Physiol. Meas. 29 (2008) 1247–1254
Retrievable micro-inserts containing oxygen sensors
for monitoring tissue oxygenation using EPR oximetry
M Dinguizli, N Beghein and B Gallez1
Louvain Drug Research Institute, Biomedical Magnetic Resonance unit, Avenue Mounier 73,
Universit´ e catholique de Louvain, B-1200 Brussels, Belgium
Received 2 April 2008, accepted for publication 9 September 2008
Published 9 October 2008
Online at stacks.iop.org/PM/29/1247
Tissue oxygenation is a crucial parameter in various physiopathological
EPR oximetry, lithium phthalocyanine (LiPc) is a serious candidate for in vivo
applications because of its narrow linewidth and its high signal-to-noise ratio.
To enhance the biocompatibility of the sensors, fluoropolymer Teflon AF2400
was used to make cylindrical micro-inserts containing LiPc crystals. This new
micro-pellet design has several advantages for in vivo studies, including the
possibility of being able to choose the implant size, a high sensor content,
the facility of in vivo insertion and complete protection with preservation of
the oxygen sensor’s characteristics. The response to oxygen and the kinetics
of this response were tested using in vivo EPR: no differences were observed
between micro-inserts and uncoated LiPc crystals. Pellets implanted in vivo
in muscles conserved their responsiveness over a long period of time (∼two
months), which is much longer than the few days of stability observed using
LiPc crystals without protection by the implant. Finally, evaluation of the
biocompatibility of the implants revealed no inflammatory reaction around the
Keywords: EPR, oximetry, oxygen, biocompatibility
(Some figures in this article are in colour only in the electronic version)
Evaluation of the oxygen pressure in tissues is critical in physiology, pathophysiology and
during certain therapies. Monitoring oxygen levels in different tissues can help in adapting
1Author to whom any correspondence should be addressed.
0967-3334/08/111247+08$30.00 © 2008 Institute of Physics and Engineering in MedicinePrinted in the UK1247
1248 M Dinguizli et al
therapeutic protocols in various situations, such as radiotherapy, wound healing and ischaemic
areas in diabetes. Different methods exist for determining and monitoring oxygen pressure
in vivo (Gallez et al 2004), and can be classified into two groups: (i) direct measurement
of oxygen (invasive and non-invasive methods), such as the polarographic oxygen electrode
(EppendorfR ?), which has, for years, been considered as the ‘gold standard’ method; and (ii)
methods based on the correlation between parameters related to oxygen levels, such as the
EPR oximetry has been widely used over the last decade to quantify oxygen pressure in
the brain (Hou et al 2003), heart (Kuppusamy and Zweier 2004), gastrointestinal tract (He
et al 1999), skeletal muscle (Jordan et al 2004, Aragon´ es et al 2008), liver (Gallez et al
1998, James et al 2002), kidneys (James et al 1996), skin (Krzic et al 2001) and tumours
(Goda et al 1996, Gallez et al 1999, Jordan et al 2000, 2002, 2003).
is based on the introduction of an oxygen sensor into the tissue of interest a few hours
before the measurement. Several parameters, such as toxicity, oxygen sensitivity and in vivo
stability, are crucial for guiding the choice of the oxygen sensor. Oxygen sensors can be
divided into two groups: (i) soluble sensors such as trityl or nitroxides, and (ii) particulate
materials such as India Ink, charcoals and chars (Gallez et al 2004, Vahidi et al 1994, James
et al 1997, Lan et al 2004, Jordan et al 1998). In spite of their interesting properties, including
their rapid diffusion and distribution in physiological compartments, soluble sensors have a
serious disadvantage for pO2monitoring due to their metabolism and elimination. In contrast,
particulate paramagnetic materials are generally very inert. Several studies have described
repeated in vivo pO2measurements with these materials over long periods of time, from days
to months, including in humans (Swartz et al 2004).
