Irreversible catalyst activation enables hyperpolarization and water solubility for NMR signal amplification by reversible exchange.

Activation of a catalyst [IrCl(COD)(IMes)] (IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene; COD = cyclooctadiene)] for signal amplification by reversible exchange (SABRE) was monitored by in situ hyperpolarized proton NMR at 9.4 T. During the catalyst-activation process, the COD moiety undergoes hydrogenation that leads to its complete removal from the Ir complex. A transient hydride intermediate of the catalyst is observed via its hyperpolarized signatures, which could not be detected using conventional nonhyperpolarized solution NMR. SABRE enhancement of the pyridine substrate can be fully rendered only after removal of the COD moiety; failure to properly activate the catalyst in the presence of sufficient substrate can lead to irreversible deactivation consistent with oligomerization of the catalyst molecules. Following catalyst activation, results from selective RF-saturation studies support the hypothesis that substrate polarization at high field arises from nuclear cross-relaxation with hyperpolarized (1)H spins of the hydride/orthohydrogen spin bath. Importantly, the chemical changes that accompanied the catalyst's full activation were also found to endow the catalyst with water solubility, here used to demonstrate SABRE hyperpolarization of nicotinamide in water without the need for any organic cosolvent--paving the way to various biomedical applications of SABRE hyperpolarization methods.

Introduction

NMR hyperpolarization can temporarily increase nuclear spin polarization by 4–8 orders of magnitude, translating into corresponding increases in NMR and MRI detection sensitivity.[1,2] The sensitivity gains improve the detection limit, which enables new imaging methodologies—including hyperpolarized MRI for molecular imaging of dilute metabolites[3] and functional contrast agents such as hyperpolarized 129Xe.[4,5]

Parahydrogen-based hyperpolarization techniques are special in that they transform the spin order of the parahydrogen gas singlet state into observable nuclear spin polarization (up to order unity) of other molecules. Conventional parahydrogen induced polarization (PHIP)[6] has relied on the molecular addition of parahydrogen to an unsaturated molecular precursor across C=C or C≡C bonds.[7] This technique was further advanced by preserving the high nuclear spin order of nascent parahydrogen pairs on 13C sites with long T1 relaxation times exceeding 1 min.[8,9] Recent advances in catalysis have also enabled PHIP hyperpolarization of 13C contrast agents in biologically compatible aqueous media in seconds[10,11] with successful application in living organisms.[12−14]

One of the main fundamental limitations of conventional PHIP is the requirement for an unsaturated precursor for molecular addition of parahydrogen (p-H2). This limitation has been lifted by the introduction of the signal amplification by reversible exchange (SABRE) hyperpolarization technique in 2009.[15,16] As with conventional PHIP, SABRE involves p-H2 and substrate exchange on metal complexes; however, instead of hydrogenation, hyperpolarization of the substrate molecule is achieved by transfer of spin order originating from p-H2 during the lifetime of the transient complex.[15−17] Despite the rapid progress of this technique with an expanding list of substrate molecules[18,19] amenable to SABRE and recent demonstration of heterogeneous SABRE (HET-SABRE, where the solubilized substrate exchanges on the metal complex tethered to solid-phase support[20]), the chemistry and the mechanism of the SABRE process are still not fully understood. Moreover, despite promising efforts[19,21−i] to mitigate the poor solubility of SABRE catalysts in water, this method has not been demonstrated in pure aqueous media suitable for biomedical applications—the main drivers behind the development of most hyperpolarized magnetic resonance techniques.[24−ii]

The work presented here takes advantage of in situ high-resolution NMR detection of SABRE at 9.4 T,[26] which unlike low-field in situ SABRE detection[17] offers an advantage of exquisite chemical shift dispersion and high-resolution for distinguishing chemical compounds and their sites. Specifically, we demonstrate in situ detection of a hyperpolarized intermediate during activation steps of the most potent SABRE catalyst, [IrCl(COD)(IMes)][27,28] (IMes = 1,3-bis(2,4,6-trimethyl-phenyl)imidazol-2-ylidene; COD = cyclooctadiene). It is found that the SABRE effect is activated only when the COD moiety is removed via its hydrogenation to cyclooctene and cyclooctane in agreement with previous studies.[29] Furthermore, such activation requires the presence of excess substrate in order to avoid side reactions that irreversibly deactivate the catalyst (likely involving catalyst oligomerization). In situ studies of high-field SABRE with the activated catalyst are consistent with substrate enhancements resulting from cross-relaxation with hyperpolarized hydride spins in a manner akin to the spin-polarization induced nuclear Overhauser effect (SPINOE).[26,40] Finally, because of the catalyst’s chemical changes that occur during activation, it is possible to solubilize the activated catalyst in aqueous media, which previously has not been demonstrated without the use of organic cosolvents.[19,22] SABRE hyperpolarization of nicotinamide (the amide of vitamin B3) in water is shown as a proof of principle for potential biological applications.

