A 18f-labeled saxitoxin derivative for in vivo pet-mr imaging of voltage-gated sodium channel expression following nerve injury

Article abstract of DOI:10.1021/ja408300e

Both chronic and neuropathic pain conditions are associated with increased expression of certain voltage-gated sodium ion channel (Na_V) isoforms in peripheral sensory neurons. A method for noninvasive imaging of these channels could represent a

Full text of DOI:10.1021/ja408300e

Communication
pubs.acs.org/JACS
18
A F‑Labeled Saxitoxin Derivative for in Vivo PET-MR Imaging of
Voltage-Gated Sodium Channel Expression Following Nerve Injury
Aileen Hoehne,∥,† Deepak Behera,∥,† William H. Parsons,∥,‡ Michelle L. James,† Bin Shen,†
Preeti Borgohain,† Deepika Bodapati,† Archana Prabhakar,† Sanjiv S. Gambhir,† David C. Yeomans,§
Sandip Biswal,*,† Frederick T. Chin,*,† and J. Du Bois*,‡
‡
§
†Department of Radiology, Department of Chemistry, and Department of Anesthesia, Stanford University, Stanford, California
94305, United States
S
* Supporting Information
Scheme 1. De Novo Synthesis of 19F- and 18F-Labeled
Saxitoxins
ABSTRACT: Both chronic and neuropathic pain con-
ditions are associated with increased expression of certain
voltage-gated sodium ion channel (NaV) isoforms in
peripheral sensory neurons. A method for noninvasive
imaging of these channels could represent a powerful tool
for investigating aberrant expression of NaV and its role in
pain pathogenesis. Herein, we describe the synthesis and
evaluation of a positron emission tomography (PET)
radiotracer targeting NaVs, the design of which is based on
the potent, NaV-selective inhibitor saxitoxin. Both auto-
radiography analysis of sciatic nerves excised from injured
rats as well as whole animal PET-MR imaging demonstrate
that a systemically administered [18F]-labeled saxitoxin
derivative concentrates at the site of nerve injury,
consistent with upregulated sodium channel expression
following axotomy. This type of PET agent has potential
use for serial monitoring of channel expression levels at
injured nerves throughout wound healing and/or following
drug treatment. Such information may be correlated with
pain behavioral analyses to help shed light on the complex
molecular processes that underlie pain sensation.
urrent clinical approaches for the diagnosis of chronic pain
C
rely on subjective methods of analysis, which include
patient self-reporting and physical exams.1 Various imaging
modalities such as X-ray, computed tomography (CT), and
conventional magnetic resonance imaging (MRI) have limited
utility for objectively identifying hypersensitive ‘pain-generating’
neural structures and accurately assessing pain syndromes.2−5
Accordingly, in vivo methods to pinpoint discrete molecular and
cellular changes associated with certain pain pathologies could
greatly enhance our understanding of the molecular etiology of
nociception and potentially offer impartial data to better inform
patient care.
Figure 1. Concentration−response curves and IC50 values for STX
(circles) and [19F]STX (squares) tested against differentiated PC12
cells endogenously expressing TTX-sensitive NaV1.2 and NaV1.7
(filled), and CHO cells transiently expressing TTX-insensitive NaV1.5
(hollow) [n ≥ 3, error bars represent standard deviations; I/I0
=
remaining current/initial current]. Current versus time plots are shown
in Figure S2 in the Supporting Information.
Voltage-gated sodium channels (NaVs), which play a
fundamental role in the formation and conduction of action
potentials in excitable tissue,6 are of particular interest as
molecular markers of pain. Certain NaV isoforms are known to be
upregulated in peripheral sensory neurons and dorsal root
ganglia following neuronal or soft tissue injury, and have been
causally linked to enhanced nociceptive activity.7 Dysregulation
of sodium channels can result in ectopic neuronal firing and
neuronal hyperexcitability associated with neuropathic pain.8 A
method for noninvasive detection of NaVs could enable imaging
experiments in living subjects that correlate changes in protein
expression with sensitization to pain stimuli. Herein, we describe
Received: August 10, 2013
Published: November 21, 2013
© 2013 American Chemical Society
18012
dx.doi.org/10.1021/ja408300e | J. Am. Chem. Soc. 2013, 135, 18012−18015

Journal of the American Chemical Society
Communication
Figure 2. Analytical HPLC chromatogram of formulated [18F]STX
coinjected with [19F]STX.
