Biomaterials 160 (2018) 82e91

Contents lists available at ScienceDirect

Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials

Layer-by-layer cell coating technique using extracellular matrix

facilitates rapid fabrication and function of pancreatic b-cell spheroids
Yasunari Fukuda a, Takami Akagi b, Tadafumi Asaoka a, Hidetoshi Eguchi a, Kazuki Sasaki a,
Yoshifumi Iwagami a, Daisaku Yamada a, Takehiro Noda a, Koichi Kawamoto a,
Kunihito Gotoh a, Shogo Kobayashi a, Masaki Mori a, Yuichiro Doki a, Mitsuru Akashi b, *
a
Department of Gastroenterological Surgery, Graduate School of Medicine, Osaka University, Osaka, Japan
b
Graduate School of Frontier Biosciences, Osaka University, Osaka, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Tissue engineering of insulin-secreting cells using alternatives to islet transplantation has been fueled by
Received 22 November 2017 the development of available materials and fabrication techniques. We have established a cell coating
Received in revised form technique that enables the cell surface to be coated with extracellular matrix based on the concept of a
11 January 2018
layer-by-layer (LbL) assembly. The present study evaluated whether this technique is beneficial for
Accepted 13 January 2018
fabricating pancreatic b-cell spheroids using a mouse b-cell line. The well-structured and dense spher-
Available online 16 January 2018
oids could immediately be constructed by the LbL-coated cells. In the functional analysis, spheroids with
the LbL-coated cells had greater insulin secretion ability with increased expression of the insulin and
Keywords:
Cell coating technique
glucose transporter 2 genes versus spheroids with non-coated cells. In addition, we found that the
Layer-by-layer expression of connexin 36, a gap junction molecule, was upregulated by the LbL cell coating. When
Spheroid spheroids with the LbL-coated cells were syngeneically transplanted in diabetic mice, blood glucose
Islet transplantation levels immediately decreased and glucose sensitivity significantly improved after intraperitoneal glucose
stimulation compared to spheroids with non-coated cells. This cell coating technique would be a clini-
cally applicable approach for fabricating pancreatic b-cell spheroids and treating type 1 diabetes mellitus.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction immunotherapies and islet isolation techniques [3,4]. In addition,

Type 1 diabetes mellitus (T1D) is characterized by the selective
destruction of pancreatic b-cells, mainly due to an autoimmune
attack, resulting in an absolute insulin deficiency. Approximately
one hundred and thirty thousand young patients have new onset
T1D every year, and the incidence has been increasing yearly
worldwide [1]. The therapeutic strategy for T1D is based on an
external supply of insulin. However, a subgroup of T1DM patients
consistently suffer from intractable hypoglycemia despite intense
insulin-based management, predisposing them to impaired quality
of life and related mortality. Islet transplantation (Tx) has emerged
as an attractive option to prevent these adverse effects [2]. How-
ever, long-term survival is still poorer with islet Tx compared to
pancreas Tx regardless of the remarkable advances in
* Corresponding author. Graduate School of Frontier Biosciences, Osaka Univer-
sity, 1-3 Yamadaoka, Suita, 565-0871, Japan.
E-mail address: akashi@fbs.osaka-u.ac.jp (M. Akashi).
because it is highly challenging to achieve successful single-donor
islet engraftment and two or more islet infusions are still neces-
sary to achieve insulin independence [3,5], this therapy is often
confronted with the risk of a donor shortage. Thus, tissue engi-
neering of insulin-secreting cells is expected to be a promising
therapeutic alternative to islet Tx.
There are two principal requirements for tissue engineering:
provision of insulin-secreting cells and the fabrication technique
for these cells. Regarding the latter, cell spheroids constitute the
bedrock of three-dimensional (3D) construction with broad utility
in clinical settings. However, cell spheroids are a mere aggregate of
single cells and, therefore, leaves much tissue to be solved for
clinical applications, including size heterogeneity, structural
fragility, and insufficient functional competence. Previous re-
searchers have shown the conceivable ways to overcome these
weaknesses of pancreatic b-cell spheroids. Notably, the develop-
ment of spheroid culture instruments, such as a stirred-suspension
culture [6], simulated microgravity generator [7], and agarose gel-
based microwells [8], have allowed the production of a large

