Storage of hydrogen spin polarization in long-lived 13C 2 singlet order and implications for hyperpolarized magnetic resonance imaging

Article abstract of DOI:10.1021/ja404936p

Hyperpolarized magnetic resonance imaging (MRI) is a powerful technique enabling real-time monitoring of metabolites at concentration levels not accessible by standard MRI techniques. A considerable challenge this technique faces is the T₁ decay of the hyperpolarization upon injection into the system under study. Here we show that A_nA′ _nXX′ spin systems such as ¹³C₂-1,2- diphenylacetylene (¹³C₂-DPA) sustain long-lived polarization for both ¹³C and ¹H spins with decay constants of almost 4.5 min at high magnetic fields of up to 16.44 T without spin-locking; the T₁ of proton polarization is only 3.8 s. Therefore, storage of the proton polarization in a ¹³C₂-singlet state causes a 69-fold extension of the spin lifetime. Notably, this extension is demonstrated with proton-only pulse sequences, which can be readily implemented on standard clinical scanners.

Full text of DOI:10.1021/ja404936p

Communication
pubs.acs.org/JACS
13
Storage of Hydrogen Spin Polarization in Long-Lived C2 Singlet
Order and Implications for Hyperpolarized Magnetic Resonance
Imaging
Yesu Feng,‡ Thomas Theis,‡ Xiaofei Liang, Qiu Wang, Pei Zhou, and Warren S. Warren*
Department of Chemistry, Duke University, Durham, North Carolina 27708, United States
S
* Supporting Information
a specific field of 0.1 T) due to residual relaxation effects that
ABSTRACT: Hyperpolarized magnetic resonance imag-
are proportional to γ, such as intermolecular dipole−dipole
ing (MRI) is a powerful technique enabling real-time
monitoring of metabolites at concentration levels not
accessible by standard MRI techniques. A considerable
challenge this technique faces is the T1 decay of the
hyperpolarization upon injection into the system under
study. Here we show that AnA′nXX′ spin systems such as
13C2-1,2-diphenylacetylene (13C2-DPA) sustain long-lived
polarization for both 13C and 1H spins with decay
constants of almost 4.5 min at high magnetic fields of up
to 16.44 T without spin-locking; the T1 of proton
polarization is only 3.8 s. Therefore, storage of the proton
polarization in a 13C2-singlet state causes a 69-fold
extension of the spin lifetime. Notably, this extension is
demonstrated with proton-only pulse sequences, which
can be readily implemented on standard clinical scanners.
interactions. Ideally, hyperpolarized signal would be stored in
13C2/15N2 singlet states but detected later on nearby protons.
This might be conceivable with conventional polarization
transfer techniques (such as INEPT), but multiple bond X−H
scalar couplings in these systems are usually too small, giving
rise to significant T2 loss during polarization transfer, and
protons directly bonded to 13C are well known to drastically
shorten the polarization lifetime.
In the present study, we demonstrate convenient polarization
1
interconversion between a 13C2-singlet state and nearby H
magnetization with a modification to a previously demonstrated
“magnetization-to-singlet” pulse sequence (Figure 1a,b).14,15
n almost all preclinical and clinical magnetic resonance
Iimaging (MRI) experiments, the acquired signal originates
from the hydrogen nuclei in water. This is because in typical
MRI experiments polarization levels are on the order of 10−5,
and only signals from molecules at high concentrations such as
water can easily be detected. To overcome this limitation,
hyperpolarization techniques to enhance magnetization by
more than 10,000-fold have been developed.1−3 They have a
wide range of applications, including in vivo imaging of
molecular markers at millimolar concentrations. However, the
generality of this method is hindered by the T1 relaxation time,
which is on the order of a few seconds for most molecules.4 In
this context, the long-lived singlet state between a pair of
Figure 1. (a) The MSM sequence composed of M2S and S2M which
are separated by the relaxation delay τr and a combination of gradients
(GZ) and 90° pulses to suppress thermal artifacts. S2M is the inverse of
1
M2S. For 13C2-DPA, M2S and S2M can also be applied on H. (b)
The “magnetization-to-singlet” (M2S) sequence which converts bulk
magnetization to singlet order. The sequence is composed of two 180°
pulse trains where n1 = 2n2. See SI for details. (c) 13C2-
diphenylacetylene (13C2-DPA) with the J-coupling constants that
determine n1, n2, and τ.
