A stimuli-responsive insulin delivery system based on reversible phenylboronate modified cyclodextrin with glucose triggered host-guest interaction

Article abstract of DOI:10.1016/j.ijpharm.2018.07.020

Injection of insulin is an effective therapy to treat most patients with the type I diabetes and some with type II diabetes. Additionally, the release of insulin under specific conditions has attracted widespread interest. In this study, a smart drug carr

Full text of DOI:10.1016/j.ijpharm.2018.07.020

Accepted Manuscript
A stimuli-responsive insulin delivery system based on reversible phenylboro-
nate modified cyclodextrin with glucose triggered host-guest interaction
Xinyuan Xu, Hui Shang, Tianyu Zhang, Panjing Shu, Yanpeng Liu, Jing Xie,
Dongyue Zhang, Hong Tan, Jianshu Li
PII:
S0378-5173(18)30488-5
https://doi.org/10.1016/j.ijpharm.2018.07.020
IJP 17636
DOI:
Reference:
International Journal of Pharmaceutics
To appear in:
Received Date:
Revised Date:
Accepted Date:
21 March 2018
19 June 2018
4 July 2018
Please cite this article as: X. Xu, H. Shang, T. Zhang, P. Shu, Y. Liu, J. Xie, D. Zhang, H. Tan, J. Li, A stimuli-
responsive insulin delivery system based on reversible phenylboronate modified cyclodextrin with glucose triggered
host-guest interaction, International Journal of Pharmaceutics (2018), doi: https://doi.org/10.1016/j.ijpharm.
2018.07.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1
2
A stimuli-responsive insulin delivery system based on reversible phenylboronate modified
cyclodextrin with glucose triggered host-guest interaction
3
4
Xinyuan Xu,# Hui Shang,# Tianyu Zhang, Panjing Shu, Yanpeng Liu, Jing Xie, Dongyue Zhang, Hong Tan, Jianshu Li*
5
6
College of Polymer Science and Engineering, State Key Laboratory of Polymer Materials Engineering, Sichuan University, Chengdu,
610065, China.
7
8
* Corresponding author. E-mail: jianshu_li@scu.edu.cn; Tel: +86-28-85466755.
# Authors contributed equally to this work.
9
10 Abstract
11
12
13
14
15
16
17
18
19
20
21
22
23
Injection of insulin is an effective therapy to treat most patients with the type I diabetes and some with
type II diabetes. Additionally, the release of insulin under specific conditions has attracted widespread
interest. In this study, a smart drug carrier that can release insulin depending on the changes in blood
glucose levels was designed. Combining two popular molecules through facile synthetic processes, a drug
carrier of reversible phenylboronate group modified cyclodextrin (β-CD-EPDME) was fabricated. The
drug carrier is composed of cyclodextrin, which can encapsulate insulin, and phenylboronate, which is
sensitive to the cis-diols in some saccharides. Moreover, β-CD-EPDME can successfully encapsulate
insulin and almost completely release insulin in the presence of glucose. The detached phenylboronic acid
moiety triggered by glucose can attack the β-CD cavity and form a host-guest complex, which can force
out the encapsulated insulin within the cavity. In addition, the insulin released from the β-CD-
EPDME@Insulin complex retains its secondary structure, and the drug carrier has been proven to have
low cytotoxicity. Thus, this safe and glucose-responsive drug carrier shows the potential for use in the
therapy of diabetes.
24
25
Keywords: insulin, diabetics mellitus, glucose-sensitive, cyclodextrin, phenylboronic acid.

1
1. Introduction
2
3
Diabetes mellitus (DM), a kind of common chronic non-communicable metabolic disease, can
be attributed either to the pancreas not producing enough insulin or the cells of the body not
responding adequately to the insulin produced (American Diabetes, 2014). DM is mainly divided
into several categories based on the pathogenesis, including type I DM, type II DM, and
gestational DM. Type I DM, characterized by an absolute deficiency of insulin secretion, has a
very low rate of cure. Furthermore, DM may cause syndromes that damage the eyes, kidneys,
cardiovascular, and nervous systems, and can even be fatal in extreme cases (Chen et al., 2011).
Hence, this disease has become one of the world's most urgent health issues that needs
investigation in the 21st century. Currently, the therapy for the patients with type I DM is to inject
insulin where there is an increase in the blood glucose. This may cause severe discomfort to
patients, among other issues. (Sajeesh and Sharma, 2006). Therefore, the development of a
responsive insulin delivery system is of vital importance (Huang et al., 2010; Li et al., 2016;
Matuszewska et al., 2015; Xie, Li et al., 2017).
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
In the past decades, the cyclodextrins (CDs) and their derivatives have been incorporated into
novel delivery systems such as liposomes, microspheres, microcapsules, hydrogels, and nano-
sponges, in order to enhance the encapsulated drugs’ physicochemical and pharmaceutical
properties (Vyas et al., 2008). The CDs are classified into α-, β-, and γ-CDs by the number of
glucopyranose units, and their inner cavities can form inclusion complexes with hydrophobic
guest molecules (Xin et al., 2010). The β-CDs are truncated cone-shaped oligosaccharides which
have the ability to envelope different proteins due to their cavity structure (Cao et al., 2004;
Szejtli, 1998). There have been numerous studies on the insulin delivery systems based on CD
molecules, such as insulin–CD complex in PLGA microspheres (Aguiar et al., 2004), β-CD
grafted chitosan (Daimon et al., 2014), and CD complexed insulin encapsulated hydrogel
microparticles (Sajeesh et al., 2010), etc.
2