Among the particulate spin probes, the phthalocyanine group has been used most for
in vivo pO2measurements. Since the first description of lithium phthalocyanine (LiPc) as
a probe for EPR oximetry by Liu et al (1993), LiPc crystals have been used for oxygen
monitoring in the brain (Hou et al 2003) and in tumours (Ilangovan et al 2004, Bratasz et al
2007, Matsumoto et al 2008, Dunn et al 2002). Most studies used ‘uncoated’ LiPc crystals
with a Teflon AF polymer (Dinguizli et al 2006). In this context, resonators were developed
to carry out measurements deeper than 1 cm, which usually represents the limit of depth
penetration for in vivo measurements using EPR spectrometers operating at 1 GHz (L-Band),
due to non-resonant absorption by water. Although LiPc is considered as stable in tissues such
as brain and tumours, a loss of oxygen sensitivity has been described in muscles (Liu et al
1993). In this regard, the development of biocompatible inserts presents a unique possibility
of preserving the oxygen sensitivity of LiPc. Other studies have indeed demonstrated the
effectiveness of this coating for preserving responsiveness in tissues (Gallez et al 1999,
He et al 2001).
In the present study, we tested a new design of retrievable polymeric micro-inserts
containing LiPc in order to stabilize the responsiveness of the sensor, to increase the content
in oxygen sensor in the implant (compared to the films previously described, Dinguizli
et al 2006), and also to facilitate implantation into and resection from tissues. To characterize
this new type of insert, the kinetics of the response and the oxygen sensitivity of the sensor
cast in polymeric tubes were measured. We also investigated by in vivo EPR the stability
of the responsiveness to oxygen during a long in vivo residence. Histological studies were
performed after implantation of this insert in the gastrocnemius muscle of mice to check the
biocompatibility of the inserted pellet.
Micro-pellets as oxygen sensors1249
2. Materials and methods
2.1. Micro-insert preparation
Cylindrical micro-pellets were made using a TeflonR ?AF2400 tubing (ID: 0.034??) that was
filled with LiPc crystals suspended in a polymer solution (1% Teflon AF 2400 w/v in FC
75 3M solvent). LiPc crystals were kindly provided by H M Swartz. The tubing was dried.
Pellets of different lengths were cut and then coated using a polymer solution (2% Teflon AF
2400 w/v in FC 75 3 M solvent), and finally dried for 24 h in an oven at 70◦C.
2.2. X-Band EPR measurements
Calibration curves and the kinetics of response for coated and uncoated materials were
performed by monitoring the variation in the EPR linewidth when the gas content was rapidly
changed from 0 to 21% oxygen. Calibration was performed at 9.3 GHz with a Bruker EMX
EPR (Rheinstetten, Germany) spectrometer equipped with a variable temperature controller
BVT-3000. Gas with known concentrations of nitrogen and air (obtained by using a gas mixer
AalborgR ?, Orangenburg, NJ, USA), equilibrated at 37◦C, was flushed over the samples, and
the spectra were recorded every minute until equilibration was achieved. The oxygen content
was less than one third of the peak-to-peak linewidth, incident microwave power was 50 μW,
and modulation frequency was 10 kHz.
2.3. In vivo implantation and measurements
After sterilization in an autoclave (120◦C for 20 min), micro-inserts were implanted using
an 18G trocar needle in the gastrocnemius muscle of NMRI male (25 g, n = 4) mice under
anaesthesia with ketamine/xylazine (60 mg kg−1and 6 mg kg−1, respectively). The loaded
trocar was inserted in the muscle, and the micro-inserts were deposited by pushing in the
stylus. In vivo EPR spectra were recorded using an EPR spectrometer (Magnettech, Berlin,
Germany) equipped with a low-frequency microwave bridge operating at 1.2 GHz and a
surface-coil resonator. Measurements were made under anaesthesia (1.8% isoflurane in air).
In order to assess the responsiveness of sensors in vivo, spectra were recorded under normal
conditions and under hypoxic conditions obtained by a transitory restriction in blood flow.
of the linewidth of recorded spectra to avoid peak distortion. Linewidth was measured after
10 min of animal stabilisation and was the average of ten accumulations (20 s scans) in the
same experimental conditions.
2.4. Histological studies
The animals were sacrificed by cervical dislocation 7 days after implantation. Muscles were
excised, and the implant was removed. The tissues were fixed with 10% formalin, embedded
in paraffin, and stained with haematoxylin–eosin. Observations were made using optical
microscopy (YS2-H, Nikon Corporation, Tokyo, Japan).
3. Results and discussion
The picture of a typical cylindrical insert is presented in figure 1.The length of the
1250 M Dinguizli et al
Figure 1. Picture of a micro-pellet composed of a Teflon AF2400 tube filled with an LiPc sensor.
