
into the femur of the right limb of mice. In the following, tumor-bearing
mice received injection of different drug formulations and DOX content
in some important tissues including spleen, heart, liver, kidney, lung,
metastatic bone, and contralateral tumor-free bone at 1 h and 12 h post-
injection was measured to evaluate drug tissue distribution. As shown in
Fig. 4B, all drug formulations were widely distributed in all tissues and
the drug distribution in spleen, heart, liver, kidney, and lung was rela-
tively high at 1 h post-injection, due to non-specific uptake by macro-
phages in the reticuloendothelial system (RES) of normal tissues, which
subsequently decreased over time. DOX content of ALN-free DOX and
PMs/DOX remained at a relatively low level in the tumor-free and
metastatic bones, indicating the difficulty in delivering therapeutic
agents to bone without a targeting effect. Increasingly accumulated DOX
content in the metastatic bone was only observed in ALN-PMs/DOX and
ALN-HR-PMs/DOX groups, showing a significant difference compared
with DOX and PMs/DOX (p < 0.01, ALN-HR-PMs/DOX versus DOX and
PMs/DOX at 1 h and 12 h post-injection; p < 0.05, ALN-PMs/DOX versus
DOX and PMs/DOX at 12 h post-injection). Metastatic bone is of cor-
rosive bone surface consisting of highly crystalline HA, while healthy
bone is mainly composed of amorphous HA [17]. ALN is more easily
combined with highly crystalline HA, resulting in a higher accumulation
of ALN-PMs/DOX and ALN-HR-PMs/DOX in the metastatic bone than in
the tumor-free bone. Moreover, although there was no significant dif-
ference found, ALN-HR-PMs/DOX showed a generally higher accumu-
lation in the metastatic bone than ALN-PMs/DOX, which was due to
hypoxia-responsive release of ALN-HR-PMs/DOX. DOX content in tis-
sues determined by HPLC comprised released and unreleased DOX and
in light of this, ALN-PMs/DOX and ALN-HR-PMs/DOX should be of
approximate DOX content in the metastatic bone. This unexpected
discrepancy might result from the fact that unreleased DOX was not
detected totally by HPLC method, offering an explanation that the mi-
celles did not show a relatively high drug distribution in the normal
tissues in comparison to DOX, although the former had improved
pharmacokinetic performance. This also offered an explanation that
ALN-functionalized micelles did not have a higher drug accumulation in
the normal tissues than ALN-free micelles, although ALN was easily
internalized by macrophages [47]. Similar phenomena were also
described by others [17,40,48].
ranging from 1 to 100 μg/mL exhibited a slight cytotoxicity comparable
to that of PMs and ALN-PMs under normoxic and hypoxic conditions.
This slight cytotoxicity should originate from highly positive surface of
the micelles since positively charged nanoparticle always have a rela-
tively high cytotoxicity in contrast to negatively charged and neutral
ones [44]. Importantly, based on drug loading capacity shown in sup-
plementary Table S1 and MTT assay shown in Fig. 3G, the concentration
of blank ALN-HR-PMs was far less than 100 μg/mL at the used DOX
concentrations and thus, the cytotoxicity of blank ALN-HR-PMs under
hypoxic condition was ignorable. Blank ALN-HR-PMs also exhibited
ignorable cytotoxicity against MDA-MB-231 cells (supplementary
Fig. S7B). Taken together, the above results suggested that ALN-HR-PMs
were biocompatible and showed improved anti-cancer efficacy under
hypoxic condition, as a result of hypoxia-responsive drug release.
3.3. Pharmacokinetics, tissue distribution, and hypoxia-responsive drug
release in vivo
There is substantial evidence showing that encapsulating a drug into
nanocarriers with PEGylated surface can considerably improve its
pharmacokinetic profile through their protective effect against degra-
dation, protein adsorption, and subsequent systemic clearance [45].
PMs/DOX should be of improved pharmacokinetic profile compared
with DOX and however, determining whether their pharmacokinetic
profile will be impaired after ALN conjugation and AZO insertion is still
necessary. To this end, the mice received intravenous injection of DOX,
PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at equivalent DOX
dosage and the time-dependent plasma drug concentration was deter-
mined by HPLC. As shown in Fig. 4A, the plasma drug concentration of
all drug formulations peaked very quickly following injection and
almost bottomed out at 24 h post-injection. Nevertheless, the concen-
tration of PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX generally
decreased more slowly than DOX. To better compare pharmacokinetic
profile, the plasma drug concentration-time curves were then fitted into
a two-compartment model using PK Solver software and the main
pharmacokinetic parameters such as the distribution half-life (t1/2α),
elimination half-life (t1/2β), the area under the curve (AUC0→∞), clear-
ance rate (CL), elimination rate constant (K10), and mean residence time
(MRT) of the drug formulations were obtained (Table 1). It could be seen
that the pharmacokinetic parameters of DOX were improved after
encapsulated; PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX were
distributed and eliminated more slowly than DOX with significant dif-
ference in t1/2α, t1/2β, AUC0→∞, CL, and K10, indicating their prolonged
blood circulation time. Moreover, there was no significant difference
observed among DOX-loaded micelles, indicating ALN conjugation and
AZO insertion had no significant role in the pharmacokinetic profile of
the micelles.
