Access to this full-text is provided by Wiley.
Content available from Analysis & Sensing
This content is subject to copyright. Terms and conditions apply.
Design and Testing of Diagnostic MRI/MRS Applications
Based On Signal Enhancement by Parahydrogen-Induced
Polarization
Francesca Reineri*[a]
Parahydrogen-induced polarization is a hyperpolarization meth-
od that exploits the spin order of hydrogen enriched in the
para-isomer, by means of a chemical reaction. Recently, its field
of application has been extended significantly, through the
introduction of non-hydrogenative PHIP (i. e. SABRE) and of
innovative h-PHIP strategies that allowed to increase the
intensity of the MR signals in molecules relevant for biological
applications. This Concept article aims to show the potentialities
of this hyperpolarization method in the field of diagnostics,
through the discussion of some of the reported applications of
parahydrogen polarized substrates. A section is also dedicated
to the methods that have been introduced for the purification
of parahydrogen polarized products, in order to make them
suitable for biological studies.
1. Introduction
The introduction of methods for hyperpolarizing nuclear spins
in solution, in particular of parahydrogen induced polarization
(PHIP)[1] and dissolution-dynamic nuclear polarization (d-
DNP),[2,3] increased the sensitivity of the 13C (and other nuclei)
NMR detection by several thousands of times. This allowed the
detection of molecules at millimolar concentration in biological
systems, on the timescale of a few seconds. PHIP polarized
substrates have been investigated since the early 2000s, when
the first in-vivo studies using hyperpolarized liquid substrates
have been published.[4,5] Then, the diagnostic applications of
hyperpolarized (HP) molecules in solution have been devel-
oped quickly thanks to d-DNP.
D-DNP is a versatile and powerful hyperpolarization
technique that can increase the sensitivity of, in principle, any
molecule. The wealth of chemical information that MRI and
MRS can provide, in the characterization of biological tissues,
non-invasively, is unique and 13C MRS is potentially the most
useful technique for monitoring tissues metabolism and its
alterations in diseases. The access to d-DNP hyperpolarized
metabolites has made possible the investigation of metabolic
pathways in-vivo, in real time. Nevertheless, its use is limited
due to the high costs of purchase and maintenance, to the
technical complexity and the long hyperpolarization cycles of
this technology.
PHIP is a cheap hyperpolarization method, less technically
demanding than d-DNP and the basic equipment needed to
carry out experiments using para-enriched hydrogen can be
easily found in a chemistry laboratory.[6] It allows hyperpolariza-
tion cycles at significantly faster pace, but its field of application
appeared quite limited, at least in the first two decades since
its discovery.[7,8] This is due to the fact that hydrogenative-PHIP
(h-PHIP) relies on the addition of a hydrogen molecule to an
unsaturated substrate, therefore, only those molecules for
which an unsaturated precursor (alkyne or alkene) exists,
seemed suitable for this hyperpolarization method.
In 2009, the scope of parahydrogen based hyperpolariza-
tion widened significantly thanks to the introduction of the
non-hydrogenative method SABRE (Signal Amplification by
Reversible Exchange), that does not imply any chemical
modification of the substrate. Recently, the use of h-PHIP has
been expanded too, making possible the hyperpolarization of
metabolites, such as pyruvate and fumarate[9,10] and the interest
in PHIP based hyperpolarization is increasing.
In the following paragraphs, the fundamentals of para-
hydrogen hyperpolarization will be presented and the potential
applications of PHIP polarized molecules to diagnostics in-vivo
and in-vitro will be described.
2. Hyperpolarization from parahydrogen
The combination of two magnetically equivalent nuclear spins
1
=
2leads to the formation of the four nuclear spin states (Eq. 1)
aaij ,bbij ,1ffiffiffi
2
pðabij þ baij Þ and 1ffiffiffi
2
pðabij baij Þ (1)
The first three states have a total spin angular momentum
I= +1 and form a triplet state, while the fourth state has a
total spin angular momentum I =0 and is a singlet state.
[a] Prof. F. Reineri
Department of Molecular Biotechnology and Health Sciences
University of Torino
Via Nizza 52
10126 Torino (Italy)
E-mail: francesca.reineri@unito.it
Homepage: https://www.cim.unito.it/website/PI/Reineri/home.php
© 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH. This
is an open access article under the terms of the Creative Commons Attri-
bution Non-Commercial License, which permits use, distribution and re-
production in any medium, provided the original work is properly cited and
is not used for commercial purposes.
Analysis & Sensing
www.analysis-sensing.org
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (1 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 5/13] 1
These states are often referred to using the singlet-triplet
notation, (Eq. 2)
Tþij ¼ aaij ,(2a)
T0ij ¼ 1ffiffiffi
2
pðabij þ baij ð2bÞ
Tij ¼ bbij (2c)
S0ij ¼ 1ffiffiffi
2
pðabij baij (2d)
where the subscript refers to the eigenvalue of the total
square spin angular momentum operator (I2¼I2
xþI2
yþI2
z) of
each state, for example I2S0
j i=0.
The singlet state is not coupled with the triplet states due
to different symmetry, in fact the triplet states are symmetric,
while the singlet state is antisymmetric upon exchange of the
two spins. In the case of hydrogen gas, the para and ortho spin
isomers are so well isolated that they form two physically
different substances, named parahydrogen and orthohydrogen,
respectively.[11]
At room temperature, all the four nuclear spin states are
equally populated, giving 25% of the para-form. Since the para
state is the most stable one, the fraction of parahydrogen
increases when the temperature is lowered being almost 100 %
at ~ 20 K. Nevertheless, symmetric interactions of the two spins
(e. g. intramolecular dipole-dipole interactions) cannot couple
the singlet and the triplet states, bringing into equilibrium the
population of the two forms of hydrogen, due to their different
symmetry. Therefore inhomogeneous interactions, such as
those on the surface of a catalyst, are needed to reach the
thermal equilibrium.[12] It follows that the enrichment in the
para-isomer obtained at low temperature can be maintained at
high temperature (room temperature or higher) provided that
the conversion catalyst is removed. This non-equilibrium
mixture is termed parahydrogen and contains a high spin order
that can be converted into hyperpolarization.[7,8,13]
Parahydrogen is magnetically inactive (the magnetic mo-
mentum is zero) and hyperpolarization can be obtained if its
symmetry is broken by means of a chemical transformation
that leads to chemically or magnetically inequivalent spins.
