¹⁵N Hyperpolarization of Dalfampridine at Natural Abundance for MRI

Abstract

Signal Amplification by Reversible Exchange (SABRE) is a promising method for NMR signal enhancement and production of hyperpolarized molecules. As nuclear spin relaxation times of heteronuclei are usually much longer than those of protons, SABRE-based hyperpolarization of heteronuclei in molecules is highly important in the context of biomedical applications. In this work, we demonstrate that the SLIC-SABRE technique can be successfully used to hyperpolarize ¹⁵N nuclei in dalfampridine. The high polarization level of ca. 8% achieved in this work allowed us to acquire ¹⁵N MR images at natural abundance of the ¹⁵N nuclei for the first time.

Graphical Abstract

A 32,000 - fold enhancement of the ¹⁵N NMR signal (7.8% spin polarization) was achieved for the biologically active molecule dalfampridine by using a combination of the Signal Amplification By Reversible Exchange and Spin Lock Induced Crossing methods in a strong magnetic field. This major boost of the ¹⁵N NMR signal allowed for the first time to obtain a ¹⁵N MR image of FAM with a natural abundance of ¹⁵N isotope.

Dalfampridine (FAM), or 4-aminopyridine, is a drug that can significantly reduce the symptoms of multiple sclerosis.^[1] FAM does not affect the course of the disease itself, but it is employed for symptomatic improvement of walking in adults with several variations of the disease.^[2,3] Its action is based on blocking of potassium channels. It leads to the increase of the action potential conduction in the demyelinated nerve fibers.^[2,3] This drug is approved by the regulating agencies in the US, Canada and Europe, and it is marketed as Ampyra and Fampyra. The distribution of the drug inside the body, the biochemical pathways of its transformation, as well as its metabolic profile are important information for determining the effectiveness of the drug. Magnetic resonance imaging (MRI), being a non-invasive method with a high information content, allows one to effectively obtain these unique data. However, nuclear magnetic resonance (NMR) and MRI have a relatively low sensitivity.^[4] This is due to the fact that population difference of nuclear spin states (termed nuclear spin polarization P) is small. For ¹H nuclei, thermal polarization is only 1.02·10⁻⁵ in a 3 T magnetic field,^[4] and for other NMR-active spin-1/2 nuclei (e.g. ¹³C, ¹⁵N, ¹⁹F, ³¹P) with their lower gyromagnetic ratios it is even less. While this polarization level is sufficient for performing in vivo MRI based on the ¹H NMR signal detection of tissue water, performing in vivo ¹H MRI for any other molecules against the strong background of water signal is highly unfavorable. The use of heteronuclei (e.g. ¹⁵N) mitigates this problem, because their concentration in tissues is significantly lower in comparison with ¹H. More importantly, these nuclei often have very long T₁ relaxation times up to ten minutes,^[5] which is very useful when transient non-equilibrium spin polarizations are involved (see below).

A wide range of biologically relevant molecules, such as nitrogenous bases, DNA, RNA, amino acids, peptides, proteins, vitamins, hormones, and drugs, contain nitrogen atoms. The direct use of a thermally polarized signal of ¹⁵N nuclei for in vivo MRI applications is very challenging because of the very low natural abundance of the ¹⁵N isotope (0.365%) and the lower gyromagnetic ratio of ¹⁵N nuclei compared to ¹H (~1:10), resulting in low detection sensitivity. As a result, a limited number of ¹⁵N MRI studies with non-isotopically enriched molecules were reported to date.^[6] Thus, the use of heteronuclei in MRI requires additional efforts to boost the detection sensitivity. NMR hyperpolarization (HP) allows one to significantly increase P to the order of unity, with the corresponding gains in the detection sensitivity. The Signal Amplification By Reversible Exchange (SABRE) technique employs a simultaneous chemical exchange of parahydrogen and to-be-hyperpolarized substrate at a metal complex.^[7–11] Parahydrogen serves as a source of hyperpolarization in this method, and two SABRE approaches have been developed. The first approach relies on the static magnetic field to create levels anti-crossings (LACs) for spontaneous nuclear spin order transfer from parahydrogen to the to-be-hyperpolarized substrate.^[12] In case of ¹⁵N nuclei, this process is most efficient at sub-microtesla magnetic fields (ca. 0.2–0.4 μT).^[6,13] Signal Amplification by Reversible Exchange in SHield Enables Alignment Transfer to Heteronuclei (SABRE-SHEATH) has been used to hyperpolarize a wide range of biologically relevant ¹⁵N-labelled molecules, including biologically active ones, with ¹⁵N polarization exceeding 24%.^[6,14–18] In some cases, the achieved enhancement was sufficient to perform spectroscopic detection at a natural abundance of ¹⁵N nuclei.^[19] The need to use a magnetic shield and the necessity to transfer the sample from microtesla magnetic fields to the MRI scanner for the signal readout are the key disadvantages of this approach. Although SABRE of ¹⁵N nuclei has been demonstrated at a high magnetic field of 9.4 T (HF-SABRE strategy),^[20,21] the attained polarization levels were very low (<0.1%), and therefore not attractive for biomedical applications. The alternative approach is based on the use of the RF-pulses instead of static magnetic fields for nuclear spin order transfer from parahydrogen to the to-be-hyperpolarized substrates at high magnetic fields.^[22,23]