https://doi.org/10.1016/j.biomaterials.2018.01.020
0142-9612/© 2018 Elsevier Ltd. All rights reserved.
 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91
amount of pancreatic b-cell spheroids at a well-controlled uniform
size. However, these reported techniques require time-consuming
processes, and few studies have attempted to fabricate well-
structured and functional pancreatic b-cell spheroids rapidly,
which is of great importance when transplanting the engineered
spheroids without decreasing cell survival and function.
In order to facilitate the structural stability and functionality of
the fabricated spheroids, the surface modification of living cells (i.e.,
cell surface engineering) is a conceivable method and previous
investigations have succeeded in immobilizing cell surface with
their own unique bioactive substances [9e12]. We have established
an original cell coating technique based on the concept of a layer-
by-layer (LbL) assembly [13], the LbL cell coating technique, using
two fundamental components of extracellular matrix (ECM),
fibronectin (FN) and gelatin (G), a collagen derivative [14e17].
Briefly, by alternatively dipping a dispersed single cell into the FN
and G solution, nanometer-thick FN-G composite films form on the
cell surface. This FN-G nanofilm promotes integrin-mediated
adhesion between neighboring cells, which confer rapid fabrica-
tion of multilayered tissues [15]. In addition, the use of ECM for the
cell coating has prominent advantages with regard to low cell
cytotoxicity and high biocompatibility. Thus far, we have attempted
to construct various types of 3D tissue models [18e22] and recently
succeeded in constructing functional vascularized human liver
tissue in terms of albumin production and metabolic activity using
our technique [23].
The present study addresses the hypothesis that the LbL cell
coating technique facilitates the rapid fabrication of well-
structured pancreatic b-cell spheroids, thereby improving func-
tion by enhancing intercellular interactions. The aim of this study
was to evaluate the effectiveness of the LbL cell coating technique
for rapid fabrication and the function of pancreatic b-cell spheroids.
2. Materials and methods
2.1. Materials
All chemicals were used without further purification. FN from
human plasma and bovine serum albumin (BSA) were purchased
from Sigma-Aldrich (MO, USA). G, 2-mercaptoethanol (2-Me), 10%
formalin neutral buffer solution, and skim milk powder were pur-
chased from Wako Pure Chemical Industries (Osaka, Japan). Dul-
becco's modified eagle medium (D-MEM), phosphate-buffered
saline (PBS), D-PBS (þ) Preparation Reagent (Ca, Mg Solution), and
proteinase K were purchased from Nacalai Tesque (Kyoto, Japan).
ImmunoSaver was purchased from Nissin EM (Tokyo, Japan). Rabbit
polyclonal anti-mouse connexin 36 (Cx 36) was purchased from
Abcam (Cambridge, UK). Rabbit monoclonal anti-mouse E-cadherin
(E-cad) was purchased from Cell Signaling Technology (Tokyo,
Japan). Rabbit polyclonal anti-mouse insulin was purchased from
Santa Cruz Biotechnology (CA, USA). Rabbit polyclonal anti-mouse
actin was purchased from Sigma Aldrich (MO, USA). Goat anti-
rabbit Alexa Fluor® 546-conjugated IgG secondary antibody, Pro-
long® Gold antifade reagent with 40 ,6-diamidino-2-phenylindole
(DAPI), and fetal bovine serum (FBS) were purchased from
Thermo Fisher Scientific (MA, USA).
2.2. Cell culture
Insulin-secreting MIN6 cells were provided by Dr. M. Konno
(University of Osaka, Osaka, Japan). The cells were cultured in D-
MEM supplemented with 10% FBS, 100 mg/ml streptomycin, 100
units/ml penicillin, 0.25 mg/ml amphotericin B, and 0.1 mM 2-Me,
and incubated in 5% CO2 at 37  C. The cells were subcultured with
0.05% trypsin and 0.02% EDTA in PBS. Passage 33-35 MIN6 cells
83
were used for the present study.
2.3. Formation of FN-G nanofilms on the MIN6 cell surface using the
LbL cell coating technique
MIN6 cells were coated with FN and G by the filtration LbL
process as described previously [15,23]. FN solution (0.2 mg/ml),
PBS, and G solution (0.2 mg/ml) were added into three wells of a 6-
well cell culture plate (Corning, NY, USA) and dispersed MIN6 cells
seeded into a 6-well culture insert with 3-mm pore polycarbonate
membranes (Corning). As a step in coating cells with FN or G, a
MIN6-seeded culture insert was placed on the well containing the
FN or G solution and stirred for 1 min on a shaking incubator (SI-
300, AS ONE, Japan). During each process, the culture insert was
placed on the well containing PBS and washed for 1 min to shake off
the non-coated FN or G. By alternatively repeating these coating
processes a total of nine times (five steps for FN coating and four
steps for G coating), FN-G nanofilms formed on the cell surface
(Fig. 1A). The time required for the entire process was approxi-
mately 30 min.
We evaluated whether MIN6 cells were appropriately coated
with FN and G in two ways. First, rhodamine-labeled FN (Rh-FN)
and fluorescein isothiocyanate-labeled G (FITC-G) were used for the
LbL coating process and the cells observed using by confocal laser
scanning microscopy (CLSM; FV10i, Olympus, Tokyo, Japan). Sec-
ond, FN labeled with the Alexa Fluor® 488 Protein Labeling Kit
(Thermo Fisher Scientific) was used for the LbL coating process and
stained cells counted by flow cytometry using the BD FACS Canto™
II system (BD Biosciences).
2.4. Fabrication of MIN6 spheroids
To clarify the usefulness of the LbL cell coating technique, we
principally compared spheroids created without the LbL cell
coating technique (LbL－ spheroids) and spheroids created with the
LbL cell coating technique (LbLþ spheroids) (Fig. 1A). The conven-
tional two-dimensional (2D) cultured cells were concurrently used
as a control. During the spheroid fabrication process, MIN6 cells
were seeded at a density of 4  103 cells per well into the low cell
attachment 96-well U-bottom plate (Sumitomo Bakelite, Tokyo,
Japan) with 200 ml of the culture medium and incubated in 5% CO2
at 37  C for 72 h without any changes of the culture medium. MIN6
cells from 2D culture were seeded in the 96-well flat-bottom
microplate (AGC Techno Glass, Shizuoka, Japan) and incubated in
a similar manner.
2.5. Cell damage assay
Cell viability and cytotoxicity assays were performed to evaluate
cell damage accompanying the LbL cell coating process. Cell
viability was evaluated by staining cells with trypan blue. Cell
cytotoxicity was evaluated by measuring the amount of lactate
dehydrogenase (LDH) released after the LbL coating process using
the Cytotoxicity LDH Assay Kit-WST (Dojindo Molecular Technol-
ogies, Kumamoto, Japan).
2.6. Histological analysis
For sample fixation in vitro, MIN6 spheroids were carefully
collected from each well and jellified using iPGell (GenoStaff, Tokyo,
Japan). In vitro jellified spheroids and in vivo tissues were fixed with
10% formalin and embedded in paraffin; 2.5-mm sectioned slides
were used for hematoxylin-eosin (HE) staining and immunofluo-
rescence (IF). For IF, after deparaffinization of the slides, the anti-
gens were retrieved using one of the following agents:
84 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91

Fig. 1. Fabrication of MIN6 spheroids using the LbL cell coating technique. A) Schematic illustration of the fabrication of spheroids using FNeG coated MIN6 cells. B) Representative
images of MIN6 cells coated with Rh-FN (red) and FITC-G (green) using CLSM. Scale bar ¼ 10 mm. C) The number of MIN6 cells stained with Alexa 488-labeled FN was counted by
flow cytometry after the LbL cell coating process. Nearly all MIN6 cells (98.3%) were properly coated with FN-G. D) Cell viability of MIN6 cells with or without the LbL cell coating
process (n ¼ 6). E) LDH leakage of MIN6 cells with or without the LbL cell coating process (n ¼ 4). Cell viability and LDH leakage were not influenced by the LbL cell coating process,
indicating that this process was intact for MIN6 cells. All data are presented as mean ± SD. Abs, absorbance; CLSM, confocal laser scanning microscopy; FITC, fluorescein iso-
thiocyanate; FN, fibronectin; G, gelatin; LbL, layer-by-layer; LDH, lactate dehydrogenase; Rh, rhodamine; SD, standard deviation; NS, not significant.

