Gene Targeting MRI: Nucleic Acid-Based Imaging
Philip K. Liu and Christina H. Liu
Gene action plays a role in neural cell migration, learning processes, stress response, drug addiction, can-
cer, mental health, psychiatric and neurological disorders, as well as neurodegenerative diseases. Studies
also show that upregulation of certain gene activities in neurons may contribute to the development
of Alzheimer’s disease and other progressive cognitive disorders many decades after the alteration itself
occurs. Endogenous, environmental stress-related, or drug-induced chemical imbalances in the brain
affect the homeostasis of gene activities in neurons in specific brain regions and contribute to the comor-
bidity of mental illness and substance dependence. On the other hand, altered gene activities are also
a necessary part of repair processes after brain injury. Our general well-being is governed by the highly
regulated gene activities in our brains. A better understanding of gene activities and their relationship to
the progression of neurological disease can help the research and medical communities develop neces-
sary measures for early intervention, as well as plan more appropriate interventions or new therapeutic
approaches that can benefit a broad spectrum of patients who will be or have been affected by brain
diseases. We developed a non-invasive imaging technique that allows real-time assessment of gene tran-
scription profiles in live brains. This imaging method has the potential to provide first-hand information
about the progression of neurological disorders by gene targeting and cell typing, and it could elucidate
a surrogate marker for therapeutic efficacy for future planning of treatments for human diseases. We have
established a workable and reproducible MRI technique in live rodent brains.
Key words: Amphetamine, antisense, cell typing, cerebral ischemia, drug abuse, gene targeting,
heart arrest, molecular imaging, MRI, nanoparticles, prognosis, siDNA.
Procedures to evaluate gene activity in the brain at the transcrip-
tion level are not performed for clinical purposes, because current
techniques to detect altered gene transcription rely on the use of
M. Modo, J.W.M. Bulte (eds.), Magnetic Resonance Neuroimaging, Methods in Molecular Biology 711,
DOI 10.1007/978-1-61737-992-5_18, © Springer Science+Business Media, LLC 2011
364 Liu and Liu
biopsy samples. The difficulty in obtaining biopsies of brain tis-
sue severely limits the utility of these methods to monitor gene
activities in vivo. Commonly brain biopsies are only permitted in
patients with late-stage brain tumor. Currently, the diagnosis of
progressive cognitive disorders, such as Alzheimer’s disease, must
rely on the assessment of clinically defined symptoms, anatomical
hallmarks in the brain, and peripheral biomarkers obtained from
blood, urine, or cerebral spinal fluid samples. These biomarkers
are typically only evident after irreversible damage to the brain has
already occurred. For example, the use of blood tests to detect key
peripheral biomarkers expressed as a result of neuronal death is
too delayed to be useful to plan a therapeutic intervention. More-
over, it cannot provide an indication as to where cell death occurs
in the brain.
However, imaging tools have shown great promise as alter-
native approaches to these more conventional methods. Mag-
netic resonance (MR) spectroscopy and several MR imaging
(MRI) methods – tools that are routinely used for non-invasive
detection of abnormal function and structure in patients suf-
fering from neurological disorders, neurodegenerative diseases,
and mental illness – are emerging as powerful tools for detec-
tion of gene action in brains. Ligand (isotope)-guided positron
emission tomography imaging also has great utility for detect-
ing changes in the distribution of brain receptors associated with
chronic drug abuse and mental illness as well as neurodegen-
erative diseases. These imaging techniques have helped tremen-
dously to advance neuroscience research by enabling direct or
indirect measurement of gene actions, so as to decipher gene tran-
scription events that occurred before symptoms emerge. How-
ever, there remains a gap between our scientific understand-
ing of gene activity in in vitro experiments compared to what
takes place in the living brain during the evolution of diseases or
Our technique uses a gene targeting MR contrast agent for
the detection of intracellular mRNA in live brains using MRI.
This technique combines two well-established research platforms,
namely mRNA targeting by oligonucleic acids and iron-based
contrast-enhanced MRI for in vivo applications. Specifically, this
design creates a targeting MR contrast conjugate consisting of a
modular short nucleic acid probe and a MR contrast agent that
is covalently connected via a biotin–avidin linkage (Fig. 18.1).