Experimental Section

All solutions were placed in 9 in. long, 5 mm medium-wall (0.77 mm-thick) NMR tubes. The end of the NMR tube (∼3/4 in. long) was placed inside a piece of Teflon tubing (∼2 in. long, 1/4 in. OD, 3/64 in. wall thickness) to provide high-pressure connection to a wye (for "Y" shape connection) 0.25 in. push-to-connect fitting (McMaster Carr, P/N 5779K44) (Figure 1). The top two ends of the wye fitting were connected to ∼1 in. long pieces of the same 0.25 in. OD Teflon tubing. On one end, a push-to-connect reducing connector (McMaster Carr, P/N 5779K352) was used to connect a 0.125 in. OD Teflon tubing line to provide an exhaust path for “used” p-H2 gas exiting the solution. On the other top end, connectors (McMaster Carr, P/N 9087K121 and Western Analytical, P/N U-510–01) were used to allow 0.125 in. OD (1/16 in. ID) tubing, connected via a short flangeless nut (Western Analytical, P/N XP-208X), to extend all the way through the length of 5 mm NMR tube to reach to the bottom of 5 mm NMR tube to deliver fresh p-H2 gas to the bottom of NMR tube for efficient bubbling through the entire content of the NMR tube. High-pressure experiments were conducted by attaching a safety valve (one-way, calibrated to the exact pressure rating) to the exhaust gas line to allow p-H2 pressure of up to 5.1 bar. All NMR experiments were conducted using a 9.4 T (400 MHz) Bruker Avance III spectrometer.

Figure 1

Experimental setup for SABRE with controlled parahydrogen (p-H2) bubbling through a catalytic solution in a 5 mm NMR tube. The solution sits in a medium-wall NMR tube where p-H2 is delivered via 1/16 in. OD Teflon tubing. A wye push-to-connect adapter allows parahydrogen to flow into the NMR tube, while allowing for the expended gas to leave the detection volume via the exhaust line. The exhaust line can be capped with a pressure-calibrated safety valve to conduct SABRE at a higher p-H2 pressure.

The iridium [IrCl(COD)(IMes)][28] (Ir-IMes) catalyst was synthesized as described previously.[26,28] All SABRE studies of pyridine (Py) at high magnetic field were conducted with the Ir-IMes catalyst dissolved in either methanol-d4, ethanol-d6, or deuterated water. For the methanol-d4-based solutions, the complex consisted of 8 mM Ir-Imes and 47 mM Py in 1 mL of methanol-d4. The ethanol-d6-based solutions contained 8 mM Ir-IMes and 32 mM Py in 1 mL of ethanol-d6. All NMR experiments with Py were performed as high-field SABRE[26] experiments wherein p-H2 was bubbled through the solution and NMR detection was performed at 9.4 T. For in situ high-field SABRE detection with Py, the NMR tube was placed into the magnet similar to a typical solution NMR experiment. P-H2 gas (>90% para-state prepared using previously described instrumentation[30]) was bubbled through a given solution at a flow rate of 0.7 mL/s under atmospheric pressure for durations ranging from 30 to 60 s, depending upon the experiment. 1H NMR spectra were acquired immediately (3 ± 2 s) after the bubbling was stopped.