ImmunoSaver, citrate buffer (pH 6), or proteinase K. Next, the slides
were blocked with 1% BSA for 1 h at room temperature (RT) and
incubated with primary antibodies (1:400 for Cx 36, E-cad and
insulin) overnight at 4  C. The slides were incubated with a sec-
ondary antibody (1:200) for 1 h at RT and mounted using Prolong®
Gold antifade reagent with DAPI. The mounted slides were
observed by CLSM.
2.7. TUNEL assay
For TdT-mediated dUTP nick end labeling (TUNEL) staining of
spheroids, 2.5-mm sectioned slides were pretreated with proteinase
K after deparaffinization and incubated with the TUNEL reaction
mixture from the In Situ Cell Death Detection Kit (Sigma-Aldrich)
for 1 h at RT. The slides were then observed by CLSM. Subsequently,
the slides were stained with Dako DABþ substrate buffer (Agilent,
CA, USA) for 10 s. Finally, the slides were counterstained with he-
matoxylin and observed using the EVOS® FL Auto Cell Imaging
System (Life Technologies).
2.8. Quantitative assessment of cell aggregation
To quantitatively evaluate the degree of cell aggregation, we
calculated the porosity of spheroids. The entire area and area of the
intracellular space of HE-stained spheroids were quantified using
Image J 1.50i (National Institutes of Health, USA) (Supplementary
Fig. 1). The porosity was calculated as the ratio of the area of the
intracellular space to the entire area.
2.9. qRT-PCR analysis of mRNA expression
The total RNA was extracted using the RNeasy Mini Kit (Qiagen,
CA, USA) according to the manufacturer's instructions. The integrity
and concentration of RNA were measured using NanoDrop Lite
(Thermo Fisher Scientific). Contaminating genomic DNA (gDNA)
(DNase I reaction) was removed and reverse transcription (syn-
thesis of complementary DNA) performed using ReverTra Ace®
qPCR RT Master with gDNA Remover (TOYOBO, Osaka, Japan) ac-
cording to the manufacturer's instructions. The quantitative real-
time polymerase chain reaction (qRT-PCR) was performed with
specific oligonucleotide primers and the LightCycler 480 Real-Time
PCR system (Roche Diagnostics, Mannheim, Germany). THUNDER-
BIRD® SYBR® qPCR MIX (TOYOBO) was used to detect the amplifi-
cation products. Actin was chosen as a reference gene. The PCR
conditions were as follows: pre-denaturation for 30 s at 95  C and
amplification for 45 cycles of denaturation for 15 s at 95  C,
annealing for 15 s at 60  C (for actin, Insulin 1 and Insulin 2) or 56  C
(for Cx 36, E-cad, glucose transporter 2 (GLUT 2), and glucokinase),
and extension for 30 s at 72  C. The PCR primers are listed in
Supplementary Table 1.
2.10. Western blot analysis of protein expression
Total protein was extracted using Pierce® RIPA Buffer adding
 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91 85

Halt™ Protease Inhibitor Cocktail and Halt™ Phosphatase Inhibitor
Cocktail (Thermo Fisher Scientific) according to the manufacturer's
instructions. Aliquots of proteins were electrophoresed on 7.5%
Extra PAGE One Precast Gel (Nacalai Tesque) and transferred to
iBlot® 2 PVDF membrane (Thermo Fisher Scientific) using the iBlot®
2 Gel Transfer Device (Thermo Fisher Scientific). After blocking
with 5% skim milk powder for 1 h at RT, the membranes were
incubated with the primary antibodies (1:1000 for E-cad and 1:250
for Cx36) overnight at 4  C. Actin (1:3000) was chosen as a refer-
ence protein. Subsequently, the membranes were incubated with
an HRP-linked anti-rabbit IgG secondary antibody (1:100,000; GE
Healthcare Biosciences, NJ, USA) for 1 h at RT and enveloped with
Chemi-Lumi One Super (Nacalai Tesque). The ChemiDoc™ Touch
imaging system (Thermo Fisher Scientific) was used for band
detection.
2.11. In vitro functional analysis
Insulin secretion levels were measured using the Mercodia
Mouse Insulin ELISA kit (Mercodia AB, Uppsala, Sweden) according
to the manufacturer's instructions. To evaluate insulin secretion
with time, the culture supernatants were collected from each well
after 24 and 72 h of culture for the assay.
For the glucose stimulating test, constructed spheroids culti-
vated for 72 h were pre-incubated in D-PBS (þ) without glucose at
37  C for 3 h. Subsequently, the cells were exposed to D-PBS (þ)
supplemented with two different concentrations of glucose (3.3
and 20.0 mM) at 37  C for 2 h and the culture supernatants
collected from each well for the assay. Because the proliferation of
MIN6 cells could differ across each construct, insulin secretion
levels were standardized by adenosine triphosphate (ATP) using
the KATAMARI ATP assay Kit (Wako Pure Chemical Industries).
2.12. Syngeneic transplantation in diabetic mice
The animal experiment protocol was approved by the Animal
Experiments Committee at Osaka University (no. 28-022-001). The
animal experiments were conducted according to the National In-
stitutes of Health guidelines for the use of experimental animals.
This study used 8 to 10-week-old male C57BL/6 mice purchased
from Charles River (Kanagawa, Japan) as syngeneic recipients.
Diabetes was induced though a single intraperitoneal injection of
180 mg/kg streptozotocin (STZ) (Nacalai Tesque), 9 days prior to Tx.
The mice with non-fasting blood glucose levels 400 mg/dl were
considered diabetic. The diabetic mice were randomly divided into
five groups: (I) sham-operated mice (n ¼ 4), (II) mice transplanted
with 60 LbL－ spheroids (60 LbL－) (n ¼ 5), (III) mice transplanted
with 60 LbLþ spheroids (60 LbLþ) (n ¼ 5), (IV) mice transplanted
with 100 LbL－ spheroids (100 LbL－) (n ¼ 5), and (V) mice trans-
planted with 100 LbLþ spheroids (100 LbLþ) (n ¼ 5). In each group,
MIN6 cells were seeded at a density of 4  103 cells per well and the
transplanted spheroids fabricated after 72 h of culture in vitro. The
mice were anesthetized by an intraperitoneal injection of 4 mg/kg
midazolam, 0.75 mg/kg medetomidine, and 5 mg/kg butorphanol,
and a 1.5-cm skin incision was made at a right lateral decubitus
position, followed by peritonectomy to approach the left kidney.
Spheroids were transplanted under the left kidney capsule.