1
2
Meanwhile, some other reports indicate that stimuli-responsive drug delivery systems can
efficiently release drugs at the target disease sites and/or in special pathological environments
(Gao et al., 2017; Yao et al., 2011). Hence, the environment of high blood glucose in the DM
patients has been used as the stimulus for insulin delivery. Currently, phenylboronic acid (PBA),
characterized by its highly glucose sensitivity, has aroused extensive attention (Chai et al., 2016)
because it can react with cis-diols in many saccharides to generate a reversible ester bond.
Nakamura et al. successfully synthesized phenylboronic acid-modified α- and β-CDs (PBA-
CyDs), which can transform between the supramolecular structure and the dissolved bond form at
different sugar concentrations (Nakamura et al., 2013). Additionally, the work mentioned that
PBA moiety of PBA-CyDs can act as guest molecules to react with the CD moiety of other PBA-
CyDs, which act as host molecules, in the absence of sugar, since the CD cavity has an affinity for
the phenyl group (Harada et al., 2009). On the other hand, in the presence of sugar, the
supramolecular structure will be inhibited because the phenyl group blocked by sugar can no
longer attack the cavity.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Different from the above studies, in this work, a reversible ester bond was firstly grafted on the
CDs molecules. Subsequently, the ester bond was broken down in the presence of glucoseto form
a dissociative phenyl group, which could attack the hydrophobic cavity (Scheme 1). Briefly, we
combined (5-ethyl-2-phenyl-1,3,2-dioxaborinan-5-yl) methyl 2-bromoacetate (EPDMB) with β-
CD via a bromination reaction to synthesize a glucose-responsive insulin carrier (β-CD-EPDME).
Next, we purified the carrier and mixed it with insulin to realize the drug encapsulation. The final
complex was named as β-CD-EPDME@Insulin. The structure of β-CD-EPDME@Insulin was
1
demonstrated by H-NMR, IR, and Time-of-flight mass spectra. The insulin-releasing capability
of this system was proved by fluorescence spectrophotometry. The results showed that
fluorescence intensity gradually decreased with an increase in the contacting time between β-CD-
EPDME@Insulin and glucose. The structure of the released-insulin was also illuminated by

1
2
3
4
circular dichroism spectra. Finally, the cytotoxicity assay provided evidence that both the carrier
and the complex had good cell viability in the measured range. Through this work, we have
provided insight into exploring a smart responsive insulin delivery system, which can release
insulin via a glucose triggered host-guest interaction.
5
6
2. Materials and methods
7
8
2.1. Materials
1,1,1-Trimethylolpropane (99.5%, TMP) and phenylboronic acid (PBA) were purchased from
J&K Scientific Ltd (Beijing, China). β-cyclodextrin (β-CD) and methylebnzene was purchased
from Keshi Chemical Reagent Ltd (Chengdu, China). Bromoacetyl bromide was purchased from
TCI Development corporation (Shanghai, China). Dichlormethane (DCM) was purchased from
Zhiyuan Chemical Reagent Ltd (Tianjin, China). NaH was purchased from Hainachuan scientific
and development Ltd (Tianjin, China). Triethylamine (Et3N, TEA) was purchased from Kermel
Chemical Reagent Ltd (Tianjin, China). Dimethyl Formamide (DMF) was purchased from Xilong
chemical engineering Ltd (Sichuan, China). All the materials are directly used except TEA. In the
synthesis process, TEA was distilled to remove any residual primary and secondary amines.
Briefly, the TEA was filtered through a Buchner funnel after mixing and stirring with potassium
hydroxide (KOH) for 12 h. Subsequently, TEA with a small amount of phthalic anhydride and 4A
molecular sieve were poured into a reflux device for 30 minutes. Finally, the pure TEA was
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
collected from the cut fraction at 87-89
℃ in atmospheric distillation unit.
2.2. Synthesis of β-CD-EPDME and preparation of β-CD-EPDME@Insulin
2.2.1. Synthesis of EPDMB
The group with the property of glucose sensitivity was synthesized by a two-step method
(Scheme 2). The first step was to form (5-ethyl-2-phenyl-1,3,2-dioxaborinan-5-yl) methanol
(EPDM) with a pentacyclic moiety utilizing the reaction between phenylboronic acid and cis-diol
4

