MULTI-ANALYTE DETECTION FOR BIOLOGICAL FLUIDS

Towards Continous Monitoring of Glucose, Ionized Calcium and pH

using a Viscometric Affinity Biosensor

Christophe Boss, Eric Meurville, Peter Ryser

Laboratory of Microengineering for Manufacturing, EPFL, Lausanne, Switzerland

Frédéric Schmitt, Lucienne Juillerat-Jeanneret

Pathology Institute, University Hospital, Lausanne, Switzerland

Pablo Dosil-Rosende, Desdemona De Souza

Baxter Healthcare, Zurich, Switzerland

Keywords: Affinity sensor, Chemico-mechanical sensor, Glucose sensing, Concanavalin A, Dextran, Viscosity.

Abstract: We present a viscometric affinity biosensor that can potentially allow continuous multi-analyte monitoring

in biological fluids like blood or plasma. The sensing principle is based on the detection of viscosity

changes of a polymeric solution which has a selective affinity for the analyte of interest. The chemico-

mechanical sensor incorporates an actuating piezoelectric diaphragm, a sensing piezoelectric diaphragm and

a flow-resisting microchannel for viscosity detection. A free-standing Anodic Alumina Oxide (AAO)

porous nano-membrane is used as selective interface. A glucose-sensitive sensor was fabricated and

extensively assessed in buffer solution. The sensor reversibility, stability and sensitivity were excellent

during at least 65 hours. Results showed also a good degree of stability for a long term measurement (25

days). The sensor behaviour was furthermore tested in fetal bovine serum (FBS). The obtained results for

glucose sensing are very promising, indicating that the developed sensor is a candidate for continuous

monitoring in biological fluids. Sensitive solutions for ionized calcium and pH are currently under

development and should allow multi-analyte sensing in the near future.

1 INTRODUCTION

Continuous detection overtime of analyte levels in

complex biological fluids such as blood or plasma is

a tricky task since strong interferences from other

biomolecules may occur during the measurements.

A successful system for continuous monitoring of

physiologically relevant parameters would afford

great benefits in numerous pathologies such as in the

thrombosis with the assessment of blood calcium

levels, in diabetes with glucose concentrations or pH

in acidosis/alkalosis disorders. In addition, these

parameters are also of primary importance for

critically ill patients. Today, these controls are

performed by hand and continuous monitoring could

contribute to reduce the risk of mortality in the

intensive care units.

Existing detection methods are currently based

on electrochemical principles, which have

limitations for in vivo monitoring. Electrochemical

measurements depend on the analyte diffusion rate.

Consequently, biofouling affects the sensitivity and

frequent calibrations are required. Furthermore, the

presence of other electrochemically active solutes

often produces inaccuracies. An alternative approach

which could overcome these limitations is affinity

sensing. Affinity sensing is more tolerant to

biofouling, which results only in an increased

stabilization time, and is intrinsically not subjected

to electroactive interferences. For these reasons,

intensive investigations on affinity binding sensors

have been carried out using different technics such

as fluorescence (Ballerstadt, 2004) or viscosity

measurements (Huang, 2009). Recently, hydrogel-

based sensors have emerged as promising materials

295

Boss C., Meurville E., Ryser P., Schmitt F., Juillerat-Jeanneret L., Dosil-Rosende P. and De Souza D..

MULTI-ANALYTE DETECTION FOR BIOLOGICAL FLUIDS - Towards Continous Monitoring of Glucose, Ionized Calcium and pH using a Viscometric

Afﬁnity Biosensor.

DOI: 10.5220/0003350902950298

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2011), pages 295-298

ISBN: 978-989-8425-37-9

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

for affinity sensing (Tierney, 2009). Despite many

efforts towards the development of continuous

biosensors for medical and biological applications,

long term reversibility and stability remains a

challenge.

In this context, we propose a novel chemico-

mechanical method which aims at detecting

viscosity changes of a solution which has a selective

affinity for the analyte of interest (Fig. 1). A semi-

permeable membrane ensures that the analyte

concentration in the biosensor is the same as in the

patient blood or plasma. The viscosity detection of

the sensitive solution is based on a microchannel

which exhibits a resistance to the flow circulating

through it. The sinusoidal actuation of a

piezoelectric diaphragm generates a flow through

the microchannel which results in a deflection of the

sensing piezoelectric diaphragm, inducing a voltage

which can be recorded. The phase shift between the

applied voltage and the sensing piezoelectric

diaphragm deflection is a measurement of the fluid

viscosity.



Analyte

Actuating

diaphragm

Sensing diaphrag

Free-standing AAO

porous nano-membrane

Sensitive solution

Microchannel

Anti-biofouling coating

Figure 1: Schematic illustration of the biosensor principle.

(Cross view).