Fig. 2. Morphological differences between LbL－ spheroids and LbLþ spheroids. A) Representative images of the aggregating process of the non-coated MIN6 cells (left panels) and
LbL-coated MIN6 cells (right panels) after 3, 6, 24, and 72 h of culture using phase-contrast microscopy. The LbL-coated MIN6 cells were more rapidly and firmly aggregated than the
non-coated MIN6 cells. Scale bar ¼ 200 mm. B) Representative images of HE-stained LbL－ spheroids (left panels) and LbLþ spheroids (right panels). The LbL-coated MIN6 cells
fabricated denser spheroids than the non-coated MIN6 cells after 24 h and 72 h of culture. Scale bar ¼ 100 mm. C) The porosity was calculated using Image J. The LbL－ spheroids
were significantly more porous than the LbLþ spheroids after 24 and 72 h of culture (n ¼ 8). All data are presented as mean ± SD. **P < .01. HE, hematoxylin-eosin; LbL, layer-by-
layer; SD, standard deviation.
86 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91

2.13. In vivo functional analysis considered significant.

Non-fasting blood glucose levels were monitored daily for the
first 7 days after Tx and every 3 days thereafter. Murine serums
were collected before and 7 and 21 days after Tx to measure serum
insulin levels using the Mercodia Mouse Insulin ELISA kit. An
intraperitoneal glucose tolerance test (IPGTT) was performed 7
days after Tx in groups (IV) and (V). After the mice fasted overnight,
a 1.5 g/kg glucose solution was infused intraperitoneally. The blood
glucose level was measured 0, 15, 30, 60, 90, and 120 min after
glucose infusion. In addition, the histology of the transplanted
spheroids was evaluated 7 days after Tx. The grafted kidney was
removed 35 days after Tx to evaluate whether the transplanted
spheroids under the kidney capsule were liable for the glucose
control.
2.14. Statistical analysis
3. Results
3.1. MIN6 cells were appropriately coated with FN and G without
any cellular damage
First, we evaluated whether the FN-G nanofilms were appro-
priately formed on the MIN6 cell surface by the LbL cell coating
process. MIN6 cell surfaces were stained for fluorescence with Rh-
FN and FITC-G, indicating that each cell was properly coated with
FN and G following the coating process (Fig. 1B). We ascertained by
flow cytometry that almost all cells (98.3%) were coated with FN
and G (Fig. 1C). Next, we showed that this process did not increase
the LDH release (Fig. 1E) and maintained the cell viability (Fig. 1D),
indicating that the LbL cell coating process did not have detrimental
effects on pancreatic b-cells.

Statistical analysis was carried out using JMP® software (SAS
Institute, Cary, NC). All continuous variables were expressed as
mean ± standard deviation (SD). The Student's t-test or Welch's t-
test was used to compare the two variables, and the Kruskal-Wallis
test was used to compare three or more variables. P < .05 was
3.2. Well-structured and denser spheroids were rapidly fabricated
using LbL-coated MIN6
To evaluate the morphological differences between the LbL
and LbLþ spheroids, we first investigated the morphological

Fig. 3. Comparison of insulin secretion ability and related gene expression. A) Chronological insulin secretion levels normalized by ATP (n ¼ 4). The insulin secretion level was
highest in the LbLþ spheroids, followed by the LbL－ spheroids, after 24 and 72 h of culture. B) Insulin secretion level normalized by ATP responsive to the glucose stimulation at low
glucose (3.3 mM glucose) and high glucose (20.0 mM glucose) (n ¼ 7). Glucose responsiveness was better in the LbLþ spheroids than the LbL－ spheroids. C) qRT-PCR analysis of
pancreatic b-cell-specific mRNA expression (n ¼ 4). All results are shown as the fold-change relative to the expression of 2D cultured cells. The expression of insulin 1 and 2 was
evaluated after 24 and 72 h of culture, and the expression of GLUT 2 and glucokinase was evaluated after 72 h of culture. Insulin 1 and 2 expression was highest in the LbLþ
spheroids, followed by the LbL－ spheroids, after 24 and 72 h of culture. Although GLUT 2 expression was significantly higher in the LbLþ spheroids compared to the LbL－
spheroids, glucokinase expression was similar between the LbL－ and LbLþ spheroids. All data are presented as mean ± SD. **P < .01. ATP, adenosine triphosphate; GLUT 2, glucose
transporter 2; LbL, layer-by-layer; SD, standard deviation; 2D, two-dimensional; NS, not significant.
 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91 87