This implant was 2 mm long.
micro-inserts can be adapted as a function of the tissue size and the site of implantation.
The choice of a cylindrical design is well adapted to easy in vivo implantation through trocar
needles in various sites, such as the brain or other specific areas. Filling the tubing with
LiPc crystals in polymer solution helps to avoid development of microbubbles, which can
form micro gas containers and disturb pO2measurements or the kinetics of equilibrium.
All micro-inserts were systematically checked with a microscope to confirm the absence of
microbubbles; inserts with bubbles were not used to avoid the potential problem of oxygen
reservoirsdisturbingthemeasurement. Interestingly, weobservedarapidequilibriumwiththe
gas phase and significant responses to small changes in oxygen level in vivo. Figure 2 shows
the responsiveness of LiPc after inclusion in the polymeric cylinder. Comparison between
the kinetics of response of LiPc in pellets and simple crystals (used without any coating)
demonstrated that there was no significant difference between these two forms of sensors
under in vitro conditions (data not shown). It should be noted that the polymer material was
carefully chosen for its oxygen permeability. Polymeric coatings should indeed be completely
permeable to oxygen as is the case for Teflon AF2400. Teflon AF2400 has a high oxygen
permeability compared to other fluoropolymers, and is used in contact lens technology where
oxygen permeability is a crucial parameter (Legeay et al 1998).
Another very important issue is the stability of the oxygen sensor. As already stated (Liu
et al 1993, Gallez and M¨ ader 2000), LiPc is known to lose its responsiveness to oxygen in
muscles. Interestingly, we observed a significant difference between uncoated LiPc crystals
and the LiPc embedded in micro-inserts that we designed. Figure 3 shows an example of the
loss of responsiveness of uncoated LiPc in mice muscle over time. During the first 10 days,
pO2values were around 20 mm Hg, which is in agreement with other studies. The apparent
oxygen pressure then decreased below 10 mm Hg. After 20–30 days, the sensors were still
able to respond to induced ischaemia, but the degree of responsiveness was significantly
reduced compared to the initial response just after implantation in the tissue. In contrast, LiPc
containing micro-inserts gave repetitive and comparative values without any loss of sensitivity
for up to two months after implantation (duration of the study) as shown in figure 4. The value
recorded after four weeks of residence in the muscle (12 mm Hg) was a little bit lower than
after one or two weeks (around 20 mm Hg). However, this measurement could probably be
considered as an outlier as the pO2recorded two months after implantation came back to the
initial values of 20 mm Hg. This could also be due to the limited number of animals.
The reason for this preservation of responsiveness is still speculative. However, we can
suggest several hypotheses: (i) since the sensitivity of LiPc to oxygen is strongly dependent
Micro-pellets as oxygen sensors1251
-3 -113579 11
Figure 2. Kinetics of the response of LiPc inserted in the micro-pellet to variations in oxygen.
The challenge consisted of switching nitrogen to air, and then air to nitrogen in the cavity of an
X-Band EPR spectrometer. Note the rapid response of the micro-inserts to changes in the oxygen
010 2030 40
Figure 3. In vivo responsiveness of uncoated LiPc after implantation in the gastrocnemius muscle.
EPR measurements were carried out with a 1 GHz EPR spectrometer on normal muscles, and
on ischaemic muscles after a transient interruption of the blood flow. The pO2values were
obtained in four different animals (mean ± standard deviation). Note the decrease in the degree of
responsiveness over time.
on the crystal structure (Bensebaa et al 1992), it is possible that minor changes or damage to
the crystal architecture might occur after a few days in vivo in tissues with a high mechanical
stress,suchasmuscles, thusaffectingresponsiveness; (ii)theimplantationofLiPccrystalscan
induce an inflammatory reaction leading to the recruitment of a high number of inflammatory
cells consuming high levels of oxygen that might interfere with the oxygen measurement
(Crokart et al 2005); (iii) the inflammatory response after insertion of oxygen sensors may
be responsible for a tissue reaction leading to release of substances that ‘poison’ the response
of the sensors or to the formation of impermeable capsules surrounding the sensors. Our
protocol induces minimum tissue damage and trauma during sensor implantation, which
means rapid stabilization of the implant in tissues, which is required before any measurement.