A proof of concept of continuous glucose

monitoring was previously reported using a

macroscopic demonstrator (Boss, 2009). The present

paper investigates the sensitivity, reversibility and

reproducibility in a buffer solution of a novel

glucose sensor fabricated by stereolithography. The

sensor behavior in biological fluids was also

assessed using fetal bovine serum (FBS). The

glucose-sensitive fluid used inside the sensor was

based on Concanavalin A (ConA), a protein which

specifically binds to glucose, and a high-molecular-

weight dextran. When glucose concentration

increases, dextran is partially replaced by glucose at

the binding sites of ConA. As a result, the network

ConA-dextran is weakened, and the viscosity of the

sensing fluid decreases.

2 MATERIALS AND METHODS

2.1 Sensor Fabrication

The sensor was fabricated by stereolithography

using a biocompatible resin specially dedicated to

medical applications (Proform). The actuating

diaphragm was made up of a 50 μm thick and 3 mm

in diameter lead zirconate titanate (PZT) disc

(Audiowell Electronics) glued on a 10 μm thick

brass foil (Goodfellow). The sensing diaphragm was

made up of a 28 μm thick and 3 mm in diameter

Polyvinylidene fluoride (PVDF) disc (Measurement

Specialties) glued on a 10 μm thick brass foil. The

semi-permeable membrane was a 50 μm thick

Anodic Aluminium Oxide (AAO) membrane with

4-6 nm in diameter pores (Synkera Technologies). A

100×100 μm

in section glass capillary was used as

microchannel. The sensor assembly was realized

using medical adhesive epoxy (Loctite M-21HP).

The sensor was 200 μm thick and the volume of

sensitive fluid encapsulated inside was 4 μL. The

sensitive fluid was prepared using the protocol

described by Kuenzi et al. (2000). The sensitive

fluid was composed of 2% [w/w] dextran 3200

(PSS) and 0.4% [w/w] ConA (Sigma). The viscosity

of the sensitive fluid ranged from 5.9-16.7 mPas

(30-2 mM glucose) at 25°C and from 4.2-9.4 mPas

(30-2 mM) at 37°C. A low viscosity was selected to

keep the glucose diffusion as fast as possible.

0 10203040506070

Phase shift [deg]

Time [hour]

Figure 2: Phase shift response to multiple glucose concentrations (2, 6, 12, 20 mM). Measurement performed in reference

solution at 25°C.

20 mM

12 mM

6 mM

2 mM

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

296

0 5 10 15 20 25

Phase shift [deg]

Time [day]

Figure 3: Long term glucose concentration measurement in reference solution at 37°C, between two physiologically

relevant glucose concentrations, 2 mM and 12 mM.

2.2 Experimental Setup

Glucose measurements were performed in a

reference solution which is an isotonic solution of

the sensitive fluid, but without dextran and ConA.

Stock solutions with different glucose concentrations

were subsequently pumped into the test cell using a

computer controlled syringe pump. The whole setup

was located in a thermally regulated chamber

(±0.01°C), as the viscosity of the sensitive fluid is

strongly temperature dependant.

Fetal bovine serum (FBS) was preserved frozen.

Before use, 0.1% of sodium azide (NaN

) was added

as preservative. FBS was heated at 56°C during

45 min to inactivate the complement system. FBS

was then filtered (Millex HV 0.45 μm syringe filter)

to remove aggregates which may clog the semi-

permeable membrane. The glucose concentration of

FBS was measured using a standard glucose meter

(Accu-Chek). Concentrated (2M) D-glucose solution

was added to increase the FBS glucose

concentration.

3 RESULTS AND DISCUSSION

3.1 Reversibility of Glucose-induced

Viscosity Change

In Figure 2, a sensor is exposed successively to

increasing and decreasing glucose concentrations (2,

6, 12, 20 mM). Ten full cycles were performed

during 65 hours showing an excellent reversibility

and stability. The response time of the sensor (time

to reach 90% of the final value) was 4.8 min for

increasing glucose concentration and 19.3 min for

decreasing glucose concentration. The longer time

constant for decreasing glucose concentration is

explained by the smaller mobility of glucose

molecules in the viscous sensitive fluid than in the

reference solution. The semi-permeable membrane

porosity affects therefore more glucose molecules

diffusing out of the sensor than glucose molecules

entering inside the sensor. At this stage,

miniaturization was not the primary focus and the

response time is over the 10 min required from a

medical point of view. The response time could be

shorted by reducing the sensor thickness or sensitive

fluid viscosity. The sensor sensitivity in the

physiological and hypoglycemic ranges of glucose

concentration (2-6 mM) was 0.1 mM, which is

accurate enough for patients monitoring.

3.2 Long Term Stability

The long term stability of the sensor at 37°C was

investigated (Fig. 3). Glucose concentrations were

changed every 12 hours during 25 days. The sensor

showed a remarkable stability over time, but a

progressive loss of sensitivity was nevertheless

observed. After 25 days, the sensor sensitivity

dropped to 73% of the initial sensitivity. The loss of

sensitivity was likely due to ConA leakage through

the biggest pores and defects of the 4-6 nm pores of

the semi-permeable membrane. We are currently

assessing new AAO porous nano-membranes with

reduced pores size (2-4 nm) which should improve

the long term stability of the sensor.