Figure 5. (a) Autoradiography of injured left sciatic and uninjured right
sciatic nerves of SNI animals harvested at different time points
postinjection. (b) Time plot of normalized signal in the neuroma over
60 min, determined by autoradiography, and coplotted with uptake in
the surgical wound as determined in biodistribution studies. (c)
Autoradiography signal in sciatic nerves of SNI and sham-operated
animals at 60 min (n = 3; error bars represent standard errors; *p <
0.05). Normalized to control areas in the same nerve.
cell surface.10 Second, prior studies have demonstrated that
functional group substitutions at N21 in STX can be
accommodated with minimal perturbation of the ligand−
receptor binding interaction.11 In particular, N-(aminohexyl)-
Figure 3. Biodistribution of [18F]STX in sham-operated versus SNI rats.
Tissues were collected at 60 min postinjection (p.i.), and associated
radioactivity was measured with a gamma-counter (n = 3, error bars
represent standard errors). Wound is defined as surgical damage to non-
neural structures, including muscle and skin. Differences in uptake in
nontarget tissues were not statistically significant.
+
STX (STX-NH3 ), a compound prepared previously by our lab,
can be conjugated to a variety of groups through selective amide
bond formation.12 Such a method enables final-step introduction
of an appropriate radiolabel for PET imaging. Our radiolabel of
choice, fluorine-18, offers both a half-life (109.8 min) compatible
with the anticipated distribution kinetics of an STX-based
radiotracer and a low positron energy, which affords optimal
imaging resolution.13
Coupling of STX-NH3+ with the N-hydroxysuccinimide ester
of 4-fluorobenzoic acid ([19F]SFB) furnished the corresponding
benzamide ([19F]STX) in 87% yield when 3.0 equiv of the ester
were employed (Scheme 1). To optimize this coupling protocol
for radiosynthesis by minimizing the amount of ester, this
+
reaction was also examined using a 4-fold excess of STX-NH3
relative to [19F]SFB. Under these conditions, [19F]STX was
obtained in 1 h in 88% yield; the uncoupled STX-NH3 was easily
+
separated from the desired product by reversed-phase HPLC.
The affinity of [19F]STX for NaV was determined by
measuring block of peak Na+ current in whole-cell voltage-
clamp experiments (Figure 1, Figure S2). Against nerve growth
factor (NGF)-differentiated PC12 cells, which endogenously
express two TTX-sensitive isoforms (rNaV1.2 and rNaV1.7),14
the measured potency of the benzamide-modified toxin was only
5-fold less than that of the natural product (IC50 = 10.6 0.6 nM
for [19F]STX and 2.1 0.2 nM for native STX). Importantly,
both STX and [19F]STX were substantially less potent inhibitors
of recombinant NaV1.5 (CHO cells), a TTX-insensitive isoform
expressed principally in heart tissue (IC50 = 256 20 nM and
Figure 4. Uptake of [18F]STX in the neuroma of SNI rats compared to
background signal in representative nontarget tissues (*p < 0.05). Data
collected at 60 min p.i. (n = 3, error bars represent standard errors).
our initial efforts to develop a positron emission tomography
(PET) radiotracer for imaging peripheral nerve injury in an
animal model of neuropathic pain by targeting overexpression of
NaVs.
A number of small molecule and peptide ligands are known to
bind NaVs selectively and could potentially serve as PET probes.
Among this collection, saxitoxin (STX), a unique bis-
guanidinium natural product, offers two salient features. First,
STX binds reversibly and with low nanomolar affinity to the
extracellular pore of tetrodotoxin (TTX)-sensitive NaV isoforms9
and, thus, has the advantage of a readily accessible target on the
1.40
0.17 μM, respectively). These results indicate that
[19F]STX binds to TTX-sensitive NaVs with high affinity and
with a channel subtype selectivity profile similar to the parent
molecule.15
Radiolabeled STX was prepared from [18F]SFB, a reagent
commonly used for generating 18F PET probes (Scheme 1). A
18013
dx.doi.org/10.1021/ja408300e | J. Am. Chem. Soc. 2013, 135, 18012−18015

Journal of the American Chemical Society
Communication
Figure 6. Representative cross-sectional MR, PET, and fused PET-MR (left to right) image of the thighs showing the transverse plane through the
neuroma of a spared nerve injury (SNI) rat comparing [18F]STX uptake in the injured left sciatic nerve (yellow arrow) versus the uninjured right nerve
(white arrowhead). Red arrowheads: uptake in the wound. PET signal (normalized to signal from adjacent muscle) in the left nerve was 1.56 0.29, n =
5, and 1.11 0.14 on the control side; p < 0.05. As expected, increased PET signal is also observed in the urinary and excretory systems as well as the knee
joints.
semiautomated procedure for this synthesis was developed from
a conventional protocol16,17 (Figure S3) and provided [18F]SFB
in 31.1 12.4% radiochemical yield (n = 26), decay-corrected
from end of bombardment. Following treatment of [18F]SFB
with STX-NH3+ (Scheme 1), the product [18F]STX was purified
by semipreparative HPLC and formulated in PBS (Figure 2).