1
strongly coupled spin-1/2 nuclei (mostly, H, 13C, or 15N) has
drawn considerable attention,5−8 since hyperpolarization stored
as singlet order might enable tracking of in vivo imaging agents
and their metabolism on time scales of minutes.
This sequence was first introduced to excite a pair of slightly
inequivalent spins,14,16 but we recently showed that a similar
approach works with chemically equivalent spins, as long as
they are asymmetrically coupled to other out-of-pair spins.15 In
this first demonstration for the CC′H2H2′ spin system in
diethyl oxalate,15 the interconversion between 13C magnet-
ization and 13C2-singlet order was shown. An important
difference between chemically inequivalent and chemically
Singlet states located on low-γ spin pairs such as 13C2 and
15N2 have been shown to sustain long-lived signal up to tens of
minutes.9,10 However, for hyperpolarized imaging, the signal-to-
noise ratio (SNR) can be estimated to scale linearly with γ in
body-noise-dominated MRI experiments, and with γ7/4 in coil-
noise-dominated NMR experiments.11,12 Although the 1H2
7
1
singlet state can also extend H magnetization lifetime, the
absolute singlet lifetime TS is generally relatively short (an
exception is presented by Buljubasich et al.,13 where the singlet
has to be unlocked by shuttling out of the high-field magnet to
Received: May 16, 2013
Published: June 20, 2013
© 2013 American Chemical Society
9632
dx.doi.org/10.1021/ja404936p | J. Am. Chem. Soc. 2013, 135, 9632−9635

Journal of the American Chemical Society
Communication
a
Table 1. Summary of Measured and Simulated T1 and Singlet Relaxation Time (TS) at Varied Field Strengths
1
field strength, B0/T
experiment type
theoretical T1(13C)/s
theoretical Ts/s
experimental T1(13C; H) /s
experimental TS/s
1
8.45
8.45
MSM(13C)
12.2
−
−
274.7 6.1
274.7 6.1
273.2 7.5
252.0 5.0
276.7 8.5
252.5 15.3
309.3 8.2
13C, 13.9; H, 3.7
288.4 3.7
267.1 5.4
282.7 3.8
244.7 1.4
223 9.1
261.7 7.3
−
MSM(13C) with H decoupling
−
−
1
8.45
M2S(1H), S2M(13C)
1
16.44
16.44
16.44
1.5
MSM(13C)
4.6
13C, 4.9; H, 3.8
M2S(13C), S2M(1H)
MSM(1H)
MSM(13C)
−
−
29.1
−
−
−
a
Detection always occurs on the S2M channel; for instance, S2M(1H) detects on the proton channel. Errors are obtained during the fitting
procedure of the data which also arise for the theoretical values because of oscillations generated by coherent effects in the beginning of the decay
traces as apparent from Figure S3 and explained in more detail in the SI.
equivalent spin systems is that, in the latter case, with spin
systems such as CC′H2H2′, the sequences can be selectively
1
pulsed on 13C or H. In the current study, this is the key to
explore the previously untapped 1H magnetization, transferring
1
it into 13C2-singlet polarization. Therefore, either 13C or H
magnetization can be hyperpolarized and converted into 13C2-
singlet polarization. Likewise, the 13C2-singlet polarization can
1
be converted back into either 13C or H magnetization for
detection.