In vivo NIR imaging was further carried out to substantiate bone-
targeted drug delivery and subsequent hypoxia-responsive intra-
tumoral drug release of ALN-HR-PMs. DOX is not suitable for in vivo
imaging as its fluorescence in visible range is severely interfered by auto-
fluorescence from biological tissues. Cy5.5 is thus loaded into the mi-
celles, forming ALN-HR-PMs/Cy5.5 as well as PMs/Cy5.5 and ALN-
PMs/Cy5.5 for the whole-body NIR imaging. As revealed in Fig. 4D,
all Cy5.5-loaded micelles were quickly distributed throughout the body
following injection. Although Cy5.5 accumulation in the metastatic
bone, facilitated by the EPR effects, was found in all groups, the fluo-
rescence in the metastatic bone in PMs/Cy5.5 group almost disappeared
at 24 h post-injection, indicating that the EPR effects was not effective
enough for drug delivery to bone using traditional nanocarriers. For
ALN-PMs/Cy5.5 group, although bone-targeted delivery was achieved,
as validated by their relatively strong fluorescence signal in the meta-
static bone, accumulated Cy5.5 was shielded inside the micelles,
resulting in the weak fluorescence signal. A stronger fluorescence signal
in the metastatic bone was observed in ALN-HR-PMs/Cy5.5 group, as a
result of hypoxia-responsive drug release following accumulation. Ex
vivo tissue-specific imaging also confirmed the above results, where
ALN-HR-PMs/Cy5.5 displayed the strongest fluorescence signal in the
metastatic bone (Fig. 4E). No high Cy5.5 accumulation in the tumor-free
bone was observed in all groups, also indicating that ALN had no strong
affinity to healthy bone. These results were well consistent with drug
tissue distribution results. NIR imaging result was then quantitatively
There are several in vitro and in vivo bone metastasis models.
Injecting tumor cells such as breast, prostate, or multiple myeloma, into
Table 1
Pharmacokinetic parameters of DOX after intravenous injection of DOX, PMs/
DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at 5 mg DOX/kg body weight (n =
3). *p < 0.05, **p < 0.01, and ***p < 0.001 versus DOX.
Parameters
T1/
T1/
AUC0→∞
g/mL⋅h)
CL
K10
MRT
(h)
α(h)
2β(h)
(
μ
(mL/
g⋅h)
(1/h)
2
DOX
0.24
11.73
97.78
8.47
0.48
14.51
PMs/DOX
ALN-PMs/
DOX
1.04***
0.98**
19.08*
21.40*
265.87***
282.05***
3.14*
3.07*
0.18***
0.18***
21.21*
25.36*
ALN-HR-
PMs/DOX
0.77*
21.97*
232.14**
3.34*
0.23**
26.97*
312

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
analyzed by calculating the fluorescence ratio of metastatic bone site to
tumor-free bone site. As shown in Fig. 4F, the ratio of ALN-HR-PMs/
Cy5.5 was generally higher than that of AZO-free groups, showing a
significant difference at 24 h post-injection (p < 0.01 versus PMs/Cy5.5
and ALN-PMs/Cy5.5), while no significant difference was found be-
tween PMs/Cy5.5 and ALN-PMs/Cy5.5. Moreover, in comparison to
PMs/Cy5.5 and ALN-PMs/Cy5.5, ALN-HR-PMs/Cy5.5 showed a stron-
ger fluorescence signal in the metastatic bone than in the normal tissues
such as liver and kidney. It is reported that only a very small proportion
of injected drugs even encapsulated in nanocarriers can be accumulated
in tumor and most of them are non-specifically accumulated in some
normal tissues, resulting from RES recognition. Therefore, the higher
fluorescence signal in the metastatic bone than in the liver and kidney
suggested that ALN-HR-PMs could retain their payload in normal tissues
for reduced side effects, while specifically releasing it in the metastatic
bone to achieve an effective therapeutic dosage.