Experiments with nicotinamide utilized a solution of 3.5 mM Ir-IMes combined with 35 mM nicotinamide in 0.7 mL of ethanol-d6, D2O, or a mixture of both. Solutions of 33% and 50% D2O in ethanol-d6 were obtained by diluting the activated Ir-Imes/nicotinamide solution in ethanol-d6 (∼0.7 mL volume) with either one or two aliquots of D2O (∼0.35 mL each), respectively. All SABRE experiments with nicotinamide were conventional (low-field) SABRE,[15,16] where p-H2 was bubbled in the fringe field of the 9.4 T magnet (the solution in the NMR tube experienced 6 ± 4 mT SABRE polarization field; this field strength was measured using a portable gauss meter). Once p-H2 bubbling was stopped, the sample was quickly transferred inside a 9.4 T NMR magnet for 1H spectra acquisition with only a 5 ± 2 s delay between the end of p-H2 bubbling and the beginning of the NMR spectra acquisition. During activation, p-H2 bubbling was performed under atmospheric pressure with a flow rate of 0.7 mL/s. Once the nicotinamide solution was fully activated, p-H2 bubbling occurred under 5.1 bar of pressure with a flow rate of 1.2 mL/s.

All experiments were conducted at room temperature (∼300 K). The experiments conducted in situ (within an 9.4 T NMR magnet) utilized no active temperature control, with the room-temperature air supply providing temperature stabilization.

Results and Discussion

Activation of Ir-IMes-Py with p-H2

The SABRE activation of Ir-IMes with pyridine (Py) in methanol-d4 was investigated using high-resolution 1H NMR spectroscopy at 9.4 T (Figure 2). The spectrum of the nonactivated catalytic Ir complex (Figure 2a) clearly displays the 1H chemical shift signatures of the ortho-, meta-, and para-protons of Py, and the methanol solvent (4.8 ppm). Upon hydrogenation by p-H2 bubbling (Figure 2b), the formation of three resonances (at −12.3, −17.4, and −22.8 ppm) from hyperpolarized hydride species are observed, in addition to that from hyperpolarized orthohydrogen (o-H2)[26] at 4.5 ppm. During catalyst activation the high-field SABRE signal of Py was monitored in situ. As the catalyst is subjected to additional p-H2 bubbling, the hydride intermediate is quickly depleted, leaving a single hyperpolarized hydride signal at −22.8 ppm (Figure 2c).[16,26] Once the Ir-IMes-Py complex is fully activated, high-field SABRE signal enhancement is observed on the ortho-protons of pyridine (manifested by the large “negative” or emissive signals), o-H2 (manifested by “positive” or absorptive enhanced signals), and the hydride peaks. Therefore, the introduction of p-H2 to the catalyst solution initiates the hydrogenation of the COD moiety followed by its complete removal as detected here by high-field SABRE. This conclusion is further supported by the disappearance of COD proton NMR resonances in the spectra of the activated catalyst producing SABRE effect (Figures S1a, S1c-d, and 1d in ref (26)) A single intermediate hyperpolarized hydride species (seen as NMR resonances at −12.3 and −17.4 ppm) is likely to be short-lived (both chemically and hyperpolarization-wise with low 1H T1) and would be difficult to detect by performing the conventional SABRE hyperpolarization procedure in low magnetic field followed by sample transfer to the high magnetic field in a high-resolution NMR spectrometer (as well as by conventional NMR without p-H2). Indeed, Figure 3b shows that after ∼12 min of p-H2 bubbling, all signatures of the hyperpolarized hydride intermediate species are absent from the 1H NMR spectrum. The hyperpolarized intermediate hydride species is likely the result of p-H2 pair being added in the axial and equatorial positions (vs both p-H2 exchanging in equatorial position only as shown in structure 3, Figure 3a, of the fully activated Ir–catalyst complex). This observation is in agreement with pioneering work by Bowers and Weitekamp,[31] who detected similar dispersive signatures with Wilkinson’s catalyst after addition of p-H2 in equatorial and axial positions of hexacoordinate Rh(I) complex with similar geometry. Moreover, the phase of these dispersive resonances uniquely reports on the sign of homonuclear J-coupling between two hydride protons in this intermediate species, and it is negative.[31]

Figure 2

Series of 1H NMR spectra illustrating the hyperpolarization of the Ir-IMes-Py complex in methanol-d4 by high-field SABRE at 9.4 T.[26] (a) Thermally polarized spectrum of the catalytic complex and Py before p-H2 is introduced to initiate high-field SABRE. (b) Spectrum recorded immediately after the introduction of p-H2 (first point in Figure 3c) to the catalyst/Py solution. Note the dispersive peaks from the intermediate hydride species at −12.3 ppm and −17.4 ppm. (c) Spectrum of the Ir-IMes-Py reaction mixture after the catalyst has been completely activated through p-H2 bubbling. Note that only one hydride species is seen, manifested by an absorptive peak at −22.8 ppm. Additionally, the selective high-field SABRE[26] hyperpolarization of ortho-H-Py protons is apparent with the opposite phase of these proton peaks.