1
2
(Yao et al., 2012). Typically, phenylboronic acid (48.4 mmol, 5.9 g) and 1,1,1-trimethylolpropane
(48.4 mmol, 6.5 g) were dissolved into 80 mL toluene, and magnetically stirred at 120 for 4 h in
the condensate return equipment. After evaporation at 60 , EPDM was obtained as a white solid
℃
3
℃
4
(9.0 g, 84.5%). The second step was a reaction of EPDM with bromoacetyl bromide to produce
EPDMB through a simple acylation reaction. Briefly, EPDM (20 mmol, 4.4 g) and triethylamine
(24 mmol, 3.4 mL) were dissolved in 25 mL DCM in a waterless and oxygen-less argon
atmosphere. After magnetic stirring for 30 min in an ice bath, the DCM solution of bromoacetyl
bromide (24 mmol, 2.1 mL) was added dropwise into a flask, following which, the solution was
5
6
7
8
9
magnetically stirred at 0 ℃ for 8 h. At the end of the reaction, the solution was vacuum filtered to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
obtain an orange-yellow and transparent liquid, which was further evaporated to remove DCM.
The EPDMB (3.6 g, 52%) as an orange, syrupy liquid was purified by column chromatography
with a solution of petroleum ether and ethyl acetate (the ratio is 8 to 2).
2.2.2. Synthesis of β-CD-EPDME
The insulin carrier (β-CD-EPDME) was synthesized based on a halogenation reaction (Scheme
3). First, β-CD was modified by sodium treatment according to a method used in previous studies
(Ding and D'Souza, 1990). Briefly, the β-CD (0.88 mmol, 1 g) dissolved in the DMF was added
dropwise into the DMF solution with NaH (2.92 μmol, 70 mg) Subsequently, the mixed solution
was magnetically stirred overnight at room temperature. After the product was vacuum filtered, a
transparent liquid (L1) was obtained, which was used in the following reaction. The above liquid,
EPDMB (1.76 mmol, 0.6 g) was dissolved in 5 mL DMF, followed by dropwise addition of L1.
After magnetic stirring at 80
℃ for 12h, the faint yellow liquid (L2) obtained was rotary
evaporated, recrystallized in acetone, filtered, and vacuum dried in sequence to obtain β-CD-
EPDME as a white solid.
2.2.3 Preparation of β-CD-EPDME@Insulin

1
2
A solution of insulin in phosphate buffer (pH 7.0, 0.44 mg/mL) was mixed with β-CD-EPDME
at different concentrations (0.5, 1.0, and 1.5 mg/mL) at room temperature and the mixed solutions
were magnetically stirred for 30 min. Subsequently, the solutions were kept for another hour at
room temperature. Some insulin aggregates formed in the oversaturated solutions werer separated
3
4
5
by centrifugation. Finally, the supernatant was freeze-dried to obtain the
complex in the form of a white solid.
β-CD-EPDME@Insulin
6
7
2.3. Characterizations
1
8
To demonstrate the structure of the obtained β-CD-EPDME, H NMR, IR spectroscopy, and
1
9
time-of-flight mass spectrometry (TOFMS) were utilized in this work. The H NMR spectroscopy
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
of EPDMB was performed using CDCl3 as solvent and the spectra were recorded at Advanced II-
400 MHz (Bruker Biospin Corporation, USA). All the molecules except EPDMB were mixed
with KBr powder and pressed into pellets, which were characterized by FT-IR spectroscopy
(Spectrum One, PerkinElmer, USA). The transmission IR spectrum of EPDMB was recorded by
the KBr smear method because of substance’s state. TOFMS (Bruker autoflex III smartbeam mass
spectrometer, USA) provided further information about the molecular weight of β-CD-EPDME.
The matrix was α-cyan-4-hydroxycinnamic acid (CHCA), which was dissolved in the THF at a
concentration of 10 mg/mL.
The encapsulation of β-CD-EPDME for insulin was mainly demonstrated by IR and fluorescence
spectroscopy. Briefly, the insulin was diluted in PBS buffer to a concentration of 0.44 mg/mL, which was
used as the control substance of the fluorescence assay. Then the β-CD-EPDME was added to the insulin
solution to prepare the β-CD-EPDME@Insulin solutions of various concentrations (0.5, 1.5, and 2.5
mg/mL). The natural insulin and lyophilized β-CD-EPDME@Insulin were separately mixed with KBr
powder and pressed into pellets, which were characterized by FT-IR spectroscopy (Spectrum One,
PerkinElmer, USA).
6

