of 16
ENGINEERING
A stretchable
wireless
wearable
bioelectr
onic
system for
multiple
xed monitoring
and combina
tion treatment
of
infected
chronic
wounds
Ehsan
Shirzaei
Sani
1
, Changhao
Xu
1
, Canr
an Wang
1
, Yu Song
1
, Jihong
Min
1
, Jiaobing
Tu
1
,
Samuel
A. Solomon
1
, Jiahong
Li
1
, Jaminelli
L. Banks
2
, David G. Arms
trong
2
, Wei Gao
1
*
Chronic
nonhealing
wounds
are one of the major
and rapidly
growing
clinical
complica
tions
all over the world.
Current therapies
frequently
requir
e emergent
surgical
interv
entions,
while
abuse
and misapplica
tion of ther-
apeutic
drugs
often
lead
to an increased
morbidity
and mortality
rate. Here, we introduce
a wearable
bioelec-
tronic
system that wirelessly
and continuously
monitors
the physiological
conditions
of the wound
bed via a
custom-dev
eloped
multiple
xed multimodal
electr
ochemical
biosensor
array and performs
noninvasiv
e combi-
nation therapy through
contr
olled
anti-inflamma
tory antimicr
obial
treatment
and electrically
stimula
ted tissue
regener
ation.
The wearable
patch is fully
biocompa
tible,
mechanically
flexible,
stretchable,
and can conformally
adher
e to the skin wound
throughout
the entire healing
process.
Real-time
metabolic
and inflamma
tory mon-
itoring
in a series
of preclinical
in vivo experiments
showed high
accuracy and electr
ochemical
stability
of the
wearable
patch for multiple
xed spatial and tempor
al wound
biomark
er analy
sis. The combina
tion
therapy
enabled
substantially
acceler
ated cutaneous
chronic
wound
healing
in a rodent
model.
Copyright
© 2023 The
Authors,
some
rights
reserved;
exclusive licensee
American
Associa
tion
for the Advancement
of Science.
No claim to
original
U.S. Government
Works. Distributed
under
a Creative
Commons
Attribution
License
4.0 (CC BY).
INTR
ODUCTION
Chronic
wounds
are characterized
by impair
ed or stagnant
healing,
prolonged
and uncontr
olled
inflamma
tion,
as well as compr
omised
extracellular
matrix (ECM)
function
(
1
3
).
Over 6.7 million
people
in the United
States alone
suffer
from chronic
nonhealing
wounds
including
diabetic
ulcers,
nonhealing
surgical
wounds,
burns,
and
venous-r
elated ulcer
ations
(
4
,
5
), causing
a staggering
financial
burden
of over $25 billion
per year on the health
care system (
6
).
Chronic
wound
healing
is a highly
comple
x biological
process
con-
sisting
of four
integr
ated and overlapping
phases:
hemos
tasis,
in-
flamma
tion,
prolifer
ation,
and
remodeling
(
1
3
).
Curr
ent
therapies
including
skin
grafts, skin
substitutes,
negative pressur
e
wound
therapy, and
others
can
be beneficial
but
frequently
requir
e procedur
es or surgical
interv
ention
(
7
). Micr
obial
infection
at the wound
site can severely prolong
the healing
process
and lead
to necrosis,
sepsis,
and even death (
3
). Both
topical
and systemic
antibiotics
are increasingly
prescribed
to patients
suffering
from
chronic
nonhealing
wounds,
but the overuse,
abuse,
and misappli-
cation
of antibiotics
often
lead
to an escala
ting drug
resistance
in
bacteria,
causing
a drastic increase
in morbidity
and
mortality
rates (
8
). As an alterna
tive therapeutic
approach, electrical
stimula-
tion
has shown to have a substantial
effect
on the wound
healing
process,
including
stimula
ting fibroblas
t prolifer
ation and differ
en-
tiation
into
myofibr
oblas
ts and collagen
forma
tion,
keratinocyte
migr
ation,
angiogenesis,
and
attracting
macrophages
(
9
,
10
).
However, currently
reported
electrical
stimula
tion
devices
usually
requir
e bulky
equipment
and
wire connections,
making
them
highly
challenging
for practical
clinical
use. More effectiv
e, fully
contr
ollable,
and
easy-to-implement
therapies
are critically
needed
for personalized
treatment
of chronic
wounds
with
minimal
side effects.
At each stage of healing
process,
the chemical
composition
of the
wound
exuda
te changes
substantially
, indica
ting the stage of healing
and even the presence
of an infection
(
11
13
).