In Hydrogenative PHIP, a hydrogen molecule is added to an
unsaturated substrate (usually an alkene or alkyne) in a pairwise
manner, i. e. the spin correlation between the two hydrogen
nuclei is maintained. The hydrogenation reaction has also to be
fast, in the timescale of the relaxation processes, because the
spin states population tends to the thermal equilibrium, once
that the symmetry of the hydrogen molecule is broken and the
singlet state is lost.
In non-hydrogenative PHIP (SABRE),[14] the latent magnet-
ization of parahydrogen is catalytically transferred to a
substrate in reversible exchange on a metal complex (usually
an iridium Crabtree-like complex) (Figure 1). The metal center
works as a molecular scaffold that brings together the hydro-
gen molecule and the substrate. It follows that the main
requirement, for a SABRE substrate, is the capability of binding
to the iridium complex. The most commonly used substrates
are heterocycles containing an electro-donating atom, typically
nitrogen.[15] The process is completed without any chemical
change of the substrate, is continuous and refreshable.[16]
The easiest way to describe how parahydrogen can lead to
hyperpolarization of the 1H-NMR signals, in an h-PHIP experi-
ment, relies on the calculation of the population of the nuclear
spin states[18] of a molecule formed upon the addition of
parahydrogen to an unsaturated bond (Figure 2).
Following to the addition of hydrogen to the substrate, the
spin states of hydrogen (Eq. 2) S0ij ,Tþij ;T0ij ;Tij ) change
instantaneously into the states of the product molecule. For
sake of completeness, it must be said that this is an ideal
situation, because the hydrogenation catalyst has an effect on
the singlet state and singlet-triplet mixing occurs on hydro-
genation intermediates. Nevertheless, this so-called sudden
approximation is often used in the description of h-PHIP
experiments. The spin states of the two protons on the product
can be conveniently written as a combination of the singlet-
triplet states as reported in Eq. 3:[20]
y1¼Tþij ¼ aaj i (3a)
y2¼cos q
2
abij þ sin q
2
baij (3b)
Figure 1. Top: a generic example of hydrogenative-PHIP. Parahydrogen is
added, by means of a hydrogenation catalyst (Rh(I)diphos complexes are
usually applied), to an alkyne. Spin order is transferred to the alkene thus
obtained. Bottom: example of SABRE hyperpolarization of a generic
substrate (Subs) through the formation of a ternary adduct between with an
Ir(I) complex (L=PCy3or L=Imes=1,3-bis(2,4,6-trimethylphenyl)imidazol-2-
ylidine[17]) and parahydrogen. Notice that, in SABRE, the chemical structure
of the substrate is not changed, while in h-PHIP the hyperpolarized product
incorporates the parahydrogen molecule.
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (2 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 6/13] 1
y3¼sin q
2
abij cos q
2
baij (3c)
y4¼Tij ¼ bjbi:(3d)
The mixing angle θdepends on the scalar coupling
between the two protons and on their frequency difference
tan qð Þ ¼ JH1H2
nH1nH2 .
When the frequency difference( nH1 nH2 ) tends to zero,
the y3state (Eq. 3c) corresponds to the parahydrogen state,
therefore this is the only state that becomes more populated.
This is the ALTADENA condition,[19] in which the reaction occurs
at low magnetic field (usually at the geomagnetic field), where
the chemical shift difference between protons is neglectable
and they resonate at the same Larmor frequency. Following an
adiabatic (i. e. slow) transport of the sample into the high field
of the NMR spectrometer, the population of the states is
maintained and hyperpolarization results into in-phase signals
(hyperpolarized net magnetization) (Figure 2).
Vice-versa, when the frequency difference between the two
protons is significantly larger than their mutual scalar coupling
(( nH1 nH2 )@JHH), i. e. in the weak coupling condition, the
states y2¼abij and y3¼baij are formed, which are equally
populated because they both connect with the parahydrogen
state. This situation occurs in the PASADENA experiments and
hyperpolarized antiphase signals are obtained for the two
protons (Figure 2).
In practice, 1H hyperpolarization is not usually applied for
in-vivo MRI studies, due to the large background signal of
water protons and to the relaxation rate of 1H, that limits the
lifetime of the HP signals to few seconds. Heteronuclear signals,
and in particular 13C resonances, are usually preferred, because
of the large chemical shift range, the lack of background signals
Figure 2. The two variants of the hydrogenative-PHIP experiments (PASADENA and ALTADENA). A) Hydrogen addition to a triple bond, on an asymmetric
substrate (R1¼6 R2), leading to two chemically different protons (H1and H2). B) in a PASADENA experiment, the hydrogenation reaction takes place at high
field (such that nAnX
j j JAX ) and an AX system is “suddenly” formed, having the spin states αα,αβ,βα and ββ. Only those states that overlap with the
para-state are more populated (thick lines) while the other states are less populated (thin lines). Consequently, four intense NMR transitions are possible from
these states, as shown by the red and blue arrows and schematized in D. B’) In an ALTADENA experiment the reaction takes place at low field, in the fringe
field of the NMR magnet (i. e. nAnX
j j J) and the spin states are little changed, so the parahydrogen state adds predominantly into the product state with
the singlet character. After the hydrogenation reaction, the parahydrogenated sample is transferred into the high field of the NMR magnet for rf irradiation
and detection and the AX system is obtained. The passage from low to high field must be rapid with respect to the spin-lattice relaxation to preserve the spin
order, but also sufficiently slow that the population of the singlet-like low field eigenstate is transferred adiabatically into the high-field eigenstate with which
it correlates with continuity,[19] as shown in the correlation diagram. As a result, in-phase signals are obtained for each proton, while antiphase signals were
observed in the PASADENA experiment (see D and D’).
