Extending the Lifetime of Hyperpolarized Propane Gas through Reversible Dissolution.

Hyperpolarized (HP) propane produced by the parahydrogen-induced polarization (PHIP) technique has been recently introduced as a promising contrast agent for functional lung magnetic resonance (MR) imaging. However, its short lifetime due to a spin-lattice relaxation time T1 of less than 1 s in the gas phase is a significant translational challenge for its potential biomedical applications. The previously demonstrated approach for extending the lifetime of the HP propane state through long-lived spin states allows the HP propane lifetime to be increased by a factor of ∼3. Here, we demonstrate that a remarkable increase in the propane hyperpolarization decay time at high magnetic field (7.1 T) can be achieved by its dissolution in deuterated organic solvents (acetone-d6 or methanol-d4). The approximate values of the HP decay time for propane dissolved in acetone-d6 are 35.1 and 28.6 s for the CH2 group and the CH3 group, respectively (similar values were obtained for propane dissolved in methanol-d4), which are ∼50 times larger than the gaseous propane T1 value. Furthermore, we show that it is possible to retrieve HP propane from solution to the gas phase with the preservation of hyperpolarization.

Introduction

Lung diseases have remained one of the major causes of human death during the past several decades. In particular, the most common lung disease is chronic obstructive pulmonary disease, which is responsible for ∼2.75 million deaths annually according to the Global Burden of Disease estimates.[1] Currently, the chest radiography and X-ray computed tomography (CT) methods find wide practical application for functional lung imaging.[2] Nevertheless, magnetic resonance imaging (MRI) might be attractive for pulmonary visualization, which, unlike CT, does not suffer from limitations caused by maximum permissible radiation doses. However, despite the widespread application of MRI in daily medical practice, pulmonary MRI is still challenging. In brief, problems with lung MR imaging can be divided into two groups: (a) problems related to the nature of lung tissue (fast signal decay due to the magnetic susceptibility differences at air–tissue interfaces[3] and loss of signal related to respiratory motion) and (b) problems related to the inherent properties of gas, namely, a low proton density and the resulting low signal-to-noise ratio. With regard to the first point, recent improvements in MRI technology have addressed these issues.

First, these problems can be addressed by using fast single-shot imaging with very short acquisition times. For this purpose, different pulse sequences such as steady-state free precession (SSFP) or half-Fourier acquisition single-shot turbo spin-echo (HASTE) sequences have been employed.[4] Through the use of ultrafast MRI sequences, subsecond image acquisition can be achieved.[5] The second approach, based on the respiratory triggering/gating[6,7] of fast spin-echo sequences, increases the acquisition time but provides better spatial resolution.

Promising opportunities are revealed by a new rapidly developing sampling method in MRI: compressed sensing[8−10] or sparse MRI. It is claimed that, if the sampling grid of k-space is randomized, then the image of an object that is sampled below the Nyquist criterion can be recovered to the original quality by using a sparsifying nonlinear transform that is built by the minimization of the l1 norm of the image data. This method was successfully applied for in vivo three-dimensional magnetic resonance spectroscopic imaging (MRSI) experiments[11,12] with the t–k–k undersampling technique demonstrating significant reduction of the acquisition time.

Nevertheless, the low signal-to-noise ratio of proton nuclear magnetic resonance (1H NMR) spectroscopy remains the main challenge for lung 1H MRI. The most efficient way to significantly improve sensitivity in pulmonary MRI is to use contrast agents with nonequilibrium nuclear spin polarization or hyperpolarization.[13,14] The first approach for the hyperpolarization of gases is the spin-exchange optical pumping (SEOP)[15] of 129Xe or 3He, which enables signal enhancement by 5 orders of magnitude.[16] The possibility of using HP noble gases (129Xe and 3He) for MRI has been widely explored.[17] In fact, currently, HP noble-gas MRI is approved for use as a clinical research imaging modality for patients with pulmonary disease.[18] However, despite these promising results, the widespread application of HP noble gases in pulmonary MR imaging seems exceedingly complex because of the requirement of having heteronuclear radio-frequency (rf) channels and rf probes, which are not included in standard equipment in commercially available MRI scanners.