Figure 3

(a) Schematic showing the tentative mechanisms underlying activation and SABRE hyperpolarization with IMes-Ir catalyst and Py. The hydrogenation of COD containing catalyst forms intermediate species corresponding to transient NMR resonances (1) and (2), before forming the hyperpolarized catalyst-pyridine complex corresponding to NMR resonance (3). Additionally, without the presence of pyridine, the activation mechanism reverts to forming an inactive species, possibly a catalyst dimer (CAT:CAT)/oligomer, see Supporting Information) that does not hyperpolarize via SABRE. (b) NMR spectra demonstrating transient dispersive NMR resonances (1 and 2) of intermediate hydride species during the activation process. (c) Plot showing the decay (measured as the integrated NMR signal in magnitude mode) of the transient NMR resonances (1 and 2)] and the rise of hyperpolarized Ir–hydride resonance (3) at −22.8 ppm corresponding to the activated catalyst. The trend lines are added to guide the eye.

It should be pointed out that the high-field SABRE of Py ortho-protons effect correlates well with formation of hyperpolarized hydride at −22.8 ppm (estimated qualitatively through correlation of peak intensities of SABRE enhancement of ortho-H-Py peak with the rise of the Ir–dihydride peak), likely indicating that this species is responsible for the high-field SABRE effect. The proposed mechanism for catalyst activation/formation[16] is shown in Figure 3a, where the catalyst proceeds through hydrogenation steps (intermediate Ir–hydride species corresponding to transient NMR resonances 1 and 2) before the fully activated species (3) is formed to yield high-field SABRE hyperpolarization of Py. Additional evidence is provided by detection of cyclooctane/cyclooctene vs cyclooctadiene in the 1H NMR spectra[26] before and after full catalyst activation, respectively (data not shown). The in situ high-field SABRE may also be a potentially useful tool for studies related to identification and discrimination of classical and nonclassical hydrides in the context of SABRE.[32] We also note that while the enhancements of in situ high-field SABRE described here and previously[26] are relatively small, much larger enhancements can be observed with the application of appropriate pulse sequences.[33,34]

The proper activation of Ir-IMes requires not only H2, but also the presence of sufficient substrate to form a fully coordinated Ir–hydride complex with SABRE properties.[16] In the absence of the ligating substrate (Py), addition of H2 to the Ir-IMes catalyst likely results in (i) COD removal via hydrogenation and (ii) irreversible formation of inactive species (potentially dimer or trimer species). Our attempts to observe SABRE hyperpolarization with such catalyst states were unsuccessful—even with subsequent additions of Py and p-H2. Additional evidence of possible dimer/trimer formation is provided in the Supporting Information. The potential oligomerization would be in agreement with previous observation of dimer formation of Crabtree’s catalyst[35] (similar to the Ir-IMes catalyst used here) as well as observation of oligomerization in catalysts from the same family.[36] From the perspective of practical consideration, deactivation of this catalyst occurs when the substrate (Py or others) is either absent or present in insufficient concentrations, yielding a solution that remains yellow at the concentrations studied, while the properly activated system turns clear upon complete activation.