1
2
The stoichiometry between β-CD-EPDME and insulin was determined by continuous variation method
(Job plot) (Dotsikas et al., 2000; Job, 1928). Briefly, the difference in fluorescence intensity (ΔF=F-F0)
was calculated by measuring the intensity of insulin in the absence (F0) and presence (F) of β-CD-
EPDME. The molar fraction of insulin (γ) varied from 0.1 to 0.9, while the total concentration of β-CD-
EPDME ([β-CD-EPDME]) and insulin ([Insulin]) keep constant (M=1×10-4 mol/L, [β-CD-
EPDME]t+[Insulin]t=M, γ=[Insulin]t/M). Then the value of ΔF*[Insulin]t was plotted against the molar
fraction of insulin (γ). The final stoichiometry was obtained in the value γ with the maximum of
ΔF*[Insulin]t.
3
4
5
6
7
8
9
2.4. Insulin release assays in vitro
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
2.4.1. Fluorescence spectroscopic investigation
The interaction between β-CD-EPDME and glucose was analyzed by fluorospectrophotometry (RF-
5301PC, Daojin, China). The insulin solution (0.44 mg/mL) in PBS buffer (pH=7.0) was magnetically
stirred for 2 h, and the solution was used to record the fluorescence intensity. As mentioned above, the β-
CD-EPDME@Insulin obtained from a solution of 0.5 mg/mL β-CD-EPDME and glucose (1.5 mg/mL)
was adequately mixed and the fluorescence intensity of the liquid was recorded at different time points.
The in vitro release was evaluated by detecting the fluorescence intensity of released insulin
permeated through the dialysis membrane (MW=12000-14000). 9 mg β-CD-EPDME@Insulin
was dissolved in 6 mL PBS buffer (pH=7.4), then the glass containers with 3 mL complex
solution were putted into bigger bottles with glucose PBS buffer and pure PBS buffer,
respectively. At different time points, 3 mL samples were withdrawn to detect the fluorescence
intensity. The 3 mL related buffer was added into the bottles to maintain the total volume.
2.4.2. Circular dichroism
Circular Dichroism spectroscopy (JASCO-1500-150, Easton, MD) was used to measure the
secondary structure change of natural insulin and the insulin released from the complex β-CD-
EPDME@Insulin under the stimulation of glucose. The natural insulin was dissolved in the PBS

1
2
buffer (pH=7.0) at a concentration of 0.44 mg/mL, and the solution was stirred completely for 2h.
The circular dichroism (CD) assay of the released insulin was conducted as follows. After stirring
the complex (0.5 mg/mL) and glucose (1.5 mg/mL) for 2 h, the liquid was centrifuged and the
precipitate was freeze-dried to detect its secondary structure. The start of the measurement was
considered as t = 0. Spectra were recorded at wavelengths between 195 and 260 nm at a resolution
of 0.5 nm and a scanning speed of 10 nm/s.
3
4
5
6
7
2.4.3 Determination of binding constant
8
The binding constants of different substances with β-CD were calculated via the ultraviolet absorption
(UV-1800PC Spectrophotometer, China) of various concentration solutions (Daimon et al., 2014). A
Benesi-Hildebrand type equation (Agbaria et al., 1996) was employed:
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
where the A, A0 and ɑ means the ultraviolet absorption in the presence and absence of β-CD (1.14
mg/mL), and a constant, respectively. The three substances include insulin, β-CD-EPDME and mixture of
β-CD-EPDME and glucose. The mixture of β-CD-EPDME and glucose was used to simulate the glucose-
modified phenyl group.
For all the fluorescence test, the laser excitation wavelength was set at 280 nm and the peak
value or spectra were recorded between the emission wavelength from 290 to 350nm.
2.5. CCK-8 Cytotoxicity
The cytotoxicity of the samples was determined using the Cell Counting Kit-8 (CCK-8) assay
to detect the viability of L929 fibroblasts. CCK-8 is widely used in cytotoxicity studies to
measure the water-soluble formazan product, which is reduced by dehydrogenase existing in
viable cells. Hence, the number of living cells is direct proportionate to the amount of formazan
product. The L929 cells (passage 3) were suspended in a growth medium (DMEM, High Glucose
containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/mL streptomycin) and 100
μL cell suspension (at a concentration of 5×104/mL) was seeded in 96-well cell culture plates,
8

1
2
3
4
5
6
7
8
9
then the 96-well was cultivated in a humidified 5% CO2 atmosphere at 37 ℃ for 24 h. The
confluence reached 80% before treatment. The growth medium was then replaced by 100 μL of
fresh medium with various concentrations (0, 0.075, 0.125, 0.25, 0.5, and 1 mg/mL) of samples.
After another 24 h of incubation, 10 μL of CCK-8 was added to each well. In addition, 10 μL of
CCK-8 was added to 100 μL of pure growth medium with samples of various concentrations (0,
0.075, 0.125, 0.25, 0.5, and 1 mg/ml) in order to eliminate the effect of sample to CCK-8. The
plates were cultured in a CO2 incubator for 3 h. The optical density (OD) was measured at 450nm
using a microplate reader (KHB ST-360, China). The cell viability was calculated from the
following equation:
10
11 3. Results and discussion
12
13
14
15
16
17
18
19
20
21
22
23
3.1. The synthesis of β-CD-EPDME
3.1.1 EPDMB
The preparation of β-CD-EPDME@Insulin included the synthesis of a drug carrier and the
1
encapsulation of insulin. As showed in Figure 1, the H-NMR spectrum of EPDMB exhibited
integrals matching the theoretical values of the predicted molecules, as well as the peaks of
similar molecules in previous work (Yao et al., 2011). Chemical shifts between 7.80-7.20 mainly
indicated the presence of benzene rings, and the the newly generated boric acid ester
(CH2OBOCH2), which was expressed in the peaks 5 - 9, especially peaks 5 and 6. Additionally,
the peak at δ4.20 ppm correlated with the H in the CH2Br group, the integral of which is
apparently two times of that of the peak at δ 7.44 ppm. Moreover, the FT-IR spectrum (Figure 2)
1
of EPDMB corresponded with the result of H NMR. Compared with the reactive substances,
EPDMB had a new ester bond and an existing original phenylboronic ester bond, with