3.3 Determination of Glucose in Fetal

Bovine Serum

The sensor behaviour in complex biological fluids

was evaluated using fetal bovine serum (FBS). Fig.

4 shows the sensor response to glucose variation (2.6

and 30 mM) in FBS at 37°C. Twelve full cycles

were performed during 24 hours, showing a good

stability. The response time did not increase with

time indicating that biofouling due to protein

adsorption on the semi-permeable membrane is not

an issue. AAO porous nano-membranes are

therefore well-suited as selective interface for

biosensors intended to be used in biological fluids.

2 mM

12 mM

MULTI-ANALYTE DETECTION FOR BIOLOGICAL FLUIDS - Towards Continous Monitoring of Glucose, Ionized

Calcium and pH using a Viscometric Affinity Biosensor

297

30.0

30.2

30.4

30.6

30.8

31.0

31.2

0 5 10 15 20 25

Time [hour]

Phase shift [deg]

Figure 4: Measurement in fetal bovine serum at 37°C between 2.6 mM and 20 mM.

The sensor sensitivity in FBS dropped to 10% of the

sensitivity in buffer solution. After FBS testing, the

sensor recovered its initial sensitivity in buffer

solution. The loss of sensitivity is therefore

reversible, which leads to the following hypothesis.

Small glycosylated peptides may enter inside the

sensor and competitively interfere with the binding

reaction of ConA to glucose. This hypothesis was

confirmed by conducting similar experiments with

dialysed FBS. A loss of sensitivity of 30% and 40%

were observed in 12 kDa and 3.5 kDa dialysed FBS,

respectively. The pores size should therefore be

significantly reduced to prevent glycosylated

peptides from entering the sensor. We are currently

assessing 2-4 nm AAO nano-membranes, coated

with Al

by atomic layer deposition, to minimize

the pores size.

4 CONCLUSIONS

A viscosity-based affinity sensor was developed for

continuous monitoring in biological fluids. The

sensor was extensively tested in buffer solution,

showing an excellent reproducibility and stability

over 65 hours at 25°C. The sensor sensitivity

matched well within the hypoglycemic and

physiological ranges (2-6 mM) with a resolution of

0.1 mM. The response time of the sensor was higher

for decreasing glucose concentration due to the

conjugated effect of both the reduced mobility of

glucose molecules in the viscous sensitive solution

and the membrane porosity. The sensor showed also

remarkable long term stability (25 days) at 37°C. A

limited loss of sensitivity was nevertheless observed,

which may be explained by ConA leakage through

defects of the AAO porous nano-membrane. In FBS,

the response time did not increase with time,

indicating that biofouling due to protein adsorption

is not an issue. The sensitivity in non-dialysed and

12 kDa dialysed FBS were 10% and 60% of the

sensitivity in buffer solution, respectively.

Glycosylated perptides may enter inside the sensor

and interfere with the ConA-glucose reaction. The

loss of sensitivity in FBS should be solved by

reducing the membrane pores size.

These measurements show good promise for the

sensor to be applied as in vivo monitoring system.

We are currently assessing Al

coated 2-4 nm

AAO nano-membranes for pores size reduction,

which should allow sensitive measurements in FBS.

Sensitive fluids for ionized calcium and pH are also

under development and should allow multi-analyte

sensing in the near future.

REFERENCES

Ballerstadt, R., Polak, A., Beuhler, A., Frye, J., 2004. In

Vitro long-term performance study of a near-infrared

fluorescence affinity sensor for glucose monitoring.

Biosensors & Bioelectronics. Elsevier.

Boss, C., Meurville, E., Sallese, J.M., Ryser, P., 2009.

Novel chemico-mechanical approach towards long-

term implantable glucose sensing. Procedia

Chemistry. Elsevier.

Huang, X., Li, S., Schultz, J., Wang, Q., Lin, Q., 2009. A

MEMS affinity glucose sensor using a biocompatible

glucose-responsive polymer. Sensor and Actuators B:

Chemical. Elsevier.

Huang, X., Li, S., Schultz, J., Wang, Q., Lin, Q., 2009. A

Capacitive MEMS Viscometric Sensor for Affinity

Detection of Glucose. J Micromech Microeng. IEEE.

Kuenzi, S., Meurville, E., Ryser, P., 2010. Automated

characterization of dextran/concanavalin A mixtures –

A study of sensitivity and temperature dependence at

low viscosity as basis for an implantable glucose

sensor. Sensor and Actuators B: Chemical. Elsevier.

Tierney, S., Falch, B., Hjelme, D., Stokke, B., 2009.

Determination of Glucose Levels Using a

Functionalized Hydrogel-Optical Fiber Biosensor:

Toward Continuous Monitoring of Blood Glucose in

Vivo. Anal. Chem. ACS.

20 mM

2.6 mM

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

298