A number of RF pulse sequences have been developed for this purpose, including Alternating Delays Achieve Polarization Transfer SABRE (ADAPT-SABRE),^[24] SABRE Insensitive Nuclei Enhanced by Polarization Transfer (SABRE-INEPT),^[25] Low Irradiation Generation of High-Tesla SABRE (LIGHT-SABRE),^[26] Spin Lock Induced Crossing SABRE (SLIC-SABRE),^[27] Quasi-resonance SABRE (QUASR-SABRE)^[28] and others. However, the achieved polarization levels, P_15N, have not exceeded 1% with any of the RF-based approaches to date.

In the present work, we have employed the SLIC-SABRE approach to transfer hyperpolarization from p-H₂ to the ¹⁵N nuclei in dalfampridine (FAM) and 4-dimethylaminopyridine (DMAP) to detect the ¹⁵N MR images of FAM and DMAP at natural abundance of ¹⁵N (Scheme I).

Scheme 1.

Generation of free HP substrate in SLIC-SABRE approach.^[29]

We have utilized the previously developed IrIMes catalyst, the experimental setup and the SLIC-SABRE technique for the hyperpolarization of ¹⁵N DMAP and FAM in a strong magnetic field (see experimental section, SI).^[27,29] The parameters of SLIC-SABRE pulse sequence such as the continuous wave (CW) pulse duration, the number of cycles and the flow rate were previously optimized (SI).^{[29] 15}N NMR spectra and ¹⁵N MR images (see below) were recorded after SLIC pumping with optimized parameters (SI).

First, DMAP with natural abundance of ¹⁵N was used as a substrate. Bubbling p-H₂ through the solution of DMAP and SABRE catalyst in a strong magnetic field (7.1 T) resulted in ¹⁵N NMR spectrum with HP signal at 216 ppm corresponding to DMAP bound to the SABRE complex (Figure S1). This signal was used for optimization of SLIC pulse sequence (Table S1). ¹⁵N signal enhancement of the free DMAP was ~29,800-fold at 7.1 T corresponding to 7.2 % ¹⁵N polarization (P_15N) under optimal conditions (Figure 1A). This level of ¹⁵N polarization is approximately 7-fold greater than any previously achieved P_15N by any RF-SABRE technique.

Figure 1.

¹⁵N NMR spectra of HP DMAP (A) and FAM (B) with natural abundance of isotope ¹⁵N recorded by using SLIC-SABRE approach at room temperature and overpressure of 4.4 bar. For the SLIC-SABRE pulse sequence, CW pulse duration was 1.17 s, v₁ = 5 Hz, v_slic = 13 Hz, and the number of cycles was 30 (A) and 40 (B). The 0.1 M solutions of DMAP and FAM with the natural abundance of isotope ¹⁵N in a 5 mm NMR tube were used. The p-H₂ flow rate was 80 sccm. The ¹⁵N NMR spectrum of 4.9 M DMAP solution recorded with 512 accumulations was used as a signal reference (C). The detailed evaluation of the enhancement factors is presented in SI.

It was previously shown that the substrate change from pyridine-¹⁵N to 3-fluoropyridine-1-¹⁵N leads to a significantly decreased P_15N (by ca. 30-fold).^[30,31] We attribute this effect to the electron-withdrawing inductive effect on the pyridine heterocycle resulting in the decreased efficiency of hyperpolarization in the context of SABRE-SHEATH^[30,32,33]. At the same time, the IrIMes-DMAP system demonstrates the highest level of ¹⁵N SABRE. We attribute this to the presence of an electron donating group, likely increasing the electron density at the nitrogen atom by positive mesomeric effect, and thereby increasing the efficiency of SABRE hyperpolarization transfer.

FAM, being a drug, is a promising compound for use in biomedical applications. The signal from the catalyst-bound FAM was observed in the ¹⁵N NMR spectrum at ~218 ppm while p-H₂ was bubbled through the solution (Figure S1), and this signal was employed for optimization of the SLIC-SABRE sequence. The ¹⁵N NMR signal of free FAM at ca. 260 ppm was boosted by ~32,100-fold corresponding to P_15N of 7.8% (Figure 1B) under optimized conditions (Table S1). The P_15N of FAM was greater in comparison to DMAP likely because the NH₂- group is a stronger electron donor than N(CH₃)₂- substituent.