To afford contrast-enhanced MRI, we use dextran-coated super-
paramagnetic iron oxide nanoparticles (SPION, a T2 suscepti-
bility agent). This modular probe construct can be modified
from antisense DNA (ODN) to various RNA platforms (siRNA,
microRNA, ribozyme, peptide DNA, etc.) for use with different
imaging modalities (fluorescence probes, isotopes, and MR con-
trast agents) and for different targets by changing the sequences
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications 365
Fig. 18.1. Molecular basis of MRI for gene targeting: nucleic acids (sODN) with a charged backbone are taken up via
membrane-bound receptors, the sequence in the sODN determines its retention in the cells. The same mechanism is
proposed for SPION-sODN.
of the nucleic acids (1).1Our design using nucleic acid and an
iron oxide-based MR visible probe was independently reproduced
using short interfering RNA (2).
This approach improves mRNA imaging, showing measur-
able improvement for a range of imaging applications, from
biopsy or postmortem samples (as shown in Fig. 18.2) to living
organs (as shown in Fig. 18.3). Still, many hurdles of working
with nucleic acids as brain probes in vivo remain, and we have
focused on improving our methods through rigorous research.
We expect to see modifications of our methods as investigation in
this area progress.
The nucleic acid-based MR probe used in MRI contains
a sequence that can bind to a specific target mRNA, forming
a hybrid that lasts long enough to create a transient window
for imaging. In addition, the dose should not block translation
if gene knockdown is not the objective; such a gene knock-
down is often seen in methods that use antisense DNA or short
interfering RNA. Because gene transcript targeting and report-
ing are based on specific binding of the reporter probe to its
target, the probe must have the highest reporting sensitivity
when its loading capacity is one; that is, maximum reporting
sensitivity is achieved with one targeting ODN to one con-
trast agent. Four targeting ODNs per contrast agent (loading
capacity of 4) will reduce reporting sensitivity by 75%. This
concept of probe design for gene targeting MRI is somewhat
366 Liu and Liu
Fig. 18.2. In vivo hybridization of FITC-sODN-c-fos in the neuronal formation of the
dentate gyrus (arrow) demonstrates DNA transfection in vivo via ICV delivery (by passing
BBB), target hybridization, clearance of nonspecific sODN-mRNA hybrids in live brain
and detection in postmortem samples.
opposite to those used in current molecular imaging which may
call for a maximum loading capacity. Since the first description
of our methods, a few modifications for ODN-based probes in
SPION MR contrast conjugate have been described (3). Addi-
tional considerations and modifications are being made to achieve
maximal MR signal changes to improve signal to noise and
to delineate mechanism of this novel technique. One sugges-
tion that has been made, and which we have carried out, is
to validate our methods in neurons grown in culture. Transfec-
tion of ODN probes to cultured primary neurons in the rest-
ing state, unlike primary astroglia, has been very difficult as
the primary neurons may not express adequate mRNA tran-
scripts after adaptation to the culture condition. Neuronal trans-
fection by fluorescent-labeled short DNA in vivo has been
detected in postmortem samples (4, 5). For this reason, we have
always validated sODN specificity by transfection in vivo (6).
Below, we present our current methods for making our gene
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications 367
Fig. 18.3. DNA transfection in vivo: FITC-sODN or SPION-sODN is delivered to cerebrospinal fluid, the probe is distributed
through the Virchow–Robin space, and uptake and retention for MRI in live brains, or postmortem histology (Copyright:
J Neurosci, 2009, Reproduced with permission).
targeting probes with examples of conjugations between sODN
and SPION to achieve a sufficient contrast-to-noise ratio for gene
All magnetic resonance contrast agents are characterized (e.g.,
particle sizes and magnetic relaxivities) before use. All enzymes
were tested before use for optimal concentration, time of incuba-
tion, and temperature according to vendor’s specification.