When 200−250 μg of STX-NH3+ were employed in the coupling
step, the desired product was generated in radiochemical yields
In order to gain an understanding of the distribution kinetics of
[18F]STX, tissue-associated radioactivity levels were determined
at serial time points following intravenous injection of the
radiotracer (Table S1). Animals were euthanized at 5, 10, 15, 30,
45, and 60 min postinjection of [18F]STX, and tissues were
collected, weighed, and analyzed for radioactivity with a gamma-
counter. Organs with higher levels of absolute uptake (above 1%
ID/g) included heart, small intestine, kidney, and liver. Low
tracer uptake in muscle (0.07−0.10%ID/g) and rapid blood
clearance (0.19%ID/g at 5 min p.i. to 0.05%ID/g at 60 min p.i.)
were noted, both of which favor minimal background signal in
PET imaging. The uptake in brain was quite low (∼0.1%ID/g)
and consistent with our expectation that the dicationic [18F]STX
should not cross the blood−brain barrier. Radioactivity levels in
bone were also ∼0.1%ID/g and constant over 60 min, indicating
negligible defluorination of the benzamide moiety. Comparison
of the results from both SNI and sham-operated animals at 60
min p.i. confirmed that nerve injury does not alter radiotracer
distribution in nontarget tissues (Figure 3).
of 6.6
3.7% (from [18F]SFB, n = 8); however, when the
+
amount of STX-NH3 was elevated to 500 μg, radiochemical
yields of [18F]STX increased to 16.4 9.5% (from [18F]SFB, n =
12). Using this latter procedure, the overall synthetic efficiency
and the specific activity of the radiotracer (1.63 1.14 Ci/μmol;
60.3 42.2 GBq/μmol; n = 12) were improved. Ultimately,
[18F]STX was obtained after a total synthesis time of 3−3.5 h
from end-of-bombardment with a radiochemical purity of ≥93%.
The successful synthesis of [18F]STX represents a rare example
of a complex natural product derivative incorporating a fluorine-
18 isotope.18
Evaluation of [18F]STX for PET imaging of neuromas was
performed with rats having undergone a spared-nerve injury
(SNI) surgical procedure. SNI is a validated experimental model
of neuropathic pain that involves transection of the tibial and
common peroneal branches of the sciatic nerve, leaving the sural
branch intact. This procedure introduces a neuroma at the site of
nerve injury and produces chronic pain lasting several months.19
In our studies, an allodynic response in the left hind paws
(operated side) of SNI animals was confirmed four weeks
postsurgery using von-Frey filaments (Figure S5). Immunohis-
topathology data of the surgically injured nerve of SNI animals
show elevated levels of protein expression of NaV1.3 and NaV1.7
as compared to the control.20,21
The metabolic stability of our STX tracer was confirmed
through HPLC analysis of urine samples taken from SNI animals
60 min postinjection (p.i.) of [18F]STX. In this experiment, 968
μCi [18F]STX with a specific activity of 1.1 Ci/μmol (40.4 GBq/
μmol) at the time of injection was administered to the rat (300
g). As expected given the small amount of tracer injected [0.47 μg
(0.90 nmol)] and the reduced potency of [19F]STX compared to
native STX (LD50 7 μg/kg, 23 nmol/kg),22 no signs of toxicity
were observed in either this experiment or subsequent studies
using lower mass doses. Analysis of the rat urine showed that 97%
of the radioactive material had a retention time identical to that of
[18F]STX (Figure S6), indicating that the radiotracer suffers little
if any degradation systemically or in the renal system.23
Consistent with our expectation, increased uptake of
[18F]STX was observed in the neuroma (0.36 0.18%ID/g at
60 min postinjection). Adjacent wound tissue, on the other hand,
retained only 0.13 0.02%ID/g (Figure 4). Further, tracer levels
in the neuroma were 3.8-fold higher than in the normal (or
uninjured) right sciatic nerve (0.09 0.03%ID/g). Radiotracer