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (3 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 7/13] 1
and the longer T1, especially in some functional groups such as
the carboxylate group.
Hyperpolarization can be transferred spontaneously from
parahydrogen protons to heteronuclei through the scalar
couplings. Let’s consider the simplest case in which a
heteronucleus (X) is added to the two protons spin states and
an A2X system is formed. The eigenstates are the product of
the singlet/triplet states for the two protons and the αand β
state for the heteroatom (Eq. 4)
Tþaj i ¼ aaaj i;Tþbj i ¼ aabj i;(4a)
S0
jai ¼ 1ffiffiffi
2
pðab baÞaij ;S0
jbi ¼ 1ffiffiffi
2
pðab baÞbij ;(4b)
T0ai ¼ 1ffiffiffi
2
p
ðab þbaÞaij ;T0bij ¼ 1ffiffiffi
2
pðab þbaÞbij ;(4c)
Taj i ¼ bbaj i;Tbj i ¼ bbbj i (4d)
In this case, only the states S0
jaiand S0
jbiare populated
and, being transitions from these states forbidden, hyper-
polarization is not observed. Vice-versa, when the two protons
are added to magnetically different sites, an AA’X spin system
is formed, thanks to asymmetric coupling of the two protons
with 13C (Figure 3).[21] In this case, the spin states can be written,
using the singlet-triplet-Zeeman basis[22] as in Eq. 5:
y1¼Tþaj i,y2¼Tþbj i (5a)
y3¼S00aij ¼ cos qG
2
baaij sin qG
2
abaij ,(5b)
y4¼T00aij ¼ cos qG
2
abaij þ sin qG
2
baaij ,(5c)
Figure 3. A) Parahydrogen is added to dimethyl-[1-13C] acetylene dicarboxylate. The cis-addition is catalyzed by the [Rhdppb] +complex (dppb =bis-
diphenylphosphino butane) that is more usually applied for h-PHIP. In the product molecule dimethyl-maleate, a three-spins system is formed (H1H2X). B)
Population of the nuclear spin states before the hydrogenation reaction (H2enriched in para-isomer): the bold-faced line represent the more populated state
(para-state). B’) After hydrogen addition to the triple bond, the H1H2X spin system is formed and two states are more populated (bold-faced lines). Transitions
from these two states to the other, thermally populated states, are hyperpolarized (blue arrows): an antiphase signal is observed for 13C (C). B’’) Population of
the nuclear spin states after the application of MFC,[23] now the hyperpolarized transitions (blue arrows) are in-phase (emission signal, C’). C and C’ are the 13C
hyperpolarized signals before and after the application of MFC, respectively.
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (4 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 8/13] 1
y5¼S00bij ¼ cos qG
2
abbij sin qG
2
babij ,(5d)
y6¼T00bij ¼ cos qG
2
babij þ sin qG
2
abbij ,(5e)
y7¼bbaj i;y8¼bbbj i (5 f)
Of the eight states thus obtained, four derive from mixing
of the states S0aj i and T0aj i (total spin -1
=
2), S0bj i and T0bj i
(total spin 1
=
2) and the mixing angle is
qG¼arctan 2wH1H2
wD
H1H2X
with wH1H2¼2pJH1H2and wD
H1H2X¼
2pJH1XJH2X
(6)
The percentage of singlet of the states y3,y4,y5and y6is
a function of the J-couplings involved.
When the two HX coupling are equal, the system is
symmetric and the states correspond to pure singlet-Zeeman
and triplet-Zeeman states. Otherwise, when the two J couplings
(JH1X and JH2X) are different, the heteronuclear transitions
y3!y6and y4!y5have non-null probability and are
hyperpolarized (Figure 3), leading to hyperpolarized anti-phase
signals on 13C (i. e. hyperpolarized 1H-13C spin order) and the net
magnetization on the heteroatom is zero.(Figure 3) Spin order
can be converted into net magnetization (Figure 3) through the
application of magnetic field cycling,[5,23] as reported by K.Gol-
man in the first in-vivo application of a hyperpolarized
substrate. Other methods have been reported to achieve this
goal, based on the application of rf pulses.[24–27]
3. Advances in PHIP-based methods
3.1. PHIP based agents for angiography and perfusion
imaging
Magnetic Resonance Angiography is an important method for
diagnosing various medical conditions and it may take
advantage from the use of vasculature contrast agents such as
blood pool agents.[28] MR-based approaches to quantify
perfusion make use of injection of relaxation-based MR contrast
media or use magnetic spin labeling techniques. These
approaches suffers from limitations, in particular, some Gd-
based CAs have been suspended in Europe because of major
concerns associated to long-term accumulation of Gd contain-
ing species in the patients’ bodies.[29] On the other hand, the
arterial spin labeling approaches are limited by 1H longitudinal
relaxation.
The strong signal of hyperpolarized spins allows the direct
detection of hyperpolarized molecules and, in the first in-vivo
application of hyperpolaried liquids, hydroxyethyl propionate
(HEP) polarized by means of h-PHIP was used for catheter
tracking and angiography.[5,30] In that case, the polarization level
was in the order of 25–30%, an aqueous solution was obtained
directly from the hydrogenation reaction carried out in water
and the concentration of the agent was 0.5–1 M. Although the
toxicity profile of HEP did not make it suitable for clinical
translation, the contrast capacity of this new method for
angiographic examination was successfully demonstrated and
also it showed to be suitable to provide perfusion data about
the myocardium in pigs.[5]
Succinate is a metabolite that takes part into the TCA cycle,
i. e. it is central in the metabolic energy production. However,
since transport of dicarboxylic acids across biological mem-
branes is quite limited, real time metabolism has not been
observed using this hyperpolarized substrate. Anyhow, its
biocompatibility should make it suitable as a tracer for
angiography and perfusion. Hyperpolarized [1-13C]succinate has
been obtained, first, by means of parahydrogenation of sodium
acetylene dicarboxylate,[31] and later from parahydrogen addi-
tion to 1-13C-fumaric acid-d2.[32] In the second case, the choice
of a deuterated precursor maximizes the T1for the hyper-
polarized product. The polarization level was 15–20% and the
concentration of the hyperpolarized molecule was 3–4 mM.