1
2
characteristic peaks of νC=O 1735.7 cm-1 and the νB-O 1317.4 cm-1, respectively. The other
characteristic peaks included νC-H 2971.9 cm-1, νB-Ar 1444.5 cm-1, νC-O 1160.9 cm-1, and νC-Br 696.7
cm-1. All the analyses provided the evidence of the successful synthesis of the glucose-sensitive
molecule (EPDMB).
3
4
5
3.1.2. β-CD-EPDME
6
The reaction between EPDMB and β-CD was designed according to a previous work by Ding
Rong et al (Ding and D'Souza, 1990). The ionization of the hydroxyl group enhanced the reaction
efficiency of CD with an electrophilic agent through the addition of sodium hydride (NaH).
Compared with the IR curve of β-CD, the IR spectrum of β-CD-EPDME showed an adsorption
peak at 1741.5 cm-1, which demonstrated the existence of an ester bond (Figure 2). Moreover, the
difference of a high adsorption peak at the 3335 cm-1 between β-CD and β-CD-EPDME also
proved the reaction between β-CD and EPDMEB. The very obvious hydroxyl characteristic peak
was related to the association state of hydroxyl group. The peak in β-CD was wider and stronger
than that in β-CD-EPDME, which proved that the hydroxyl group in β-CD-EPDME was in the
free state. The result was attributed to the decrease of a hydroxyl group and the decline of
hydrogen bond in β-CD after reacting with EPDMB. Therefore, this data proved the successful
preparation of β-CD-EPDME.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
To better confirm the degree of substitution in β-CD-EPDME, TOFMS was used to detect the
molecular weight (MW) of β-CD-EPDME. Theoretically, the MW of β-CD is 1145 and the
expectant molecular weight of β-CD-EPDME is 1395 (mono-) and 1655 (bi-). As shown in Figure
3, the peak in 1394.562 and 1654.671 m/z explained the presence of mono- and bi-
functionalization. Additionally, the intensity of the 1394.562 peak was much higher than that of
the 1654.671 peak, Hence, the synthesized β-CD-EPDM mainly exhibited mono-substituted
molecules with a few bi-substituted molecules.
3.2. The construction of the β-CD-EPDME@Insulin complex
10

1
2
The FT-IR spectrum was used to confirm the encapsulation capability of β-CD-EPDME to
insulin (Sajeesh and Sharma, 2006). The β-CD-EPDME and natural insulin were chosen as
reference materials to compare with the complex. The IR curve of β-CD-EPDME and β-CD-
EPDME@Insulin was almost the same, except the peak at 1535 cm-1. This peak is one that is
characteristic in natural insulin, in addition to the amide I at 1654 cm-1. Compared with the
spectra of the two reference substances, the special peak seen in the complex provided the
primary evidence to explain the presence of β-CD-EPDME@Insulin (Figure 4). In addition, more
powerful evidence about the construction of a complex would be shown in the fluorescence
studies.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
As shown in Figure 5, the fluorescence intensity increased with the enhancement of the β-CD-
EPDME concentration (with a same amount of insulin) and all of the fluorescence intensities were
higher than that of natural insulin. The possible reason for this phenomenon was that the cavity of
β-CD had an affinity for tyrosine (Tyr) of insulin, which played a more important role in the
adsorption spectrum (Teale, 1960). Compared with the natural insulin, the Tyr in the complex was
encapsulated into the cavity of the CDs and the water was excluded from the immediate Tyr
neighborhood, which induced an increase of fluorescence intensity (Sajeesh and Sharma, 2006).
Moreover, the continuous variation plots (Figure 6) indicated that the ΔF*[Insulin]t maximum had
a molar fraction of 0.8, which means that β-CD-EPDME@Insulin complex had 1:4 stoichiometry.
All of these results demonstrated the successful construction of the β-CD-EPDME@Insulin
complex.
3.3. Glucose triggered insulin release
The β-CD-EPDME@Insulin interaction of complex and glucose was confirmed firstly through
the change of fluorescence intensity. The longer the β-CD-EPDME@Insulin complex acts with
glucose, the lower is the fluorescence intensity of the solution, a value that grows closer to the
value of natural insulin (Figure 7A). This indicated the increased destruction of the complex