retention in the neuroma was confirmed by ex vivo auto-
radiography of the injured and uninjured sciatic nerve segments
excised at multiple time points over a 1 h period. At 60 min
following injection, the normalized tracer uptake (determined
from autoradiography images) in the neuroma of SNI animals
was higher than that in sham-operated animals (3.07 0.54 vs
1.56 0.28; F(2,6) = 13.6; p < 0.01, Figure 5c). In addition,
[18F]STX retention in the neuroma remained constant over time
(p > 0.05, Figure 5a). Collectively, these data provide support for
specific [18F]STX uptake in the neuroma compared to the
surrounding tissue (Figure 5b).24
Encouraged by the results of our ex vivo experiments, we
performed preliminary PET scans of SNI rats with [18F]STX
using MRI for anatomic correlation. Since current small-animal
PET scanners are limited by their relatively poor spatial
resolution,25 it is difficult to identify peripheral nerves using
only PET. MRI offers greater spatial resolution and has been used
in conjunction with PET for anatomic localization. By
coregistering images from both modalities, the excellent spatial
and soft-tissue contrast resolution of MRI can be leveraged with
18014
dx.doi.org/10.1021/ja408300e | J. Am. Chem. Soc. 2013, 135, 18012−18015

Journal of the American Chemical Society
Communication
(5) Srinivas, S. V.; Deyo, R. A.; Berger, Z. D. Arch. Intern. Med. 2012,
172, 1016−1020.
the high sensitivity of PET to locate radiotracer signal in the
sciatic nerve.
Results from the PET-MRI studies revealed increased
[18F]STX uptake in the left injured nerve of SNI animals (Figure
6, PET signal (normalized to signal from adjacent muscle) 1.56
0.29, n = 5). By contrast, the uninjured sciatic nerve showed
(6) Hille, B. Ion Channels of Excitable Membranes, 3rd ed.; Sinauer
Associates: Sunderland, MA, 2001.
(7) Rogers, M.; Tang, L.; Madge, D. J.; Stevens, E. B. Sem. Cell Devel.
Biol. 2006, 17, 571−581.
(8) Dib-Hajj, S. D.; Black, J. A.; Waxman, S. G. Pain Med. 2009, 10,
1260−1269.
(9) Fozzard, H. A.; Lipkind, G. M. Mar. Drugs 2010, 8, 219−234.
(10) Morse, D. L.; Gillies, R. J. Biochem. Pharmacol. 2010, 80, 731−738.
(11) Andresen, B. M.; Du Bois, J. J. Am. Chem. Soc. 2009, 131, 12524−
12525.
(12) Ondrus, A. E.; Lee, H. D.; Iwanaga, S.; Parsons, W. H.; Andresen,
B. M.; Moerner, W. E.; Du Bois, J. Chem. Biol. 2012, 19, 902−912.
́
(13) Lasne, M.-C.; Perrio, C.; Rouden, J.; Barre, L.; Roeda, D.; Dolle,
F.; Crouzel, C. Top. Curr. Chem. 2002, 222, 201−258.
(14) D’Arcangelo, G.; Paradiso, K.; Shepherd, D.; Brehm, P.;
Halegoua, S.; Mandel, G. J. Cell Biol. 1993, 122, 915−921.
(15) [19F]STX also binds with nanomolar potency to NaV1.4 (Figure
S2). We anticipate that it should bind with nanomolar affinity to NaV1.1,
1.3, and 1.6, which are TTX-sensitive.
reduced signal (normalized PET signal 1.11
0.14 on the
control side; p < 0.05). These initial data demonstrate the
feasibility of using [18F]STX PET to investigate sodium channel
levels and distribution in disease states.
We have successfully prepared a radiolabeled form of a potent
sodium channel antagonist to generate a novel PET tracer,
[18F]STX. Elevated concentrations of [18F]STX are detected at
the location of nerve injury in an animal model of neuropathic
pain. Our work represents one of the first attempts to mark NaVs
in an intact, living animal model.26 We envision using this type of
PET agent for serial monitoring of changes in channel expression
levels that occur at injured nerves as a function of wound healing
and/or drug treatment. Such information may be correlated with
pain behavioral analyses to help shed light on the complex
molecular processes that underlie pain sensation.
(16) Tang, G.; Zeng, W.; Yu, M.; Kabalka, M. J. Label. Compd.
Radiopharm. 2008, 51, 68−71.
(17) Tang, G.; Tang, X.; Wang, X. J. Label. Compd. Radiopharm. 2010,
53, 543−547.