Fig. 5. In vivo anti-cancer efficacy. (A) Protocol of drug treatment for cancer-induced spontaneous pain evaluation. When cancer-induced pain was observed on the
7th day, the mice were intravenously injected with PBS, DOX, PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at 5 mg DOX/kg body weight every four days. (B
and C) Number of flinches, (D and E) lifting time, and (G and F) score for the use of tumor-bearing limb over a 4-min period. Four, three, two, one, and zero scores
represented normal use, occasional limping, partial non-use, substantial non-use, and non-use of tumor-bearing limb, respectively. (H) Survival situation analysis of
tumor-bearing mice. Data were presented as mean ± SD (n = 8); *p < 0.05, **p < 0.01, and ***p < 0.001 versus ALN-PMs/DOX; #p < 0.05, ##p < 0.01, and ###p <
&
&&
&&&
0.001 versus PMs/DOX; p < 0.05, p < 0.01, and
p < 0.001 versus DOX; $p < 0.05, $$p < 0.01, and $$$p < 0.001 versus PBS.
313

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
After Ex vivo tissue-specific imaging, the tumor tissue was stripped
from the metastatic bone for immunofluorescence staining. HIF-1
DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX groups were 19, 22, 19,
24, and 31 days, respectively. ALN-HR-PMs/DOX significantly pro-
longed survival time of the mice (p < 0.05 versus DOX and ALN-PMs/
DOX; p < 0.01 versus PBS and PMs/DOX). Notably, non-targeted PMs/
DOX even promoted death compared with DOX. Although the accu-
mulation of PMs/DOX in metastatic bone was at least not less than that
of DOX, according to tissue distribution data, their slow and sustained
drug release behavior might make local therapeutic dosage even lower
than free DOX. This result also indicated that for nanocarriers, effective
drug accumulation in metastatic bone followed by rapid drug release to
an effective therapeutic dosage was very crucial for treating bone
metastasis.
α
(hypoxia inducible factor 1α), oxygen-labile α subunit of HIFs, is a key
regulator of cellular responses to hypoxia and its overexpression is
highly correlated with hypoxia. Immunofluorescence staining revealed
that HIF-1
α was over-expressed in all bone metastatic tumor tissues and
in the HIF-1
α
overexpression region, a weak Cy5.5 fluorescence was
observed in PMs/Cy5.5 and ALN-PMs/Cy5.5 groups, while a relatively
strong fluorescence was observed in ALN-HR-PMs/Cy5.5 group
(Fig. 4G). The immunofluorescence staining result fitted with NIR im-
aging result well, which was also a validation that ALN-HR-PMs could
sensitively respond to hypoxic bone metastasis.
In another experiment, tumor-bearing mice were treated (Fig. 6A)
and metastatic bones from tumor-bearing mice as well as healthy bone
from healthy mice were harvested on the 21st day for Micro-CT and
histopathological analysis. Digital photos revealed that the metastatic
bone in PBS group was heavily burdened with tumor and tumor cells had
expanded out of bone, forming a solid tumor around bone. After treated
with DOX, PMs/DOX, and ALN-PMs/DOX, tumor burden was reduced in
some extent. Encouragingly, the metastatic bone in ALN-HR-PMs/DOX
group appeared to be same with the healthy bone in shape (Fig. 6B).
Micro-CT scanning provided more information about the metastatic
bones. Micro-CT 2D images demonstrated that compared with the
healthy bone, bone morphology and structure of metastatic bones was
severely damaged, showing severe facture in PBS group. DOX and PMs/
DOX seemed to have no beneficial effect on protecting metastatic bone
against fracturing, although reduced tumor burden was observed in
these two groups; bone loss could still be seen in ALN-PMs/DOX group.
Desirable therapeutic outcome was observed in ALN-HR-PMs/DOX
group, showing no fracture (Fig. 6C). The 3D reconstruction further
indicated the surface of the healthy bone was smooth and flat, while it
became incomplete and rough, once metastasized. Especially, 3D im-
aging results demonstrated that compared with the other groups, ALN-
HR-PMs/DOX could effectively inhibit bone destruction, maintaining
its shape and integrity well, although bone surface of this group was still
a little rough, due to tumor erosion (Fig. 6D). Bone metrics such as BMD
and Tb.N were also evaluated (Fig. 6E and F). Bone metrics in ALN-HR-
PMs/DOX group were significantly better than those in PBS, DOX, PMs/
DOX, and ALN-PMs/DOX groups, which were comparable to those in the
healthy bone. No significant difference was found among DOX, PMs/
DOX, and ALN-PMs/DOX groups and between healthy and ALN-HR-
PMs/DOX groups.