For example,
in-
creased
temper
ature is associa
ted with
bacterial
infection,
and
changes
in temper
ature can
provide
informa
tion
on various
factors
relevant
to healing,
inflamma
tion,
and oxygena
tion
in the
wound
bed; acidity
(pH)
indica
tes a healing
state with
balanced
pro-
tease
activities
and effectiv
e ECM
remodeling,
moreover, eleva
ted
pH in wound
envir
onment
can be a sign
of infection;
eleva
ted
uric
acid (UA)
indica
tes wound
severity
with
excessiv
e reactive
oxygen
species
and inflamma
tion
and shows immune
system re-
sponding
to inflamma
tory
cytokines
(
14
);
lactate and ammonium
are crucial
mark
ers for soft-tissue
infection
diagnosis
and angiogen-
esis in diabetic
foot ulcers
(
15
);
wound
exuda
te glucose
has a strong
correlation with
blood
glucose
and bacterial
activities
(
16
),
provid-
ing crucial
therapeutical
guidance
for clinical
diabetic
wound
treatment.
Recent
advances
in digital
health
and flexible
electr
onics
have
transformed
conv
entional
medicine
into
remote
at-home
health
care (
17
23
).
Wearable
biosensors
could
allow real-time
and con-
tinuous
monitoring
of physical
vital
signs
and physiological
bio-
mark
ers in various
biofluids
such
as sweat, saliva,
and inters
titial
fluids
(
18
21
,
24
30
).
In gener
al, an ideal
wound
dressing
should
provide a mois
t wound
envir
onment,
offer
protection
from second-
ary infections,
remove wound
exuda
te, and promote
tissue
regener-
ation.
Despite
the promising
prospects
opened
by the wearable
technologies
(
31
37
),
major
challenges
exist to realize
their
full po-
tential
toward practical
chronic
wound
management
applica
tions:
the chronic
nonhealing
wounds
pose
high
requir
ement
on the flex-
ibility
, breathability
, and biocompa
tibility
of the wearable devices
to
protect
the wound
bed from bacterial
infiltr
ations
and infection
and
1
Andrew and Peggy Cherng
Department
of Medical
Engineering,
Division
of Engi-
neering
and Applied
Science,
California
Institute
of Technology
, Pasadena,
CA
91125,
USA.
2
Keck School
of Medicine,
University
of Southern
California,
Los
Angeles,
CA 90033,
USA.
These
authors
contributed
equally
to this work.
*Corresponding
author.
Email:
weigao@caltech.edu.
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
1 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
modula
te wound
exuda
te level; the comple
x wound
exuda
te matrix
could
substantially
affect
the biosensor
performance,
and thus,
there are few reports
on prolonged
evalua
tion
of biosensors
in
vivo (
13
,
31
); personalized
wound
management
demands
both
ef-
fectiv
e wound
therapy and
close
monitoring
of crucial
wound
healing
biomark
ers in the wound
exuda
te; the absence
of minia
tur-
ized
user-inter
active fully
integr
ated closed-loop
wearable
systems
and the evalua
tion
of such
systems
in vivo impede
their
practi-
cal use.
To address these
challenges,
here, we introduce
a fully
integr
ated
wireless
wearable bioelectr
onic
system that effectiv
ely monitors
the
physiological
conditions
of the wound
bed via multiple
xed and
multimodal
wound
biomark
er analy
sis and performs
combina
tion
therapy through
electr
o-responsiv
e contr
olled
drug
deliv
ery for
anti-inflamma
tory
antimicr
obial
treatment
and exogenous
electri-
cal stimula
tion for tissue
regener
ation (Fig.
1, A and B). The wear-
able patch is mechanically
flexible,
stretchable,
and can conformally
adher
e to the skin
wound
throughout
the entir
e wound
healing
process,
preventing
any undesir
ed discomfort
or skin
irrita
tion.
Because
of the wound
s
comple
x pathophy
siological
envir
onment,
compar
ed to previously
reported
single-analyte
sensing,
multiple
x-
ing analy
sis of wound
exuda
te biomark
ers can provide
more com-
prehensiv
e and
personalized
informa
tion
for effectiv
e chronic
wound
management.
In this regard,
a panel
of wound
biomark
ers
including
temper
ature, pH, ammonium,
glucose,
lactate, and UA
were chosen
on the basis
of their
importance
in reflecting
the infec-
tion,
metabolic,
and inflamma
tory
status of the chronic
wounds.
Real-time
selectiv
e monitoring
of these
biomark
ers in comple
x
wound
exuda
te could
be realized
in situ using
custom-engineer
ed
electr
ochemical
biosensor
arrays (Fig.
1C).
The wearable
system
s
capabilities
of multiple
xed monitoring,
biomark
er mapping,
and
combina
tion
therapy
were evalua
ted in vivo over prolonged
periods
of time
in rodent
models
with
infected
diabetic
wounds.
The multiple
xed biomark
er informa
tion
collected
by the wearable
patch revealed
both
spatial and tempor
al changes
in the microenvi-
ronment
as well as inflamma
tory
status of the infected
wound
during
differ
ent healing
stages.
In addition,
the combina
tion
of
electrically
modula
ted antibiotic
delivery with
electrical
stimula
tion
on the wearable
technology
enabled
substantially
acceler
ated
chronic
wound
closur
e.