changes in aggregating cells using phase-contrast microscopy. The
LbL-coated cells were more highly aggregated than the non-coated
cells after just 3 h of culture, and the LbLþ spheroids had lower
optical transparency after 24 h of culture, indicating that the LbL-
coated cells were firmly piled up in the culture well. No clear dif-
ferences were detected between the two spheroids after 72 h of
culture (Fig. 2A). Histological examination of HE-stained spheroids
revealed that well-structured and denser spheroids were con-
structed with the LbL-coated cells after 24 and 72 h of culture
(Fig. 2B), which was quantitatively assessed by measuring the
porosity of spheroids; the porosity was significantly lower in the
LbLþ spheroids than in the LbL－ spheroids after 24 and 72 h of
culture (Fig. 2C). In addition, TUNEL-positive cells were scattered
throughout, despite the higher aggregation of the LbL-coated cells
after 72 h of culture (Supplementary Fig. 2A), and the ratio of
TUNEL-positive cells to whole cells was similar between the LbL－
and LbLþ spheroids (Supplementary Fig. 2B), indicating that
apoptosis or necrosis was not induced in the LbLþ spheroids.
3.3. LbLþ spheroids exhibited higher insulin secretion in vitro
We compared the insulin secretion ability of 2D cultured cells,
LbL－ spheroids, and LbLþ spheroids. As cell proliferation can differ
across each construct, we estimated the cell proliferation by ATP
(Supplementary Fig. 3). The insulin secretion levels, normalized by
a total amount of ATP, were highest in the LbLþ spheroids, followed
by the LbL－ spheroids, after 24 and 72 h of culture (Fig. 3A).
Regarding the glucose stimulating test, although the insulin
secretion levels were comparable between the LbL－ and LbLþ
spheroids when exposed to a low glucose concentration (3.3 mM),
the levels were significantly higher in the LbLþ spheroids than the
LbL－ spheroids when exposed to a high glucose concentration
(20.0 mM), indicating that the glucose responsiveness was better in
the LbLþ spheroids than in the LbL－ spheroids (Fig. 3B).
Subsequently, the mRNA expression was evaluated by qRT-PCR
(Fig. 3C). Insulin 1 and 2 expression was highest in the LbLþ
spheroids, followed by LbL－ spheroids, after 24 and 72 h of culture.
With respect to the expression of GLUT2 and glucokinase, which
promote glucose intake and glucose phosphorylation in pancreatic
b-cells, respectively, GLUT2 expression was significantly higher in
the LbLþ spheroids compared to the LbL－ spheroids, whereas
glucokinase expression was comparable between the LbL－ and
LbLþ spheroids after 72 h of culture.
3.4. LbL cell coating upregulated Cx 36 expression
To clarify the underlying mechanisms of better insulin secretion
with the LbL cell coating, we focused on the intercellular in-
teractions and assessed the expression of two types of cell junction
molecules, E-cad (adherence junction) and Cx 36 (gap junction),
which were previously found to play several roles in pancreatic b-
cells [24e27]. IF revealed that the expression of E-cad was similar
between LbL－ and LbLþ spheroids, whereas the expression of Cx
36 was distinctly higher in the LbLþ spheroids than in the LbL－
spheroids (Fig. 4A). These results were confirmed by qRT-PCR and
Western blot (Fig. 4B and C).
3.5. LbLþ spheroids had superior antidiabetic effects in syngeneic
recipient mice
We elucidated the anti-diabetic effect of LbLþ spheroids in vivo.
The blood glucose levels and body weights of mice prior to Tx were
well balanced between each group (Supplementary Fig. 4). When
60 or 100 spheroids were transplanted in STZ-induced diabetic
mice, the non-fasting blood glucose level more immediately
decreased in the groups of mice transplanted with LbLþ spheroids.
Finally, the blood glucose levels of all transplanted mice stabilized
owing to the reproductive integrity of MIN6 cells and re-elevated
after nephrectomy (Fig. 5A). To further examine graft function
in vivo, an IPGTT was performed 7 days after Tx for the groups (IV)

Fig. 4. Comparison of the expression of E-cadherin (E-cad) and connexin 36 (Cx 36) between LbL－ spheroids and LbLþ spheroids. A) Representative images of the LbL－ spheroids
(left panels) and LbLþ spheroids (right panels) immunostained with E-cad (red; upper panels), Cx 36 (red; lower panels), and DAPI (blue). Scale bar ¼ 100 mm for low magnification
and 10 mm for high magnification. B) qRT-PCR analysis of E-cad and Cx 36 mRNA expressions (n ¼ 4). Results are shown as the fold-change relative to the expression of 2D cultured
cells. C) Western blot analysis of E-cad and Cx 36 protein expression. At both the mRNA and protein levels, E-cad expression by the LbLþ spheroids was similar to LbL－ spheroids,
whereas that of Cx 36 was significantly higher in LbLþ spheroids. All data are presented as mean ± SD. **P < .01. DAPI, 40 ,6-diamidino-2-phenylindole; LbL, layer-by-layer; qRT-PCR,
quantitative real-time polymerase chain reaction, SD standard deviation; 2D, two-dimensional; NS, not significant.
88 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91

Fig. 5. Antidiabetic effects in a syngeneic transplantation model. A) Changes in the non-fasting blood glucose levels after Tx. When 60 or 100 spheroids were transplanted in STZ-
induced diabetic mice, the blood glucose level more immediately decreased in the groups of LbLþ spheroids than the groups of LbL－ spheroids. In all transplanted mice, the blood
glucose level stabilized and re-elevated after nephrectomy. B) Intraperitoneal glucose tolerance test in mice transplanted with 100 LbL－ or 100 LbLþ spheroids 7 days after Tx. Non-
diabetic mice were used as controls. The elevation of blood glucose was significantly prevented in the groups of 100 LbLþ spheroids compared to 100 LbL－ spheroids 15 and 30 min
after glucose infusion. C) Serum insulin levels before and 7 and 21 days after Tx. The insulin level was highest in the groups with LbLþ spheroids, followed by those with LbL－
spheroids 7 days after Tx. However, the levels were similar between the LbL－ and LbLþ spheroids 21 days after Tx. All data are presented as mean ± SD. *P < .05, **P < .01. LbL, layer
by layer; ope., operation; SD, standard deviation; STZ, streptozotocin; Tx, transplantation; NS, not significant.