1. Superparamagnetic iron oxide nanoparticles (SPION) with
a core diameter of 10 nanometers (nm) and a hydrodiam-
eter of 18–30 nm or less were homemade, but the same
material can be purchased from Biophysics Assay Labora-
tory, Inc. [Molday ION (No. CL-30Q02-2), colloidal size
of 30 nm, Worchester, MA].2
2. Sodium citrate, sodium hydroxide, BupHTMPhosphate
Buffered Saline (PBS), AmionLink Reductant sodium
cyanoborohydride, and NeutrAvidinTM
368 Liu and Liu
protein were purchased from Thermo Scientific (St Louis,
MO). Prepared BupH-PBS in 500 ml distilled water
(0.1 M, pH 7.2).
ACROS ORGANICS (Geel, Belgium).
4. Centrifugal filter devices (Amico Ultra-4, Ulturacel-100 K)
were from Millipore (Billerica, MA).
5. Biotinylated single-stranded phosphorothioate-modified
cein isothiocyanate (FITC), Cy3, or rhodamine are
homemade.3We have ordered the same material from
commercial sources such as ThermoScientific, Invitrogen,
Sigma, or Amitof (Cambridge, MA).
6. AntisenseDNA design:
wasused as example–
sequence no. 1,925–1,946 of FosB mRNA of the
mouse [mmFosB, accession no. X14897]) can be obtained
from the mouse genome library. The sequences of sODN-
c-fos, sODN-actin, and sODN-Ran have been reported
previously (7, 8). Single-stranded ODNs were synthesized
with protection from the use of non-specific nucleases by
phosphorothioate modification of all nucleotide bridges,
and the resulting sODNs were purified by polyacrylamide
7. Sense DNA (UPS, 5?-GATCGCCGAGCTGCAAAAAG-
3?), 146 nucleotides upstream of sODN-fosB, from the
same mouse FosB mRNA, was used to determine the
specificity of antisense sODN-fosB in the polymerase chain
8. Mouse brain cDNA library is homemade, but can be pur-
chased from Stratagene (La Jolla, CA). Polymerase chain
reaction buffer and Taq polymerase were also purchased
9. Antibodies were from Abcam (Cambridge, MA).
10. All solutions are sterilized by filtration using Nalgene filter
units (Nalge Nunc International Corp., Rochester, NY).
11. Inhalation anesthesia for in vivo MRI of rodents: halothane
Inc.) or forane (isoflurane, USP, Baxter Healthcare Corp.,
12. Toothpaste (preferably alcohol free to minimize repeat irri-
tation in the case of chronic survival studies) to fill rodents’
ear canals during MRI to reduce artifacts that occurs at the
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications 369
of the SPION-NA
1. Attaching functional groups to SPION: Incubate 20 g
of Molday ION in a 15 ml of 2 N sodium hydrox-
ide with 2 M 2-chloroethylamine with gentle stirring
overnight at room temperature. The reaction generates
hydrochloride and should be performed in a well-ventilated
2. Dialysis filter the solution in 10× the volume of 25 mM
filter-sterilized sodium citrate (pH 8.0), then continue dial-
ysis filtering in additional 10× the volume of 0.1 M BupH-
PBS using an Ulturacel-30 K filter (30 KD cutoff).
3. Incubate NeutrAvidin (NA, 1–20 mg) overnight in the pres-
ence of sodium cyanoborohydride (1 M) in BupH-PBS
(0.1 M, pH 7.2) in an amber bottle (avoid light) at a
total volume of 10 ml at room temperature.1The resulting
covalently linked product, SPION-NA, is filtered and dia-
lyzed against a 20X volume of sodium citrate buffer solution
(25 mM, pH 8.0).
4. Concentrate the volume to 5–6 ml or less.
5. Iron concentration is measured in hydrogen peroxide and
the optical density is measured at 410 nm.
6. Store the activated SPION (SPION-NA) at 4◦C in an
amber-colored bottle with rubber stopper to minimize
oxygen contact, at a concentration of 3–4 mg iron per
milliliter sodium citrate buffer (25 mM, pH 8.0).
in Total Cerebral
1. Mix 10 μg of total cDNA, FITC-sODN-fosB, upstream
sense DNA (10 pmol each), and polymerase reaction buffer
in 40 μl volume.