3.4. In vivo anti-cancer efficacy
Most literature evaluating bone-targeted drug carriers for bone
metastasis focus on tumor inhibition efficacy while neglecting the other
important clinical aspects. Cancer-induced bone pain is the most prev-
alent symptom of patients with bone metastasis. Bone pain relief is
particularly important for improving patients' quality of life and also
emerges as an index of anti-cancer efficacy. To evaluate in vivo anti-
cancer efficacy, tumor-bearing mice began to receive drug treatment
every four days when cancer-induced spontaneous bone pain, such as
number of flinches and lifting time, was observed (Fig. 5A). As shown in
Fig. 5B and C, the mice quickly recovered from the surgery for bone
metastasis model establishment, showing no problem in walking on the
1st, 3rd, and 5th days. Slight pain with a small number of flinches over a
4-min period was observed on the 7th day in all groups. The number of
flinches in PBS, DOX and PMs/DOX groups quickly increased over time,
reaching peak on the 17th day, and subsequently decreased, probably
due to aggravated tumor burden. Compared with PBS, DOX, and PMs/
DOX groups, ALN-PMs/DOX group showed slightly reduced number of
flinches with considerable difference on the 17th day (p < 0.05 versus
DOX and PMs/DOX; p < 0.01 versus PBS), but not at any other time
points. Remarkably reduced number of flinches was observed in ALN-
HR-PMs/DOX group on the 15th, 17th, and 19th days. For example,
on the 17th day, ALN-HR-PMs/DOX group was significantly different
from PBS, DOX, PMs/DOX, and ALN-PMs/DOX groups (p < 0.001).
Moreover, the number of flinches in ALN-HR-PMs/DOX group generally
remained at a relatively low level. Spontaneous lifting time exhibited
nearly the same tread as number of flinches and similarly, ALN-HR-PMs/
DOX group showed significantly short lifting time in comparison to PBS,
DOX, PMs/DOX, and ALN-PMs/DOX groups (Fig. 5D and E). The level of
pain was further scored by the use of tumor-bearing limb. As depicted in
Fig. 5F and G, the limb use in ALN-HR-PMs/DOX group appeared normal
with occasional limping, showing a significant difference compared with
the other groups, especially on the 15th, 17th, and 19th days. Taken
together, PBS, DOX, and PMs/DOX groups exhibited almost the same
pain behaviors without any significant differences among them
throughout the experiment, corroborating the fact that traditional
chemotherapy and nanomedicine were not effective in treating bone
metastasis. Although a bone-targeted effect was achieved for ALN-PMs/
DOX, as validated by drug tissue distribution data, their efficacy in
relieving pain was unsatisfactory. ALN-HR-PMs/DOX could significantly
relieve pain, as a result of bone-targeted accumulation followed by
hypoxia-responsive release to achieve an effective therapeutic dosage in
the metastatic bone. Additionally, it was worth pointing out that ALN in
ALN-PMs/DOX and ALN-HR-PMs/DOX acted as a bone-targeted moiety
more than as an anti-resorptive agent for pain relief. Bioavailability loss
when its amino group was used for conjugation with copolymer and/or
insufficient dosage might contribute to its restricted therapeutic
outcome. On the other hand, as mentioned above, its binding with bone
relied on hydroxyl and phosphonate groups, and thus its bone-targeted
ability was not affected.
After stripping tumor tissue, the harvested bones were decalcified for
H&E, TRAP, ALP, and TRAP/ALP staining (Fig. 6G). H&E staining
showed that almost all cells in bone marrow were necrotic in PBS group.
Although alive cells in bone marrow increased, a lot of necrotic cells
were still observed in DOX, PMs/DOX, and ALN-PMs/DOX groups.
Noting that the region with many necrotic cells in bone marrow was also
denoted as tumor in the literature [49]. As expected, the metastatic bone
in ALN-HR-PMs/DOX group was histologically similar to the healthy
bone, indicating that tumor growth was dramatically inhibited. There is
a dynamic balance between osteoclast-mediated bone resorption and
osteoblast-mediated bone formation for bone turnover in healthy bone
[50]. TRAP and ALP are two biomarkers used to detect osteoclast and
osteoblast activities in bone, respectively. TRAP, ALP, and TRAP/ALP
staining revealed that osteoclasts stained as wine red and osteoblasts
stained as dark brown were mostly concentrated on bone/bone marrow
interface and around trabecular bone [51], respectively. Bone turnover
was imbalanced and osteoclast activity increased, while osteoblast ac-
tivity decreased in PBS, DOX, PMs/DOX, and ALN-PMs/DOX groups. It
indicated that bone resorption was predominant over bone formation in
these groups, thus leading to bone destruction, as revealed in Micro-CT
imaging (Fig. 6C and D). Moreover, although bone metastatic prostate
cancer is reported to be preponderantly osteoblastic [52], it is osteolytic
in this work, inferring that bone metastasis phenotype is mainly deter-
mined by tumor cell type. After ALN-HR-PMs/DOX treatment, bone
Pain behaviors of the mice reflected their survival situation well. As
demonstrated in Fig. 5F, the mean survival times of PBS, DOX, PMs/
314

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
Fig. 6. Bone histological evaluation. (A) Protocol of drug treatment for histological evaluation. On the 7th day after RM-1 cells injection, the mice were intravenously
injected with PBS, DOX, PMs/DOX, ALN-PMs/DOX, and ALN-HR-PMs/DOX at 5 mg DOX/kg body weight every four days, and metastatic bones as well as healthy
bone were harvested for histological evaluation on the 21st day. The bone from healthy mice without bearing tumor and drug treatment was denoted as healthy bone
and was also evaluated for comparison. (B) Digital photos, (C) Micro-2D imaging, and (D) Micro-CT 3D reconstruction of metastatic bones as well as healthy bone. (E)
BMD and (F) Tb.N parameters of metastatic bones as well as healthy bone acquired from (D). Data were presented as mean ± SD (n = 3); *p < 0.05, **p < 0.01, and
***p < 0.001. (G) H&E, TRAP, ALP, and TRAP/ALP staining of metastatic bones as well as healthy bone. B: bone; BM: bone marrow. Black arrows indicated os-
teoclasts stained as wine red and red arrows indicated osteoblasts stained as dark brown. (For interpretation of the references to color in this figure legend, the reader
is referred to the web version of this article.)