The spin system we use here is 13C2-1,2-diphenylacetylene
(13C-DPA, Figure 1c). Recently, molecules containing 13C2-
Figure 2. Observed 13C2-DPA singlet state relaxation at 8.45 T with
the MSM sequence. The singlet relaxation is preceded by a fast triplet
relaxation on the order of T1 (12 s), and after ∼15 s a single
exponential decay (singlet lifetime, TS) of 288.4 3.7 s is observed
acetylene substructures have been reported to sustain long-lived
carbon singlet states.10 In this moiety the scalar coupling
between the 13C spins is consistently large (∼180 Hz), which
can suppress the singlet−triplet mixing caused by either small
chemical shift differences or small out-of-pair J-coupling
differences. Moreover, the linear configuration of 13C2-acetylene
can result in highly correlated chemical shift anisotropy (CSA)
perturbation on the two 13C spins17 and therefore largely
extended singlet lifetimes compared to 13C-T1 were ob-
served.10,18 It is worth noting that in the previous study
samples were degassed, nearby protons were deuterated, and
polarization could only be transferred from the singlet-bearing
spins (13C) to the 13C2-singlet state.
1
without H decoupling (black) and a TS of 267.1 5.4 is observed
1
with H decoupling during the delay between M2S and S2M (red).
Signal is normalized against equilibrium magnetization obtained by a
separate 90°-acquire experiment.
effect is proportional to (ΔJCH/JCC);28 thus, we expect a trivial
effect for 13C2-DPA. This is verified by a second measurement
(Figure 2, red curve), where the coherent mixing due to ΔJCH is
1
suppressed by H decoupling during the relaxation delay (τr in
In contrast, in our study with 13C2-DPA, the sample is
dissolved in CDCl3 without degassing, and the phenyl groups
are not deuterated. The small out-of-pair J-couplings (3JCH and
4JCH in Figure 1c) between 13C and the closest aromatic
protons provide access to the 13C2-singlet state. More
importantly, magnetization from the aromatic protons can be
transferred into 13C2-singlet polarization and vice versa. We
show that this enables proton-only experiments with observed
relaxation lifetimes of almost 4.5 min at 16.44 T despite the
very short proton T1 times of 3.8 s at the same field strength.
All relaxation lifetime measurements and theoretical predictions
are summarized in Table 1.
Figure 1a). Indeed, no significant change of TS was observed; in
fact, a slightly shorter lifetime was measured, which may be due
1
to heating effects from the extended irradiation during H
decoupling.
Because the resonance condition of MSM sequence is
independent of field strength, the same measurement can be
conveniently transferred to a different field strength (16.44 T).
This can help distinguish the major relaxation effects (dipole−
dipole interactions between the 13C spins is expected to be field
independent within the extreme narrowing regime,19−21
whereas CSA relaxation has a quadratic dependence on B021).
For 13C2-DPA, the nearly doubled magnetic field reduces 13C
T1 to about 35% of its original value, whereas TS is only
reduced to about 90% of its original value (compare row 1 and
4 in Table 1). Ignoring any other relaxation effects, the relative
contribution of CSA vs dipolar interactions can be estimated.
CSA contributes 66% to the 13C T1 relaxation at 8.45 T and
88% at 16.44 T, whereas the relative CSA contribution to
singlet-state relaxation is only 4% at 8.45 T and 14% at 16.44 T,
confirming that the CSA at the 13C sites must be highly
correlated. On the other hand, CSA has minimum relaxation
For 13C2-DPA, the difference of the J-couplings of either 13
C
3
4
spin to the same aromatic proton, ΔJCH = JCH − JCH, is
around 6.1 Hz, much smaller than JCC (182 Hz), resulting in
small singlet−triplet mixing. As described previously15 and
detailed in the Supporting Information (SI), the MSM
sequence parameters can be adjusted to resonate with this
mixing effect and thus achieve interconversion between 13C
bulk magnetization and 13C2-singlet state population. In a field
of 8.45 T, the singlet state lifetime TS is measured to be 288.4
3.7 s (Figure 2, black curve). This singlet state relaxation rate
has contributions from multiple mechanisms, including the
coherent mixing between singlet and triplet states. Previous
studies have shown that the relaxation rate due to the coherent
1
1
effects on H; thus, H-T1 remains unchanged. The theoretical
predictions22 of relaxation lifetimes agree well with exper-
imental results at 8.45 and 16.44 T (Table 1 and Figure S3).