RESUL
TS
Design
of the fully
integr
ated stretchable
wearable
bioelectr
onic
system
The disposable
wearable
patch consis
ts of a multimodal
biosensor
array for simultaneous
and multiple
xed electr
ochemical
sensing
of
wound
exuda
te biomark
ers, a stimulus-r
esponsiv
e electr
oactive hy-
drogel
loaded
with
a dual-function
anti-inflamma
tory and antimi-
crobial
peptide
(AMP),
as well as a pair
of voltage-modula
ted
electr
odes
for contr
olled
drug
release
and electrical
stimula
tion
(Fig.
1, B and C). The multiple
xed sensor
array patch is fabrica
ted
via standard
micr
ofabrica
tion
protocols
on a sacrificial
layer of
copper
followed by transfer
printing
onto
a poly[s
tyrene-
b
-(ethyl-
ene-
co
-butylene)-
b
-s
tyrene]
(SEBS)
thermoplas
tic elastomer
sub-
strate (figs.
S1 and S2).
The
serpentine-lik
e design
of electr
onic
inter
connects,
and the highly
elastic nature of SEBS
enables
high
stretchability
and resilience
of the sensor
patch agains
t undesir
able
physical
deforma
tions
(Fig.
1, D and
E). The
flexible
bandage
seamlessly
interfa
ces with
a flexible
printed
circuit board
(FPCB)
for electr
ochemical
sensor
data acquisition,
wireless
communica-
tion,
and programmed
voltage
modula
tion for contr
olled
drug
de-
livery and electrical
stimula
tion (Fig.
1, F to H, and figs. S3 to S5).
The wireless
wearable device
can be attached
to the wound
area with
firm
adhesion,
allowing
the animals
to move freely over a prolonged
period
(movie S1 and figs. S6 and S7).
Design
and characteriza
tion
of the soft sensor
array for
multiple
xed biomark
er analy
sis
The array of flexible
biosensors
was custom developed
to allow real-
time
multiple
xed monitoring
of the biomark
ers in comple
x wound
exuda
te. The
continuous
and selectiv
e measur
ement
of glucose,
lactate, and UA is based
on amper
ometric
enzyma
tic electr
odes
with
glucose
oxidase,
lactate oxidase,
and uricase
immobilized
in
a highly
permeable,
adhesiv
e, and biocompa
tible
chitosan
film,
re-
spectiv
ely (Fig.
2A).
Electr
odeposited
Prussian
blue
(PB)
serves as
the electr
on-tr
ansfer
redox
media
tor for the enzyma
tic reaction,
which
allows the biosensors
to operate at a low potential
(~0.0
V)
to minimize
the interfer
ences
of oxygen
and other
electr
oactive
molecules.
Because
of the comple
x and heter
ogeneous
composition
of wound
fluid
(e.g.,
high
protein
levels, local
and migr
ated cells,
and exogenous
factors
such
as bacteria)
(
13
),
previously
reported
enzyma
tic sensors
suffer
from severe matrix effects
and fail to accu-
rately measur
e the target
metabolite
levels in untreated wound
fluid
(figs.
S8 and S9 and note
S1). Moreover, high
levels of metabolites
in
diabetic
wound
fluid,
especially
glucose
(up to 50 mM),
pose
another
major
challenge
to obtain
linear
sensor
response
in the
physiological
concentr
ation
ranges.
To address these
issues
and
achiev
e accurate wound
fluid
metabolic
monitoring,
increase
sensor
range,
and minimize
biofouling
effects,
we explor
ed the
use of an outer
porous membr
ane that serves as a diffusion
limiting
layer to protect
the enzyme,
tune
response,
increase
operational
stability
, as well as enhance
the linearity
and sensitivity
magnitude
of the sensor.
We fabrica
ted our enzyma
tic glucose
oxidase/chito-
san/single-w
alled
carbon
nanotubes
(GOx/CS/MWCNT)
glucose
sensor
with
additional
porous membr
ane coatings
including
CS,
poly(ethylene
glycol)
diglycidyl
ether
(PEGDGE),
Nafion,
and poly-
urethane
(PU)
(fig. S9). As expected,
the addition
of diffusion
layers
indeed
impr
oves the sensor
s linear
range
in simula
ted wound
fluid
(SWF).
However, CS-,
PEGDGE-,
and Nafion-coa
ted sensors
did
not show reliable
responses
in wound
fluid
upon
the addition
of
glucose.
The PU-based
enzyma
tic sensors
showed the highes
t line-
arity
over the wide
physiological
concentr
ation range
as well as high
reproducibility
in comple
x wound
fluid
matrix (fig. S10).