1
2
structure and an almost complete release of insulin. As shown in the fitting curve (Figure 7B), the
release rate slowly decreases over time because of the reduced concentration of complex. More
release data was shown in Figure 8. The quantity of released insulin in the presence of glucose
was significantly higher than that of without glucose. The insulin was completely released after 80
h. It is noted that the different release time in Figure 7B and Figure 8 was caused by the solution
concentration. The data in Figure 7 was obtained from the mixed solution of complex and glucose,
and the data in Figure 8 was plotted basing on the release liquids through dialysis membrane.
Both results proved that insulin could release from the complex while being triggered by glucose.
The mechanism of insulin release triggered by the glucose includes two aspects. The first is that
the phenylboronic acid (PBA) can react with cis-diols to form a reversible ester bond. Moreover,
the PBA and cis-diols in most of the saccharides can introduce more stable boric acid ester than
PBA and acyclic diols (Lorand and Edwards, 1959). In other words, the glucose will cause the
breakdown of the structure of the insulin carrier due to the production of a new ester bond
(Scheme 1). Secondly, the affinity of various samples with phenyl groups to the β-CD cavity are
different (Meng et al., 2010). Hence, the measurement of binding constant is of importance to
explain the release mechanism. The phenylboronate derivate (EPDMB) conjugated on the β-CD
structure can fall off to combine with glucose, which named as glucose-modified phenyl group.
The binding constant of insulin, glucose-modified phenyl group and phenylboronate group of β-
CD-EPDME to β-CD cavity were calculated by the data shown in Figure 9, which were 30.69 M-1,
46.04 M-1, and 7.63 M-1, respectively. The different binding constants demonstrated that the
formaiton of β-CD-EPDME@Insulin complex is more stable than that of the self-inclusion
complexation of β-CD-EPDME. Moreover, the glucose-modified phenyl group could act as the
guest molecule to attack the β-CD cavity, leading to the final release of the encapsulated insulin.
Additionally, circular dichroism spectroscopy was used to detect the secondary structure of
natural insulin and the released insulin triggered by glucose (Figure 10). The curve of natural
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
12

1
2
insulin and the released-insulin is almost the same, both have two minima at around 209 and
222nm, which are seen in classical predominant α-helix structure proteins (Sarmento et al., 2007).
Hence, these results indicated that the β-CD-EPDME@Insulin complex could successfully release
insulin under the stimulation of glucose and the released insulin could still maintain its secondary
structure.
3
4
5
6
3.4. Cytotoxicity analysis
7
The effective controlled delivery of β-CD-EPDME@Insulin triggered by glucose indicated that it
may have potential in diabetes therapy. Thus, it is of significance to research the cytotoxicity of β-CD-
EPDME and β-CD-EPDME@Insulin. In this work, CCK-8 was used for the cytotoxicity assay. As
shown in Figure 11, various concentrations of the complex exhibited cell viability above 80%,
suggesting the acceptable safety of the complex in potential biomedical applications. Compared
with β-CD-EPDME@Insulin, β-CD-EPDME exhibited lower cell viability (around 70%) at high
concentrations (>0.5 mg/mL). The reason for high cell viability of complex may be ascribed to the
cell growth promotion property of insulin (Straus, 1984). In addition, the cytotoxicity evaluation
could be generally regarded as less toxic since the final concentration used in this work was 0.5
mg/mL for the complex.
8
9
10
11
12
13
14
15
16
17
18 4. Conclusion
19
20
21
22
23
24
25
The insulin delivery system is of great interest to treat diabetes mellitus. Stimuli-responsive is
one of the challenges faced by current research. CDs, a popular substance used for drug delivery
systems, have a natural hydrophobic hole, which can carry a hydrophobic drug, like insulin.
Phenylboronic acid is always used as the glucose-responsive molecule due to its the high affinity
with the cis-diols in glucose. The combination of these two key materials will produce an
intelligent drug delivery carrier. In this work, the β-CD-EPDME is composed of β-CD and a
modified phenylboronate derivative. The former has a great affinity for hydrophobic drugs, and

1
2
3
4
5
6
7
8
hence, can encapsulate insulin. On the other hand, the later can detach from the β-CD-EPDME,
after being triggered by glucose. Subsequently, the phenyl group in the dissociated molecules will
attack the β-CD cavity to cause the release of insulin via a host-guest interaction. Moreover, the
insulin can be completely released from the complex after stimulation by glucose and can retain
its secondary structure as that of natural insulin. In addition, the β-CD-EPDME@Insulin in the
measured range shows low cytotoxicity. Hence, the β-CD-EPDME exhibits the potential to be
utilized as a safe glucose-responsive drug carrier for diabetes therapy.
9
Acknowledgements
10
11
12
13
14
15
Financial support from the National Natural Science Foundation of China (21534008,
51703018 and 51773128), National Key Research and Development Program of China
(2016YFC1100404), State Key Laboratory of Polymer Materials Engineering (Grant No.
sklpme2017-2-02) and the MOE Key Laboratory of Macromolecular Synthesis and
Functionalization, Zhejiang University (2017MSF04) are gratefully acknowledged.
16 References
17
18
19
20
21
22
23
Agbaria, R.A., Butterfield, M.T., Warner, I.M., 1996. Use of cyclodextrins and fluorescence
spectroscopy to probe the dual fluorescence of 9-anthroic acid. J. Phys. Chem. 100, 17133-17137.
Aguiar, M.M., Rodrigues, J.M., Silva Cunha, A., 2004. Encapsulation of insulin-cyclodextrin complex in
PLGA microspheres: a new approach for prolonged pulmonary insulin delivery. J. Microencapsulation
21, 553-564.
American Diabetes, A., 2014. Diagnosis and classification of diabetes mellitus. Diabetes care 37 Suppl 1,
S81-S90.
14

