A colon-targeted podophyllotoxin nanoprodrug: synthesis, characterization, and supramolecular hydrogel formation for the drug combination

Article abstract of DOI:10.1039/d0tb02719g

Making full use of the undeveloped bioactive natural product derivatives by selectively delivering them to target sites can effectively increase their druggability and reduce the wastage of resources. Azo-based prodrugs are widely regarded as an effective targeted delivery means for colon-related disease treatment. Herein, we report a new-type of azo-based nanoprodrug obtained from bioactive natural products, in which the readily available podophyllotoxin natural products are connected with methoxy polyethylene glycol (mPEG)viaa multifunctional azobenzene group. The amphiphilic prodrug can form nanosized micelles in water and will be highly selectively activated by azoreductases, leading to thein situgeneration of anticancer podophyllotoxin derivatives (AdP) in the colon after the cleavage of the azo bond. To satisfy the demand of drug carriers for cancer combination therapy in clinics, α-CD is further introduced into this nanoprodrug micelle system to form a supramolecular hydrogelviaa cascade self-assembly strategy. Using imaging mass spectrometry (IMS), the colon-specific drug release ability of the hydrogel after oral administration is demonstrated at the molecular level. Finally, the nanoprodrug hydrogel is further used as a carrier to load a hydrophilic anti-cancer drug 5-FU during the hierarchical self-assembly process and to co-deliverAdPand 5-FU for the drug combination. The combination use ofAdPand 5-FU provides enhanced cytotoxicity which indicates a significant synergistic interaction. This work offers a new way to enhance the therapeutic effect of nanoprodrugsviadrug combination, and provides a new strategy for reusing bioactive natural products and their derivatives.

Full text of DOI:10.1039/d0tb02719g

Journal of
Materials Chemistry B
View Article Online
View Journal | View Issue
PAPER
A colon-targeted podophyllotoxin nanoprodrug:
synthesis, characterization, and supramolecular
hydrogel formation for the drug combination†
Cite this: J. Mater. Chem. B, 2021,
9, 3200
Wei Ha,‡a Xiao-Bo Zhao,‡a Wei-Hua Zhao,a Jiang-Jiang Tangb and
a
Yan-Ping Shi
*
Making full use of the undeveloped bioactive natural product derivatives by selectively delivering them
to target sites can effectively increase their druggability and reduce the wastage of resources.
Azo-based prodrugs are widely regarded as an effective targeted delivery means for colon-related
disease treatment. Herein, we report a new-type of azo-based nanoprodrug obtained from bioactive
natural products, in which the readily available podophyllotoxin natural products are connected with
methoxy polyethylene glycol (mPEG) via a multifunctional azobenzene group. The amphiphilic prodrug
can form nanosized micelles in water and will be highly selectively activated by azoreductases, leading
to the in situ generation of anticancer podophyllotoxin derivatives (AdP) in the colon after the cleavage
of the azo bond. To satisfy the demand of drug carriers for cancer combination therapy in clinics, a-CD
is further introduced into this nanoprodrug micelle system to form a supramolecular hydrogel via a
cascade self-assembly strategy. Using imaging mass spectrometry (IMS), the colon-specific drug release
ability of the hydrogel after oral administration is demonstrated at the molecular level. Finally, the
nanoprodrug hydrogel is further used as a carrier to load a hydrophilic anti-cancer drug 5-FU during
the hierarchical self-assembly process and to co-deliver AdP and 5-FU for the drug combination.
The combination use of AdP and 5-FU provides enhanced cytotoxicity which indicates a significant
synergistic interaction. This work offers a new way to enhance the therapeutic effect of nanoprodrugs
via drug combination, and provides a new strategy for reusing bioactive natural products and their
derivatives.
Received 21st November 2020,
Accepted 9th March 2021
DOI: 10.1039/d0tb02719g
rsc.li/materials-b
drug discovery. However, only a few of them have been successfully
launched, while a huge number of other active derivatives were
Introduction
Compounds found in natural resources have abundant wasted due to their poor physicochemical properties. Taking
chemical structures and broad biological activities, and are podophyllotoxin as an example, at least 20 000 active derivatives
the most important sources for drug discovery.1–3 The structural have been reported, but only two of them are clinically applied.8,9
modification and optimization of active natural products could Therefore, making full use of these undeveloped bioactive natural
greatly improve their druggability and pharmacological activity, product derivatives can effectively increase their druggability and
which has been considered as a crucial strategy in natural reduce the wastage of resources.
product-based drug discovery.4–7 In the past century, a large
The prodrug strategy has received extensive attention in the
number of natural product derivatives have been synthesized for structural modification of active natural products because of its
significant advantages in improving the physicochemical
properties.10–12 One of the ultimate goals of the prodrug
strategy is to deliver the drug to the target site with high
a CAS Key Laboratory of Chemistry of Northwestern Plant Resources and Key
Laboratory for Natural Medicine of Gansu Province, Lanzhou Institute of Chemical
specificity while minimizing the drug exposure of other organs
by rational design.13,14 Azo-based prodrugs provide an effective
target delivery means for colon-related disease treatment.15,16
The colonic microflora can specifically secrete azoreductase
which has the ability to reduce the azobenzene group to the
corresponding phenylamine.17–19 Based on this unique character,
many commonly used drugs, such as sulfasalazine, olsalazine and
Physics, Chinese Academy of Sciences (CAS), Lanzhou 730000
People’s Republic of China. E-mail: shiyp@licp.cas.cn
b Shaanxi Key Laboratory of Natural Products & Chemical Biology, College of
Chemistry and Pharmacy, Northwest A&F University, Yangling, Shaanxi 712100
People’s Republic of China
† Electronic supplementary information (ESI) available: General information,
Tables S1 and S2, and supplementary Fig. S1–14. See DOI: 10.1039/d0tb02719g
‡ These authors contributed equally.
This journal is © The Royal Society of Chemistry 2021
3200 | J. Mater. Chem. B, 2021, 9, 3200–3209

View Article Online
Paper
Journal of Materials Chemistry B
balsalazide, are used as a prodrug for the targeted release of
active 5-ASA inside the colon.20–22 The phenylamino group is
present in various natural product derivatives, which include
many highly active derivatives. However, most of them have not
been further investigated since they were reported with some
physicochemical defects arising from their molecular structure;
this creates enormous wastage of compound resources.23–26
We thus proposed that introducing an azobenzene group to
these compounds via rational design of the molecular structure
would provide the possibility to deliver them specifically to the
colon for colon related disease treatment, resulting in more
efficient utilization of the bioactive compounds which bear a
phenylamino group. It is undeniable that azo-based prodrugs,
especially amphiphilic nanoprodrugs, are attractive in anti-
cancer compound delivery due to their potential in improving
drug solubility, stability and accumulation at the target organ.