The am-
perometric
current signals
gener
ated from the PU-coa
ted enzyma
tic
glucose,
lactate, and UA sensors
are proportional
to the physiolog-
ically
relevant
concentr
ations
of the corresponding
metabolites
in
SWF
with
sensitivities
of 16.34,
41.44,
and 189.60
nA mM
1
, respec-
tively (Fig.
2, B to D). Continuous
monitoring
of ammonium
is
based
on a potentiometric
ion-selectiv
e electr
ode
wher
e the
binding
of ammonium
with
its ionophor
e results
in an electr
ode
potential
log-linearly
corresponding
to the target
ion concentr
ation
with
a sensitivity
of 59.7
mV decade
1
(Fig.
2, E and F). Similarly
,
the pH sensor
uses
an electr
odeposited
polyaniline
film as the pH-
sensitiv
e membr
ane and shows a sensitivity
of 59.7
mV per pH
(Fig.
2G).
For all chemical
sensors,
a polyvinyl
butyr
al (PVB)
coated Ag/AgCl
electr
ode was used
as the reference
electr
ode that
provides
a stable
voltage
independent
of the varia
tions
of wound
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
2 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
fluid
compositions
(
24
).
A gold
microwire-based
resistive temper
a-
ture sensor
is integr
ated as part of the sensor
array and shows a sen-
sitivity
of approxima
tely
0.21%
°C
1
in the
physiological
temper
ature range
of 25° to 45°C
(Fig.
2H).
Considering
that other
electr
olytes
and metabolites
present
in
wound
fluid
may negatively affect
the sensor
outputs,
we examined
the selectivity
of the sensor
array consis
ting
of all six sensors.
As
illustrated in Fig. 2I, the addition
of nontarget
electr
olytes
and me-
tabolites
did not trigger
any substantial
interfer
ence
to the sensor
response.
Moreover, all biosensors
showed high
selectivity
over
nonspecific
compounds
when
evalua
ted in SWF
(fig.
S11).
It
should
be noted
that while
temper
ature has negligible
effects
on
the potentiometric
sensors,
it substantiallyinfluences
the perfor-
mance
of the enzyma
tic sensors
due to the temper
ature-dependent
enzyme
activities
(fig.
S12).
Moreover, our data show that the
medium
pH could
also
affect
the performance
of enzyma
tic
sensors
(fig.
S13).
With pH and temper
ature sensors
integr
ated
into
the wearable
patch, we are able
to perform
real-time
adjus
t-
ments
and calibr
ation of the enzyma
tic biosensors
based
on temper-
ature and
pH varia
tions
to realize
accurate wound
metabolite
analy
sis.
Fig. 1. Awireless
stretchable
wearable
bioelectr
onic
system for multiple
xed monitoring
and treatment
of chronic
wounds.
(
A
) Schema
ticof a soft wearable patch
on an infected
chronic nonhealing
wound
on a diabetic
foot. (
B
) Schema
tic of layer assembly
of the wearable patch, showing the soft and stretchable
poly[styrene-
b
-
(ethylene-
co
-butylene)-
b
-s
tyrene] (SEBS)
substrate, the custom-engineer
ed electrochemical
biosensor
array, a pair of voltage-modula
ted electrodes for controlled drug
release
and electrical
stimula
tion, and an anti-inflamma
tory and antimicr
obial drug-loaded
electroactive hydrogel layer. (
C
) Schema
tic layout of the smart
patch con-
sisting of a temper
ature (T) sensor,
pH, ammonium
(NH
4
+
), glucose
(Glu), lactate (Lac), and UA sensing
electrodes, reference (Ref) and counter
electrodes, and a pair of
voltage-modula
ted electrodesforcontrolled drug release
and electrical
stimula
tion. (
D
and
E
) Photogr
aphs of the fingertip-sized
stretchable
and flexiblewearable patch.
Scale bars, 1 cm. (
F
and
G
) Schema
tic diagram (F) and photogr
aph (G) of the fully integrated miniaturized
wireless wearable patch. Scale bar, 1 cm. ADC, analog
to digital
converter; AFE, analog
front end; PSoC,programmable
system on chip;MUX, multiplex
er;BLE, Bluetooth
Low Energy
.(
H
) Photogr
aph of afully integrated wearable patch
on a diabetic
rat with an open wound.
Scale bar, 2 cm.
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
3 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
Owing
to the soft SEBS
substrate and the serpentine-lik
e design
of electr
onic
interconnects,
the wound
patch showed excellent
me-
chanical
flexibility
and stretchability
, which
are essential
to main-
taining
good
conta
ct with
the skin
in vivo during
the chronic
wound
healing
process.
Negligible
alterations
in the sensor
respons-
es befor
e and under
unidir
ectional
tensile
stretching
(Fig.
2I) and
after repetitiv
e mechanical
bending
(fig. S14)
were observ
ed, indi-
cating highly
consis
tent sensor
performance
under
various
physical
deforma
tions.