This gives us confidence in a theoretical extrapolation by
9633
dx.doi.org/10.1021/ja404936p | J. Am. Chem. Soc. 2013, 135, 9632−9635

Journal of the American Chemical Society
Communication
simulation down to a field of 1.5 T (Table 1). The combination
of lower field and hyperpolarization is especially appealing since
SNR in hyperpolarized MRI experiments is generally field
independent. In addition, in cases of slight chemical
inequivalence, larger chemical shift differences can be tolerated
which satisfy the near equivalent condition, so the geometrical
constraints are relaxed.
As is shown in Figure 2, around 12% of the total 13C
magnetization has been converted into 13C2-singlet state
population, and therefore has an extended lifetime. This
conversion efficiency is generally lower in chemically equivalent
cases than in the case of a single pair of inequivalent 13C spins.
However, for chemically equivalent AnA′nXX′ spin systems
such as contained in 13C2-DPA, another opportunity arises.
Proton magnetization can be used by applying the M2S part of
the pulse sequence on the proton resonant channel as
demonstrated in Figure 3. In so doing the initial proton
Figure 4. 13C2-DPA singlet state relaxation at 16.44 T observed on the
1
1H channel. (a) M2S and S2M applied on the H channel such that
the 13C channel remains entirely unused. A TS of 261
7 s is
observed. (b) M2S applied on the 13C channel and S2M applied on
1
the H channel. A TS of 223 9 s is observed.
Figure 3. Observed 13C2-DPA singlet state relaxation at 8.45 T. A TS
of 288.4 3.7 s is observed (black, the same data as in Figure 2). This
is compared to a singlet state relaxation measurement where the M2S
useful strategy, and Figure 4b demonstrates the singlet state
1
relaxation after such an experiment. This long-lived H signal
originated from 13C thermal polarization is small thus we
1
part of the MSM sequence has been applied at the H resonance
1
observe residual H thermal polarization in this experiment
frequency instead of the 13C resonance frequency (red), resulting in an
(note the ∼0.7 baseline offset in Figure 4b). This may also
explain the smaller TS (223 s) obtained from this measurement.
On the other hand, recent studies have shown a higher
1
13C
enhancement of the acquired signal of γ H/γ ≈ 4. For the singlet
1
state relaxation using the H polarization, a TS of 282.7
3.8 s is
observed. The small discrepancy in the relaxation times is likely due to
subtle differences in the contributions of thermal signal in the two
measurements.
1
polarization level can be achieved with H hyperpolarization
within a much shorter time compared with polarization buildup
of 13C or 15N;23,24 therefore, converting H polarization into
1
1
13C2-singlet polarization may also be beneficial. Transfer of H
1
13C
polarization that is larger by nearly a factor of γ H/γ can be
magnetization from the hyperpolarizer to the imager may cause
transferred into the long-lived 13C2-singlet state (Figure 3
1
a significant signal loss due to short H T1, but this might be
resolved by converting 1H magnetization into 13C2-singlet
polarization within the hyperpolarizer.
1
13C
demonstrates a signal enhancement of γ H/γ ≈ 4). This is
because the M2S sequence, resonant with ΔJCH, can convert 1H
magnetization into the 13C2-singlet, following a very similar
pathway as that of 13C magnetization.15 In contrast, it is not
possible to pursue this strategy if only one pair of slightly
inequivalent spins14 is used. Another intriguing possibility is to
convert 13C2-singlet polarization, created from initial 1H
magnetization, back into proton magnetization for detection.
Last, we want to point out that 13C2-DPA, with a long-lived
hyperpolarized signal, may potentially serve as a desirable
labeling group. The chemical installation of 1,2-diphenylacety-
lene can be achieved modularly and readily.25 Such an internal
diarylacetylene functionality is chemically26 and metabolically
stable.27 For example, diarylacetylene moieties are embedded in
certain antibiotic drugs.28,29 These are LpxC inhibitors such as
CHIR-090.28 Additionally, efforts are underway to synthesize
drugs structurally related to suberolyanilide hydroxamic acid,30
an FDA-approved anticancer drug sold under the trade name
Vorinostat. It appears that such derivatives of 13C2-DPA are
easily synthesized by appropriate aromatic substitutions giving
access to a wide range of functional groups that can be used to
link to molecules with biological activity without severely
disturbing the singlet-bearing spin system.31
1
This can be done with pulses exclusively on the H channel;
therefore, this could be readily implemented on standard
clinical imagers without a 13C channel. A demonstration of this
proton-only MSM experiment is shown in Figure 4a. A
relaxation lifetime of 262 7 s is observed at 16.44 T, where
1H-T1 is merely 3.8 s.
