Proton-electron double magnetic resonance imaging of free radical solutions

the ability to image the distribution of paramagnetic solutions in vivo has many
potential uses in biology and medicine. Free radicals are present in varying amounts
in different tissues, and changes in their concentration may be associated with carcinogenesis (I). Injected free radicals might be targeted at specific tissues (2). Oxygen
concentration in tissues might be deduced via its effect on an injected free radical (3,
4). Considerable effort has gone toward the development of electron paramagnetic
resonance imaging (2, 5-9), using methods analogous to those employed in nuclear
magnetic resonance imaging. However, EPR imaging has so far been restricted to
small samples (-5 mm) because of the strong magnetic field gradients necessary to
achieve good spatial resolution (2). We report here the first results of a new technique
which we call proton-electron double-resonance imaging (PEDRI). The method combines proton NMR imaging with nuclear-ekctron double resonance (l&12) and allows
the distribution of free radicals in large aqueous samples to be imaged.
In nuclear-electron double resonance (10-12) an NMR resonance of a solvent is
observed while an EPR resonance of a paramagnetic solute is excited. Under favorable
conditions, saturation of the solute EPR resonance causes a dramatic enhancement
of the solvent NMR signal. In particular, the effect requires that the NMR nuclei be
relaxed by their interactions with the unpaired electrons of the solute. In PEDRI, a
conventional proton NMR image is collected while the solute EPR resonance is irradiated. The NMR signal from those protons being relaxed by the paramagnetic
solute is enhanced, and these regions exhibit a greater intensity in the final image.
Subtraction of images obtained with and without EPR irradiation will provide an
image showing the distribution of the paramagnetic solute alone.
The maximum possible enhancement is given by -us/2u, where vs and vl are the
EPR and NMR frequencies, respectively; for protons, this corresponds to a factor of
-330. A negative enhancement means that the NMR signal changes phase by 180”
upon irradiation of the EPR resonance, while its magnitude increases by the numerical
factor. The observed enhancement factor depends on the degree of saturation of the
solute EPR resonance and therefore on the strength of the irradiating magnetic field.
It also depends on the concentration of the paramagnetic compound, the particular
0022-2364188 $3.00 366
Cqy@tt 0 1988 by Ademic l’ress, Inc.
All ri#tts of rcpri3duction in any form resemd.
flG. 1. Double-resonance coil assembly with sample tube in place. Split-solenoid NMR receiver coil has
diameter 85 mm, length 67 mm. EPR resonator has diameter 10 mm, length 20 mm.
Zk% 2. PEDRO magnitude image of a sample tube containing 2.5 mMTEMPOL solution, showing increase
in image intensity within EPR resonator. Saturation-recovery NMR sequence, repetition time 1 s. Projective
image (i.e., no slice selection); field ofview 10 cm square. EPR tiquency 1.123 GHz, power - 1 W. Observed
enhancement factor, -6.9.
367
368 COMMUNICATIONS
solvent and solute under study, the temperature of the sample, and the strength of
the static magnetic field (10-12).
We have implemented PEDRI on our home-built whole-body proton NMR imager
(13). This instrument has a resistive side-access magnet operating at 0.04 T, giving
NMR and EPR frequencies of 1.7 MHz and 1.12 GHz, respectively. Crossed NMR
transmit and receive coils were used, the latter being a miniature split solenoid of
diameter 85 mm and length 67 mm (Fig. 1). Proton NMR images were collected using
a standard saturation-recovery gradient echo spin-warp pulse sequence (14). Images
were acquired on a 128 X 128 matrix (interpolated to 256 X 256 for display) with a
field of view of 10 X 10 cm. Repetition times varied between 200 ms and 1 s, giving
image acquisition times between 25.6 and 128 s.
A separate synthesized frequency generator (FamelI Instruments, UK) provided the
microwave EPR excitation signal, which was amplified by a broadband amplifier (Farnell Instruments) and applied to the sample using a cylindrical segregated loop resonator
(Fig. 1). The resonator consisted of 20 loops connected in parallel and operated in
parallel resonance, generating a magnetic field along the axis of the NMR receiver
coil. Trimming capacitors allowed the resonator to be tuned to the appropriate EPR
frequency and matched to the 50 ohm coaxial cable. The EPR resonance was irradiated
FIG. 3. PEDRI magnitude image of two sample tubes containing air-equilibrated (left) and nitrogenequilibrated (right) 2.5 mM TEMPOL solution. Profile across image shows d&rent enhancement levels in
the two tubes, due to different oxygen concentration. Imaging parameters are as in Fig. 2. Observed enhancement factors, -6.9 (left) and -9.8 (right).
COMMUNICATIONS 369
continuously during the collection of a PEDRI image, the maximum power delivered
to the sample being 1 W.
PEDRI images were obtained of free radical solutions contained in glass sample
tubes (10 mm outside diameter, 70 mm length). Experiments were performed using
aqueous solutions of the nitroxide free radical TEMPOL (4-hydroxy-2,2,6,6-tetramethylpiperidine- 1-oxyl) (Aldrich Chemical Co.) at room temperature. An enhancement factor of -6.9 was observed with 2.5 mM TEMPOL solution at approximately
1 W microwave power. Much larger enhancements have been observed using free
radicals dissolved in organic solvents; however, these would not be suitable for biological
experiments.
Figure 2 shows a PEDRI image of a sample tube containing a 2.5 mi’kf aqueous
solution of TEMPOL, the increased image intensity within the resonator is clearly
visible. An interesting feature is the reduction of image intensity adjacent to the ends
of the resonator. This occurs because the enhancement factor is negative: since the
microwave magnetic field strength decreases sharply outside the resonator, it reaches
in a short distance such a value that the “enhanced” NMR signal is zero, giving a
band of zero intensity in the image. The nonuniformity of the image intensity within
the enhanced region is due to variations of the microwave magnetic field strength
inside the resonator.
The potential of PEDRI for oxygen concentration measurement was investigated
by imaging simultaneously two identical phantoms of 2.5 mM TEMPOL solution,
one of which had been equilibrated with nitrogen (by bubbling the gas through the
sample), the other with air. The nitrogen-equilibrated sample exhibited a 40% greater
enhancement than did the air-equilibrated one (Fig. 3). This effect occurs because
EPR resonances of spin labels are broadened in the presence of dissolved oxygen (3,
4,15,16). A broad EPR line is more difficult to saturate; thus the observed enhancement
factor for a given microwave magnetic field strength is reduced (10-12).
PEDRI images have been obtained with TEMPOL concentrations as low as 0.3
mM. This is nevertheless at least three orders of magnitude greater than the concentration of naturally occurring free radicals (I), which it would seem are unlikely to be
detectable using PEDRI. It is necessary to explore further the characteristics of the
enhancement and the targeting and encapsulation of free radicals if PEDRI is to be
used in vivo.
In conclusion, we have developed a new technique for imaging solutions of free
radicals. In contrast to conventional EPR imaging, our method is easily implemented
using standard NMR imaging hardware and software, requiring only the addition of
a microwave source and antenna. Furthermore, PEDRI is not restricted to small samples. The feasibility of the technique for oximetry has been demonstrated. It is likely
that the average power deposited in the sample by the microwave irradiation can be
reduced considerably by pulsing the EPR excitation; experiments to investigate this
are presently underway. We are currently working on designs for larger resonators
and microwave horns, to enable larger samples to be imaged.