The parahydrogenation was carried out in aqueous solution,
using a water-soluble rhodium catalyst and a dedicated
instrumentation for parahydrogen addition and transfer of spin
order.[33] Another route to obtain the same product has been
reported, using 13C labelled maleic anhydride as a precursor.[34]
In this case, the use of a lipophilic substrate allowed to carry
out the hydrogenation reaction in an organic hydrophobic
solvent. Following to the spin order transfer to 13C of the para-
hydrogenated anhydride, the aqueous solution of the hyper-
polarized succinic acid was obtained by means of the fast
hydrolysis of the anhydride and extraction of the product in in
the aqueous phase. The liquid-liquid extraction of the HP agent
corresponds also to the purification step, as will be discussed in
a following section.
Another molecule that has to be considered in this context
is 13C-urea, as it is not taken up and metabolized by most
tissues. It has been shown that hyperpolarized urea, obtained
by means of the d-DNP method, provides accurate assessment
of blood perfusion in animal cancer models.[35] Hyperpolariza-
tion of this substrate has been obtained also by using SABRE-
RELAY.[36] This variant of SABRE relies on the non-hydrogenative
PHIP of a transfer reagent (ammonia) and the subsequent relay
of hyperpolarization to the analyte of interest, though fast
proton exchange. This method allowed to extend the applica-
tion of PHIP quite significantly, to different classes of com-
pounds (carboxylic acids, alcohols, phosphates, glucose) further
than urea (13C and 15N labelled). However the attained polar-
ization level is still rather low (<0.5 %) and further improve-
ments of the method are necessary to make it suitable for in-
vivo applications. Another issue deals with the solvent in which
HP-urea is obtained (i.e. CD2Cl2), that is not compatible with
biological studies.
3.2. PHIP polarized substrates for metabolic imaging
The most relevant application of hyperpolarized substrates in
the diagnostic field is the assessment of metabolic processes
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (5 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 9/13] 1
in-vivo. Although many metabolites have been hyperpolarized
by means of d-DNP,[37] only a handful of them have been
successfully applied for these studies. Pyruvate is the most
widely used for the investigation of different diseases, and the
first translated to the clinic.[38,39] This molecule is at the
crossroads of different metabolic processes, as it is transformed
into lactate by the cytoplasmatic LDH (lactate dehydrogenase),
into alanine by alanine transaminase (ALT) and enters the TCA
cycle thanks to decarboxylation by means of pyruvate decar-
boxylase.
Hyperpolarization of pyruvate has been pursued by means
of both hydrogenative and non-hydrogenative PHIP based
techniques.
Using the SABRE method, the issue of the poor coordina-
tion of pyruvate to the iridium complex Ir(COD)(NHC)Cl (NHC =
IMes)[40,41] had to be solved. This task was tackled by tuning the
coordination capacity of the metal center by means of different
ligands. It was observed that the addition of DMSO (or its
derivatives) allows to obtain an unstable intermediate [Ir-
(H)2(DMSO)2Imes] that forms [Ir(H)2(η2-pyruvate)(sulfoxide)].[42]
In SABRE, the hyperpolarization level results from complex
interplay between different factors, in particular the catalyst
identity, the substrate and co-substrate concentration, temper-
ature and solvent properties.[43] The best conditions for SABRE
polarization of pyruvate brought to 1.7% 13C polarization on
the free molecule, while ~ 13 % polarization has been observed
as total polarization for free and coordinated pyruvate, at low
temperature.[44]
By using the SABRE polarized [1,2-13C2]pyruvate, it was
possible to observe its chemical transformation into CO2and
ethanolic acid, through the reaction with hydrogen peroxide.[45]
The h-PHIP approach to pyruvate hyperpolarization makes
use of the SAH (Side Arm Hydrogenation) method,[9] which
relies on the functionalization of the carboxylate group with an
unsaturated alcohol, to yield an ester derivative of pyruvate
containing the unsaturated group on the alcoholic moiety.
Hyperpolarization is obtained through the following steps: a)
hydrogenation of the unsaturated side-arm using p-H2; b) spin
order transfer from the parahydrogen protons to the 13C
carboxylate spin; c) cleavage of the hydrogenated alcohol by
means of hydrolysis. Hydrogenation is carried out in an
hydrophobic solvent (e. g. chloroform or toluene[46]) and the use
of an aqueous base for hydrolysis leads to the separation of
two phases. The catalyst is retained in the organic phase, while
the carboxylate salt is extracted in the aqueous one. The
aqueous solutions of HP metabolite thus obtained have been
used for metabolic investigations in-vivo[47] and in-cells[48,49]
(Figure 4). The SAH method can be applied, in principle, to any
substrate containing a carboxylate group such as acetate,[50]
lactate,[51] amino acids.[52,53] Concerning the hyperpolarization of
AAs, it should be mentioned that also other PHIP-based
Figure 4. A) time-resolved 13C-NMR spectra (20°flip angle, 2 s inter-scan delay) acquired following to the perfusion of HP-[1-13C]pyruvate (5 mM) through
prostate cancer cells (10 M PC3 cells) suspended in their growth medium. The build-up of the [1-13C]lactate signal can be observed, due to the metabolic
conversion of the injected pyruvate. The setup for perfusion of the HP metabolite through the cells is shown on the right: cells are placed inside the NMR
spectrometer and the temperature is set to 310 K, before the addition of the HP metabolite. The stacked plot of 13C-NMR spectra is reproduced from ref [49],
copyright (2020), with permission of the Publisher (article distributed under the terms of the Creative Commons Attribution License (CC BY)).