As the sensor
patch is designed
for long-term
in vivo use, its cy-
tocompa
tibility
and biocompa
tibility
are of great importance.
Cell
viability
and metabolic
activity
of the cells
seeded
on a multiple
xed
sensor
array were analyzed
using
a commer
cial live/dead
kit and
PrestoBlue
assay, respectiv
ely (Fig.
2, K to N, and fig. S15).
The
high
cell viabilities
shown in the representa
tive live/dead
staining
images
of human
dermal
fibroblas
ts (HDFs)
and normal
human
epidermal
keratinocytes
(NHEK)
cells
(Fig.
2, K to M, and fig.
S15),
along
with
the consis
tently
increased
cell metabolic
activities
Fig. 2. Design
and characteriza
tion of the sensor
array for multiple
xed wound
analy
sis.
(
A
to
D
) Schema
tic(A)and chronoamper
ometric
responses
ofthe enzyma
tic
glucose
(B), lactate (C), and UA (D) sensors
in SWF. Insets
in (B) to (D), the calibration plots with a linear fit. PB, Prussian
blue; Sub, substrate; Prod, product;
CE, counter
electrode; WE, working
electrode; RE, reference electrode;
I
, current. (
E
and
F
) Schema
tic (E) and potentiometric
response
(F) of an NH
4
+
sensor
in SWF. Insets
in (F), the
calibrationplotwithalinearfit.ISE,ion-selectiv
eelectrode;PEDOT,poly(3,4-ethylenedioxythiophene);
U
, potential.
(
G
) Potentiometric
response
ofapolyaniline-based
pH
sensor
in McIlvaine
buffer.
Insets,
the calibration plot with a linear fit. (
H
) Resistive response
of an Au microwire
based
temper
ature sensor
under
temper
aturechanges
in
physiologically
relevant
range in SWF. Insets,
schema
tic of a temper
ature sensor
and the calibration plot with a linear fit. All error bars in (A) to (H) represent the SD from
three sensors.
(
I
) Selectivity
studyof the multiple
xed sensor
array in SWF. Ten millimolar
glucose,
50
μ
M UA, 1 mM lactate, and 1 mM NH
4
+
were added
sequentially
to the
SWF. (
J
) Responses
of the multiple
xed sensor
array before and during
mechanical
stretching
(15%) in SWF (pH 8) containing
10 mM glucose,
50
μ
M UA, 1 mM lactate, and
0.25 mM NH
4
+
. (
K
and
L
) Representa
tive live (green)/dead
(red) images
of human
dermal
fibroblasts (HDFs)
(K) and normal
human
epidermal
keratinocytes
(NHEKs)
(L)
cells seeded
on the multiplex
ed sensor
arrayand in PBS (control) after 1-dayand 7-day culture. Scale bars, 200
μ
m. (
M
and
N
) Quantita
tive analysis of cell viability
images
(M) and cell metabolic
activity
(N) over a 7-day period
after culture. RFUs, relative fluorescence
units. Error bars represent the SD (
n
= 4).
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
4 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
(Fig.
2N) over multida
y cultur
e periods,
indica
te the high
cytocom-
patibility
of the soft sensor
patch.
Char
acteriza
tion
of the therapeutic
capabilities
of the
wearable
patch in vitro
In addition
to the multiple
xed and multimodal
biosensing,
the
wearable
patch is able
to perform
combina
tion
treatment
of
chronic
wounds
through
drug
release
from an electr
oactive hydro-
gel layer and electrical
stimula
tion under
an exogenic
electric
field,
both
contr
olled
by a pair of voltage-modula
ted electr
odes
(Fig.
3, A
to C). The electr
oactive hydrogel
consis
ts of chondr
oitin
4-sulfa
te
(CS),
a sulfated glycosaminoglycan
composed
of units
of glucos-
amine,
cross-link
ed with
1,4-butanediol
diglycidyl
ether
(fig. S16).
Because
of the shear-thinning
behavior of the prepolymer
solution,
the hydrogel can be precisely
fabrica
ted via three-dimensional
(3D)
printing
(fig. S17).
The negatively charged
CS hydrogel
is an ideal
choice
for loading
and contr
olled
release
of positiv
ely charged
large
biological
drug
molecules
based
on an electrically
modula
ted
on/
Fig. 3. Char
acteriza
tion
of the therapeutic
capabilities
of the wearable
patch in vitro.
(
A
to
C
) Schema
tic illustration of the therapeutic
modules
of the wearable
patch (A) and the working
mechanisms
of the controlled drug delivery for antimicr
obial treatment
(B) and electrical
stimula
tion for tissue
regener
ation (C). (
D
) Loading
efficiency
of dual-functional
TCP-25
anti-inflamma
tory and AMP into CS electroactive hydrogel after 0.5- to 24-hour
incuba
tion. (
E
) Release
amount
of AMP from the
hydrogel under
programmed
on-off
electrical
voltage
(1 V, 10 min each step). (
F
) Long-term
cumula
tive release
of the AMP under
programmed
electrical
modula
tion. (
G
and
H
) In vitro antimicr
obial tests including
zone of inhibition
(G) and colony
forming
units (H) assays for electroactive hydrogels with and without
TCP-25
AMP agains
t
multidrug-r
esistant
Escherichia
coli
(MDR
E. coli
),
P. aeruginosa
,
and methicillin-r
esistant
Staphylococcus
aureus
(MRSA).