and (V). The LbLþ spheroids significantly prevented the elevation of
blood glucose 15 and 30 min after glucose infusion (Fig. 5B). In
addition, serum insulin levels were highest in the groups with the
LbLþ spheroids, followed by the LbL－ spheroids, 7 days after Tx,
though it became similar between the LbL－ and LbLþ spheroids 21
days after Tx (Fig. 5C). Histological examination confirmed the
presence of transplanted spheroids by immunostaining with in-
sulin (Fig. 6A) and revealed that Cx 36 expression was higher in the
LbLþ spheroids than the LbL－ spheroids 7 days after Tx (Fig. 6B).
4. Discussion
In the current study, we demonstrated that the LbL cell coating
technique using two basic components of ECM benefits the rapid
fabrication of well-structured pancreatic b-cell spheroids and in-
sulin secretion ability. In addition, the spheroids with the LbL cell
coating could effectively treat STZ-induced diabetic mice in a syn-
geneic Tx model.
Since Decher first advocated the principle of LbL assembly
technology [13], it has been drastically innovated by available
substances and involved in a broad variety of fields, including tissue
engineering. Our group developed the LbL cell coating technique,
which can form FN-G nanofilms on cell surfaces with little damage,
and this technique is one of the leading approaches to fabricating
3D tissues [14e23,28]. The predominant property of our technique
different from ECM-coated culture surfaces or gels is the potential
to control the cellular spatial arrangement freely and readily at a
single cell level. In this respect, we considered that our technique
could be applied not only to culture on polyester membrane, but
also to the spheroid culture. Indeed, our technique could regulate
the architecture and function of pancreatic b-cell spheroids in just 3
days. In addition, considering that the pancreatic ECM mainly in-
cludes FN and collagens [29,30], the application of FN and G for the
LbL cell coating is quite reasonable.
On the other hand, we did not focus on the size of the spheroids.
Previous studies have reported that the function and survival of
small islet or b-cell spheroids (<250 mm) is superior to that of large
spheroids because smaller spheroids can be supplied more suffi-
cient oxygen and nutrients [6e8,31,32]. In the present study, our
fabricated spheroids were approximately 300 mm in size, regardless
of the presence or absence of the LbL cell coating, which was larger
than the ideal spheroid size. However, importantly, apoptotic or
necrotic cells did not increase in the LbLþ spheroids
(Supplementary Fig. 2), suggesting that the LbL cell coating did not
inhibit the supply of oxygen and nutrients into the spheroids. In
addition, constructing a large amount of spheroids at a time was
not taken into account in our series. Combination of the LbL cell
coating technique with previously reported spheroid culture
techniques that enable the creation of a large amount of size-
controlled spheroids [6e8] will allow the construction of ideal
pancreatic b-cell spheroids rapidly and abundantly.
We showed that the LbL cell coating technique enhances the
insulin secretion ability in vitro. Insulin was highly expressed in the
LbL-coated cells after beginning the culture and GLUT 2 expression
 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91 89

Fig. 6. Histological findings of transplanted spheroids 7 days after transplantation. A) Representative images of transplanted LbL－ spheroids (left panel) and LbLþ spheroids (right
panel) immunostained with insulin (red) and DAPI (blue). Scale bar ¼ 100 mm. B) Representative images of transplanted LbL－ spheroids (left panels) and LbLþ spheroids (right
panels) immunostained with Cx 36 (red) and DAPI (blue). Scale bar ¼ 100 mm for low magnification and 10 mm for high magnification. The Cx 36 expression was still higher in the
LbLþ spheroids than the LbL－ spheroids 7 days after Tx. Cx 36, connexin 36; DAPI, 40 ,6-diamidino-2-phenylindole; LbL, layer-by-layer. (For interpretation of the references to color
in this figure legend, the reader is referred to the Web version of this article.)