Therefore, the development of gaseous contrast agents for MR imaging with proton detection is highly desired. The easiest way to obtain proton-hyperpolarized gases is to use parahydrogen-induced polarization (PHIP).[19,20] PHIP is based on the generation of nonequilibrium magnetization through the use of parahydrogen, the spin isomer of the hydrogen molecule with zero nuclear spin. Parahydrogen itself cannot be detected by NMR spectroscopy because of its zero nuclear spin, so to obtain observable nonequilibrium magnetization, it is necessary to break the symmetry of the parahydrogen molecule, for example, through the use of a hydrogenation reaction.[21] If both atoms from a parahydrogen molecule are added to the same substrate molecule (pairwise hydrogen addition) and their positions in the product molecule are magnetically nonequivalent, the NMR signals of such product molecule become significantly enhanced (in theory, up to ∼104–105-fold).[19]

The first demonstration of the in vivo use of an HP contrast agent produced by the PHIP technique was a 13C MR angiography study published in 2001.[5] Despite its several shortcomings (e.g., the presence of a toxic metal complex used as a homogeneous hydrogenation catalyst, the use of an organic solvent, and 13C MRI acquisition), this work served as a catalyst for a large number of studies devoted to the development of the PHIP technique for MR imaging[22−100] through the use of HP contrast agents in solution. The pioneering works[26,27] that demonstrated the production of hyperpolarized gas (in that case, propane) through the use of PHIP technique were based on the utilization of heterogeneous hydrogenation catalysts. This work motivated further research on MR imaging with HP propane. Recently, a number of studies have been conducted that demonstrated successfully the feasibility of HP propane in the two-dimensional[28−32] and three-dimensional[33,34] imaging of various model objects.

However, despite the exciting results on HP propane utilization in MR imaging, its biomedical use is complicated because of the relatively short lifetime of hyperpolarization governed by the spin–lattice relaxation time T1. Typically, T1 values are less than 1 s at standard temperature and pressure.[35]

The fast T1 relaxation of gaseous propane is caused by the spin-rotation interaction.[36] In a rapidly rotating molecule (or part of a molecule, such as a methyl group), the local magnetic field fluctuations are generated by the rotational motion, leading to fast T1 relaxation. The spin-rotation relaxation is important for nuclei of small molecules in the gas phase and for nuclei in highly mobile environments in large molecules in solution (e.g., protons of CH3 groups in proteins).

In solution, the T1 times of small molecules can often be increased by reducing the number of protons by deuterating the molecule. However, for gases this effect is quite small: for example, the T1 value of the CHD2 group of propane-d6 at 9.4 T is 684 ± 27 ms, whereas the T1 value of the CH3 group of propane is 616 ± 16 ms.[37] This is consistent with the predominance of the spin-rotataion mechanism for nuclear spin relaxation in gaseous propane.

Another possibility for slowing down the decay of hyperpolarization is to store the molecules in a slowly relaxing singlet state (called a long-lived spin state, or LLSS).[38−40] The lifetime-limiting factors of singlet-state relaxations were investigated recently.[41,42] Furthermore, a relatively long hyperpolarization lifetime, TLLSS, was achieved for HP propane-d6 (∼6.0 s)[37] and for HP propane (4.7–13.1 s)[33,43] at low magnetic fields.

Therefore, this work logically follows from a number of our previous studies[30,33,37] that concentrated on the development of 1H magnetic resonance imaging of propane, including hyperpolarized (HP) propane, for potential application in functional lung imaging. Here, we address the issue of the short lifetime of HP propane and demonstrate that it is possible to extend the lifetime of the hyperpolarized state of propane at high magnetic fields by dissolving the propane in organic solvents. We found that the estimated values of hyperpolarization decay times for propane are 35.1 and 28.6 s for the CH2 group and the CH3 group, respectively, which are about 1–2 orders of magnitude greater than the T1 times for gaseous propane.[37]