(
I
to
K
) In vitro cytocompa
tibility
assessment
of
TCP-25
loaded
electroactive hydrogels using live/dead
staining
(I) and quantifica
tion of cell viability
(J) and metabolic
activity
(K) for HDF and NHEK cells cultured in the
presence
of hydrogels.
Scale bar, 100
μ
m. (
L
and
M
) Fluorescence
images
(L) and quantita
tive wound
closure analysis (M) to evalua
te the wearable patch
s therapeutic
capability
viaelectrical
stimula
tionusinganinvitrocircularwound
healing
assaycreatedbyHDFcells.ES,electrical
stimulat
ion.Apulsed
voltage
wasapplied
forelectrical
stimula
tion (1 V at 50 Hz, 0.01 s voltage
on for each cycle).
Scale bar, 500
μ
m. (
N
) Numerical
simula
tion of the electrical
field generated by the custom-designed
electrical
stimula
tion electrodes during
operation.
E
, electrical
field. Scale bar, 500
μ
m. Error bars represent the SD (*
P
< 0.05, **
P
< 0.01, ***
P
< 0.001,
and ****
P
< 0.0001;
n
3). ns,
not significant.
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
5 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
off
drug
release
mechanism
(Fig.
3B).
Here, an AMP
, thrombin-
deriv
ed c-terminal
peptide-25
(TCP-25)
(
38
),
was loaded
within
the CS hydr
ogel
network
through
the electr
ostatic inter
actions
with
the polymer
backbone,
with
up to 15%
loading
efficiency
(Fig.
3D).
The highly
porous hydrogel
network
under
equilibrium
swelling
could
further
enhance
the drug
loading
efficiency
(fig.
S18).
Under
an applied
positiv
e voltage,
the electr
oactive hydrogels
will be rapidly
protona
ted, resulting
in anisotr
opic
and microscopic
contr
action
followed by syneresis/e
xpelling
of water from the gel
(
39
)
and consequently
allowing
a contr
olled
release
of the TCP-25
AMP
(Fig.
3, E and F, and figs. S19 and S20).
In addition,
the elec-
trical
field
will also
facilita
te the diffusion
of positiv
ely charged
AMP
out of the stimuli-sensitiv
e CS hydrogel
toward the cathode
due to electr
ophor
etic flow (
40
).
The antimicr
obial
activity
of the TCP-25
AMP
loaded
hydrogel
was evalua
ted agains
t Gram-positiv
e methicillin-r
esistant
Staphylo-
coccus
aureus
(MRSA)
and
Pseudomonas
aeruginosa
,
and Gram-
nega
tive multidrug-r
esistant
Escherichia
coli
(MDR
E. coli
)
and
Staphylococcus
epidermidis
,
the most common
pathogenic
bacteria
associa
ted with
micr
obial
coloniza
tion
of chronic
nonhealing
wounds
(Fig.
3, G and H, and fig. S21).
The
zone
of inhibition
assay indica
tes the susceptibility
of the MDR
E. coli
,
P. aeruginosa
,
and
MRSA
toward TCP-25
AMP
(Fig.
3G),
while
the standard
colony-forming
units
(CFU)
showed that the drug-loaded
hydrogel
was effectiv
ely protected
from all pathogenic
coloniza
tion (Fig.
3H).
For cells
cultur
ed on antimicr
obial
peptide
(AMP)-loaded
hydro-
gels,
the viability
of HDF
and NHEK
cells
remained
>90%,
and
their
metabolic
consis
tently
increased
during
the 7-day cultur
e
(Fig.
3, I to K, and fig. S22),
indica
ting
that the TCP-25
loaded
gels are highly
cytocompa
tible
and support
cell prolifer
ation.
The wearable
patch
s
therapeutic
capability
toward enhanced
tissue
regener
ation
via electrical
stimula
tion was assessed
using
an
in vitro wound
healing
assay (Fig.
3, L and M, and fig. S23).
The
model
wound
treated with
electrical
stimula
tion
showed substan-
tially
faster and more consis
tent
migr
ation
of HDF
cells
toward
the wound
area for four
consequent
days after wounding
as com-
pared to the
contr
ol group
without
electrical
stimula
tion
(Fig.
3L).
Quantita
tive analy
sis of the model
wound
closur
e indi-
cates higher
wound
closur
e rates in the wounds
treated with
electri-
cal stimula
tion
(Fig.
3M).