However, although anticancer compounds can be delivered to
the organ or tumor via the nanoprodrug strategy, it is still
difficult to produce a satisfactory therapeutic effect by admin-
istration of a single type of drug for cancer treatment.27–29
Therefore, the development of an effective platform for controlled
combination drug therapy has received wide attention as a new
approach to antitumor treatment.30–35
Scheme 1 Schematic illustration of the design of the azo-based
podophyllotoxin nanoprodrug and its hydrogel form for drug combination.
Experimental
The synthesis route to the azo-based podophyllotoxin prodrug
is shown in Fig. 1a.
Synthesis of compound 243
In brief, a solution of podophyllotoxin natural products 40-O-
Podophyllotoxin is an important anticancer lead compound demethylepipodophyllotoxin (compound 1a, 2.5 mmol) or
isolated from the Podophyllum genus. Using podophyllotoxin as podophyllotoxin (compound 1b, 2.5 mmol) in 25 mL of dry CH2Cl2
a starting compound, a large number of derivatives have been was bubbled with HBr gas at 0 1C for 45 min. The unreacted HBr
synthesized, which include two anticancer drugs etoposide was removed through bubbling with N2 and the solution was
(VP16) and teniposide (VM26).36–38 However, the overwhelming evaporated to obtain the crude intermediate compounds.
majority of podophyllotoxin-derived active compounds were
gradually abandoned. 40-O-Demethyl-4b-(400-aminoanilino)-4- dry CH2Cl2 (10 mL), and K2CO3 (3 mmol, 414 mg) was added
desoxy-podophyllotoxin (AdP) is one of the phenylamino- under N2 atmosphere.
min later, 4,40-azodianiline
The intermediate compounds (1 mmol) were dissolved in
a
5
containing podophyllotoxin derivatives which was first reported (1.2 mmol, 254 mg) was added and allowed to react for 4 h.
in 1990.39 This compound exhibits better DNA topoisomerase II The reaction solution was filtered, evaporated and purified to
inhibitory activity than its structural analogue GL331,
a
obtain compound 2 as a gray solid (53–61%). Compound 2a,
podophyllotoxin-derived candidate anticancer drug, but no 1H NMR (400 MHz, d6-DMSO) d: 3.05–3.08 (m, 1H, 3-H), 3.28–
in-depth research has been reported.40,41 Therefore, we proposed 3.36 (dd, 1H, J = 5.2 Hz, 14.4 Hz, 2-H), 3.66 (br, 6H, 30,50-OCH3),
a novel nanoprodrug system in which the readily available 3.70 (m, 1H, 11-H), 4.39 (t, 1H, J = 7.6 Hz, 11-H), 4.54 (d, 1H, J =
podophyllotoxin natural products are connected with methoxy 5.2 Hz, 1-H), 4.98 (m, 1H, 4-H), 6.01 (d, J = 10.4 Hz, 2H, 13-H),
polyethylene glycol (mPEG) via a multifunctional azobenzene 6.29 (s, 2H, 20,60-H), 6.57 (s, 1H, 8-H), 6.64 (d, 2H, J = 8.8 Hz,
group. The resulting amphiphilic prodrug can effectively improve 800,1200-H), 6.82 (m, 3H, 5,300,500-H), 7.56 (d, 2H, J = 8.4 Hz, 900,1100-
the water solubility of podophyllotoxin by forming nanosized H), 7.62 (d, 2H, J = 8.4 Hz, 200,600-H); 13C NMR (100 MHz, d6-DMSO)
micelles. As a consequence of the cleavage of the azo bond in d: 175.1, 151.6, 150.5, 147.7, 147.6, 147.6, 147.0, 144.0, 143.6,
the presence of azoreductases, active AdP will be in situ generated 135.2, 132.3, 131.4, 130.7, 124.4, 124.3, 113.9, 109.9, 109.7, 109.0,
in the colon, in which a phenylamino group could be modified on 101.7, 68.9, 60.2, 56.5, 50.8, 43.3, 41.4, 38.7; HRMS: C33H30N4O7Na
the parent compound (Scheme 1). Furthermore, the colon- for [M + Na]+, calculated 617.2007, found 617.2014. Compound 2b,
targeted nanoprodrug can be used as a building block to construct 1H NMR (400 MHz, CDCl3) d: 3.05–3.06 (m, 1H, 3-H), 3.11–3.16 (dd,
a biocompatible supramolecular hydrogel for loading another 1H, J = 4.7 Hz, 12.0 Hz, 2-H), 3.77 (br, 6H, 30,50-OCH3), 3.82 (s, 3H,
water-soluble anticancer drug just by introducing a-CD to the 40-OCH3), 4.00 (t, 1H, J = 9.6 Hz, 11-H), 4.42 (t, 1H, J = 7.9 Hz, 11-H),
system (Scheme 1). The strong hydrogen bonding interaction 4.62 (d, 1H, J = 4.6 Hz, 1-H), 4.79 (m, 1H, 4-H), 5.97 (d, 2H, J =
among adjacent a-CD pseudopolyrotaxanes (PPRs) provides a 6.8 Hz, H-13), 6.33 (s, 2H, 20,60-H), 6.54 (s, 1H, 8-H), 6.63 (d, 2H, J =
shell–shell interaction for driving the self-assembly of the prodrug 8.2 Hz, 800,1200-H), 6.74 (d, 2H, J = 7.8 Hz, 300,500-H), 6.81 (s, 1H, 5-H),
nanomicelles into a supramolecular hydrogel.42 Combination of 7.75 (d, 2H, J = 7.7 Hz, 900,1100-H), 7.80 (d, 2H, J = 8.0 Hz, 200,600-H);
the colon-targeted podophyllotoxin nanoprodrug and its hydrogel 13C NMR (100 MHz, CDCl3) d: 174.5, 152.6, 148.8, 148.7, 148.4,
form provides a new strategy to reuse bioactive natural products 147.7, 145.6, 135.0, 131.9, 130.0, 124.6, 124.5, 114.8, 112.1, 109.9,
and offers a powerful carrier for co-delivery of hydrophobic and 109.3, 108.3, 101.6, 68.8, 60.8, 56.3, 52.4, 43.6, 41.8, 38.7; HRMS:
hydrophilic drugs for drug combination.