was also increased. These results were supported by a previous
study demonstrating that adding ECM increased the expression of
insulin and GLUT 2 during the differentiation of insulin-secreting
cells from mesenchymal stem cells [29]. The LbL cell coating
enhanced the uptake of glucose into the pancreatic b-cells, leading
to better insulin secretion in response to glucose stimulation.
Cell-to-cell interactions via cell junctions is implicated in cell
differentiation, proliferation, and function. Because we used FN and
G for the LbL cell coating, they can act as mediators for adjacent
cells to communicate with each other. In pancreatic b-cells, the gap
junction communication made of Cx 36 is closely linked with the
biosynthesis and release of insulin [24e26], as demonstrated by Cx
36 regulating the synchronization of Ca2þ oscillations, which
directly influences the exocytosis of insulin [26]. In addition, Guo
et al. reported that connexin expression by alveolar epithelial cells
is accelerated in FN-rich culture [33]. In the present study, the LbL
cell coating promoted the expression of Cx 36, leading to better
insulin secretory responsiveness of pancreatic b-cells. In addition, a
better IPGTT response and higher blood insulin levels were found in
mice with the LbLþ spheroids because upregulated Cx 36 expres-
sion was still maintained 7 days after Tx. In contrast, the cell
adhesion molecule E-cad has been reported to be associated with
pancreatic b-cell proliferation rather than its function [24,27].
Carvell et al. reported that E-cad expression was conversely asso-
ciated with b-cell proliferation by assessing MIN6 cells transfected
with the E-cad sense or anti-sense expression vector [27]. This
90 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91
finding was consistent with the results of the present study
showing that the E-cad expression of 2D cultured cells and LbL－
and LbLþ spheroids conversely correlated with these proliferations
(Fig. 4 and Supplementary Fig. 3). Furthermore, the LbL cell coating
did not increase E-cad expression, indicating that adjacent cells in
the LbLþ spheroids were indirectly aggregated via integrin-
mediated binding between the cells and surrounding ECM.
The current study has several limitations. First, we selected
MIN6 cells as the insulin-secreting cells because the glucose-
stimulated insulin secretion ability of MIN6 cells is similar to that
of normal islets [34]. However, due to the reproductive integrity of
MIN6 cells, the cells are not suitable for clinical application and we
could not show the survival advantage of LbLþ spheroids after Tx.
Second, we determined the conditions of the LbL cell coating pro-
cess, including concentration of the FN and G solution, stirring time
in the FN and G solution and step number, according to our pre-
vious reports [15,23]. However, the optimal conditions, in which
the function of MIN6 cells was most effectively enhanced, were not
evaluated in our series. Third, we identified the functional benefit of
the LbL cell coating by evaluating genes involved in glucose intake
and cell junctions. However, the LbL cell coating technique could
regulate mitochondrial or transcriptional programs, which also
play pivotal roles in b-cell function [35,36]. Clarifying these un-
derlying mechanisms is of great importance to better under-
standing the significance of this technique.
5. Conclusion
We showed that the LbL cell coating technique accelerates the
rapid fabrication and function of pancreatic b-cell spheroids. This
technique can provide a cellular microenvironment capable of
regulating the spheroid architecture and exerting its function. In
addition, this technique is clinically feasible in terms of the
simplicity and biological safety of the process. However, it is
necessary to elucidate whether our fabrication technique is effec-
tive in stem cell replacement therapy using embryonic or plurip-
otent stem cell-derived b-cells. We expect our technique to
contribute to creating functional stem cell-derived b-cell spheroids
and treating T1D patients.
Grant support
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
Author contributions
Y.F. designed the study and drafted the manuscript. T.A., T.A.,
H.E., K.S., K.K. and M.A. contributed to analysis and interpretation of
the study. Y.I., D.Y., T.N., K.G., S.K., M.M. and Y.D. were the super-
vising surgeons. All authors read and approved the final
manuscript.
Acknowledgements
The authors wish to thank Michiaki Unno and Naoaki Sakata
(Tohoku University) and Gunpei Yoshimatsu (Fukuoka University)
for their insightful comments and suggestions in this study.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.biomaterials.2018.01.020.
References
[1] International Diabetes Federation, IDF Diabetes Atlas, eighth ed., 2017. http://
www.diabetesatlas.org/. (Accessed 8 January 2018).
[2] H. Colvin, T. Mizushima, H. Eguchi, S. Takiguchi, Y. Doki, M. Mori, Gastroen-
terological surgery in Japan: the past, the present and the future, Ann. Gas-
troenterol. Surg. 1 (2017) 5e10.
[3] A.M. Shapiro, Islet transplantation in type 1 diabetes: ongoing challenges,
refined procedures, and long-term outcome, Rev. Diabet. Stud. 9 (2012)
385e406.
[4] B.J. Hering, W.R. Clarke, N.D. Bridges, T.L. Eggerman, R. Alejandro, M.D. Bellin,
K. Chaloner, C.W. Czarniecki, J.S. Goldstein, L.G. Hunsicker, D.B. Kaufman,
O. Korsgren, C.P. Larsen, X. Luo, J.F. Markmann, A. Naji, J. Oberholzer,
A.M. Posselt, M.R. Rickels, C. Ricordi, M.A. Robien, P.A. Senior, A.M. Shapiro,
P.G. Stock, N.A. Turgeon, Phase 3 trial of transplantation of human islets in
type 1 diabetes complicated by severe hypoglycemia, Diabetes Care 39 (2016)
1230e1240.
[5] A.M. Shapiro, C. Ricordi, B.J. Hering, H. Auchincloss, R. Lindblad, R.P. Robertson,
A. Secchi, M.D. Brendel, T. Berney, D.C. Brennan, E. Cagliero, R. Alejandro,
E.A. Ryan, B. DiMercurio, P. Morel, K.S. Polonsky, J.A. Reems, R.G. Bretzel,
F. Bertuzzi, T. Froud, R. Kandaswamy, D.E. Sutherland, G. Eisenbarth, M. Segal,
J. Preiksaitis, G.S. Korbutt, F.B. Barton, L. Viviano, V. Seyfert-Margolis,
J. Bluestone, J.R. Lakey, International trial of the Edmonton protocol for islet
transplantation, N. Engl. J. Med. 335 (2006) 1318e1330.
[6] L.T. Lock, S.G. Laychock, E.S. Tzanakakis, Pseudoislets in stirred-suspension
culture exhibit enhanced cell survival, propagation and insulin secretion,
J. Biotechnol. 151 (2011) 278e286.
[7] H. Tanaka, S. Tanaka, K. Sekine, S. Kita, A. Okamura, T. Takebe, Y.W. Zheng,
Y. Ueno, J. Tanaka, H. Taniguchi, The generation of pancreatic b-cell spheroids
in a simulated microgravity culture system, Biomaterials 34 (2013)