turnover balance was re-established, showing decreased osteoclast ac-
tivity and increased osteoblast activity, and resultant bone microenvi-
ronment was similar to the healthy bone. Quantitative analysis further
demonstrated that osteoclast activity in ALN-HR-PMs/DOX group was
significantly low, while osteoblast activity was significantly high,
compared with DOX, PMs/DOX, and ALN-PMs/DOX groups (Supple-
mentary Fig. S8). Together, histological examination corroborated that
ALN-HR-PMs/DOX were able to inhibit tumor growth and thus balance
bone turnover by inhibiting osteoclast activity and promoting osteoblast
activity for relieved bone pain and prolonged survival time.
effect of ALN and loaded DOX is an alternative to solve this issue.
Contralateral tumor-free bones of tumor-bearing mice were also
harvested for histological evaluation by H&E, TRAP, ALP, and TRAP/
ALP staining. As shown in supplementary Fig. S9, tumor-free bones in all
drug-treated groups were histologically similar to the healthy bone from
tumor-free mice. ALN-HR-PMs/DOX were therefore safe to healthy
bone, benefiting from their affinity to metastatic bone. Moreover, some
important tissues including heart, liver, spleen, lung, and kidney were
also harvested for H&E staining (Supplementary Fig. S10). The results
showed that free DOX was toxic to heart, which was mainly character-
ized by myocardial disruption. No considerable toxicity to heart was
observed in the other groups and no considerable toxicity to liver,
spleen, lung, and kidney was observed in all groups. ALN-HR-PMs/DOX
were therefore a biocompatible drug delivery system.
Previously reported ALN-functionalized nanocarriers aim to improve
therapeutic outcome by increasing drug accumulation in metastatic
bone through bone-targeted effect of ALN. We develop ALN-
functionalized hypoxia-responsive nanocarriers, aiming to improve
therapeutic outcome by maximize drug bioavailability via hypoxia-
responsive drug release after bone-targeted effect is achieved. The re-
sults demonstrate that therapeutic outcome of ALN-functionalized
nanocarriers is significantly augmented, once endowed with hypoxia-
responsive feature, showing the outstanding advantage over conven-
tional ALN-functionalized nanocarriers. Additionally, it is worth noting
that ALN-functionalized nanocarriers always exhibit significantly
improved therapeutic outcome in comparison to non-functionalized
nanocarriers in the literature, while no significant difference between
them is observed in this work, displaying one of advantages of designed
ALN-functionalized copolymer. Conjugating ALN with abundant amino
groups of PLL to increase ALN content in the copolymers for synergistic
4. Conclusion
In summary, we prepared alendronate-functionalized hypoxia-
responsive polymeric micelles as drug carrier for effective treatment of
bone metastatic prostate cancer. In vitro studies revealed that the pre-
pared micelles were highly sensitive to hypoxia and were more toxic to
hypoxic tumor cells. In vivo studies showed that after a targeted accu-
mulation in metastatic bone, the micelles could rapidly release their
payload to an effective therapeutic dosage in responding to hypoxic
bone metastasis microenvironment. As a result, the micelles could
effectively suppress tumor growth and preserve bone structure by
315

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
[19] W.J. Zhang, I. Bado, H. Wang, H.C. Lo, X.H.F. Zhang, Bone metastasis: find your
niche and fit in, Trends Cancer 5 (2019) 95–110.
inhibiting osteoclast activity and promoting osteoblast activity,
achieving an enhanced therapeutic outcome with relieved bone pain and
prolonged survival time. The micelles therefore emerged as a reliable
drug delivery system for treating bone metastasis.
[20] R. Kumari, D. Sunil, R.S. Ningthoujam, Hypoxia-responsive nanoparticle based
drug delivery systems in cancer therapy: an up-to-date review, J. Control. Release
319 (2020) 135–156.
[21] Y. Li, J. Jeon, J.H. Park, Hypoxia-responsive nanoparticles for tumor-targeted drug
delivery, Cancer Lett. 490 (2020) 31–43.