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (6 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 10/13] 1
methods have been used, from the direct h-PHIP of de-
hydrogenated precursors (dehydro-AAs)[54] to SABRE.[55]
D-DNP studies showed that HP [1,4-13C2]fumarate can be
used as a reporter of necrosis, being transformed into
[1,4-13C2]malate by fumarase, in those cells that had lost plasma
membrane integrity.[56]
HP [1-13C]fumarate has been obtained from the addition of
p-H2to [1-13C]acetlylenedicarboxylic acid (ADC).[10] The attain-
ment of E-alkenes, from catalytic hydrogenation of alkynes, is
unusual and their formation derives, mostly, from isomerization
of the product of the cis-addition to the triple bond. The
complex [Cp*Ru]+catalyzes the unorthodox formation of the E
isomer, thus leading to fumarate, instead of malate, which is
the product obtained using conventional hydrogenation com-
plexes. The reaction is quite efficient in aqueous solution,
leading to high polarization level (>24 %) on the 13C-carbox-
ylate signal and high concentration of the agent. Its metabolic
transformation into [1-13C] and [4-13C]maleate was observed in
cells lysates[57] and in-vivo.[58]
Several other metabolites, or their derivatives, were hyper-
polarized by means of PHIP based methods, as reported in
different reviews.[59–61] Although the polarization level reported
on some of them is good, very few were tested in-vivo. In fact,
although the signal enhancement may be sufficient for the in-
vitro detection, this is only the first step for identifying a HP
molecule as a potential candidate for acting as a diagnostic
probe and a sensor for altered metabolic processes.
3.3. Biocompatibility of the solutions of the HP products
The use of a metal complex to exploit the parahydrogen spin
order and transforming it into enhanced MR signals is
mandatory in both h-PHIP and SABRE. In h-PHIP experiments,
rhodium(I) complexes, having the general formula [Rh-
(diphos)(diene)]+(diphos =chelating phosphine) are com-
monly used, while Crabtree-like Ir(I) complexes are the systems
selected for SABRE experiments. A few cases have been
reported in which the activation of parahydrogen is obtained
by means of organic molecules (i.e. frustrated Lewis pairs),[62]
nevertheless their use as hydrogenation catalyst has not been
reported yet.
As a prerequisite for clinical trials, the safety profile of any
candidate tracer reagent, catalyst and other substances in the
product solution has to be carefully assessed. In order to be
suitable for in-vivo studies, the osmolarity must be physiolog-
ical and the concentration of the agent has to be relatively
high, in the order of 80–100 mM.
Although it has been shown that the Rh(I) complex used for
parahydrogen activation in h-PHIP may be compatible with
pre-clinical studies,[63] its removal from the solution of the
hyperpolarized product is necessary to obtain solutions suitable
for clinical applications. Different methods have been pro-
posed, that can be classifies as filtration, liquid-liquid extraction
and precipitation.
When the hydrogenation reaction is carried out in aqueous
solution, using a cationic metal complex, filtration of the
solution of the hyperpolarized product through a cation-
exchange filter[63] appears the most straightforward way to
reduce the concentration of the catalyst well below the toxicity
level.
Obviously, heterogeneous catalysts are much easier to
separate from a reaction mixture than homogeneous ones,
therefore routes based on the use of heterogeneous-PHIP (het-
PHIP) or het-SABRE has also been considered with the aim of
generating catalyst-free hyperpolarized fluids. Metal complexes
supported on solids (het-PHIP[64] or het-SABRE[65,66]) and metal
particles[67] have been tested. The observed polarization were
good in the organic solvents,[68] but significantly lower (<0.1 %
on 13 C) when the reactions are carried out in aqueous
solution.[50]
Metal scavenging agents have also been proposed to
completely eliminate metals from the aqueous solutions of the
hyperpolarized products. It was shown that silica particles,
functionalized with thiol groups, can precipitate the iridium
complexes used in the SABRE experiments, thus allowing to
filter off the catalyst.[69]
The liquid-liquid phase extraction is a purification method
that has been applied both to h-PHIP[34,9] and SABRE.[70] Hydro-
gen activation on the metal complex occurs in a lipophilic
solvent, such as chloroform, dichloromethane and toluene.[71]
The use of an organic solvent facilitates the hydrogenation
reaction thanks to the higher solubility of hydrogen and to the
better efficiency of the catalyst. In the h-PHIP experiments, the
lipophilic hyperpolarized product, obtained in the organic
phase, undergoes a reaction (usually a hydrolysis of an
anhydride or an ester) that makes it water-soluble, and the
hydrophilic molecule thus obtained is extracted in the aqueous
phase. The metal catalyst, being lipophilic, is almost completely
retained in the organic phase, while traces of organic solvents
may be found in the aqueous phase. Anyhow, these organic
solvent residues (toxic) residuals can be removed by means of
filtration[46] on a lipophilic resin. It was reported that this latter
purification passage causes a decrease of the hyperpolarization
which resulted to be about 20 % of that observed before
filtration.
The liquid-liquid extraction SABRE experiments, named
Catalyst Separated Hyperpolarization (CASH) SABRE, rely on the
repartition of the substrate between the organic and the
aqueous phase.[70] The hyperpolarization substrate is dissolved
in the aqueous phase, while the metal complex is in the
organic, and the two phases are emulsified, through vigorous
shaking, thus allowing an exchange of the substrate between
them. Hyperpolarization of the substrate (N-heterocyclic com-
pound) occurs in the organic phase, and hyperpolarization is
maintained when it is transferred into the aqueous one. The
separation of the two phases takes a few tens of seconds and,
being the entire process cyclic, hyperpolarization can be
restored, on the product, by successive shakings.
Precipitation of the hyperpolarized molecules from the
reaction mixture is another efficient way to remove all the side-
compounds (solvents, by-products, catalyst and other reagents)
from the hyperpolarized product. This method has been
applied for the purification of hyperpolarized [1-13C] fumaric
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (7 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 11/13] 1
acid, thanks to its very low solubility in the acidic form, (7 mg/
ml, 0.06 M)[72,73] in water. Therefore it was precipitated at acidic
pH and then redissolved using an aqueous base. A very
important aspect of these experiments was the non-trivial fact
that nuclear spin polarization is maintained through phase
transitions, provided that the passages from the liquid to the
solid state (and vice-versa) takes place in a sufficiently strong
magnetic field. In the experiments carried out with HP
fumarate, the field provided by a ~ 100 mT Halbach magnet
was sufficient. However it must be noticed that the applicability
of this purification strategy is limited by the physicochemical
characteristics of the substances.