NMR Radio Frequency (RF) Saturation Studies

In the high-field SABRE effect, hyperpolarization occurred within the 9.4 T field of the high-resolution NMR spectrometer, and the following signal enhancements occur: Ir–hydride (at −22.8 ppm) and o-H2 (both with absorptive signals), and the selective (emissive) enhancement of the ortho-protons of pyridine (ortho-H-Py) in free and catalyst-bound forms, Figure 2c.[26] Here, the activated catalytic complex (3) was further probed to study chemical and polarization exchange in three hyperpolarized species by applying frequency selective RF irradiation during p-H2 bubbling (i.e., RF irradiation was applied throughout the entire 60 s-long bubbling step) during the build-up phase of the high-field SABRE effect. Application of a frequency-selective saturation RF pulse at the hyperpolarized hydride resonance frequency during SABRE polarization (Figure 4a) revealed that the hyperpolarized o-H2 signal (at 4.5 ppm) was suppressed along with the hydride signal (at −22.8 ppm) even with very soft (B1 ∼ 7.5 Hz) RF saturation. Importantly, the ortho-H-Py hyperpolarized signal intensities were suppressed such that the high-field SABRE enhancements were significantly reduced under conditions of a relatively strong (B1 ∼ 2.4 × 102 Hz) RF saturation on the hydride spectral band (Figure 4a, black trace). When RF saturation is applied at the o-H2 frequency (Figure S1, Supporting Information), a similar trend is observed, with nearly full depletion of the hyperpolarized Ir–hydride peak, and partial depletion of ortho-H-Py peaks as in Figure 4a, black trace. Shifting the RF saturation to the ortho-H-Py (corresponding to free Py in solution, 1H resonance at ∼8.55 ppm) resonances leads to a different trend (Figure S1c, where only the ortho-H-Py resonances of free Py and exchangeable Py (∼8.33 ppm) are clearly suppressed. RF saturation of free ortho-H-Py has little observable effect on the hyperpolarization enhancements for ortho-H2, Ir–hydride, and catalyst-bound (nonexchangeable in cis- position) ortho-H-Py.

Figure 4

High-resolution proton NMR RF-saturation study via in situ detection of high-field SABRE. (a) NMR spectra were recorded 3 ± 2 s after p-H2 bubbling was stopped inside the 9.4 T magnet of the NMR spectrometer. The color-coded NMR spectra were recorded under conditions of variable RF-power saturation (“soft” (green trace) with B1 ∼ 7.5 Hz; “hard” (black trace) with B1 ∼ 2.4 × 102 Hz; and “no” (red trace) RF saturation pulses) applied at the Ir–hydride resonance frequency during in situ p-H2 bubbling. Continuous RF irradiation on the hydride peak diminishes the polarization enhancements at both the o-H2 peak (4.5 ppm), and the ortho-H-Py peaks (8.05–8.55 ppm) in addition to destroying the hydride hyperpolarized signal. The same trend is observed for RF-saturation at o-H2 (see Supporting Information) as depicted in the scheme shown in part b. However, RF saturation on the ortho-H-Py has little observable effect on the polarization of Ir–hydride and o-H2 (see Supporting Information). Frequency-selective RF saturation was applied during high-field SABRE hyperpolarization (p-H2 bubbling for 60 s at 9.4 T).

The one-H-PHIP[37] mechanism is a potential candidate to explain the current and previous[26] observation of high-field SABRE effect, when at least one of the hydrogen atoms from p-H2 molecule is incorporated via hydrogen exchange. However, while such exchange is indeed present, and is concurrent with the SABRE processes, it occurs on a much longer time scale of days and therefore cannot explain high- or low-field SABRE effects.[26]

On the basis of the RF saturation experimental results described above the following high-field SABRE model is proposed (Figure 4b). Hyperpolarization of exchangeable and free ortho-H-Py (the residence time of exchangeable Py sites is on the order of 0.1–0.2 s[38]) is induced by the hyperpolarization pool of o-H2 and Ir–hydride, but not the other way around. This model is in agreement with previously proposed spin polarization-induced nuclear Overhauser effect (SPINOE):[4,26,39,40] The catalyst’s interaction with p-H2 produces high z-magnetization in the form of hyperpolarized hydride and o-H2. The close proximity of hyperpolarized hydride spins then drives the bound ortho-H-Py spins out of thermal equilibrium via nuclear dipolar cross-relaxation (which then results in enhanced spectra of free ortho-H-Py as well; note that emissive enhancements would be consistent with expectation that the double-quantum term of the NOE should dominate in the extreme-narrowing regime). On the other hand, reduction or destruction of the hydride/o-H2 z-magnetization suppresses the high-field SABRE enhancement—consistent with the conclusion that p-H2 itself is not the direct source of spin order for high-field SABRE.