2. Prepared polymerase and four deoxynucleotide triphos-
phates (dNTP, 20 mM) in 10 μl.
3. The reaction mix is incubated at 95◦C for 30 s; the mix is
then maintained at 70◦C when a bolus of 10 μl of poly-
merase and four dNTPs is added.
4. Polymerase chain reaction is carried out for 25 cycles at 90◦C
(30 s), 39◦C (30 s), and 65◦C (45 s) followed by 68◦C for
5. The resulting product is resolved by electrophoresis on
agarose gel (1%) for 100 V-h (100 V in 1 h or 50 V in 2 h).
6. A good antisense sODN should yield only one product
of 146 base pairs according to the mouse fosB mRNA
370 Liu and Liu
1. We mix 10 μl of SPION-NA (~30 μg, duplicate tubes) to
increasing concentration of FITC-sODN-fosB-biotin, from
0, 1, 2, 9, 15, or 30 pmol of FITC-sODN-biotin per micro-
gram of SPION (lanes 1–6, lane 7 free FITC-sODN-fosB,
Fig. 18.4a), incubate 30 min at room temperature (or
overnight at 4◦C). The coupling reaction is resolved by elec-
trophoresis on agarose gel (1%) at room temperature for
2. A successful coupling reaction is shown by a fragment upshift
of FITC signal: Lane 7 shows unbound sODN (arrow)
that traveled 4 cm from the loading well, SPION-sODN-
FITC hardly traveled (or only a short distance from the
loading well), as shown in lanes 3–6. Lane 2 with a dupli-
cate sample shows saturation binding at 30 pmol of FITC-
sODN-biotin per microgram of SPION, especially when
Fig. 18.4. Binding capacity of SPION-NA by FITC-sODN-biotin. Upon binding between
NeutrAvidin (NA) on SPION and biotinylated-sODN-FITC, there is an upshift from low
molecular weight sODN to high molecular sODN (Panel a, 50 V-h). The mobility of sODN,
as little as 1 pmol of sODN per microgram of SPION-NA (Fe) and increasing with bind-
ing at 30 pmol per microgram Fe (lanes 2 and 3), is capable of pulling bound SPION
(dark spot below the well) to the opposite direction of SPION-NA (lane 1, Panel a). This
pulling mobility is used to detect the binding capacity of SPION-NA by increasing the
amount of FITC-sODN-biotin (Panel b, 110 V-h). The presence of unbound FITC-sODN
(arrows)indicates thesaturation ofbinding. FITC-sODN is pulling the SPIONalong during
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications 371
free (unbound) FITC-ODN appears as excess FITC-sODN-
fosB-biotin. Lane 3 shows near saturation because there is
no unbound FITC-sODN-fosB-biotin.
3. We observed the pulling of sODN on SPION with high
sODN binding (lane 2).
4. To narrow the saturation binding, we incubated FITC-
sODN-gfap-biotin and SPION-NA at a ratio of 28,
26, 24, 22, 20, 18, 16 pmol per microgram SPION.
Figure 18.4b shows all mixtures contain increasing
pmol/μg SPION. Indeed, free FITC-sODN-gfap appeared
at high ratio (28 pmol/μg), at the highest intensity, and
reduces the intensity as binding capacity is lowered to 16
pmol FITC-sODN-fosB-biotin per microgram iron oxide.4
5. Each NA has four binding sites for biotin; we calculated no
more than 4 pmol of NA per microgram of iron oxide in this
batch of SPION-NA.
of Biotinylated sODN
1. SPION-NA and biotinylated sODN are mixed and incu-
bated at 4◦C overnight (Section 3.4 describes how the ratio
is determined). Generally, we add 3–4 pmol per microgram
SPION-NA (16 pmol divided by 4, as there are four biotin
binding sites per molecule of avidin).5
2. The optimal MR contrast is achieved when one molecule of
SPION reports one copy of mRNA. This can be achieved by
reducing the sODN-biotin to a constant amount of SPION-
NA and testing the signal reduction frequency (R2∗value)
in vivo (Section 2.4.8).