C32H29N4O7 for [M + H]+, calculated 609.2344, found 609.2335.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. B, 2021, 9, 3200–3209 | 3201

View Article Online
Journal of Materials Chemistry B
Paper
Fig. 1 Synthetic routes to the azoreductase-responsive amphiphilic prodrug (a). 1H NMR of prodrug 4a (b) in CDCl3.
Synthesis of mPEG-Ts (compound 3)
evaporated and purified using a dextran gel column to obtain
the prodrug. Compound 4a, 1H NMR (400 MHz, CDCl3) d: 7.72–
7.84 (m, 6H,200,300,500,600,900,1100-H), 7.72–7.74 (8’’, 12’’-H), 6.80
(s, 1H, 5-H), 6.63–6.65 (d, 2H, J = 8.8 Hz, 800,1200-H), 6.54 (s, 1H,
8-H), 6.32 (s, 2H, 20,60-H), 5.96 (s, 1H, 13-H), 5.98 (s, 1H, 13-H),
5.29 (s, NH), 4.81 (m, 1H, 4-H), 4.61 (d, 1H, J = 4.6 Hz, 1-H), 4.42
(t, 1H, J = 7.7 Hz, 11-H), 4.23 (m, 4H, -NH-CH2CH2O- of PEG
chain), 3.46–4.01 (m, 11, 30,50-OCH3, –OCH2CH2O- units of PEG
chain), 3.36 (s, 3H, –OCH3 of PEG chain), 3.28–3.36 (dd, 1H, J =
4.6 Hz, 14.0 Hz, 2-H), 3.06 (m, 1H, 3-H); 13C NMR (100 MHz,
CDCl3) d: 174.6, 152.9, 150.3, 148.3, 147.8, 147.6, 146.5, 146.4,
144.8, 134.2, 132.0, 130.5, 130.1, 124.5, 124.4, 112.5, 112.1,
109.9, 109.3, 108.0, 101.6, 71.9, 71.3, 70.5, 70.3, 69.4, 68.8,
59.0, 56.5, 56.3, 56.0, 52.3, 43.4, 43.2, 42.7, 41.9, 38.6.
Compound 4b, 1H NMR (400MHz, CDCl3) d: 7.83 (m, 6H,
In brief, to a solution of monomethoxy polyethylene glycol
(mPEG, Mw = 1900, 2.63 mmol, 5.0 g) in CH2Cl2 (10 mL),
triethylamine (7.89 mmol, 1.11 mL) was added. After being
stirred under a N2 atmosphere for 5 min, p-toluenesulfonyl
chloride (5.26 mmol, 1.0 g) was added and allowed to react for
4 h. The crude product was purified using a dextran gel column
to obtain intermediate mPEG-Ts (4.5 g, compound 3). 1H NMR
(400 MHz, CDCl3) d: 7.75 (d, 2H, J = 8.2 Hz, 2, 6-H), 7.30 (d, 2H,
J = 8.0 Hz, 3, 5-H), 4.11–4.13 (m, 2H, 8-H), 3.50–3.77 (m,
9,10,11-H), 3.34 (s, 3H, H-12), 2.41 (s, 3H, H-7).
Synthesis of the azo-based podophyllotoxin prodrug
(compound 4)
The compound 2 (1 mmol) was dissolved in dry CH2Cl2 (10 mL), 200,300,500,600,900,1100-H), 7.75 (800,1200-H), 6.80 (s, 1H, 5-H), 6.64
and K2CO3 (3 mmol, 414 mg) was added under a N2 atmo- (d, 2H, 800, 1200-H), 6.54 (s, 1H, 8-H), 6.32 (s, 2H, 20,60-H), 5.98
sphere. 5 min later, mPEG-Ts (1.5 mmol, 3.1 g) was added and (s, 1H, 13-H), 5.98 (s, 1H, 13-H), 5.29 (s, NH), 4.81 (m, 1H, 4-H),
allowed to react for 24 h. The reaction solution was filtered, 4.62 (1H, 1-H), 4.42 (1H, 11-H), 4.26 (m, 4H, –NH–CH2CH2O- of
This journal is © The Royal Society of Chemistry 2021
3202 | J. Mater. Chem. B, 2021, 9, 3200–3209

View Article Online
Paper
Journal of Materials Chemistry B
PEG chain), 3.46–3.99 (m, 11, 30,40,50-OCH3, –OCH2CH2O- units PBS buffer, and then 10 mL of rat cecal content solution were
of PEG chain), 3.36 (s, 3H, –OCH3 of PEG chain), 3.13 (dd, 1H, added by bubbling nitrogen and incubated for 12 h at 37 1C.
2-H), 3.06 (m, 1H, 3-H); 13C NMR (100 MHz, CDCl3) d: 174.4, At predetermined time intervals, 2.0 mL of release medium
152.6, 149.4, 148.8, 148.4, 147.7, 146.2, 146.0, 145.5, 144.2, was removed and analysed by UPLC-MS/MS to determine the
137.3, 131.9, 129.8, 125.1, 123.3, 119.6, 114.9, 112.0, 110.0, concentrations of the AdP released from prodrug 4a.
109.2, 108.3, 101.6, 71.9, 70.5, 70.3, 68.9, 63.6, 60.7, 59.0,
In vitro cytotoxicity assays
56.3, 52.3, 43.6, 41.8, 38.6, 32.1.
AdP, prodrug 4a, 5-FU and a mixture of AdP/5-FU with 1 : 1 mole
ratio were used to evaluate the in vitro cytotoxicity against
HepG2 cells. Hep G2 cells were incubated with 0, 3.13, 6.25,
12.50, 25.00, 50.00 and 100.00 mM of test compounds for 72 h
and MTT assay was performed to determine the cell viability
according to the reported method.44 The CI values with the
combination use of AdP and 5-FU (ESI,† Tables S1 and S2) were
calculated by the computerized Chou-Talalay method.45,46
Degradation selectivity of the prodrug
To the solution of prodrug 4a in deoxygenated PBS buffer (pH =
7.4, 10 mM, 2 mL), 1 mL of PBS solution, proteinase K solution,
lipase solution (10 mg mLꢀ1), Na2S2O4 solution (10 mg mLꢀ1
)
and rat cecal content solution were added by bubbling
nitrogen, respectively. The mixture was incubated for 12 h at
37 1C and detected by HPLC or UPLC-MS/MS to monitor the
degradation of the prodrug and the release of AdP.