For a constant fractional hyperpolarization (e.g., 10% carbon
hyperpolarization or 10% proton hyperpolarization) converted
into the singlet, the signal is identical starting from either
nucleus; however, the detection frequency out of the singlet
gives SNR proportional to γ or γ7/4 as mentioned earlier so
In summary, we have shown the M2S−S2M sequence can be
1
applied selectively on either 13C frequency or H frequency,
1
1
detection of H signal is certainly favorable (x4−x11 SNR for
leading to a 69-fold lifetime extension (∼4.5 min) for H
13C, x10−x56 for 15N). Accordingly, M2S(X)−S2M(1H) is a
magnetization. This is especially significant given the relative
9634
dx.doi.org/10.1021/ja404936p | J. Am. Chem. Soc. 2013, 135, 9632−9635

Journal of the American Chemical Society
Communication
proximity of the aromatic protons to the singlet bearing 13C
spins. The convenient polarization transfer between 13C/1H
magnetization and 13C2-singlet state population can further
improve the SNR in hyperpolarized MRI/NMR experiments.
For lower γ nuclei such as 15N2-singlet state, this will be even
more significant. On the other hand, MSM sequence entirely
(13) Buljubasich, L.; Franzoni, M. B.; Spiess, H. W.; Munnemann, K.
J. Magn. Reson. 2012, 219, 33.
(14) Pileio, G.; Carravetta, M.; Levitt, M. H. Proc. Natl. Acad. Sci.
U.S.A. 2011, 107, 17135.
(15) Feng, Y.; Davis, R. M.; Warren, W. S. Nat. Phys. 2012, 8, 831.
(16) Tayler, M. C. D.; Levitt, M. H. Phys. Chem. Chem. Phys. 2011,
13, 5556.
(17) Pileio, G. Prog. Nucl. Magn. Reson. Spectrosc. 2010, 56, 217.
(18) Pileio, G.; Bowen, S.; Laustsen, C.; Tayler, M. C. D.; Hill-
Cousins, J. T.; Brown, L. J.; Brown, R. C. D.; Ardenkjaer-Larsen, J. H.;
Levitt, M. H. J. Am. Chem. Soc. 2013, 135, 5084.
(19) Ernst, R. R.; Bodenhausen, G.; Wokaun, A. Principles of nuclear
magnetic resonance in one and two dimensions; Clarendon Press/Oxford
University Press: New York, 1987.
(20) Abragam, A. The principles of nuclear magnetism; Clarendon
Press: Oxford, 1961.
(21) Kowalewski, J.; Maler, L. Nuclear Spin Relaxation in Liquids:
Theory, Experiments, and Applications; Taylor & Francis: Independ-
ence, KY, 2006.
(22) Hogben, H. J.; Krzystyniak, M.; Charnock, G. T. P.; Hore, P. J.;
Kuprov, I. J. Magn. Reson. 2011, 208, 179.
1
resonant with the H frequency is very promising given the
1
faster and larger absolute polarization levels achieved with H
hyperpolarization. Even the potentially large water and fat
background is expected to be well separated from the signal of
the aromatic protons. In such case, an experiment signal would
1
be prepared and detected on H (high γ nucleus) while it is
stored on the 13C2/15N2-singlet state (low γ nucleus) to
diminish relaxation effects. Such an experiment does not even
require heteronuclear capabilities on the spectrometer/imager
and is readily achievable on standard clinical imagers.
ASSOCIATED CONTENT
* Supporting Information
■
S
(23) Jannin, S.; Bornet, A.; Melzi, R.; Bodenhausen, G. Chem. Phys.
Lett. 2012, 549, 99.