5785e5791.
[8] Y. Ichihara, R. Utoh, M. Yamada, T. Shimizu, Y. Uchigata, Size effect of engi-
neered islets prepared using microfabricated wells on islet cell function and
arrangement, Heliyon 2 (2016), e00129.
[9] R. Georgieva, S. Moya, M. Hin, R. Mitlo €hner, E. Donath, H. Kiesewetter,
H. Mo € hwald, H. Ba€umler, Permeation of macromolecules into polyelectrolyte
microcapsules, Biomacromolecules 3 (2002) 517e524.
[10] A. Diaspro, D. Silvano, S. Krol, O. Cavalleri, A. Gliozzi, Single living cell
encapsulation in nano-organized polyelectrolyte shells, Langmuir 18 (2002)
5047e5050.
[11] C.A. Custo €dio, J.F. Mano, Cell surface engineering to control cellular in-
teractions, ChemNanoMat 2 (2016) 376e384.
[12] M.R. Haque, J.H. Jeong, Y. Byun, Combination strategy of multi-layered surface
camouflage using hyperbranched polyethylene glycol and immunosuppres-
sive drugs for the prevention of immune reactions against transplanted
porcine islets, Biomaterials 84 (2016) 144e156.
[13] G. Decher, Fuzzy nanoassemblies: toward layered polymeric multicomposites,
Science 277 (1997) 1232e1237.
[14] M. Matsusaki, K. Kadowaki, Y. Nakahara, M. Akashi, Fabrication of cellular
multilayers with nanometer-sized extracellular matrix films, Angew Chem.
Int. Ed. Engl. 46 (2007) 4689e4692.
[15] A. Nishiguchi, H. Yoshida, M. Matsusaki, M. Akashi, Rapid construction of
three-dimensional multilayered tissues with endothelial tube networks by
the cell-accumulation technique, Adv. Mater. 23 (2011) 3506e3510.
[16] M. Matsusaki, H. Ajiro, T. Kida, T. Serizawa, M. Akashi, Layer-by-layer as-
sembly through weak interactions and their biomedical applications, Adv.
Mater. 24 (2012) 454e474.
[17] M. Matsusaki, C.P. Case, M. Akashi, Three-dimensional cell culture technique
and pathophysiology, Adv. Drug Deliv. Rev. 74 (2014) 95e103.
[18] A. Matsuzawa, M. Matsusaki, M. Akashi, Construction of three-dimensional
liver tissue models by cell accumulation technique and maintaining their
metabolic functions for long-term culture without medium change, J. Biomed.
Mater. Res. A 103 (2015) 1554e1564.
[19] Y. Amano, A. Nishiguchi, M. Matsusaki, H. Iseoka, S. Miyagawa, Y. Sawa,
M. Seo, T. Yamaguchi, M. Akashi, Development of vascularized iPSC derived
3D-cardiomyocyte tissues by filtration Layer-by-Layer technique and their
application for pharmaceutical assays, Acta Biomater. 33 (2016) 110e121.
[20] Y. Tsukamoto, T. Akagi, F. Shima, M. Akashi, Fabrication of orientation-
controlled 3D tissues using a layer-by-layer technique and 3D printed a
thermoresponsive gel frame, Tissue Eng. C Meth. 23 (2017) 357e366.
[21] M. Matsusaki, K. Fujimoto, Y. Shirakata, S. Hirakawa, K. Hashimoto, M. Akashi,
Development of full-thickness human skin equivalents with blood and
lymph-like capillary networks by cell coating technology, J. Biomed. Mater.
Res. A 103 (2015) 3386e3396.
[22] T. Akagi, M. Nagura, A. Hiura, H. Kojima, M. Akashi, Construction of three-
dimensional dermo-epidermal skin equivalents using cell coating technol-
ogy and their utilization as alternative skin for permeation studies and skin
irritation tests, Tissue Eng Part A 23 (2017) 481e490.
[23] K. Sasaki, T. Akagi, T. Asaoka, H. Eguchi, Y. Fukuda, Y. Iwagami, D. Yamada,
T. Noda, H. Wada, K. Gotoh, K. Kawamoto, Y. Doki, M. Mori, M. Akashi, Con-
struction of three-dimensional vascularized functional human liver tissue
using a layer-by-layer cell coating technique, Biomaterials 133 (2017)
263e274.
[24] A. Calabrese, D. Caton, P. Meda, Differentiating the effects of Cx36 and E-
 Y. Fukuda et al. / Biomaterials 160 (2018) 82e91
cadherin for proper insulin secretion of MIN6 cells, Exp. Cell Res. 294 (2004)
379e391.
[25] A.C. Hauge-Evans, P.E. Squires, S.J. Persaud, P.M. Jones, Pancreatic beta-cell-to-
beta-cell interactions are required for integrated responses to nutrient stim-
uli: enhanced Ca2þ and insulin secretory responses of MIN6 pseudoislets,
Diabetes 48 (1999) 1402e1408.
[26] A. Calabrese, M. Zhang, V. Serre-Beinier, Connexin 36 controls synchroniza-
tion of Ca2þ oscillations and insulin secretion in MIN6 cells, Diabetes 52
(2003) 417e424.
[27] M.J. Carvell, P.J. Marsh, S.J. Persaud, P.M. Jones, E-cadherin interactions
regulate beta-cell proliferation in islet-like structures, Cell. Physiol. Biochem.
20 (2007) 617e626.
[28] A. Matsuzawa, M. Matsusaki, M. Akashi, Effectiveness of nanometer-sized
extracellular matrix layer-by-layer assembled films for a cell membrane
coating protecting cells from physical stress, Langmuir 29 (2013) 7362e7368.
[29] H.Y. Lin, C.C. Tsai, L.L. Chen, S.H. Chiou, Y.J. Wang, S.C. Hung, Fibronectin and
laminin promote differentiation of human mesenchymal stem cells into in-
sulin producing cells through activating Akt and ERK, J. Biomed. Sci. 17 (2010)
56.
[30] M. Riopel, R. Wang, Collagen matrix support of pancreatic islet survival and
function, Front Biosci. (Landmark Ed). 19 (2014) 77e90.
[31] Z. Su, J. Xia, W. Shao, Y. Cui, S. Tai, H. Ekberg, M. Corbascio, J. Chen, Z. Qi, Small
91
islets are essential for successful intraportal transplantation in a diabetes
mouse model, Scand. J. Immunol. 72 (2010) 504e510.
[32] R. Lehmann, R.A. Zuellig, P. Kugelmeier, P.B. Baenninger, W. Moritz, A. Perren,
P.A. Clavien, M. Weber, G.A. Spinas, Superiority of small islets in human islet
transplantation, Diabetes 56 (2007) 594e603.
[33] Y. Guo, C. Martinez-Williams, C.E. Yellowley, H.J. Donahue, D.E. Rannels,
Connexin expression by alveolar epithelial cells is regulated by extracellular
matrix, Am. J. Physiol. Lung Cell Mol. Physiol. 280 (2001) L191eL202.
[34] H. Ishihara, T. Asano, K. Tsukuda, H. Katagiri, K. Inukai, M. Anai, M. Kikuchi,
Y. Yazaki, J.I. Miyazaki, Y. Oka, Pancreatic beta cell line MIN6 exhibits char-
acteristics of glucose metabolism and glucose-stimulated insulin secretion
similar to those of normal islets, Diabetologia 36 (1993) 1139e1145.
[35] E. Yoshihara, Z. Wei, C.S. Lin, S. Fang, M. Ahmadian, Y. Kida, T. Tseng, Y. Dai,
R.T. Yu, C. Liddle, A.R. Atkins, M. Downes, R.M. Evans, ERRg is required for the
metabolic maturation of therapeutically functional glucose-responsive b cells,
Cell Metabol. 23 (2016) 622e634.
[36] E. Bader, A. Migliorini, M. Gegg, N. Moruzzi, J. Gerdes, S.S. Roscioni, M. Bakhti,
E. Brandl, M. Irmler, J. Beckers, M. Aichler, A. Feuchtinger, C. Leitzinger,
H. Zischka, R. Wang-Sattler, M. Jastroch, M. Tscho € p, F. Machicao, H. Staiger,
H.U. H€aring, H. Chmelova, J.A. Chouinard, N. Oskolkov, O. Korsgren, S. Speier,
H. Lickert, Identification of proliferative and mature b-cells in the islets of
Langerhans, Nature 535 (2016) 430e434.