Preparation of a nanoprodrug based supramolecular hydrogel
Results and discussion
Synthesis and characterization of the nanoprodrug
The colon-targeted nanoprodrug was further used as a building
block to interact with a-CD for preparing a supramolecular
hydrogel according to our previously reported method.42 Podophyllotoxin natural product compounds 1a and b are the
The concentration of the prodrug was fixed at 20 mg mLꢀ1 main secondary metabolites isolated from the Podophyllum
and the concentration of a-CD ranged from 40 to 120 mg mLꢀ1
.
genus. As shown in Fig. 1a, these two readily available natural
The lyophilized supramolecular hydrogels were analysed by compounds were used as a starting compound and further
scanning electron microscopy (SEM) and X-ray diffraction connected with an azobenzene linker, 4,40-azodianiline, to
(XRD). The 5-FU loaded nanoprodrug hydrogel was obtained afford an important podophyllotoxin-derived intermediate
with the same approach just by mixing 5-FU (2 mg mLꢀ1) with (compounds 2) bearing an azo bond. Then, the target
the prodrug solution before introducing a-CD to the mixture.
azoreductase-responsive amphiphilic prodrug (prodrug 4) was
prepared by the reaction of compounds 2 with mPEG-Ts (Mn =
Imaging mass spectrometry (IMS) analysis
1900). The structures of the intermediate products were con-
1
Normal diet mice were sacrificed after gavage with the prodrug firmed using H and 13C NMR spectra (Fig. S1–S6, ESI†). The
(compound 4b) hydrogel (0.1 mL, 100 mg kgꢀ1) for 16 h; chemical structures of the prodrug were confirmed by 1H NMR
different organs were dissected and frozen at ꢀ80 1C. As a (see Fig. 1b for prodrug 4a and Fig. S7 (ESI†) for prodrug 4b)
control, cecal and colorectal tissues of mice were also dissected and 13C NMR (Fig. S8 and S9, ESI†). Taking prodrug 4a as an
after gavage with PBS (0.1 mL) at 16 h. The frozen tissues were example, besides the characteristic H signals that belong to the
fixed on a holder and sliced at 20 mm thickness at ꢀ20 1C using molecular framework of podophyllotoxin, the signals of
a cryomicrotome. The tissue sections were placed on ITO- H-200,300,500,600,800,900,1000, and 1100 belonging to the azobenzene
coated glass slides (sheet resistivity r5 O, Shenzhen, China). linker and H signals of a–g in the PEG chain can be clearly
The glass slides were dried and the surface was deposited with observed in 1H NMR spectra (Fig. 1b), which indicated that the
matrix CHCA using an iMLayer (Shimadzu, Kyoto, Japan).
ingeniously designed prodrug has been successfully synthesized.
IMS analysis was performed using an imaging mass spectro- It should be noted that colon tissue contains a large number of
meter (iMScope TRIO Shimadzu, Kyoto, Japan). Positive-ion azoreductases specifically expressed by colonic microflora,
detection mode (m/z range of 200–650) was used in MS which can selectively reduce the azobenzene linker to form the
and MS/MS analysis, respectively. The target compound was corresponding aniline group. Therefore, as a consequence of the
confirmed with selected fragment ion m/z 397.13 in MS/MS cleavage of the azobenzene linker in the prodrug reduced by
analysis.
azoreductase, one of the corresponding aniline groups will be
retained on the parent molecular structure to in situ form the
target compound AdP (Scheme 1).
In vitro drug release
A cuvette containing 1 mL of 5-FU loaded nanoprodrug hydrogel
Due to the poor water solubility of parent podophyllotoxin
was immersed in 30.0 mL of PBS (pH 7.4), PBS containing molecules, the resultant amphiphilic prodrug will aggregate in
10 mg mLꢀ1 Na2S2O4 or PBS containing 10 mg mLꢀ1 amylase water to form typical nanomicelles (Fig. 2a), in which the
in
a 37 1C water bath, respectively. At predetermined hydrophobic precursor molecule and hydrophilic PEG chains
time intervals, 5 mL of release medium was changed. The are located in the core and surface of nanomicelles via the azo
concentrations of 5-FU, prodrug 4a and AdP released from linkage. It should be noted that the loading efficiency of active
hydrogels were determined by HPLC. In the second stage, AdP in the nanoprodrug is constant, which is conducive to
4 mg of prodrug 4a was dissolved in 10 mL of deoxygenated regulating the proportion of the drug for combination therapy.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. B, 2021, 9, 3200–3209 | 3203

View Article Online
Journal of Materials Chemistry B
Paper
Fig. 2 The TEM images of prodrug 4 micelle solution before (a) and after
(b) introducing 10 mg mLꢀ1 Na2S2O4 to the solution. (c) Accumulative
release of AdP from prodrug 4a toward PBS, lipase, protease and Na2S2O4.
(d) The release behavior of AdP upon addition of different volumes of rat
cecal contents.
Fig. 3 (a) Photos of the supramolecular hydrogels made of prodrug
4
nanomicelles and different amounts of a-CD. For gels 1–5, the
The activation of AdP from the prodrug was investigated by
HPLC upon addition of an alternative to azoreductase (sodium
dithionite, Na2S2O4), and other common enzymes in the digestive
system, such as proteinase and lipase. Remarkable AdP release
was only observed when prodrug 4a reacted with Na2S2O4 while
no notable change in the distinct peak of the prodrug 4a was
observed in the presence of digestive enzymes (Fig. 2c and
Fig. S10, ESI†). Meanwhile, the TEM images of prodrug 4 nano-
micelle solution after introducing Na2S2O4 indicated that the
micelle structures were totally disrupted (Fig. 2b). Furthermore,
the activation of AdP from prodrug 4a was investigated by
UPLC-MS/MS upon addition of different amounts of rat cecal
contents. The amount of released AdP was also determined by
UPLC-MS/MS using the equation for the calibration curve
(Fig. S11, ESI†). When the volume of rat cecal content solution
was increased from 200 mL to 900 mL, the accumulated release of
AdP from prodrug 4a reached above 95% at an almost constant
rate (Fig. 2d). This result can be ascribed to the fact that there
were a large number of azoreductases present in rat cecal
contents.47–49 Therefore, we can conclude that the target AdP
molecules can be released in the colon from the proposed
nanoprodrug with high selectivity.
concentrations of a-CD are 40, 60, 80, 100 and 120 mg mLꢀ1, respectively.