The
enhanced
tissue
regener
ation
is
attributed
to the directional
electrical
field
gener
ated from our
custom-designed
electrical
stimula
tion
electr
odes
(Fig.
3N),
which
plays a crucial
role in cell behavior
modula
tion
including
cell-cell
junctions,
cell division
orienta
tion,
and cell migration
trajectories
(galvanotaxis
or electr
otaxis)
(
41
43
).
The electrical
potential
was
applied
directly
to a pair of insula
ted electr
odes
to gener
ate electrical
field
for electrical
stimula
tion.
It should
also be noted
that contin-
uous
electrical
stimula
tion did not cause
substantial
temper
ature in-
crease
(fig. S24).
Evalua
tion
of the wearable
patch in vivo for multiple
xed
wound
biomark
er monitoring
To valida
te the capability
and effica
cy of our wearable patch, in vivo
preclinical
evalua
tions
are essential.
In this regard,
the in vivo bio-
compa
tibility
of the wearable patch was assessed.
The immunohis-
tofluor
escent
staining
of subcutaneously
implanted
hydrogel
and
electr
odes
in rats showed negligible
signs
of leukocyte
(CD3)
and
macrophage
(CD68)
antigens
after 56 days, indica
ting the high
bio-
compa
tibility
of the wearable patch (fig. S25).
All custom-dev
eloped
biosensors
on the wearable
patch displa
yed consis
tent
sensitivity
during
a 6-hour
continuous
measur
ement
in SWF
, indica
ting the
high
electr
ochemical
stability
of the sensors
for wound
analy
sis
(fig.
S26).
In vivo multiple
xed sensing
study
was then
performed
using
an infected
excisional
wound
model
in diabetic
mice.
The
wound
fluid
composition
was assessed
by the wearable
patch
befor
e infection
(day 1), after infection
(day 4), and after treatment
(day 7) (Fig.
4A).
Subs
tantially
eleva
ted UA,
temper
ature, pH,
lactate, and
ammonium
levels were observ
ed as compar
ed to
those
befor
e infection.
The increase
in temper
ature can be poten-
tially
linked to inflamma
tion
(
44
).
The eleva
ted levels of UA after
infection
can be due to up-regula
tion
of xanthine
oxidase,
a com-
ponent
of the innate immune
system responding
to inflamma
tory
cytokines
in chronic
ulcers
that plays a key role in purine
metabo-
lism to produce
UA (
45
).
pH, lactate, and ammonium
are all acidity
related, and their
eleva
tion
during
the bacteria
infection
has also
been
widely
reported
(
46
).
In contr
ast, the glucose
level in infected
wound
fluid
showed >35%
decrease
after infection,
attributing
to
the increased
glucose
consumption
of bacteria
activities
(
16
).
Upon
wound
treatment,
the temper
ature, pH, lactate, UA, and am-
monium
decreased
toward the levels befor
e the infection,
while
the
glucose
level increased
significantly
after treatment,
indica
ting the
successful
bacterial
elimina
tion (Fig.
4A).
Considering
that dietary
intak
e may have major
impa
ct on the
composition
of diabetic
wound
fluid,
we evalua
ted the metabolic
changes
in wound
fluid
in response
to tail vein glucose
adminis
tra-
tion
(Fig.
4B) and food
feeding
(Fig.
4C).
Glucose
adminis
tration
via tail vein into the 24-hour
fasted mice
spark
ed ~10 mM
increase
in the blood
glucose
level. The
in vivo sensor
readings
from the
wearable
patch were recorded
from 30 min
befor
e injection
and
continued
until
270 min
after injection
(Fig.
4B).
The
glucose
level in wound
fluid
showed a gradual
increase
throughout
the 4
hours
after injection,
indica
ting a protracted
delay with
respect
to
blood
glucose.
A similar
trend
was observ
ed for temper
ature
values,
attributing
to an increased
metabolic
rate to facilita
te diges-
tion.
No appar
ent change
in UA level after injection
was detected
because
of the absence
of purine
intak
e in the glucose
adminis
tra-
tion.
For the food
feeding
study, the wearable
patch was tested
befor
e fasting,
after 24-hour
fasting,
and
6 hours
after feeding
(Fig.
4C).
The lactate and ammonium
levels increased
substantially
after fasting,
while
glucose
and UA levels decreased
after fasting,
consis
tent with
the trend of observ
ed blood
level changes
(
47
).
In
the meantime,
temper
ature decreased
due to the fasting-induced
hypothermia
(
48
).
As expected,
6 hours
after feeding,
the glucose
and UA levels increased
from 11.9
to 20.3
mM
and from 45.9
to
60.3
μM,
respectiv
ely. These
results
indica
te that wearable
patch-
enabled
wound
fluid
analy
sis could
be a promising
approach to
realize
continuous
and personalized
metabolic
monitoring.