1
2
Cao, S., Su, Y., Yang, X., Chen, X., Zhao, Y., 2004. The investigation of β‐ cyclodextrin noncovalent
complex with protein or dipeptide by electrospray lonization mass spectrometry. Anal. Lett. 37, 1871-
1883.
3
4
Chai, Z., Ma, L., Wang, Y., Ren, X., 2016. Phenylboronic acid as a glucose-responsive trigger to tune the
insulin release of glycopolymer nanoparticles. J. Biomater. Sci., Polym. Ed. 27, 599-610.
Chen, X., Wu, W., Guo, Z., Xin, J., Li, J., 2011. Controlled insulin release from glucose-sensitive self-
assembled multilayer films based on 21-arm star polymer. Biomaterials 32, 1759-1766.
Daimon, Y., Izawa, H., Kawakami, K., Żywicki, P., Sakai, H., Abe, M., Hill, J.P., Ariga, K., 2014. Media-
dependent morphology of supramolecular aggregates of β-cyclodextrin-grafted chitosan and insulin
through multivalent interactions. J. Mater. Chem. B 2, 1802-1812.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Ding, R., D'Souza, V.T., 1990. A convenient method for functionalization of the 2-position of
cyclodextrins. Tetrahedron Lett. 31, 4275-4278.
Dotsikas, Y., Kontopanou, E., Allagiannis, C., Loukas, Y.L., 2000. Interaction of 6-p-
toluidinylnaphthalene-2-sulphonate with β-cyclodextrin. J. Pharm. Biomed. Anal. 23, 997-1003.
Teale, F.W., 1960. The ultraviolet fluorescence of proteins in neutral solution. Biochem. J. 76, 381-388.
Gao, L., Wang, T., Jia, K., Wu, X., Yao, C., Shao, W., Zhang, D., Hu, X.Y., Wang, L., 2017. Glucose-
responsive supramolecular vesicles based on water-soluble pillar[5]arene and pyridylboronic acid
derivatives for controlled insulin delivery. Chem. 23, 6605-6614.
Harada, A., Hashidzume, A., Takashima, Y., 2009. Cyclodextrin-based supramolecular polymers. Chem.
Soc. Rev. 38, 875-882.
Huang, L., Xin, J., Guo, Y., Li, J., 2010. A novel insulin oral delivery system assisted by cationic β-
cyclodextrin polymers. J. Appl. Polym. Sci. 115, 1371-1379.
Job, P., 1928. Job’s method of continuous variation. Ann. Chim. 9, 113-203.

1
2
Li, L., Jiang, G., Yu, W., Liu, D., Chen, H., Liu, Y., Huang, Q., Tong, Z., Yao, J., Kong, X., 2016. A
composite hydrogel system containing glucose-responsive nanocarriers for oral delivery of insulin.
Mater. Sci. Eng., C 69, 37-45.
3
4
Lorand, J.P., Edwards, J.O., 1959. Polyol complexes and structure of the benzeneboronate ion. J. Org.
Chem. 24, 769-774.
5
6
Matuszewska, A., Uchman, M., Adamczyk-Wozniak, A., Sporzynski, A., Pispas, S., Kovacik, L.,
Stepanek, M., 2015. Glucose-responsive hybrid nanoassemblies in aqueous solutions: ordered
phenylboronic acid within intermixed poly(4-hydroxystyrene)-block-poly(ethylene-oxide) block
copolymer. Biomacromolecules 16, 3731-3739.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Meng, H., Xue, M., Xia, T., Zhao, Y.-L., Tamanoi, F., Stoddart, J.F., Zink, J.I., Nel, A.E., 2010.
Autonomous in vitro anticancer drug release from mesoporous silica nanoparticles by pH-sensitive
nanovalves. J. Am. Chem. Soc. 132, 12690-12697.
Nakamura, K., Seki, T., Egawa, Y., Miki, R., Oda, Y., Yamanoi, T., Seki, T., 2013. Sugar-sensitive
supramolecular structures based on phenylboronic Acid-modified cyclodextrins. Chem. Pharm. Bull.
61, 1188-1191.
Sajeesh, S., Bouchemal, K., Marsaud, V., Vauthier, C., Sharma, C.P., 2010. Cyclodextrin complexed
insulin encapsulated hydrogel microparticles: An oral delivery system for insulin. J. Control. Release
147, 377-384.
Sajeesh, S., Sharma, C.P., 2006. Cyclodextrin-insulin complex encapsulated polymethacrylic acid based
nanoparticles for oral insulin delivery. Int. J. Pharm. 325, 147-154.
Sarmento, B., Ferreira, D.C., Jorgensen, L., van de Weert, M., 2007. Probing insulin's secondary structure
after entrapment into alginate/chitosan nanoparticles. Eur. J. Pharm. Biopharm. 65, 10-17.
Straus, D. S., 1984. Growth-stimulatory actions of insulin in vitro and in vivo. Endocr. Rev. 5, 356.
Szejtli, J., 1998. Introduction and General overview of cyclodextrin chemistry. Chem. Rev. 98, 1743.
16