Water-Soluble SABRE Catalytic Complex

Prior to activation, the Ir-IMes catalyst (with its COD ligand intact) is insoluble in water, even with the presence of SABRE substrates such as Py or nicotinamide added to the solution. Therefore, organic solvents such as methanol-d4, ethanol-d6 or DMSO are typically used to solubilize the Ir-IMes catalyst for SABRE hyperpolarization.[18] However, it was found that when the activated, hexacoordinate Ir-IMes complex generated in methanol or ethanol is dried, the resulting activated complex could be dissolved in pure water. This highly desirable characteristic is likely endowed by the chemical changes that accompany catalyst activation, including the loss of the hydrophobic COD moiety.

Motivated by the biomedical utility of hyperpolarization techniques, SABRE hyperpolarization of nicotinamide in aqueous media was tested owing to the molecule’s biological relevance (it belongs to the vitamin B group). Here, Ir-IMes was activated with nicotinamide in ethanol-d6, achieving hyperpolarization on the four aromatic protons located on the pyridine ring via SABRE at 6 ± 4 mT detected at 9.4 T (Figure 5a). The SABRE polarization enhancements for free nicotinamide protons approach ε ∼ 100 and are shown in Table 1. Two sequential aliquots of D2O were added that resulted in reductions of SABRE hyperpolarization enhancements (Table 1) for all protons except for Hd, which exhibited low polarization enhancements. A separately prepared (after full activation with H2 in ethanol) catalyst/nicotinamide sample was dried and then reconstituted in pure D2O. The structure of this complex is expected to be similar to that described by Cowley and co-workers for this catalyst and pyridine.[27] SABRE polarization enhancements approaching ε ∼ 30 were observed for most nicotinamide aromatic proton peaks in 100% D2O (Figure 5c). The preparation procedure for the solution corresponding to the spectrum shown in Figure 5c was different from those reported earlier by Zeng and co-workers,[22] Mewis and coworkers,[i] and Hövener and co-workers[19] in that once the SABRE catalyst activation is achieved in ethanol, the activated complex solution is dried on a rotovap first, and the catalyst/nicotinamide mixture is then reconstituted in 100% D2O. This approach offers multiple advantages. First, the water-insoluble COD and its hydrogenation products do not migrate into the aqueous phase. Moreover, the use of 10% alcohol in previous works[19] prevents potential in vivo injections, while 100% D2O solutions prepared using the method described here potentially allows for in vivo use of SABRE hyperpolarized contrast agents via intravenous administration.

Figure 5

Conventional (low-field) SABRE hyperpolarization of nicotinamide in ethanol-d6 (a) and in D2O (c) yields signal enhancement of four aromatic proton peaks of interest (Ha, Hb, Hc, and Hd), with the position of each proton labeled in the structure inset. The enhancement (ε) values for SABRE polarization of nicotinamide using the Ir-IMes catalyst can be found in Table 1. The respective thermally polarized reference 1H NMR spectra of nicotinamide in ethanol-d6 and D2O are shown in parts b and d; these thermal spectra are magnified by 20-fold relative to the corresponding hyperpolarized spectra. Note that the apparent “lower” peak heights in thermal spectrum of aqueous sample vs thermal spectrum of the sample in ethanol-d6 is likely the result of partial sample reconstitution from the dried solid into aqueous medium. Note that SABRE was conducted conventionally at low field (6 ± 4 mT), and it was detected by high-resolution proton NMR spectroscopy (at 9.4 T) using the activated Ir-IMes catalyst in ethanol-d6 and D2O.

Table 1

Polarization Enhancement Values (ε) for Four Aromatic Protons of Nicotinamide (Figure 5) Achieved via Conventional (Low-Field) SABRE and Detected by High-Resolution Proton NMR Spectroscopy Using the Activated Ir-IMes Catalyst in Ethanol-d6, D2O, and Their Mixturesa

medium	ε(Ha)	ε(Hb)	ε(Hc)	ε(Hd)
ethanol-d6 (100%)	–88.1 ± 0.3	–71.2 ± 0.8	–64.7 ± 2.9	9.1 ± 0.1
ethanol-d6 (50%), D2O (50%)	–74.1 ± 11.5	–54.8 ± 4.3	–40.0 ± 0.9	9.3 ± 0.4
ethanol-d6 (33%), D2O (67%)	–56.3 ± 4.7	–42.6 ± 0.2	–33.3 ± 4.1	10.0 ± 0.6
D2O (100%)	–33.3 ± 6.1	–29.5 ± 1.8	–24.0 ± 1.2	–11.0 ± 0.4

SABRE hyperpolarization was conducted at 6 ± 4 mT and detected at 9.4 T.