4. Conclusions and outlook
Hyperpolarization of small molecules in solution provides a
powerful tool for in-vivo MRI applications, in particular for the
detection of altered metabolic processes in various diseases.
D-DNP is the gold-standard methodology to obtain hyper-
polarized molecules in the liquid state, but PHIP can take
advantage of the inherent low cost, the higher rate of the
polarization cycles and the ease of implementation.
Recently, the advancements in the PHIP-based methods
have significantly extended the range of molecules that can be
hyperpolarized by means of parahydrogen. Particularly relevant
to widen the scope of PHIP appears to be the application of
emerging methods such as SABRE-RELAY and PHIPX.
The routine application of the HP probes obtained by
means of PHIP to biological investigations still needs a
commercial, dedicated polarizer, although different instrumen-
tations have been developed for PHIP, as reviewed recently.[74]
Limitations to the application of PHIP polarized substrates
are still given by the hyperpolarization level and the presence,
in some cases, of toxic substances (solvents, catalyst). Therefore
further efforts have to be devoted to solve these issues.
Nevertheless, the availability of an affordable HP technique,
capable of providing doses of hyperpolarized metabolites
suitable for metabolic investigations at a preclinical stage,
would represent a new powerful tool for research laboratories
in the field of drug developments, in the investigation of cancer
and of other diseases.
Acknowledgements
I’d like to thank all the colleagues, students and professors, who
collaborated and helped in this research. The EU Horizon 2020
projects Marie Skłodowska-Curie (Grant Agreement No. 766402)
and the FETOPEN program (Grant agreement 858149, proposal
acronym Alternatives to Gd) are gratefully acknowledged. Open
Access funding provided by Università degli Studi di Torino within
the CRUI-CARE Agreement.
Conflict of Interest
The authors declare no conflict of interest.
Keywords: hyperpolarization ·metabolism ·NMR ·PHIP ·
pyruvate
[1] C. R. Bowers, Encycl. Magn. Reson. 2007,vol. 9; 750–769.
[2] J. H. Ardenkjaer-Larsen, B. Fridlund, A. Gram, G. Hansson, L. Hansson,
M. H. Lerche, R. Servin, M. Thaning, K. Golman, Proc. Natl. Acad. Sci. USA
2003,100, 10158–63.
[3] M. Goldman, Appl. Magn. Reson. 2008,34, 219–226.
[4] K. Golman, O. Axelsson, H. Johannesson, S. Månsson, C. Olofsson, J.-S.
Petersson, Magn. Reson. Med. 2001,46, 1–5.
[5] L. E. Olsson, C. M. Chai, O. Axelsson, M. Karlsson, K. Golman, J. S.
Petersson, Magn. Reson. Med. 2006,55, 731–737.
[6] F. Reineri, A. Viale, W. Dastrù, R. Gobetto, S. Aime, Contrast Media Mol.
Imaging 2011,6, 77–84.
[7] T. C. Eisenschmid, R. U. Kirss, P. P. Deutsch, S. I. Hommeltoft, R.
Eisenberg, J. Bargon, R. G. Lawler, A. L. Balch, J. Am. Chem. Soc. 1987,
109, 8089–8091.
[8] C. R. Bowers, D. P. Weitekamp, J. Am. Chem. Soc. 1987,109, 5541–5542.
[9] F. Reineri, T. Boi, S. Aime, Nat. Commun. 2015,6, 5858.
[10] B. Ripka, J. Eills, H. Kour, M. Leutzsch, M. H. Levitt, K. Munnemann, Chem
Comm 2018,54, 12246–12249.
[11] A. Farkas, Orthohydrogen, Parahydrogen and Heavy Hydrogen, Cam-
bridge University Press, London, 1935.
[12] H. S. Taylor, H. Diamond, J. Am. Chem. Soc. 1935,57, 1251–1256.
[13] C. Bowers, Phys. Rev. Lett. 1986,57, 2645–2648.
[14] R. W. Adams, J. A. Aguilar, K. D. Atkinson, M. J. Cowley, P. I. P. Elliott,
S. B. Duckett, G. G. R. Green, I. G. Khazal, J. Lopez-Serrano, D. C.
Williamson, Science 2009,323, 1708–1711.
[15] D. A. Barskiy, S. Knecht, A. V. Yurkovskaya, K. L. Ivanov, Prog. Nucl. Magn.
Reson. Spectrosc. 2019,114–115, 33–70.
[16] J.-B. Hövener, N. Schwaderlapp, T. Lickert, S. B. Duckett, R. E. Mewis,
L. A. R. Highton, S. M. Kenny, G. G. R. Green, D. Leibfritz, J. G. Korvink,
et al., Nat. Commun. 2013,4, 2946.
[17] M. J. Cowley, R. W. Adams, K. D. Atkinson, M. C. R. Cockett, S. B. Duckett,
G. G. R. Green, J. A. B. Lohman, R. Kerssebaum, D. Kilgour, R. E. Mewis, J.
Am. Chem. Soc. 2011,133, 6134–6137.
[18] D. Canet, C. Aroulanda, P. Mutzenhardt, S. Aime, R. Gobetto, F. Reineri,
Concepts Magn. Reson. Part A 2006,28, DOI 10.1002/cmr.a.20065.
[19] M. G. Pravica, D. P. Weitekamp, Chem. Phys. Lett. 1988,145, 255–258.
[20] G. Pileio, Prog. Nucl. Magn. Reson. Spectrosc. 2017,98–99, 1–19.
[21] J. Barkemeyer, M. Haake, J. Bargon, J. Am. Chem. Soc. 1995,117, 2927–
2928.
[22] G. Stevanato, J. Eills, C. Bengs, G. Pileio, J. Magn. Reson. 2017,277, 169–
178.