1
2
Vyas, A., Saraf, S., Saraf, S., 2008. Cyclodextrin based novel drug delivery systems. J. Inclusion Phenom.
Macrocyclic Chem. 62, 23-42.
3
Xie, J., Li, A., Li, J., 2017. Advances in pH-Sensitive Polymers for Smart Insulin Delivery. Macromol.
Rapid Commun. 38, 1700413.
4
5
Xin, J., Guo, Z., Chen, X., Jiang, W., Li, J., Li, M., 2010. Study of branched cationic beta-cyclodextrin
polymer/indomethacin complex and its release profile from alginate hydrogel. Int. J. Pharm. 386, 221-
228.
6
7
8
Yao, Y., Wang, X., Tan, T., Yang, J., 2011. A facile strategy for polymers to achieve glucose-responsive
behavior at neutral pH. Soft Matter 7, 7948.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Yao, Y., Zhao, L., Yang, J., Yang, J., 2012. Glucose-responsive vehicles containing phenylborate ester for
controlled insulin release at neutral pH. Biomacromolecules 13, 1837-1844.

1
2
Scheme 1. The schematic illustration of glucose-responsive insulin delivery based on the
phenylboronic acid modified β-cyclodextrin system.
3
Scheme 2. Synthetic route of EPDMB.
4
Scheme 3. Synthetic route of β-CD-EPDME.
5
Figure 1. 1H NMR spectrum of EPDMB.
6
Figure 2. FT-IR spectra of EPDMB, β-CD-EPDME, and β-CD.
7
Figure 3. Time-of-flight mass spectrum of β-CD-EPDME.
8
Figure 4. IR spectra of β-CD-EPDMB, natural insulin, and β-CD-EPDME@Insulin.
Figure 5. Fluorescence spectra of natural insulin and β-CD-EPDME@Insulin at various concentrations.
Figure 6. Continuous variation plot (Job plot) of β-CD-EPDME@Insulin.
Figure 7. A) Fluorescence curves of β-CD-EPDME@Insulin after being interacted by glucose at
different time points, B) the fitting curve of insulin release profile triggered by glucose.
Figure 8. Released curve of insulin from β-CD-EPDME@Insulin in the absence and presence of
9
10
11
12
13
14
15
16
17
18
19
20
21
glucose (1.5 mg/mL) at pH 7.4 and 37
℃ .
Figure 9. Benesi-Hildebrand plots for the effect of different samples with phenyl groups on the β-
CD UV absorption intensity.
Figure 10. Circular dichroism spectrum of natural insulin and insulin released from β-CD-
EPDME@Insulin triggered by glucose.
Figure 11. Cell viability of β-CD-EPDME and β-CD-EPDME@Insulin at different concentrations.
18

1
2
3
4
Scheme 1. The schematic illustration of glucose-responsive insulin delivery based on the phenylboronic
acid modified β-cyclodextrin system.
5
6
7
8
9
10
11
12
13
14
15

1
2
Scheme 2. Synthetic route of EPDMB.
3
4
5
6
7
8
9
10
11
12
20

1
2
Scheme 3. Synthetic route of β-CD-EPDME.
3
4
5
6
7
8
9
10
11
12
13
14

1
2
3
4
5
6
7
8
9
Figure 1. 1H NMR spectrum of EPDMB.
22

1
2
Figure 2. FT-IR spectra of EPDMB, β-CD-EPDME, and β-CD.
3
4
5
6
7
8
9
10

1
2
Figure 3. Time-of-flight mass spectrum of β-CD-EPDME.
3
4
5
6
7
8
9
10
11
12
13
14
24

1
2
Figure 4. IR spectra of β-CD-EPDMB, natural insulin, and β-CD-EPDME@Insulin.
3
4
5
6
7
8
9
10
11
12
13
14
15
16

1
2
Figure 5. Fluorescence spectrum of natural insulin and β-CD-EPDME@Insulin.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
26

1
2
Figure 6. Continuous variation plot (Job plot) of β-CD-EPDME@Insulin.
3
4
5
6
7
8
9
10
11
12
13
14
15
16

1
2
Figure 7. A) Fluorescence curves ofβ-CD-EPDME@Insulin after being interacted by glucose at
3
4
different time points, B) the fitting curve of insulin releaseprofile triggered by glucose.
5
6
7
8
9
10
11
12
13
14
15
28

1
2
Figure 8. Released curve of insulin from β-CD-EPDME@Insulin in the absence and presence of
glucose (1.5 mg/mL) at pH 7.4 and 37
3
4
℃ .
5
6
7
8
9
10
11
12
13
14
15
16