Materials and Methods

NMR and MRI Experiments

Commercially available hydrogen, propylene, propane, acetone-d6 (Astrachem, 99.7%), and methanol-d4 (Astrachem, 99.7%) were used as received. Rh/TiO2 catalyst with 1 wt % metal loading was provided by the group of Prof. V. I. Bukhtiyarov (Boreskov Institute of Catalysis, Novosibirsk, Russia) and was described in detail elsewhere.[33] For the production of hyperpolarized propane, hydrogen gas was first enriched with parahydrogen isomer up to 50% by passing it through an ortho–para conversion catalyst FeO(OH) maintained at 77 K (hereinafter, the resultant hydrogen gas is referred to as p-H2). The propene/p-H2 (1:4) mixture was passed through the Rh/TiO2 catalyst held at 200 °C to form hyperpolarized propane, which was supplied through a 1/16-in.-o.d. polytetrafluoroethylene (PTFE) capillary to a 10-mm NMR tube located in the NMR spectrometer for detection [an adiabatic longitudinal transport after dissociation engenders nuclear alignment (ALTADENA)[44] experiment]. In the case of dissolution experiments, HP propane was supplied to the bottom of the NMR tube containing 2 mL of deuterated solvent (acetone-d6 or methanol-d4). The concentration of dissolved propane was calculated through the use of an internal standard method (toluene) and was approximately 50 mM. The HP NMR spectra of liquid solutions were acquired after a rapid interruption of gas flow to avoid magnetic field inhomogeneities caused by bubbles. The NMR spectra of liquids in thermal equilibrium were acquired after the complete relaxation of hyperpolarization with the gas flow stopped.

All dissolution experiments were carried out at 25 °C. The gas flow rate was varied from 1.89 to 3.13 mL/s through the use of an Aalborg rotameter. A scheme of the experimental setup is presented in Figure S1 (Supporting Information). T1 measurements were carried out using the standard inversion–recovery sequence.

The extraction of dissolved HP propane to the gas phase was carried out as follows: First, the reaction mixture was supplied through the reactor with Rh/TiO2 catalyst to an NMR tube containing 0.5 mL of acetone-d6. The position of the NMR tube inside the NMR spinner turbine was adjusted for the detection of the gas phase just above the liquid (Figure S2). Next, the displacement of HP propane from the solution was achieved by switching the flow of the reaction mixture (propene/p-H2) from the catalyst to a bypass line using two valves (Figure S2). During this experiment, the NMR spectra of the gas phase were acquired continuously through the use of a small-angle (10°) rf excitation pulse.

NMR spectra were acquired on a 300 MHz Bruker AV 300 NMR spectrometer. All MR images were obtained on a Bruker Avance III 400 MHz NMR spectrometer with microimaging accessories using ParaVision 5.1 software based on Bruker TopSpin 3.0 and a commercial 30-mm 1H/31P birdcage rf coil (Bruker). Before MRI experiments were conducted in ParaVision, the magnetic field was shimmed in TopSpin to provide field homogeneity for the sample in a 10-mm NMR tube. The fast low-angle shot (FLASH)[45] pulse sequence was used for image acquisition; the spectral bandwidth was 50 kHz, and the pulse angle was 5°. The total time of the MRI experiment was 1.7 s, the echo time (TE) was 14.8 ms, and the repetition time (TR) was 26.8 ms. The field of view (FOV) was 5 × 5 cm for the axial slice orientation and 5 × 3 cm for the sagittal plane, and the slice thickness was equal to 50 mm for both experiments. This means that the sample was projected onto the XY and XZ planes for the axial and sagittal orientations, respectively. The matrix size was 64 × 64, so the image resolution was approximately 0.8 × 0.8 mm2 for the axial slice orientation and 0.8 × 0.47 mm2 for the sagittal slice orientation. Because the two polarized signals of propane have opposite phases in ALTADENA NMR spectra,[44] the net integral is approximately zero. Therefore, to avoid mutual cancelation of the two signals in an MRI experiment, we used a selective excitation rf pulse centered at the frequency of the ALTADENA peak of the propane CH3 group.

Results and Discussion

In this article, we address the issue of the preservation of propane hyperpolarization by its dissolution in deuterated organic solvents. Gaseous HP propane was obtained by passing a propene/p-H2 gas mixture through a reactor with a Rh/TiO2 catalyst (Figure ), which usually provides both higher catalytic activity and higher PHIP levels than other heterogeneous catalysts.[33] Indeed, it was found that utilization of the Rh/TiO2 catalyst allows ∼100% conversion of propene to be achieved. Also, as expected, pronounced PHIP effects with characteristic line patterns in the NMR spectra were observed for the propane molecule (Figure c), with one line in emission and another in enhanced absorption.

Figure 1

(a) Scheme of the experimental setup for HP propane production. (b) Reaction scheme of propene hydrogenation. (c) ALTADENA single-scan 1H NMR spectrum acquired during gas-phase hydrogenation of propene with parahydrogen over 1 wt % Rh/TiO2 catalyst.