Spatial and tempor
al monitoring
of critical-sized
wounds
using
the wearable
patch
The wearable
patch is mass
producible
and readily
reconfigur
able
for various
wound
care applica
tions.
In the case
of large
chronic
ulcers,
the wound
parameters
and micr
oenvir
onment
may vary
from site to site, making
localized
monitoring
crucial
for optimized
assessment
and treatment
of chronic
wound
infection.
As a proof of
concept,
we demons
trate customized
wearable
patches
for spatial
mapping
of physiological
conditions
of critical-sized
wounds
during
the healing
process.
As illustrated in Fig. 5 (A and B), we
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
6 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE
could
incorpor
ate a sensor
array containing
seven pH sensors
and
nine
temper
ature sensors
onto
our wearable platform
for monitor-
ing and
mapping
critical-sized
full-thickness
infected
chronic
wounds
in diabetic
rats. The
pH and temper
ature sensor
arrays
showed high
reproducibility
and stability
in SWF
solutions
befor
e
and after in vivo applica
tion (Fig.
5, C and D, and fig. S27).
On-body
valida
tion of the sensor
array for spatial and tempor
al
wound
monitoring
was conducted
on critical-sized
full-thickness
wounds
(35 mm
in diameter)
in Zuck
er diabetic
fatty (ZDF)
rats
befor
e infection,
after infection,
and after treatment.
The dynamic
changes
in pH and temper
ature values
for each biosensor
on the
wearable
patch in noninfected
and infected
critical-sized
wounds
are illustrated in Fig. 5 (E and F). For noninfected
wound
studies,
the pH and temper
ature values
did not notably
change
over the 7-
day period.
However, for infected
wound
studies,
the pH and tem-
perature values
increased
daily
upon
applying
a mixed
infection
(MRSA
and
P. aeruginosa
)
on day 1 and reached
the peak
value
on days 3 and 4. Upon
treatment
on day 4, the pH and temper
ature
values
for each sensor
decreased
substantially
and recovered toward
the levels befor
e the infection
on day 7. The spatial mapping
plots
of
pH (Fig.
5G) and temper
ature (Fig.
5H) in the chronic
wound
area
on each day over the 7-day period
were successfully
gener
ated on
the basis
of localized
sensor
readings.
These
results
are in agreement
with
previous
literature on the changes
in the pH and temper
ature
values
during
the healing
progress (
46
).
A wide
varia
tion
was ob-
served in both
pH and
temper
ature in differ
ent regions
of the
wound
upon
bacterial
infiltr
ation
on day 2, showing
a higher
bac-
teria
growth in the wound
edges.
The
infected
wound
showed a
more uniform
pH and temper
ature at differ
ent regions
2 and 3
days after infection
due
to the forma
tion
of uniform
biofilm.
Upon
treatment,
the varia
tions
increased
in the treated wounds
on days 5 and 6, indica
ting the disruption
and eventually
elimina-
tion of the biofilm
after treatment
(Fig.
5, G and H).
Evalua
tion of the therapeutic
effica
cy of the wearable
patch
in chronic
wound
healing
in vivo
The
wearable
patch-fa
cilita
ted combina
tion
therapy and wound
healing
were evalua
ted in a splinted
excisional
wound
model
in
ZDF
diabetic
rats (Fig.
6A).
Four differ
ent groups
were tested: neg-
ative contr
ol, drug
release,
electrical
stimula
tion,
and combina
tion
therapy. The drug
treatment
was primarily
used
to elimina
te bacte-
rial infections
and regula
te immune
response
in early
stages
of
healing.
The
electrical
stimula
tion
was used
to facilita
te ion
channel
up-regula
tion
and redistribution,
resulting
in acceler
ated
cell migration and wound
healing.
The wearable patch
s
high
flexi-
bility
and stretchability
provided
intact and comfortable
conta
ct
with
the animal
s
back curva
ture. Over a 14-da
y period,
the
animals
were routinely
weighed
wher
e infected
rats showed a
Fig. 4. In vivo evalua
tion of the wearable
patch for multiple
xed wound
biomark
er monitoring
in a wound
model
in diabetic
mice.
(
A
) Invivo multiple
xed analysis
of the chemical
composition
of wound
fluid using awearable patch in an infected
excisional
wound
model
in a diabetic
mouse.
Infection
and treatment
were performed
after the sensor
recording
on days 1 and 4, respectiv
ely. (
B
) In vivo continuous
and multiple
xed evalua
tion of wound
parameters
in a 24-hour
fasted mouse
before and
after glucose
adminis
tration via tail vein. (
C
) In vivo assessment
of metabolic
changes
in wound
microenvironment
in response
to fasting and food feeding
in a diabet-
ic mouse.
Shirzaei
Sani
et al.
,
Sci. Adv.
9
, eadf7388
(2023)
24 March 2023
7 of 16
SCIENCE
ADVANCES
|
RESEARCH
ARTICLE