3. Take four microvials (0.5 ml capacity) and add 30 μg of
SPION-NA to each vial, then add 120, 60, 30, and 15
pmol of biotin-sODN-gfap to each vial to achieve four dif-
ferent sODN to SPION conjugation ratios (pmol sODN per
4. Mix well by tapping and then briefly centrifuge in a bench-
top microfuge (Fisher Scientific).
5. Add sodium citrate (25 mM, pH 7.8) to a final volume of
60 μl, so that the iron oxide is 0.5 μg Fe/μl.
6. Incubate overnight at 4◦C.
7. Deliver SPION-fosB (1.6–2 μl, ICV) to mice (40 μg/kg for
20–25 g of mice); generally, two mice per dose or n=8.
8. Acquire MRI R2∗maps at 9.4 T 4 h after delivery (accord-
ing to Section 2.9) and select the best conjugate ratio
(described in Section 2.4.3) which results in maximal R2∗
increase in most brain regions (Section 2.4.3) from each
batch of SPION-NA.
372 Liu and Liu
and in Cerebrospinal
1. Animals are anesthetized using either isoflurane or halothane
(for drug stimulation studies).
2. Five milliliters of arterial blood is withdrawn from the carotid
artery with a 26 g needle on a 10-ml syringe; blood plasma
is collected after centrifugation at 500×g for 10 min.
3. We incubated SPION-sODN-FITC (200 pmol) in 0.1 ml
of saline (9 or 45 μg Fe/μl), plasma (9 μg Fe/μl), or
cerebrospinal fluid (45 μg Fe/μl) for 24 h at 37◦C, and
the resultants of reaction were resolved by electrophoresis
agarose gel (0.8%) (120 V-h). We found that SPION-sODN
remained linked in both fluids for the duration of the study
Fig. 18.5. Stability of SPION-sODN in body fluid: SPION-sODN or sODN was incubated
in body fluids [serum or cerebrospinal fluid (CSF)] at 37◦C for various hours and then
the degradation of sODN was determined in agarose gel (1%) by electrophoresis. The
sODN in SPION-sODN is protected from binding to serum proteins (compare lanes 3–9
to lanes 11 and 1) and is stable for 24 h at physiological temperature without significant
degradation as seen in lanes 2 and 10. We observed the same result of SPION-sODN in
the CSF for the same duration (not shown).
1. Animals are anesthetized with either isoflurane or halothane
(for drug stimulation studies) and placed in a stereotac-
tic device (Stoelting, Wood Dale, IL) for ventricle infu-
sion, via Hamilton syringe, to bypass the blood–brain
2. The mouse is placed in a supine position; the skin is cut open
on the midline of the head and the skull is exposed. The
needle is placed on bregma and the needle tip is adjusted
to stereotactic coordinates from bregma: LR 1.0 mm,
AP –0.2 mm.
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications 373
3. The coordinate is marked by a surgical marker and the skull
on the mark is drilled to make the skull thinner so the needle
can puncture through the skull.
4. The needle is lowered to the surface of the thin skull (zero
point), and the needle is lowered by 3.0 mm (DV–3.0 mm).
5. We deliver 2 μl of SPION-sODN to the lateral third ven-
tricle of the mouse at the rate of 0.5 μl/min. The coordi-
nates can change from strain to strain. For example, those
for C57black6 mice may differ by 0.05–0.1 mm on the AP
plane between 18 and 25 g. Furthermore, C57black6 mice
from Jackson Laboratory may show some differences from
mice of the same strain from Taconic Farm, NY.
6. The burr hole on the skull created by the infusion proce-
dure is sealed with bone wax before the incision is closed
with suture. The infusion dose of SPION-sODN we used in
mice was 0.04 mg Fe/kg body weight in 2 μl sodium citrate
solution (25 mM).
1. We conjugate SPION-NA and FITC-sODN-biotin at the
optimal mixture (0.5 μg Fe/μl at 4◦C overnight). For one
mouse, we generally deliver 0.2 ml (0.5 μg Fe/μl) to a 25 g
mouse or 4 mg/kg.