[23] H. Jóhannesson, O. Axelsson, M. Karlsson, Comptes Rendus Phys. 2004,
5, 315–324.
[24] M. Goldman, H. Jóhannesson, Comptes Rendus Phys. 2005,6, 575–581.
[25] S. Kadlecek, K. Emami, M. Ishii, R. Rizi, J. Magn. Reson. 2010,205, 9–13.
[26] J. Eills, G. Stevanato, C. Bengs, S. Glöggler, S. J. Elliott, J. Alonso-
valdesueiro, G. Pileio, M. H. Levitt, J. Magn. Reson. 2017,274, 163–172.
[27] S. Bär, T. Lange, D. Leibfritz, J. Hennig, D. Von Elverfeldt, J. B. Hövener, J.
Magn. Reson. 2012,225, 25–35.
[28] S. Kinner, S. Maderwald, N. Parohl, J. Albert, C. Corot, P. Robert, B. Jorg,
F. M. Vogt, Invest. Radiol. 2011,46, 254–259.
[29] https://www.federalregister.gov/documents/2017/01/11/2016-31262/
national-primary-drinking-water-regulations-announcement-of-the-re-
sults-of-epas-review-of-existing, 2016.
[30] M. Goldman, H. Jóhannesson, O. Axelsson, M. Karlsson, in Magn. Reson.
Imaging,2005, pp. 153–157.
[31] P. Bhattacharya, E. Y. Chekmenev, W. H. Perman, K. C. Harris, A. P. Lin,
V. A. Norton, C. T. Tan, B. D. Ross, D. P. Weitekamp, J. Magn. Reson.
2007,186, 150–155.
[32] E. Y. Chekmenev, J. Hövener, V. A. Norton, K. Harris, L. S. Batchelder, P.
Bhattacharya, B. D. Ross, D. P. Weitekamp, J. Am. Chem. Soc. 2008,130,
4212–4213.
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (8 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 12/13] 1
[33] J. B. Hövener, E. Y. Chekmenev, K. C. Harris, W. H. Perman, L. W.
Robertson, B. D. Ross, P. Bhattacharya, Magn. Reson. Mater. Phys. 2009,
22, 111–121.
[34] F. Reineri, A. Viale, S. Ellena, T. Boi, V. Daniele, R. Gobetto, S. Aime,
Angew. Chem. Int. Ed. 2011,50, 7350–7353; Angew. Chem. 2011,123,
7488–7491.
[35] C. Von Morze, P. E. Z. Larson, S. Hu, K. Keshari, D. M. Wilson, J. H.
Ardenkjaer-Larsen, A. Goga, R. Bok, J. Kurhanewicz, D. B. Vigneron, J.
Magn. Reson. Imaging 2011,33, 692–697.
[36] W. Iali, P. J. Rayner, S. B. Duckett, Sci. Adv. 2018,4, 1–7.
[37] K. R. Keshari, D. M. Wilson, Chemistry and Biochemistry of 13 C Hyper-
polarized Magnetic Resonance Using Dynamic Nuclear Polarization.,
2014.
[38] S. J. Nelson, J. Kurhanewicz, D. B. Vigneron, P. E. Z. Larson, A. L.
Harzstark, M. Ferrone, M. Van Criekinge, J. W. Chang, R. Bok, I. Park,
et al., Sci. Transl. Med. 2013,5, 198ra108.
[39] J. H. Ardenkjaer-Larsen, J. Magn. Reson. 2019,306, 124–127.
[40] L. S. Lloyd, A. Asghar, M. J. Burns, A. Charlton, S. Coombes, M. J. Cowley,
G. J. Dear, S. B. Duckett, G. R. Genov, G. G. R. Green, et al., Catal. Sci.
Technol. 2014,4, 3544–3554.
[41] A. J. Holmes, P. J. Rayner, M. J. Cowley, G. G. R. Green, A. C. Whitwood,
S. B. Duckett, Dalton Trans. 2014,44, 1077–1083.
[42] W. Iali, S. S. Roy, B. J. Tickner, F. Ahwal, A. J. Kennerley, S. B. Duckett,
Angew. Chem. Int. Ed. 2019,58, 10271–10275; Angew. Chem. 2019,131,
10377–10381.
[43] B. J. Tickner, O. Semenova, W. Iali, P. J. Rayner, A. C. Whitwood, S. B.
Duckett, Catal. Sci. Technol. 2020,10, 1343–1355.
[44] I. Adelabu, P. Tomhon, M. S. H. Kabir, S. Nantogma, M. Abdulmojeed, I.
Mandzhieva, J. Ettedgui, R. E. Swenson, M. C. Krishna, T. Theis, et al.,
Chem Phys Chem 2022,23, 131–136.
[45] B. J. Tickner, P. J. Rayner, S. B. Duckett, Anal. Chem. 2020,92, 9095–
9103.
[46] O. Bondar, E. Cavallari, C. Carrera, S. Aime, F. Reineri, Catal. Today 2021,
397–399, 94–102, DOI 10.1016/j.cattod.2021.11.030.
[47] E. Cavallari, C. Carrera, M. Sorge, G. Bonne, A. Muchir, S. Aime, F. Reineri,
Sci. Rep. 2018,8, 8366.
[48] E. Cavallari, C. Carrera, S. Aime, F. Reineri, ChemPhysChem 2019,20,
318–325.
[49] E. Cavallari, C. Carrera, G. Di Matteo, O. Bondar, S. Aime, F. Reineri, Front.
Oncol. 2020,10, 497.
[50] K. V. Kovtunov, D. A. Barskiy, R. V. Shchepin, O. G. Salnikov, I. P.
Prosvirin, A. V. Bukhtiyarov, L. M. Kovtunova, V. I. Bukhtiyarov, I. V.
Koptyug, E. Y. Chekmenev, Chem. A Eur. J. 2016,22, 16446–16449.
[51] E. Cavallari, C. Carrera, S. Aime, F. Reineri, Chem. A Eur. J. 2017,23,
1200–1204.