(b–f) are SEM images corresponding to gels 1–5, respectively.
further aggregate and self-assemble into a supramolecular
hydrogel. The resulting supramolecular hydrogel not only has
the advantages of prodrug nanomicelles, but also increases the
space and dimensions of the micelle system during the progressive
self-assembly process, which makes it possible to load other
antitumor drugs, especially water-soluble antitumor drugs.
Therefore, to satisfy the demand of drug carriers for
combination therapy in clinics,27–29 different amounts of
a-CD were introduced into prodrug
4 micelle solution
(20 mg mLꢀ1). A turbid solution was observed when the
concentrations of a-CD ranged from 40 to 120 mg mLꢀ1, and
stable hydrogels can be obtained at a wide range of a-CD
concentrations (Fig. 3a), even as low as 60 mg mLꢀ1. The
threading and dethreading of a-CD from the PEG chain is a
dynamic equilibrium process in typical PPR formation. At
relatively lower concentration of a-CD, it tends to dethread
from both ends of the PEG chain side.52 In this PEGylated
prodrug nanomicelle system, the PEG brushes on the micellar
surface can provide only one chain end, which is beneficial for
retaining a-CD in the PPR structure. With increasing the
concentration of a-CD, the hydrogel formation time is
dramatically reduced. When the concentration is above
100 mg mLꢀ1, the hydrogel can be rapidly formed during the
ultrasonication process. The lyophilized samples were further
observed by SEM to investigate the progressive self-assembly
process after adding different amounts of a-CD (Fig. 3b–f).
With increasing the amount of a-CD, the inner morphology of
the prodrug/a-CD complex changes from typical lamellar
Progressive self-assembly of the prodrug micelles
Our previously reported work has proved that PEGylated
amphiphilic prodrug nanomicelles can be used as a building
block to further construct superstructures, such as supramolecular
hydrogels, with the introduction of a-CD.42,50,51 The strong
hydrogen bonding interaction among adjacent pseudopolyro-
taxanes (PPRs), formed between a-CD and PEG chains
distributed on the micellar surface, will provide a so-called
shell–shell coupling effect to drive the prodrug nanomicelles to
This journal is © The Royal Society of Chemistry 2021
3204 | J. Mater. Chem. B, 2021, 9, 3200–3209

View Article Online
Paper
Journal of Materials Chemistry B
aggregates to a unique morphology piled up with porous
microspheres. This can be interpreted as the strengthened
shell–shell interaction among prodrug nanomicelles with the
increased number of PPRs in this system. Lower concentration
of a-CD generated relatively few PPRs on the prodrug micellar
surface and can only provide weak shell–shell interaction,
resulting in a morphology similar to typical pure PPR supra-
molecular structures. However, at increasing concentrations
(above 80 mg mLꢀ1), sufficient PPRs could be formed and
retained onto the micellar surface. The strong hydrogen bonds
among adjacent PPRs provide an enhanced shell–shell inter-
action to induce PPR-modified prodrug micelle self-assembly
into porous microspheres. Finally, the interactions among
adjacent microspheres provide the necessary cross-links and
building blocks for supramolecular hydrogel formation
(Fig. 4a). All freeze-dried prodrug/a-CD complexes were further
analysed using complementary X-ray diffraction (XRD) to
confirm the formation of the PPR structure in self-assembled
hydrogels. In the patterns of all five gels (Fig. 4b), newly
generated diffraction peaks (around 2y = 19.91 and 22.61)
belonging to the (210) and (300) planes of crystalline columnar
a-CD that originated from a-CD PPRs were observed.53,54
Meanwhile, the peaks (2y = 19.11 and 23.21) belonging to the
prodrug disappeared in the pattern of gels 2–5, implying that
Fig. 5 (a) The cell viability of the prodrug hydrogel using MTT assay.
(b) Live/dead assay of Vero cells incubated with the prodrug/a-CD hydrogel
(1 mg mLꢀ1) for 24 h. All scale bars: 100 mm. Red and green colours
represent dead and live cells.
Fig. 4 (a) Schematic representation of the process of hydrogel formation.
(b) XRD patterns for freeze-dried gels 1–5, prodrug 4a and a-CD.
Fig. 6 Optical photos of the nanoprodrug 4a hydrogels before and after
introducing protease, lipase, amylase and sodium hydrosulfite.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. B, 2021, 9, 3200–3209 | 3205

View Article Online
Journal of Materials Chemistry B
Paper
most of the PEG chains participated in the PPR formation. The (Fig. 5a). The cell viability of both two cell lines against the
diffraction peak around 2y = 23.21 belonging to the prodrug prodrug hydrogel is higher than 90% after incubation for
was still found in the pattern of gel-1, which indicates that 3 days, even at a high concentration (1.0 mg mLꢀ1). To further
relatively few PPRs were formed at low concentration of a-CD. visualize the cell viability on the prodrug hydrogel, live/dead
These results confirm that the introduction of a-CD into assay was performed using Vero cells. As shown in Fig. 5b, after
PEGylated prodrug micelles can generate PPR suprastructures, incubation with both prodrug hydrogel and PBS (control) for
which effectively induce the aggregation of prodrug micelles 24 h (Fig. 5b), the fluorescence images showed almost no red
and make the formation of a hydrogel possible.
color in both groups. These results indicate that there are
It has been extensively proved that cyclodextrin PPR almost no cytotoxic effects of the PEGylated podophyllotoxin
hydrogels have significant biocompatibility, but the cytotoxicity nanoprodrug and its hydrogel form, which is beneficial for
of PPR hydrogels incorporated with an azo linked prodrug have colon-specific drug delivery.
not yet been verified. Human embryonic lung fibroblast cells
The response performance of the prodrug hydrogel was
(MRC-5) and kidney cells of African green monkey (Vero) were investigated upon addition of Na2S2O4, lipase, proteinase and
used to evaluate the biocompatibility by MTT assay first amylase. As shown in Fig. 6, the hydrogel form showed almost
Fig. 7 IMS of tissue sections at 16 h after gavage with the prodrug 4b hydrogel in mice. (a) Optical, IMS, and merged images of podophyllotoxin
derivatives in colon and cecal tissue. (b) Merged images of other organs.