Next, dissolution experiments were carried out. For this purpose, HP propane was supplied to the bottom of the NMR tube containing acetone-d6. As expected from our previous results,[46] we observed PHIP effects for propane dissolved in acetone-d6 (Figure ). It should be noted that the concentration of dissolved propane was virtually independent of the duration of bubbling and the average concentration was equal to 50 mM. We noticed that the hyperpolarization decay for dissolved propane was significantly slower than for propane in the gas phase. Therefore, hyperpolarization “lifetime” measurements of dissolved HP propane were carried out by recording ALTADENA signal decay curves through the acquisition of a set of 1H NMR spectra at 1-s intervals using a small-angle rf excitation pulse (α = 10°). In a first approximation, the influence of the 10° rf excitation pulse on the longitudinal magnetization was neglected. In particular, correcting for the signal reduction caused by the influence of small-angle rf pulses would result in an increase in the hyperpolarization decay time (THP) for both the CH2 and CH3 groups of propane; therefore, the values shown here represent lower estimates. To confirm this conclusion, the results of experiments with 5° rf excitation pulses are presented in Figure S3. The hyperpolarization decay time measurements for dissolved HP propane yielded the values THP(CH2) = 35.1 ± 0.1 s and THP(CH3) = 28.6 ± 0.1 s (Figure d), which are ∼50 times greater than T1 values for gaseous propane at the same 7.1 T magnetic field (∼0.6 s[35]). The obtained THP values are in a good agreement with the T1 values obtained by the inversion–recovery technique for fully relaxed dissolved propane (Figure e).

Figure 2

(a) Scheme showing the production of dissolved HP propane. (b) Single-scan 1H NMR spectrum of dissolved HP propane produced by hydrogenation of propene with parahydrogen over 1 wt % Rh/TiO2 catalyst and subsequent dissolution of HP propane in acetone-d6. The signal enhancement was ∼17-fold. (c) Corresponding 1H NMR spectrum of the same solution as in panel b acquired after complete relaxation of hyperpolarization. (d) THP measurements for HP propane dissolved in acetone-d6 through the use of 10° rf pulses. (e) T1 measurements for thermally polarized propane dissolved in acetone-d6 by the inversion–recovery sequence.

A reference inversion–recovery experiment with a concentrated solution of propane in acetone-d6 at 25 °C obtained by bubbling pure propane from a gas tank provided comparable T1 times for CH2 (41.8 s) and CH3 (35.6 s) protons (Figure S4). We note, however, that the results of inversion–recovery measurements are strongly dependent on the presence of dissolved oxygen, which is paramagnetic, with longest T1 time being obtained after a sufficiently long bubbling (2 min) of pure propane through acetone-d6. This resulted in the efficient removal of oxygen from the solution and a 5–6-fold elongation of T1 times compared to the T1 values obtained in the case of a short bubbling time (8 s). Similarly, bubbling of the reaction mixture in the experiments with hyperpolarized propane served to remove dissolved oxygen, which is important for obtaining the longest possible preservation of the hyperpolarization.

This dramatic elongation of the hyperpolarization decay time in the case of dissolved propane compared to gaseous propane is most likely due to the change in the dominant relaxation mechanism. This is confirmed by the fact that the hyperpolarization decay time is in a good agreement with the spin–lattice relaxation time in both cases. As discussed above, the main relaxation mechanism for small molecules (e.g., propane) in the gas phase is the spin-rotation mechanism,[47] the contribution of which depends on the time intervals between molecular collisions. However, in the liquid phase, molecular collisions are very frequent, leading to a dramatic decrease of the spin-rotation channel of relaxation. On the other hand, in solution, the dipole–dipole and intermolecular (including solute–solvent) interactions emerge as the prevailing mechanism of relaxation. Thus, such a significant increase of the relaxation time in solution is explained by the changes in the relaxation mechanism.

The significant signal enhancement obtained for dissolved HP propane allowed fast MRI experiments to be performed using two-dimensional FLASH image acquisition. Here, we used the conventional FLASH pulse sequence to acquire two-dimensional slices with a spatial resolution of 0.8 × 0.8 mm2 for axial slice orientations and 0.8 × 0.47 mm2 for sagittal slice orientations and a short acquisition time (∼1.7 s). Note that the signal enhancement obtained by the PHIP technique is crucial for the 1H FLASH MRI of propane in the liquid phase. The signal enhancement for HP propane versus thermally polarized propane allows for the detection of images with a spatial resolution that is sufficient to visualize a 1/16-in.-o.d. Teflon capillary (Figure a). Thermally polarized propane imaging does not yield an appreciable signal-to-noise ratio (SNR) to detect MR images for both slice orientations (Figures b and S5b).