2. Immediately before delivery, we add 2 μl lipofectamine2000
(1 mg/ml, Invitrogen Lifesciences) per 200 μl of SPION-
sODN solution. Mix well by tapping and then briefly cen-
trifuge in a bench-top centrifuge. We acquire baseline MRI
before delivery of the probe (pre-delivery), then one or more
images after delivery in a 9.4 T magnet at various time points
(generally at 2, 4, 6, and 24 h after i.p).
of Delivery (OTRD)
We prepared SPION-sODN (3 μg Fe/μl) and lipofectamine as
in step 2.7. After general anesthesia [ketamine (100 mg/kg) and
xylazine (10 mg/kg) anesthesia i.p.] we delivered SPION-sODN-
FITC via micropipette and pipette tips as eye drops to the eye
sac at a rate of 10 μl every 15 min (33.3 μl, OTRD, 3.996 mg
3.9. MRI Acquisition
(Live Brain Imaging)
The following steps refer to MRI acquisition with a 9.4 T
Bruker/Magnex horizontal bore (21 cm) animal MR scanner. A
custom-made cradle equipped with a tooth bar and nose cone
complete with gas input and exhaust lines is used for scanning.
The MR parameters described below are suitable for a surface
coil with transmit and receive capability.
1. Animals are anesthetized during the entire MRI session to
minimize motion. Inhalation anesthesia such as halothane
or isoflurane is used for most survival MRI studies, as
374 Liu and Liu
it has maintenance advantages over injectable anesthetics.
2% isoflurane or halothane in pure O2 is used to ensure
proper oxygenation of the brain throughout the entire
MRI session. Whilst isoflurane is a more acceptable anes-
thetic regimen for small animal MRI, halothane is recom-
mended for drug stimulation studies, because isoflurane has
been shown to block or reduce the cerebral hemodynamic
changes associated with pharmacological stimuli in rodents
(depicted in functional MRI studies). Interestingly, we also
observed in animals under isoflurane anesthesia reduced
SPION-sODN signal changes associated with amphetamine
2. Animals are placed in a prone position and secured with the
tooth bar. To reduce air–tissue interface artifact, toothpaste
was injected into the ear canals, using a flexible catheter,
without rupturing the ear drums. It is important that ear
drums are not punctured to prevent the leakage of tooth-
paste into the trachea, which can result in suffocation. A sur-
face coil is then placed on top of the animal’s head and the
whole setup is inserted into the MRI scanner.
3. A standard tune and match procedure is performed followed
by automatic or manual shimming before each MRI scan
series to ensure maximal homogeneity in the field strength
inside the gradient for signal sensitivity and reproducibility.
4. The MRI protocols include as follows: (1) A general localiza-
tion sequence (for example, RARE Tripilot) (2) T2anatom-
ical imaging for slice positions, TR/TE=7,000/25 ms,
117×117 μm2, 20 0.5 mm contiguous slices, RARE factor
8, number of average (NA)=2 (2) Serial, 2D gradient echo
fast imaging (GEFI) with TR/TE=500/3, 4, 6, 8, 10 ms,
flip angle=30, NA=2 with the same geometry as in (2). The
entire scan series lasts for less than 30 min.
5. For voxel-wise and region-of-interest comparison, images
should first undergo automatic and manual alignment pro-
cedures using any standard image processing software with
registration capability. Fine-tuning of alignment can be per-
formed by visual comparison to template images, focus-
ing on obvious landmarks such as the corpus callosum and
outlines of the ventricles. R2∗maps are constructed from
the aligned images (with incremental TEs). R2∗(inverse of
T2∗) maps are calculated using pixel-wise linear fitting from
the set of images with the same TR and incremental TEs
based on equation M=M0× exp(–TE/T2∗). Elevated R2∗
(or reduced T2∗) is theoretically caused by the presence of
SPION in tissue. All ROIs are outlined according to ‘The
Mouse Brain in Stereotaxic Coordinates’ (10). Averaged
R2∗values within ROIs are extracted from each animal and
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications375
we calculate the group mean and standard error of the mean
(SEM) in each group for statistical analysis.