This journal is © The Royal Society of Chemistry 2021
3206 | J. Mater. Chem. B, 2021, 9, 3200–3209

View Article Online
Paper
Journal of Materials Chemistry B
no change after incubation with lipase and proteinase for release of 5-FU, and a relatively small amount of prodrug 4a, but
30 min. However, it showed a typical gel–sol transition behavior no AdP can be detected during the entire release process (Fig. 8a).
after incubation with amylase and Na2S2O4. It could be ascribed This result indicated that the azo-linked prodrug and its hydrogel
to the fact that the degradation of a-CD in the presence of form are stable in a normal physiological environment. The release
amylase can destroy the PPR structure, resulting in the rate of both 5-FU and prodrug was significantly accelerated
dissociation of the supramolecular hydrogel further. It should (Fig. 8b) in the presence of amylase, which is mainly because of
be noted that the abundant colonic microflora can secrete the partial gel–sol transition of the prodrug hydrogel on the
corresponding enzymes, such as amylase, which are only active contact surface induced by the degradation of a-CD. Similarly,
in the colon on degradation of polysaccharides.55–57 Furthermore, no AdP was found in the presence of amylase. In the next stage, the
the breakage of the azo linkage in prodrug upon addition of released prodrug 4a can also continue to release active AdP due to
Na2S2O4 can destroy the micelle structure and supra-cross links, the breakage of the azo linkage in the presence of rat caecal
and lead to destruction of supramolecular hydrogel. Therefore, contents (Fig. 8c). When Na2S2O4 was introduced into the 5-FU
through the combination use of amylase and azoreductase which loaded hydrogel, the release of both 5-FU and AdP was observed,
are specifically secreted by colonic microflora, the proposed resulting from the disruption of prodrug nanomicelles (Fig. 8d).
prodrug hydrogel shows significant potential in co-delivering Meanwhile, the dramatically accelerated release of 5-FU is also
active natural products and other water soluble drugs to the colon directly attributable to the disruption of prodrug nanomicelles
for combination therapy purpose.
which is essential for maintaining the structural stability of the
Although the enzyme-responsive drug release and gel–sol hydrogel. These results suggest that the enzymes specifically
transition behavior of the azo linked prodrug and its hydrogel secreted by the colon can effectively accelerate the release rate of
form have been proved in vitro, the colon-specific delivery of 5-FU and targeted delivery of AdP to the colon from the prodrug
podophyllotoxin natural products in vivo still needs to be hydrogel, which is beneficial for combination therapy in the
further verified. Therefore, IMS was performed in mouse tissue treatment of colon-related diseases.
after intragastric administration of the prodrug 4b hydrogel to
Finally, the cytotoxicity of AdP, 5-FU and the combination of
map the biodistribution of released molecules precisely. As a 5-FU and AdP against Hep G2 cells was investigated using MTT
control, IMS analysis was also carried out after intragastric assay to evaluate the combination effect of AdP and 5-FU.
administration of the same volume of PBS. In the first order In addition, the cytotoxicity of prodrug 4a was also investigated.
spectra of molecules released from prodrug 4b after the As shown in Fig. S14 (ESI†), the prodrug exhibited weak
reduction of the azo bond, the quasimolecular ion at m/z cytotoxic activity against normal cell lines, normal human liver
505.19 [M + H]+ (Fig. S12b, ESI†) afforded a characteristic L-02 and mouse myoblast C2C12 cells. Moreover, prodrug 4a
fragment ion [M-ursol + H]+ in MS/MS experiments (Fig. S12c, presented less cytotoxicity than AdP for Hep G2 cells in the
ESI†), which corresponded to the loss of ursol from the ion entire range of concentrations (Fig. 9a), indicating that the
[M + H]+ under a H transfer mechanism (Fig. S12a, ESI†).58,59 resultant prodrug can effectively release AdP to the colon on
Therefore, the quasimolecular ion, m/z 505.19, was selected as the basis of reducing its toxicity in the delivery process.
the precursor ion in MS/MS analysis and the fragment ion m/z Furthermore, the combination use of AdP and 5-FU showed
397.13 was chosen as the diagnostic ion for evaluating the enhanced cytotoxicity compared to single drug treatment.
biodistribution of released podophyllotoxin derivatives in dif-
ferent tissue sections. Fig. 7 presents the biodistribution of
podophyllotoxin derivatives (MS/MS fragmentation, m/z 397.13)
from the prodrug 4b hydrogel in different organs. As seen in
Fig. 7a, the diagnostic ion belonging to target compounds
(m/z 397.13) was observed clearly in colon and cecal tissue
sections. However, in the control group, few characteristic MS
signals were found in colon and cecal tissue sections (Fig. S13,
ESI†). Meanwhile, in other organs, almost no characteristic MS
signals were found (Fig. 7b). These results presented in vivo
evidence at the molecular level that the prodrug and its hydrogel
form possess excellent ability to release podophyllotoxin
derivatives in the colon in a targeted manner.
Encouraged by the above results, the prodrug hydrogel was
further used as a carrier to load another hydrophilic anti-cancer
drug during the progressive self-assembly process. 5-FU is a
typical water-soluble drug with
a
different anticancer
mechanism from the podophyllotoxin drug. The release behavior
of 5-FU, the prodrug and AdP from the hydrogel was investigated
under different conditions. In a normal physiological environment
(PBS, pH 7.4), the 5-FU loaded hydrogel presented a sustained
Fig. 8 Release kinetics of 5-FU, prodrug 4a and AdP from the 5-FU
loaded prodrug hydrogel in PBS (a), in the presence of 10 mg mLꢀ1
amylase (b) or Na2S2O4 (d). (c) AdP release kinetics from prodrug 4a in
the presence of rat cecal contents.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. B, 2021, 9, 3200–3209 | 3207

View Article Online
Journal of Materials Chemistry B
Paper
hydrogel. The capability to deliver active podophyllotoxin
derivatives to the colon based on the resultant nanoprodrug
and its hydrogel form has been successfully demonstrated by
investigating the release behavior in the presence of a chemical
mimic of azoreductase, digestive enzymes, rat cecal contents,
and the biodistribution of podophyllotoxin derivatives in mouse
tissue was also precisely mapped at the molecular level using
IMS. Furthermore, the hydrogel form of the podophyllotoxin-
derived prodrug provides an effective platform to load another
water-soluble anticancer drug, 5-FU, for combination therapy.