[52] L. Kaltschnee, A. P. Jagtap, J. McCormick, S. Wagner, L. S. Bouchard, M.
Utz, C. Griesinger, S. Glöggler, Chem. A Eur. J. 2019,25, 11031–11035.
[53] A. N. Pravdivtsev, G. Buntkowsky, S. B. Duckett, I. V. Koptyug, J. B.
Hövener, Angew. Chem. Int. Ed. 2021,60, 23496–23507.
[54] T. Trantzschel, M. Plaumann, J. Bernarding, D. Lego, T. Ratajczyk, S.
Dillenberger, G. Buntkowsky, J. Bargon, U. Bommerich, Appl. Magn.
Reson. 2013,44, 267–278.
[55] S. Glöggler, R. Müller, J. Colell, M. Emondts, M. Dabrowski, B. Blümich, S.
Appelt, Phys. Chem. Chem. Phys. 2011,13, 13759.
[56] F. A. Gallagher, M. I. Kettunen, D.-E. Hu, P. R. Jensen, R. In’t Zandt, M.
Karlsson, A. Gisselsson, S. K. Nelson, T. H. Witney, S. E. Bohndiek, et al.,
PNAS 2009,106, 19801–19806.
[57] J. Eills, E. Cavallari, C. Carrera, D. Budker, S. Aime, F. Reineri, J. Am.
Chem. Soc. 2019,141, 20209–20214.
[58] N. J. Stewart, H. Nakano, S. Sugai, M. Tomohiro, Y. Kase, Y. Uchio, T.
Yamaguchi, Y. Matsuo, T. Naganuma, N. Takeda, et al., ChemPhysChem
2021,22, 915–923.
[59] F. Reineri, E. Cavallari, C. Carrera, S. Aime, Magn. Reson. Mater. Phys. Biol.
Med. 2021,34, 25–47.
[60] A. J. Hovener, A. N. Pravdivtsev, B. Kidd, R. Bowers, S. Glöggler, K. V.
Kovtunov, R. Katz-Brull, K. Buckenmaier, F. Reineri, T. Theis, et al.,
Angew. Chem. Int. Ed. 2018,57, 11140–11162; Angew. Chem. 2018,130,
11310–11333.
[61] G. Buntkowsky, F. Theiss, J. Lins, Y. A. Miloslavina, L. Wienands, A.
Kiryutin, A. Yurkovskaya, RCS Adv. 2022,12, 12477–12506.
[62] K. Sorochkina, V. V. Zhivonitko, K. Chernichenko, V. V. Telkki, T. Repo,
I. V. Koptyug, J. Phys. Chem. Lett. 2018,9, 903–907.
[63] H. Chan, P. Bhattacharya, A. Imam, A. Freundlich, T. Tran, W. Perman, A.
Lin, K. Harris, E. Chekmenev, M. Ingram, et al., Proc. 17th Sci. Meet. Int.
Soc. Magn. Reson. Med. 2009,Honolulu, 2448.
[64] L. S. Bouchard, S. R. Burt, M. S. Anwar, K. V. Kovtunov, I. V. Koptyug, A.
Pines, Science 2008,319, 442–445.
[65] F. Shi, A. M. Coffey, K. W. Waddell, E. Y. Chekmenev, B. M. Goodson,
Angew. Chem. Int. Ed. 2014,53, 7495–7498; Angew. Chem. 2014,126,
7625–7628.
[66] K. V. Kovtunov, L. M. Kovtunova, M. E. Gemeinhardt, A. V. Bukhtiyarov, J.
Gesiorski, V. I. Bukhtiyarov, E. Y. Chekmenev, I. V. Koptyug, B. M. Good-
son, Angew. Chem. Int. Ed. 2017,56, 10433–10437; Angew. Chem. 2017,
129, 10569–10573.
[67] K. V. Kovtunov, I. E. Beck, V. I. Bukhtiyarov, I. V. Koptyug, Angew. Chem.
Int. Ed. 2008,47, 1492–1495; Angew. Chem. 2008,120, 1514–1517.
[68] S. Korchak, S. Mamone, S. Glçggler, Chem. Open 2018,7, 672–676.
[69] D. A. Barskiy, L. A. Ke, X. Li, V. Stevenson, N. Widarman, H. Zhang, A.
Truxal, A. Pines, J. Phys. Chem. Lett. 2018,9, 2721–2724.
[70] W. Iali, A. M. Olaru, G. G. R. Green, S. B. Duckett, Chem. A Eur. J. 2017,23,
10491–10495.
[71] O. Bondar, E. Cavallari, C. Carrera, S. Aime, F. Reineri, Catal. Today 2022,
397–399, 94–102.
[72] J. Eills, J. Alonso-valdesueiro, E. S. Marcano, J. Ferreira, S. Alom, G. J.
Rees, J. V. Hanna, M. Carravetta, M. H. Levitt, ChemPhysChem 2018,19,
40–44.
[73] S. Knecht, J. W. Blanchard, D. Barskiy, E. Cavallari, L. Dagys, E. van Dyke,
M. Tsukanov, B. Bliemel, K. Münnemann, S. Aime, et al., Proc. Natl. Acad.
Sci. USA 2021,118, 1–6.
[74] A. B. Schmidt, C. R. Bowers, K. Buckenmaier, E. Y. Chekmenev, H. de
Maissin, J. Eills, F. Ellermann, S. Glöggler, J. W. Gordon, S. Knecht, et al.,
Anal. Chem. 2022, 94, 479-501DOI 10.1021/acs.analchem.1c04863.
Manuscript received: April 27, 2022
Revised manuscript received: June 27, 2022
Accepted manuscript online: July 5, 2022
Version of record online: July 29, 2022
Analysis & Sensing
Concept
doi.org/10.1002/anse.202200028
Analysis & Sensing 2022,2, e202200028 (9 of 9) © 2022 The Authors. Analysis & Sensing published by Wiley-VCH GmbH
Wiley VCH Mittwoch, 26.10.2022
2206 / 257652 [S. 13/13] 1