ASSOCIATED CONTENT
* Supporting Information
■
(18) Recent examples of other 18F-labeled natural products:
(a) Kiesewetter, D. O.; Jagoda, E. M.; Kao, C.-H.; Ma, Y.; Ravasi, L.;
Shimoji, K.; Szajek, L. P.; Eckelman, W. C. Nucl. Med. Biol. 2003, 30, 11−
24. (b) Lee, E.; Kamlet, A. S.; Powers, D. C.; Neumann, C. N.;
Boursalian, G. B.; Furuya, T.; Choi, D. C.; Hooker, J. M.; Ritter, T.
Science 2011, 334, 639−642. (c) Lee, J. H.; Peters, O.; Lehmann, L.;
Dence, C. S.; Sharp, T. L.; Carlson, K. E.; Zhou, D.; Jeyakumar, M.;
Welch, M. J.; Katzenellenbogen, J. A. Nucl. Med. Biol. 2012, 39, 1105−
1116.
S
Experimental details, supplementary figures and table. This
material is available free of charge via the Internet at http://pubs.
acs.org.
AUTHOR INFORMATION
■
Corresponding Authors
biswals@stanford.edu
chinf@stanford.edu
(19) Decosterd, I.; Woolf, C. J. Pain 2000, 87, 149−158.
(20) Black, J. A.; Cummins, T. R.; Plumpton, C.; Chen, Y. H.;
Hormuzdiar, W.; Clare, J. J.; Waxman, S. G. J. Neurophysiol. 1999, 82,
2776−2785.
jdubois@stanford.edu
Author Contributions
∥A.H., D.B., and W.P. contributed equally.
(21) Persson, A.-K.; Gasser, A.; Black, J. A.; Waxman, S. G. Exp. Neurol.
2011, 230, 273−279.
Notes
(22) Kohane, D.; Lu, N. T.; Gogkol-Kline, A. C.; Shubina, M.; Kuang,
̈
̈
The authors declare the following competing financial
interest(s): Three of the authors, Yeomans, Biswal, and Du
Bois, are cofounders and hold equity shares in SiteOne
Therapeutics, Inc., a start-up interested in developing diagnostic
probes for pain and novel medicines for pain treatment.
Y.; Hall, S.; Strichartz, G.; Berde, C. B. Reg. Anesth. Pain Med 2000, 25,
52−59.
(23) Hagenbuch, B. Clin. Pharmacol. Ther. 2010, 87, 39−47.
(24) In order to demonstrate the specificity of the observed uptake
further, we attempted to block radiotracer binding by coinjection with
an excess of [19F]STX. For details about this experiment refer to Figure
S8 and the supplementary discussion in the Supporting Information.
(25) Weber, S.; Bauer, A. Eur. J. Nucl. Med. Mol. Imaging 2004, 31,
1545−1555.
ACKNOWLEDGMENTS
■
The authors would like to thank Dr. John Mulcahy and Dr.
Stephanie Torreilles for their assistance with the animal studies as
well as the Stanford Small Animal Imaging Facility. This project
was supported in part by a Stanford University Bio-X Grant (S.B.,
J.D.B.) and NIH R01 NS045684 (J.D.B.), and by the NCI
ICMIC P50 Grant CA114747 (S.S.G.). W.H.P. was supported as
a Stanford Interdisciplinary Graduate Fellow (SIGF).
(26) Per
́
ez-Medina, C.; Patel, N.; Robson, M.; Lythgoe, M. F.; Årstad,
E. Bioorg. Med. Chem. Lett. 2013, 23, 5170−5173.
REFERENCES
■
(1) Institute of America (U.S.) Committee on Advancing Pain
Research Care and Education, in Relieving Pain in America: A Blueprint
for Transforming Prevention, Care, Education and Research, Washington
D.C., 2011.
(2) Englund, M.; Guermazi, A.; Gale, D.; Hunter, D. J.; Aliabadi, P.;
Clancy, M.; Felson, D. T. N. Engl. J. Med. 2008, 359, 1108−1115.
(3) Elfering, A.; Semmer, N.; Birkhofer, D.; Zanetti, M.; Hodler, J.;
Boos, N. Spine 2002, 27, 125−134.
(4) Vetti, N.; Krakenes, J.; Ask, T.; Erdal, K. A.; Torkildsen, M. D. N.;
Rorvik, J.; Gilhus, N. E.; Espeland, A. Am. J. Neuroradiol. 2011, 32,
1836−1841.
18015
dx.doi.org/10.1021/ja408300e | J. Am. Chem. Soc. 2013, 135, 18012−18015