In vitro assays demonstrated that the combination use of 5-FU
and podophyllotoxin derivatives produces significant synergetic
cytotoxicity to cancer cells. We believe that the colon-targeted
prodrug obtained from natural products and its hydrogel form
will not only provide an attractive strategy for reconsideration of
bioactive natural products and their derivatives, but also provide
an effective platform for co-delivery of multiple drugs for drug
combination. However, for colon tumor site-specific delivery of
these anticancer natural products, the ligand-mediated targeted
strategy needs to be considered further in a tumor-bearing
animal model.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (No. 21974145), and the Scholar Program
of West Light Project, Youth Innovation Promotion Association
of the Chinese Academy of Sciences (2017457).
Fig. 9 (a) In vitro cytotoxicity results against Hep G2 cells determined by
MTT assay. (b) CI value versus fa.
The combination index (CI) of AdP and 5-FU was further
calculated according to the Chou-Talalay method (Tables S1
and S2, ESI†).45,46 As seen in Fig. 9b, the CI was o1 for Hep G2
References
cells at
Z fa20, demonstrating a significant synergistic
1 T. Rodrigues, D. Reker, P. Schneider and G. Schneider, Nat.
Chem., 2016, 8, 531.
2 A. L. Harvey, R. A. Edrada-Ebel and R. J. Quinn, Nat. Rev.
Drug Discovery, 2015, 14, 111.
interaction between the two drugs. These results implied that
the proposed prodrug can effectively reduce the toxicity of AdP
and its hydrogel form provides a possibility to enhance the
therapeutic effect of AdP by loading another drug.
3 B. A. Shen, Cell, 2015, 163, 1297.
4 E. Patridge, P. Gareiss, M. S. Kinch and D. Hoyer, Drug
Discovery Today, 2015, 21, 204.
5 J. C. Chen, W. L. Li, H. Q. Yao and J. Y. Xu, Fitoterapia, 2015,
103, 231.
6 D. C. Blakemore, L. Castro, I. Churcher, D. C. Rees,
A. W. Thomas, D. M. Wilson and A. Wood, Nat. Chem.,
2018, 10, 383.
7 G. Li and H. X. Lou, Med. Res. Rev., 2018, 38, 1255.
8 Y. Q. Liu, J. Tian, K. Qian, X. B. Zhao, S. L. Morris-Natschke,
L. Yang, X. Nan, X. Tian and K. H. Lee, Med. Res. Rev., 2015,
35, 1.
Conclusions
In conclusion, taking an important anticancer natural product
podophyllotoxin as a proof of concept, a new-type of azo-based,
amphiphilic, and colon-targeted podophyllotoxin-derived
nanoprodrug was designed and synthesized. The readily
available podophyllotoxin natural products were used as starting
compounds to connect with PEG chains via a multifunctional
azobenzene group, and highly selectively activated by azoreduc-
tase, which resulted in the generation of active podophyllotoxin
derivatives in the colon through an ingenious in situ
modification strategy. Subsequently, host–guest inclusion based
self-assembly was performed to generate a supramolecular
9 X. Zhang, K. P. Rakesh, C. S. Shantharam,
H. M. Manukumar, A. M. Asiri, H. M. Marwani and
H. L. Qin, Bioorg. Med. Chem., 2018, 26, 340.
This journal is © The Royal Society of Chemistry 2021
3208 | J. Mater. Chem. B, 2021, 9, 3200–3209

View Article Online
Paper
Journal of Materials Chemistry B
10 Y. Mehellou, H. S. Rattan and J. Balzarini, J. Med. Chem., 34 Y. Gao, L. Jia, Q. Wang, H. Hu, X. Zhao, D. Chen and
2017, 61, 2211. M. Qiao, ACS Appl. Mater. Interfaces, 2019, 11, 16296.
11 J. Rautio, H. Kumpulainen, T. Heimbach, R. Oliyai, D. Oh, 35 R. Wang, Z. Zhao, Y. Han, S. Hu, Y. Opoku-Damoah, J. Zhou,
¨
T. Jarvinen and J. Savolainen, Nat. Rev. Drug Discovery, 2008,
L. Yin and Y. Ding, Mol. Pharmaceutics, 2017, 14, 2999.
7, 255.
36 H. Xu, M. Lv and X. Tian, Curr. Med. Chem., 2009, 16, 327.
12 S. Jana, S. Mandlekar and P. Marathe, Curr. Med. Chem., 37 Y. J. You, Curr. Pharm. Des., 2005, 11, 1695.
2010, 17, 3874.
38 M. Gordaliza, P. A. Garcia, J. M. Miguel Del Corral,
M. A. Castro and M. A. Gomez-Zurita, Toxicon, 2004, 44, 441.
39 Z. Q. Wang, Y. H. Kuo, D. Schnur, J. P. Bowen, S. Y. Liu,
F. S. Han, J. Y. Chang, Y. C. Cheng and K. H. Lee, J. Med.
Chem., 1990, 33, 2660.
13 V. Abet, F. Filace, J. Recio, J. Alvarez-Builla and C. Burgos,
Eur. J. Med. Chem., 2017, 127, 810.
14 Y. Q. Zheng, X. G. Ji, B. C. Yu, K. L. Ji, D. Gallo, E. Csizmadia,
M. Y. Zhu, M. R. Choudhury, L. K. Cruz, V. Chittavong,
Z. X. Pan, Z. N. Yuan, L. E. Otterbein and B. H. Wang, Nat. 40 T. S. Huang, C. C. Lee, Y. Chao, C. H. Shu, L. T. Chen,
Chem., 2018, 10, 787.
15 S. S. Dhaneshwar and G. Vadnerkar, Curr. Top. Med. Chem.,
2011, 11, 2318.
16 M. Roldo, E. Barbu, J. F. Brown, D. W. Laight, J. D. Smart
and J. Tsibouklis, Expert Opin. Drug Delivery, 2007, 4, 547.
17 S. H. Lee, E. Moroz, B. Castagner and J. C. Leroux, J. Am.
Chem. Soc., 2014, 136, 12868.
18 J. Rao, C. Hottinger and A. Khan, J. Am. Chem. Soc., 2014,
136, 5872.