[22] Y. Li, A.L. Lu, M.M. Long, L. Cui, Z.P. Chen, L. Zhu, Nitroimidazole derivative
incorporated liposomes for hypoxia-triggered drug delivery and enhanced
therapeutic efficacy in patient-derived tumor xenografts, Acta Biomater. 83 (2019)
334–348.
Declaration of Competing Interest
The authors declare no competing financial interests.
[23] M.M. Long, A.L. Lu, M. Lu, L.Y. Weng, Q.P. Chen, L. Zhu, Z.P. Chen, Azo-inserted
responsive hybrid liposomes for hypoxia-specific drug delivery, Acta Biomater. 115
(2020) 343–357.
Acknowledgments
[24] M. Nagae, T. Hiraga, T. Yoneda, Acidic microenvironment created by osteoclasts
causes bone pain associated with tumor colonization, J. Bone Miner. Metab. 25
(2007) 99–104.
The work was supported by the National Natural Science Foundation
of China (81401506, 81971131, and 81873924) and the Open Research
Fund of State Key Laboratory of Bioelectronics, Southeast University
(OPSKLB202007).
[25] S. Feng, Z.X. Wu, Z. Zhao, J. Liu, K. Sun, C. Guo, H. Wang, Z. Wu, Engineering of
bone- and CD44-dual-targeting redox-sensitive liposomes for the treatment of
orthotopic osteosarcoma, ACS Appl. Mater. Interfaces 11 (2019) 7357–7368.
[26] R.W. Johnson, M.E. Sowder, A.J. Giaccia, Hypoxia and bone metastatic disease,
Curr. Osteoporos. Rep. 15 (2017) 231–238.
[27] F.F. Wang, L. Chen, R. Zhang, Z.P. Chen, L. Zhu, RGD peptide conjugated liposomal
drug delivery system for enhance therapeutic efficacy in treating bone metastasis
from prostate cancer, J. Control. Release 196 (2014) 222–233.
[28] T. Thambi, S. Son, D.S. Lee, J.H. Park, Poly(ethylene glycol)-b-poly(lysine)
copolymer bearing nitroaromatics for hypoxia-sensitive drug delivery, Acta
Biomater. 29 (2016) 261–270.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.jconrel.2021.04.035.
[29] C. Lim, J. Moon, T. Sim, W.R. Won, E.S. Lee, Y.S. Youn, K.T. Oh, A nano-complex
system to overcome antagonistic photo-chemo combination cancer therapy,
J. Control. Release 295 (2019) 164–173.
References
[30] Z. Zhao, Y. Zhao, C. Xie, C. Chen, D. Lin, S. Wang, D. Lin, X. Cui, Z. Guo, J. Zhou,
Dual-active targeting liposomes drug delivery system for bone metastatic breast
cancer: synthesis and biological evaluation, Chem. Phys. Lipids 223 (2019)
104785.
[1] M.E. Sowder, R.W. Johnson, Bone as a preferential site for metastasis, JBMR Plus 3
(2019), e10126.
[2] S. D’Oronzo, R. Coleman, J. Brown, F. Silvestris, Metastatic bone disease:
pathogenesis and therapeutic options up-date on bone metastasis management,
J. Bone Oncol. 15 (2019) 100205.
[31] W. Piao, S. Tsuda, Y. Tanaka, S. Maeda, F.Y. Liu, S. Takahashi, Y. Kushida,
T. Komatsu, T. Ueno, T. Terai, T. Nakazawa, M. Uchiyama, K. Morokuma,
T. Nagano, K. Hanaoka, Development of Azo-based fluorescent probes to detect
different levels of hypoxia, Angew. Chem. Int. Ed. 52 (2013) 13028–13032.
[32] P. Verwilst, J. Han, J. Lee, S. Mun, H.G. Kang, J.S. Kim, Reconsidering azobenzene
as a component of small-molecule hypoxia-mediated cancer drugs: a theranostic
case study, Biomaterials 115 (2017) 104–114.
[3] G.D. Roodman, T.A. Guise, Chapter 22 - bone and cancer, in: D.B. Burr, M.R. Allen
(Eds.), Basic and Applied Bone Biology, Second edition, Academic Press, 2019,
pp. 411–423.
[4] C. Kan, G. Vargas, F. Le Pape, P. Clezardin, Cancer cell colonisation in the bone
microenvironment, Int. J. Mol. Sci. 17 (2016) 1–16.
[5] I.M. Adjei, M.N. Temples, S.B. Brown, B. Sharma, Targeted nanomedicine to treat
bone metastasis, Pharmaceutics 10 (2018) 205.
[33] M. Danaei, M. Dehghankhold, S. Ataei, F.H. Davarani, R. Javanmard, A. Dokhani,
S. Khorasani, M.R. Mozafari, Impact of particle size and Polydispersity index on the
clinical applications of Lipidic Nanocarrier systems, Pharmaceutics 10 (2018) 57.