The smaller polarization enhancements achieved in D2O versus the organic solvents can be largely attributed to the reduced solubility of hydrogen gas in water. Previous studies have shown that H2 is up to 14 times more soluble by molar fraction in organic solvents than in water.[41−43] The reduced solubility of H2 in aqueous media limited the effectiveness of p-H2 bubbling in our setup and hindered the ability of p-H2 to interact with the catalytic Ir complex needed for SABRE to occur. However, potential solutions such as increasing the partial p-H2 pressure may recover the enhancement reduction in water versus that observed in ethanol. Moreover, the decrease in the SABRE signal enhancements in water vs ethanol may be also caused by the difference in T1 relaxation times, which can be lower in aqueous media compared to that in organic solvent media.

Furthermore, the low-field SABRE proton signal enhancements for nicotinamide in ethanol were significantly lower than SABRE enhancements reported by other groups, when using more advanced specialized SABRE equipment.[18,27] The experimental limitations (including the lack of precise field control, introduction of p-H2 via bubbling and potentially longer NMR sample shuttling time) resulted in significantly reduced SABRE signal enhancements in the presented study. Therefore, the reported enhancements for nicotinamide in water can be potentially significantly improved using more advanced hyperpolarization hardware as well the optimization of temperature[18] and p-H2 pressure, because these experimental variables are related to nicotinamide residency/SABRE contact time on this Ir-dihydride complex.

Conclusions

The activation and mechanism for hyperpolarization of the Ir-IMes catalyst in alcoholic and aqueous media was probed using in situ high-field SABRE.[26] Introduction of p-H2 to the catalyst in the presence of pyridine substrate initiates a hydrogenation of the COD moiety, leading to intermediate species. SABRE hyperpolarization of the activated catalytic complex results in the enhancement of Ir–hydride, o-H2, and ortho-H-Py protons in free and catalyst-bound forms. RF saturation studies probing polarization and chemical exchange of the hyperpolarized species support the conclusion that substrate hyperpolarization at high field (9.4 T)[26] arises from cross-relaxation with protons belonging to the hyperpolarized hydride/o-H2 spin bath.

Once activated and dried, the Ir-IMes catalyst was found to be soluble in aqueous media with either Py and nicotinamide substrates incorporated into its hexacoordinate structure. SABRE of nicotinamide substrate in D2O was conducted using conventional low-field SABRE[15,16,44] and detected at 9.4 T. The SABRE enhancements achieved in aqueous media, while significant (ε ∼ 30), were somewhat smaller than those in ethanol-d6 (Table 1)—which likely reflects the lower solubility of hydrogen gas in water versus ethanol.[19,43] Nevertheless, the SABRE enhancements in water can be potentially increased through the use of higher p-H2 partial pressures (e.g., ≥10 bar) frequently used in conventional PHIP.[10] Furthermore, additional gains of aqueous SABRE enhancements may be realized through better mixing of catalyst/substrate solution with p-H2, which has already been shown through the use of solution spray injection in p-H2 atmosphere,[45,46] or hollow fiber membranes,[23,47] and the optimization of other conditions (e.g., temperature and magnetic field).[18,i]

In any case, the ability to conduct SABRE hyperpolarization with biologically relevant molecules in aqueous media significantly increases the value of SABRE to generate hyperpolarized contrast agents for in vivo molecular imaging—eliminating one of the major shortcomings of the SABRE hyperpolarization method. Furthermore, other substrates beyond those studied here can be potentially hyperpolarized in aqueous media, including recently reported tuberculosis drugs pyrazinamide and isoniazid.[18] Finally, it may be possible to combine aqueous SABRE hyperpolarization with HET-SABRE,[20] potentially allowing the preparation of pure aqueous hyperpolarized contrast agents with a recyclable catalyst. The latter can potentially enable not only clinical translation of this hyperpolarization technique, but also high-speed and high-throughput production of hyperpolarized contrast agents at relatively low cost.