Localization of sODN
1. Postmortem brain preparation: animals are anesthetized and
brains are quickly removed from the skull and flash frozen
in liquid nitrogen. Thin coronal tissue sections can be pre-
pared using Cryostat and mounted on glass slides; thick
tissue sections can be prepared using a Vibratom system
and stored as floating sections in 0.1% sodium azide/PBS
2. Frozen sections were incubated in 3.7% freshly prepared
paraformaldehyde for 10 min and rinsed in fresh double-
distilled water. Fluorescent signals can be directly observed
under a fluorescent microscope.6
Localization of SPION
1. Postmortem sample preparation: animals are anesthetized
and transcardially perfused with 15 ml heparinized saline at
a rate of 10 ml/min and then with 10 ml of freshly pre-
pared 4% paraformaldehyde (PFA) in 0.1 M PB, pH 7.4.
The brains are stored overnight in PFA solution at 4◦C and
stored at 4◦C in 20% sucrose/PBS solution to chase out
PFA. Samples are ready when the brains drop to the bot-
tom of the sample container, usually within 24 h. Brain
samples collected using this method can also be used for
MRI microscopy (see Section 2.12 for detailed steps). Sam-
ples are embedded in paraffin and thick coronal tissue slices
(20–100 μm thickness) are prepared and mounted on glass
2. After removing paraffin from the glass slides, tissue sam-
ples can be stained for Prussian blue (PB) with 2%
potassium ferrocyanide in 2% HCl (Perl’s method) and
counterstained with nuclear fast red (NFR) (Fisher Scien-
tific, Houston, TX). Iron stains can be viewed under a light
3.12. MRI Acquisition
The following steps refer to a 14.1 T Bruker/Magnex vertical
bore (8.9 cm) MR spectrometer. A volume coil with inner diam-
eter of 1 cm was used.
1. The entire brain can be immersed in a 1 cm diameter NMR
tube in perfluoro compound FC-40 (Fluoroinert FC-40,
Sigma) to eliminate background proton signal. Care should
be taken to minimize air bubbles as the brain is inserted into
2. A standard tune and match procedure is performed followed
by automatic or manual shimming before each MRI scan
series to ensure maximal homogeneity in the field strength
inside the gradient, for signal sensitivity and reproducibility.
376Liu and Liu
3. The MRI protocols include as follows: (1) A general local-
ization sequence (for example, RARE Tripilot) (2), a three-
dimensional (3D) fast low-angle shot (FLASH) gradient
sequence is to acquire T2∗–weighted images (TR/TE =
50/18 ms, 40 × 40 × 40 μm3, flip angle=20, NA=24).
The entire scan lasts for approximately 11 h.
1. Notes to nomenclature of MR probes: We use the
cleotide (ORN) for synthetic nucleic acids with antisense
sequence to mRNA targets because they may function differ-
ently from naturally occurring dicer-dependent and dicere-
indipendent micro RNA that exists endogenously from de
For consistency with practice of using capital letters in
the names of proteins, we use lower case characters for
our mRNA-targeting sODN linked to MR-visible agents
(SPION-cfos, SPION-fosB, sODN-fos or sODN-fosB),
and reserve upper case for future antigen-targeting probes
(SPION-ACTIN or sODN-ACTIN). We use SPION-Ractin
or SPION-Rgfap for sORN-labeled SPION. We use abbre-
viations for probes with no target (SPION-NA), or for DNA
with random sequence (SPION-Ran). Phosphorothioate-
modified nucleic acid is abbreviated as sODN or sORN, and
peptide-modified nucleic acid is referred to as peptide ODN
or peptide ORN.
2. Generally, we ask BIOPAL to send us a small 0.1 ml sample
of freshly made Molday ION (CL30Q02-2, 30 nm in diam-
eter) to validate the core size and the ability to be function-
alized to accept NeutrAvidin, a key step in probe construc-
tion. We generally start with 20 mg of SPION made within
3 weeks. The ratio of NA to SPION is essential as more NA
on one SPION will reduce the sensitivity of SPION-NA. We
have been working with 1–20 mg ratio. SPION-NA has a
shelf-life of 3 months at 4◦C (1).