[34] I.M. Adjei, B. Sharma, C. Peetla, V. Labhasetwar, Inhibition of bone loss with
surface-modulated, drug-loaded nanoparticles in an intraosseous model of prostate
cancer, J. Control. Release 232 (2016) 83–92.
[6] C. Shi, T.T. Wu, Y. He, Y. Zhang, D.H. Fu, Recent advances in bone-targeted
therapy, Pharmacol. Ther. 207 (2020) 107473.
[7] H.H. Van Acker, S. Anguille, Y. Willemen, E.L. Smits, V.F. Van Tendeloo,
Bisphosphonates for cancer treatment: mechanisms of action and lessons from
clinical trials, Pharmacol. Ther. 158 (2016) 24–40.
[35] J. Panchal, J. Kotarek, E. Marszal, E.M. Topp, Analyzing subvisible particles in
protein drug products: a comparison of dynamic light scattering (DLS) and
resonant mass measurement (RMM), AAPS J. 16 (2014) 440–451.
[36] R.Z. Zhao, S. Jiang, N.Y. Ru, B. Jiao, Z.B. Yu, Comparison of hypoxic effects
induced by chemical and physical hypoxia on cardiomyocytes, Can. J. Physiol.
Pharmacol. 97 (2019) 980–988.
[8] B.A. Gartrell, F. Saad, Managing bone metastases and reducing skeletal related
events in prostate cancer, Nat. Rev. Clin. Oncol. 11 (2014) 335–345.
´
[9] L.M.S. Boeve, M.C.C.M. Hulshof, A.N. Vis, A.H. Zwinderman, J.W.R. Twisk, W.P.
J. Witjes, K.P.J. Delaere, R.J.A.V. Moorselaar, P.C.M.S. Verhagen, G. Van Andel,
Effect on survival of androgen deprivation therapy alone compared to androgen
deprivation therapy combined with concurrent radiation therapy to the prostate in
patients with primary bone metastatic prostate cancer in a prospective randomised
Clinical Trial: Data from the HORRAD Trial, Eur. Urol. 75 (2019) 410–418.
[37] L. Cavallini, M. Valente, A. Bindoli, NADH and NADPH inhibit lipid peroxidation
promoted by hydroperoxides in rat liver microsomes, BBA-Lipid Lipid Met. 752
(1983) 339–345.
´
[10] E. Akoury, P. Ahangar, A. Nour, J. Lapointe, K.-P. Guerard, L. Haglund, D.
[38] T. Ito, L. Sun, M.A. Bevan, R.M. Crooks, Comparison of nanoparticle size and
electrophoretic mobility measurements using a carbon-nanotube-based coulter
counter, dynamic light scattering, transmission electron microscopy, and phase
analysis light scattering, Langmuir 20 (2004) 6940–6945.
H. Rosenzweig, M.H. Weber, Low-dose zoledronate for the treatment of bone
metastasis secondary to prostate cancer, Cancer Cell Int. 19 (2019) 28.
[11] H. Cheng, A. Chawla, Y. Yang, Y. Li, J. Zhang, H.L. Jang, A. Khademhosseini,
Development of nanomaterials for bone-targeted drug delivery, Drug Discov.
Today 22 (2017) 1336–1350.
[39] T.T. Nguyen, D.P. Tran, P.D.Q. Huy, Z. Hoang, P. Carloni, P.V. Pham, C. Nguyen,
M.S. Li, Ligand binding to anti-cancer target CD44 investigated by molecular
simulations, J. Mol. Model. 22 (2016).
[12] C.F. Mu, J.L. Shen, J. Liang, H.S. Zheng, Y. Xiong, Y.H. Wei, F.Z. Li, Targeted drug
delivery for tumor therapy inside the bone marrow, Biomaterials 155 (2018)
191–202.
[40] T.K. Ryu, R.H. Kang, K.Y. Jeong, D.R. Jun, J.M. Koh, D. Kim, S.K. Bae, S.W. Choi,
Bone-targeted delivery of nanodiamond-based drug carriers conjugated with
alendronate for potential osteoporosis treatment, J. Control. Release 232 (2016)
152–160.
[13] H. Cheng, A. Chewla, Y.F. Yang, Y.X. Li, J. Zhang, H.L. Jang, A. Khademhosseini,
Development of nanomaterials for bone-targeted drug delivery, Drug Discov.
Today 22 (2017) 1336–1350.
[41] E.A. Sheta, H. Trout, J.J. Gildea, M.A. Harding, D. Theodorescu, Cell density
[14] X.Z. Ke, W. Lin, X.K. Li, H.L. Wang, X. Xiao, Z. Guo, Synergistic dual-modified
liposome improves targeting and therapeutic efficacy of bone metastasis from
breast cancer, Drug Deliv. 24 (2017) 1680–1689.
mediated pericellular hypoxia leads to induction of HIF-1α via nitric oxide and
Ras/MAP kinase mediated signaling pathways, Oncogene 20 (2001) 7624–7634.