Figure 3

(a) MR image of a 10-mm NMR tube filled with solution of HP propane in acetone-d6 in axial orientation. SNR = 17.2. (b) Corresponding MR image of fully relaxed solution shown in panel a. The field of view (FOV) was 5 × 5 cm, with a 64 × 64 matrix size and a slice thickness equal to the diameter of the NMR tube. The total acquisition time was 1.7 s.

Next, we aimed to determine the applicability of our approach to other deuterated solvents. For this purpose, we dissolved HP propane in methanol-d4 (Figure ). As expected, strong PHIP signals were observed for dissolved propane (Figure S6). The THP values for propane dissolved in methanol-d4 were estimated as 34.6 ± 0.3 s for the CH2 group and 28.3 ± 0.1 s for the CH3 group (Figure b). These values are in agreement with the T1 values for propane dissolved in methanol-d4 (Figure c) and the values obtained previously for HP propane dissolved in acetone-d6. This means that the observed prolongation of the dissolved HP propane lifetime is a common feature for different deuterated organic solvents.

Figure 4

(a) Scheme showing the production of dissolved HP propane. (b) THP measurements for HP propane dissolved in methanol-d4 through the use of 10° rf pulses. (c) T1 measurements for thermally polarized propane dissolved in methanol-d4 by the inversion–recovery sequence.

For potential applications of HP propane as a contrast agent in the MRI of lungs, one needs to retrieve the HP propane from solution to the gas phase. Therefore, we performed experiments to confirm the viability of our approach. To that end, we adjusted the position of the NMR tube in the rf probe to acquire NMR spectra of the gas phase just above the liquid. First, the propene/p-H2 mixture was passed through the catalyst layer, and the resulting mixture of HP propane and residual p-H2 was bubbled through acetone-d6. Then, the gas flow was stopped for ∼1–2 s, and next, the propene/p-H2 mixture was bubbled directly into the acetone-d6 solution, bypassing the reactor (see Figure S2 for the experimental setup). During all of these procedures, NMR spectra of the gas phase were acquired continuously in 1-s intervals. Thus, in the beginning, we observed NMR signals of HP propane in the gas phase; after the interruption of the gas flow, the HP propane gas above the solution rapidly relaxed to thermal equilibrium, and after the gas flow had been restarted, we observed NMR signals of both HP propane and propene (Figure S7). Thus, HP propane was successfully retrieved to the gas phase while retaining a substantial degree of hyperpolarization. Consequently, the developed approach is a very impressive technique for overcoming the lifetime limitations of HP propane usage as a potential hyperpolarized contrast gas agent for biomedical MRI applications caused by its extremely short spin–lattice relaxation time. Prolongation of the HP propane lifetime through its dissolution, followed by retrieval of the HP propane to the gas phase, not only allows the HP propane production and MRI investigations to be separated in time but also makes these two procedures (HP gas production and MRI studies) fully independent. In this way, the dissolved state can play the role of a temporary carrier for the polarization storage that allows for the transport of HP propane gas. Nevertheless, further investigations are undoubtedly needed, for example, to optimize the conditions for maximizing the hyperpolarization in both the liquid[48,49] and gas[50] phases.

Conclusions

The presented new approach based on the dissolution of HP propane enables one to preserve the hyperpolarization obtained by the PHIP technique for ca. 50 times longer than in the gas phase. This phenomenon can be explained by the switching of the main contribution to the relaxation mechanism from spin-rotation in the gas phase to dipole–dipole interactions in the liquid phase. Moreover, we demonstrated the successful retrieval of dissolved HP propane to the gas phase with a substantial retention of its hyperpolarized state. We hope that this new approach to the prolongation of the hyperpolarization lifetime can greatly expand the application of HP propane as a contrast agent in gas-phase MR imaging. Moreover, the future development of the reported approach might have synergism with LLSS technique that, without doubt, allows for the production of hyperpolarized molecules with extremely long relaxation times.