L. L. Chen, M. H. Chen, C. C. Yuan and J. Whang-Peng,
Pharm. Res., 1999, 16, 997.
41 J. M. Liu, L. T. Chen, Y. Chao, A. F. Y. Li, C. W. Wu, T. S. Liu,
H. S. Shiah, J. Y. Chang, J. D. Chen, H. W. Wu, W. C. Lin,
C. Lan and J. Whang-Peng, Cancer Chemother. Pharmacol.,
2002, 49, 425.
42 W. Ha, X. B. Zhao, X. Y. Chen, K. Jiang and Y. P. Shi, ACS
Appl. Mater. Interfaces, 2018, 10, 21149.
43 K. H. Lee, S. A. Beers, M. Mori, Z. Q. Wang, Y. H. Kuo, L. Li,
S. Y. Liu, J. Y. Chang, F. S. Han and Y. C. Cheng, J. Med.
Chem., 1990, 33, 1364.
19 L. Zhang, Y. Wang, X. B. Zhang, X. Wei, X. Xiong and
S. B. Zhou, ACS Appl. Mater. Interfaces, 2017, 9, 3388.
20 X. M. Li, J. Y. Li, Y. Gao, Y. Kuang, J. Shi and B. Xu, J. Am. 44 I. Mickuviene, V. Kirveliene and J. B. Uodka, Toxicol.
Chem. Soc., 2010, 132, 17707. In Vitro, 2004, 18, 639.
21 O. H. Nielsen and L. K. Munck, Nat. Rev. Gastroenterol. 45 T. C. Chou and P. Talalay, Adv. Enzyme Regul., 1984, 22, 27.
Hepatol., 2007, 4, 160.
46 A. Khafif, S. P. Schantz, T. C. Chou, D. Edelstein and
22 R. B. Muijsers and K. L. Goa, Drugs, 2002, 62, 1689.
P. G. Sachs, Carcinogenesis, 1998, 19, 419.
23 I. B. Seiple, Z. Zhang, P. Jakubec, A. Langlois-Mercier, 47 X. B. Zhao, W. Ha, K. Gao and Y. P. Shi, Anal. Chem., 2020,
P. M. Wright, D. T. Hog, K. Yabu, S. R. Allu, T. Fukuzaki, 92, 9039.
P. N. Calsen, Y. Kitamura, X. Zhou, M. L. Condakes, 48 S. Saphier, A. Haft and S. Margel, J. Med. Chem., 2012,
´
F. T. Szczypinski, W. D. Green and A. G. Myers, Nature,
55, 10781.
2016, 533, 338.
49 W. Kim, J. Nam, S. Lee, S. Jeong and Y. Jung, Mol. Pharma-
ceutics, 2016, 13, 2126.
24 C. F. Martinez-Farina, A. W. Robertson, H. Yin, S. Monro,
S. A. McFarland, R. T. Syvitski and D. L. Jakeman, J. Nat. 50 W. Ha, J. Yu, X. Y. Song, Z. J. Zhang, Y. Q. Liu and Y. P. Shi,
Prod., 2015, 78, 1208. J. Mater. Chem. B, 2013, 1, 5532.
25 E. Fernandez-alvaro, W. D. Hong, G. L. Nixon, P. M. O’Neill 51 W. Ha, J. Yu, X. Y. Song, J. Chen and Y. P. Shi, ACS Appl.
´
´
´
and F. Calderon, J. Med. Chem., 2016, 59, 5587.
Mater. Interfaces, 2014, 6, 10623.
26 M. Lamblin, R. Spingarn, T. T. Wang, M. C. Burger, 52 M. Ceccato, P. L. Nostro and P. Baglioni, Langmuir, 1997,
B. Dabbas, N. Moitessier, J. H. White and J. L. Gleason,
J. Med. Chem., 2010, 53, 7461.
27 A. Petrelli and S. Giordano, Curr. Med. Chem., 2008, 15, 422.
28 B. Al-Lazikani, U. Banerji and P. Workman, Nat. Biotechnol.,
2012, 30, 679.
13, 2436.
53 I. N. Topchieva, A. E. Tonelli, I. G. Panova, E. V. Matuchina,
F. A. Kalashnikov, V. I. Gerasimov, C. C. Rusa, M. Rusa and
M. A. Hunt, Langmuir, 2004, 20, 9036.
54 X. Song, Z. Zhang, J. Zhu, Y. Wen, F. Zhao, L. Lei, N. Phan-
Thien, B. C. Khoo and J. Li, Biomacromolecules, 2020,
21, 1516.
29 X. Dong and R. J. Mumper, Nanomedicine, 2010, 5, 597.
30 N. Kolishetti, S. Dhar, P. M. Valencia, L. Q. Lin, R. Karnik,
S. J. Lippard, R. Langer and O. C. Farokhzad, Proc. Natl. 55 E. A. Gu¨nter and O. V. Popeyko, Carbohydr. Polym., 2016,
Acad. Sci. U. S. A., 2010, 107, 17939. 147, 490.
31 T. M. Sun, J. Z. Du, Y. D. Yao, C. Q. Mao, S. Dou, S. Y. Huang, 56 P. Wen, K. Feng, H. Yang, X. Huang, M. H. Zong, W. Y. Lou,
P. Z. Zhang, K. W. Leong, E. W. Song and J. Wang, ACS
Nano, 2011, 5, 1483.
32 X. Guan, Y. Li, Z. Jiao, L. Lin, J. Chen, Z. Guo, H. Tian and
X. Chen, ACS Appl. Mater. Interfaces, 2015, 7, 3207.
33 T. Li, X. Shen, Y. Geng, Z. Chen, L. Li, S. Li, H. Yang, C. Wu,
N. Li and H. Wu, Carbohydr. Polym., 2017, 169, 157.
57 N. G. Kotla, M. Gulati, S. K. Singh and A. Shivapooja,
J. Controlled Release, 2014, 178, 55.
58 L. Zhao, X. Tian, P. C. Fan, Y. J. Zhan, D. W. Shen and Y. Jin,
J. Chromatogr. A, 2008, 1210, 168.
H. Zeng and Y. Liu, ACS Appl. Mater. Interfaces, 2016, 59 L. Zhao, Z. L. Liu, P. C. Fan, Z. W. Zhang, X. Liu, Y. J. Zhan
8, 13748.
and X. Tian, J. Sep. Sci., 2009, 1323.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. B, 2021, 9, 3200–3209 | 3209