[42] V.P. Torchilin, Recent advances with liposomes as pharmaceutical carriers, Nat.
Rev. Drug Discov. 4 (2005) 145–160.
[15] T.T. Pham, H.T. Nguyen, C.D. Phung, S. Pathak, S. Regmi, D.-H. Ha, J.O. Kim, C.
S. Yong, S.K. Kim, J.-E. Choi, S. Yook, J.-B. Park, J.-H. Jeong, Targeted delivery of
doxorubicin for the treatment of bone metastasis from breast cancer using
alendronate-functionalized graphene oxide nanosheets, J. Ind. Eng. Chem. 76
(2019) 310–317.
˜
[43] D. Manzanares, V. Cena, Endocytosis: the nanoparticle and submicron
nanocompounds gateway into the cell, Pharmaceutics 12 (2020) 371.
[44] A. Sukhanova, S. Bozrova, P. Sokolov, M. Berestovoy, A. Karaulov, I. Nabiev,
Dependence of nanoparticle toxicity on their physical and chemical properties,
Nanoscale Res. Lett. 13 (2018) 44.
[16] S.B. Bai, D.Z. Liu, Y. Cheng, H. Cui, M. Liu, M.X. Cui, B.L. Zhang, Q.B. Mei, S.
Y. Zhou, Osteoclasts and tumor cells dual targeting nanoparticle to treat bone
metastases of lung cancer, Nanomed-Nanotechnol. 21 (2019) 102054.
[17] Y. Wang, Q. Huang, X. He, H. Chen, Y. Zou, Y. Li, K. Lin, X. Cai, J. Xiao, Q. Zhang,
Y. Cheng, Multifunctional melanin-like nanoparticles for bone-targeted chemo-
photothermal therapy of malignant bone tumors and osteolysis, Biomaterials 183
(2018) 10–19.
[45] J.S. Suk, Q. Xu, N. Kim, J. Hanes, L.M. Ensign, PEGylation as a strategy for
improving nanoparticle-based drug and gene delivery, Adv. Drug Deliv. Rev. 99
(2016) 28–51.
[46] F. Aielli, M. Ponzetti, N. Rucci, Bone metastasis pain, from the bench to the
bedside, Int. J. Mol. Sci. 20 (2019) 280.
[47] A.J. Roelofs, K. Thompson, F.H. Ebetino, M.J. Rogers, F.P. Coxon,
Bisphosphonates: molecular mechanisms of action and effects on bone cells,
monocytes and macrophages, Curr. Pharm. Des. 16 (2010) 2950–2960.
[18] P. Lavrador, V.M. Gaspar, J.F. Mano, Stimuli-responsive nanocarriers for delivery
of bone therapeutics - barriers and progresses, J. Control. Release 273 (2018)
51–67.
316

M. Long et al.
Journal of Controlled Release 334 (2021) 303–317
[48] Y.F. He, Y.J. Huang, Z.Y. Huang, Y.L. Jiang, X.Q. Sun, Y.F. Shen, W.J. Chu, C.
S. Zhao, Bisphosphonate-functionalized coordination polymer nanoparticles for the
treatment of bone metastatic breast cancer, J. Control. Release 264 (2017) 76–88.
[49] L. Coenegrachts, C. Maes, S. Torrekens, R. Van Looveren, M. Mazzone, T.A. Guise,
R. Bouillon, J.M. Stassen, P. Carmeliet, G. Carmeliet, Anti-placental growth factor
reduces bone metastasis by blocking tumor cell engraftment and osteoclast
differentiation, Cancer Res. 70 (2010) 6537–6547.
F. Djouad, G. Schett, C. Scholtysek, G. Kronke, PPAR delta-mediated mitochondrial
rewiring of osteoblasts determines bone mass, Sci. Rep. 10 (2020) 8428.
[51] T. Ninomiya, A. Hosoya, H. Nakamura, K. Sano, T. Nishisaka, H. Ozawa, Increase
of bone volume by a nanosecond pulsed laser irradiation is caused by a decreased
osteoclast number and an activated osteoblasts, Bone 40 (2007) 140–148.
[52] M.N. Wang, F. Xia, Y.Q. Wei, X.W. Wei, Molecular mechanisms and clinical
management of cancer bone metastasis, Bone Res. 8 (2020) 30.
[50] D.I.H. Muller, C. Stoll, K. Palumbo-Zerr, C. Bohm, B. Krishnacoumar, N. Ipseiz,
J. Taubmann, M. Zimmermann, M. Bottcher, D. Mougiakakos, J. Tuckermann,
317