3. For histological localization of intracellular sODN, we label
the sODNs with FITC, Cy3, or rhodamine.
4. SPION-NA (dark brown band) often migrates to the oppo-
site direction of FITC-sODN in agarose gel of 1%. We
observed the pulling of sODN on SPION because the dark
brown band (SPION) in SPION-sODN with high sODN
binding (lane 2) will migrate slower than the SPION in
lane 1 (no sODN). This is evidence that the charge on
Gene Targeting MRI: Nucleic Acid-Based Imaging and Applications377
FITC-sODN, that is different from peptide DNA, affects
the migration and transfection of SPION (Fig. 18.4a).
5. Saturation binding on SPION-NA will reduce probe sen-
sitivity, as more mRNA transcripts will bind to antisense
sODN per SPION. We found 1 mg or less NA with 20 mg
SPION ratio to be most suitable for our work (1).
6. Permeabilization of tissue samples for immunohistochem-
istry should be minimized to retain sODN probes. If per-
fusion is needed, try the two-buffer system (4).
We thank Drs. Charng-ming Liu and Jia Q. Ren for techni-
cal assistance and Ms. N. Eusemann for excellent editing. This
project was supported by grants from NIH (DA024235 and
DA026108 to CHL, NS057556 and DA29889 to PKL), Amer-
ican Heart Association (09GRNT2060416 to CHL), and funds
from the Stanley Medical Research Institute through the Stanley
Center for Psychiatric Research at the Broad Institute. Athinoula
A Martinos Center for Biomedical Imaging is partially supported
1. Liu, C. H., Kim, Y. R., Ren, J. Q., Eichler,
F., Rosen, B. R., Liu, P. K. Imaging cere-
bral gene transcripts in live animals. J Neu-
2. Medarova, Z., Pham, W., Farrar, C., Petkova,
V., Moore, A. In vivo imaging of siRNA
delivery and silencing in tumors. Nat Med
3. Liu, P. K., Mandeville, J. B., Guangping, D.,
Jenkins, B. G., Kim, Y. R., Liu, C. H. Tran-
scription MRI: A new view of the living brain.
4. Cui,J. K., Hsu,
Suppression of postischemic hippocampal
nerve growth factor expression by a c-fos
antisense oligodeoxynucleotide. J Neurosci
5. Hecker, J. G., Hall, L. L., Irion, V. R. Nonvi-
ral gene delivery to the lateral ventricles in rat
brain: Initial evidence for widespread distri-
bution and expression in the central nervous
system. Mol Ther 2001;3:375–384.
6. Liu, C. H., Ren, J. Q., Yang, J., Liu, C. M.,
Mandeville, J. B., Rosen, B. R., Bhide, P. G.,
C. Y., Liu,P. K.
Yanagawa, Y., Liu, P. K. DNA-based MRI
probes for specific detection of chronic expo-
sure to amphetamine in living brains. J Neu-
7. Liu, C. H., Huang, S., Cui, J., Kim,
Y. R., Farrar, C. T., Moskowitz, M. A.,
Rosen, B. R., Liu, P. K. MR contrast
probes that trace gene transcripts for cerebral
ischemia in live animals. FASEB J 2007c;21:
8. Liu, C. H., Huang, S., Kim, Y. R., Rosen,
B. R., Liu, P. K. Forebrain ischemia-
reperfusion simulating cardiac arrest in mice
induces edema and DNA fragmentation in
the brain. Mol Imaging 2007b;6:156–170.
9. Liu, C. H., Huang, S., Cui, J., Kim,
Y. R., Farrar, C. T., Moskowitz, M. A.,
Rosen, B. R., Liu, P. K. MR contrast
probes that trace gene transcripts for cerebral
ischemia in live animals. FASEB J 2007;21:
10. Paxinos, G., Franklin, K. B. J. The Mouse
Brain in Stereotaxic Coordinates. London:
Academic Press Limited; 2001.