Detailed pulse sequence parameters, synthetic routes, sample
13C spectra, and description of the simulation methods. This
material is available free of charge via the Internet at http://
pubs.acs.org
(24) Bornet, A.; Melzi, R.; Linde, A. J. P.; Hautle, P.; van den Brandt,
B.; Jannin, S.; Bodenhausen, G. J. Phys. Chem. Lett. 2013, 4, 111.
(25) Chinchilla, R.; Najera, C. Chem. Soc. Rev. 2011, 40, 5084.
(26) Davis, H. E.; Allinger, N. L.; Rogers, D. W. J. Org. Chem. 1985,
50, 3601.
AUTHOR INFORMATION
Corresponding Author
warren.warren@duke.edu
■
(27) Chen, H.; Chen, W.; Gan, L.-S.; Mutlib, A. E. Drug Metab.
Dispos. 2003, 31, 122.
(28) Lee, C.-J.; Liang, X.; Chen, X.; Zeng, D.; Joo, S. H.; Chung, H.
S.; Barb, A. W.; Swanson, S. M.; Nicholas, R. A.; Li, Y.; Toone, E. J.;
Raetz, C. R. H.; Zhou, P. Chem. Biol. 2011, 18, 38.
(29) Liang, X.; Lee, C.-J.; Chen, X.; Chung, H. S.; Zeng, D.; Raetz, C.
R. H.; Li, Y.; Zhou, P.; Toone, E. J. Bioorg. Med. Chem. 2011, 19, 852.
(30) Minucci, S.; Pelicci, P. G. Nat. Rev. Cancer 2006, 6, 38.
(31) Komaromi, A.; Novak, Z. Chem. Commun. 2008, 4968.
Author Contributions
‡Y.F. and T.T. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research is funded by the National Science Foundation
(grant CHE-1058727) and the National Institutes of Health
(grant R01EB02122 to W.S.W. grants AI055588 and AI094475
to X.L. and P.Z.). The authors also thank Ryan M. Davis for
valuable discussion and for help editing the manuscript.
REFERENCES
■
(1) Ardenkjaer-Larsen, J. H.; Fridlund, B.; Gram, A.; Hansson, G.;
Hansson, L.; Lerche, M. H.; Servin, R.; Thaning, M.; Golman, K. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 10158.
(2) Kauczor, H. U.; Surkau, R.; Roberts, T. Eur. Radiol. 1998, 8, 820.
(3) Natterer, J.; Bargon, J. Prog. Nucl. Magn. Reson. Spectrosc. 1997,
31, 293.
(4) Kurhanewicz, J.; Vigneron, D. B.; Brindle, K.; Chekmenev, E. Y.;
Comment, A.; Cunningham, C. H.; DeBerardinis, R. J.; Green, G. G.;
Leach, M. O.; Rajan, S. S.; Rizi, R. R.; Ross, B. D.; Warren, W. S.;
Malloy, C. R. Neoplasia 2011, 13, 81.
(5) Carravetta, M.; Johannessen, O. G.; Levitt, M. H. Phys. Rev. Lett.
2004, 92, 153003.
(6) Carravetta, M.; Levitt, M. H. J. Am. Chem. Soc. 2004, 126, 6228.
(7) Sarkar, R.; Vasos, P. R.; Bodenhausen, G. J. Am. Chem. Soc. 2007,
129, 328.
(8) Warren, W. S.; Jenista, E.; Branca, R. T.; Chen, X. Science 2009,
323, 1711.
(9) Pileio, G.; Carravetta, M.; Hughes, E.; Levitt, M. H. J. Am. Chem.
Soc. 2008, 130, 12582.
(10) Pileio, G.; Hill-Cousins, J. T.; Mitchell, S.; Kuprov, I.; Brown, L.
J.; Brown, R. C. D.; Levitt, M. H. J. Am. Chem. Soc. 2012, 134, 17494.
(11) Hoult, D. I.; Richards, R. E. J. Magn. Reson. 1976, 24, 71.
(12) Hoult, D. I. Encyclopedia of Magnetic Resonance; Wiley: New
York, 2007.
9635
dx.doi.org/10.1021/ja404936p | J. Am. Chem. Soc. 2013, 135, 9632−9635