We have employed HP DMAP to demonstrate the feasibility of ¹⁵N MRI at the natural abundance level of ¹⁵N isotope (Figure 2A). The total experimental time, including SLIC hyperpolarization and MRI signal detection, was 43 s. Note that relaxation times of both substrates were about 37 s (Figure S5) and therefore, 43 s total MRI experiment time is appropriate for their visualization. Thus, for the first time the application of the SLIC-SABRE approach allowed us to obtain the ¹⁵N MR image of the HP substrate at the natural abundance of ¹⁵N nuclei (¹⁵N concentration was 1.8 mM) with good signal-to-noise ratio and spatial resolution. The slightly higher ¹⁵N NMR signal enhancement of FAM molecule allowed us to obtain ¹⁵N MR image with better SNR than that for DMAP (109 vs. 49, Figure 2B) on the natural-abundant content (0.365 %) of ¹⁵N in 0.5 M substrate solutions in a 10 mm NMR.

Figure 2.

2D ¹⁵N FLASH MRI of HP DMAP and FAM: XY projection image of a 10 mm NMR tube. The void in the center corresponds to the presence of the capillary supplying p-H₂. SLIC-SABRE pulse parameters: CW pulse duration was 1.17 s, v₁ = 5 Hz, v_slic = 13 Hz, number of cycles was 30 for (A) and 40 for (B). The p-H₂ flow rate was 140 sccm for (A) and 100 sccm for (B), and p-H₂ overpressure of 4.4 bar was used. The experiments were carried out at room temperature. TR = 3.1 ms, TE = 1.5 ms. SW = 10 kHz, spatial resolution was 0.3×2.4 mm²/pixel. Field of view was 3.8×3.8 cm². Acquisition matrix was 128×16, and it was zero-filled to the matrix size of 128×128.

In conclusion, we have demonstrated a very efficient RF-SABRE for ¹⁵N hyperpolarization of a biocompatible molecule as an example. The high levels of ¹⁵N polarization (~8%) allowed us for the first time to demonstrate the feasibility of ¹⁵N MRI at the natural abundance of ¹⁵N nuclei. This is important, because SABRE-compatible motifs of promising drugs can be potentially studied for their suitability for SABRE hyperpolarization and MR imaging without the need of labor-some ¹⁵N isotopic enrichment. The levels of polarization demonstrated here (ca. 8%) can be potentially further improved through the use of nearly 100% p-H₂ (versus 87% p-H₂ employed here)^[11]. All in all, the significant advances in ¹⁵N polarization levels demonstrated in this work potentially make SLIC-SABRE technique useful for future biomedical applications. Moreover, recent advantages in ¹⁵N MRI imaging^[34] and SABRE catalyst removing^[35,36] are open up new perspectives for ultrafast contrast agents imaging which are free from the homogeneous catalyst.

Experimental Section

Iridium complexes with a N-heterocyclic carbene ligand, Ir(COD)(IMes)Cl (IMes = 1,3-bis(2,4,6-trimethylphenyl)imidazole-2-ylidene and COD = cyclooctadiene), was used as a pre-catalyst for generating SABRE hyperpolarization. This complex was synthesized according to the published procedure.^[37] Pre-catalyst was activated by bubbling p-H₂ (20 sccm at 2.7 bar overpressure) for 30 minutes through the solution containing a substrate, S. SLIC-SABRE experiments for DMAP (Aldrich, CAS: 1122−58−3, #107700) and FAM (Aldrich, CAS: 504−24−5, #A78403) were carried out using 0.6 mL of 0.1 M solution of the substrate and 5 mM of IrIMes catalyst in a 5 mm NMR tube. MRI experiments with DMAP and FAM were carried out using 3 mL of 0.5 M solution of the substrate and 5 mM of IrIMes catalyst in a 10 mm tube. ¹⁵N NMR spectra of DMAP, FAM, and MNZ-¹⁵N₂ were acquired on a 300 MHz NMR spectrometer. MR images were acquired using the 400 MHz (B₀ = 9.4 T) microimaging instrument (Bruker, Avance III) equipped with a two-channel probe (¹H, ¹⁵N) and the gradient strength of up to 150 G/cm. The SLIC pulse sequence was used to transfer nuclear spin order from p-H₂ to ¹⁵N in the strong magnetic field.

For MRI, gradient echo pulse sequence was used. The entire k-space data set was acquired after one SLIC-SABRE block. In this setting, the hyperpolarization by SLIC-SABRE took 43 s (30 cycles) and the FLASH MRI less than 1 s (repetition time (TR) 3.1 ms, echo time (TE) 1.5 ms, 16 phase encoding steps, 128 readout points, Cartesian encoding). Four scans were averaged. At the beginning, p-H₂ was flushed through the solution with a flow rate of 80 sccm and 4.4 bar overpressure. Importantly the flow of p-H₂ was stopped 4 – 6 s before the onset of the MRI to reduce magnetic field inhomogeneities and fast convection caused by the bubbles. There was a delay of approximately 2 – 4 s between SLIC-SABRE and MRI due to hardware limitations while switching between two pulse sequences. The phase encoding gradient had 8% strength of maximal value and 400 μs duration, and the readout gradient had 4% strength and 2.1 ms duration. The flip angle was 30°. No k-space filter was applied. As a result, ¹⁵N images were obtained using a 128×16 matrix that yielded the spatial resolution of 0.3×2.4 mm²/pixel. For the presentation, the k-space data sets were zero-filled to 128×128.