Late-stage Deuteration of ¹³C-enriched Substrates for T₁ Prolongation in Hyperpolarized ¹³C MRI

Abstract

A robust and selective late-stage deuteration methodology was applied to ¹³C-enriched amino and alpha hydroxy acids to increase spin-lattice relaxation constant T₁ for hyperpolarized ¹³C magnetic resonance imaging. For the five substrates with ¹³C-labeling on the C1-position ([1-¹³C]alanine, [1-¹³C]serine, [1-¹³C]lactate, [1-¹³C]glycine, and [1-¹³C]valine), significant increase of their T₁ was observed at 3T with deuterium labeling (+26%, 22%, +16%, +25% and +29%, respectively). Remarkably, in the case of [2-¹³C]alanine, [2-¹³C]serine and [2-¹³C]lactate, deuterium labeling led to a greater than four fold increase in T₁. [1-¹³C,2-²H]alanine, produced using this method, was applied to in vitro enzyme assays with alanine aminotransferase, demonstrating a kinetic isotope effect.

Table of Contents Entry

Magnetic resonance imaging employing hyperpolarized substrates (HP MRI) has recently emerged as a powerful tool for studying metabolism in cells, animal models and patients.^1–9 Polarization of substrates can be realized through a variety of mechanisms including dissolution dynamic nuclear polarization (DNP),¹⁰ parahydrogen induced polarization (PHIP),^11,12 or signal amplification by reversible exchange (SABRE).¹³ While these are versatile methods that allow for real time imaging of metabolism, the short lifetime of the hyperpolarized signal, which decays exponentially based upon the spin lattice relaxation time T₁, remains one of the key limiting factors in the implementation of this technology. The most widely used HP ¹³C probe is [1-¹³C]pyruvate, a key metabolic intermediate, which has a T₁ of 67 s at 3T. However, other ¹³C nuclei, especially those with directly attached protons, are not feasible HP ¹³C probes due to very short T₁’s (less than 5s).¹⁴ One approach to increase T₁ is the substitution of ¹H with ²H (or D), a quadrupolar nucleus with a gyromagnetic ratio γ about 6.5-fold smaller than the one for ¹H.^15–29 This use of deuterated substrates has proved particularly fruitful in the case of SABRE¹³ and PHIP^11,12 methods. This approach is effective when dipolar ¹³C-¹H coupling contributes substantially to T₁ relaxation. Fortunately, in the case of pyruvate, the incorporation of deuterium is straightforward owing to lack of stereocenters.³⁰ However, the synthesis of multiply labelled molecules containing stereocenters including both ¹³C and ²H is generally both expensive and time consuming, and most isotopically enriched molecules require multi-step syntheses. Therefore, a robust method for incorporation of deuterium in the final step of synthesis would be generally valuable in the field of HP MRI.

For the synthesis of deuterated molecules, late-stage isotopic exchange has several advantages over a synthetic pathway from enriched building blocks. Numerous methods based on homogeneous or heterogeneous catalysts for H/D exchange have already been described, but the development of a deuteration methodology with mild reaction conditions, high selectivity and deuterium incorporation is still a challenge.^31,32 In order to develop [¹³C,²H]labelled probes for HP MRI, we considered the regioselective deuteration, at the α-position of aliphatic alcohols and sugars, developed by Sajiki et al., as a straightforward way to the deuterium labelling of ¹³C-substrates with attached O or N.^33,34 In this manuscript, we report the application of this methodology to a variety of ¹³C-enriched compounds, enabling high incorporation yields with retention of configuration, and demonstrate a significant increase in T₁ of the resulting deuterated substrates. One of the probes, [1-¹³C,2-²H]alanine, was studied in an in vitro enzymatic assay with alanine aminotransferase (ALT), revealing a deuterium kinetic isotope effect.

Initially, we evaluated the performance of the labelling methodology with a variety of labelled substrates including α-amino and hydroxy acids. We performed the one-step deuterium labelling reaction on position C2 of several commercial ¹³C-labeled substrates (Scheme 1, Table 1). Reactions were incubated in D₂O in the presence of RuC 5% (40 wt%), under H₂, overnight, at 80°C (Table S1, ESI^†). Efficient deuterium incorporation on position C2 (95–97 %) was observed for aliphatic amino acids [1-¹³C,2-²H] and [2-¹³C,2-²H]alanine (1 and 6), [1-¹³C,2-²H₂]glycine 4 and [1-¹³C,2-²H]valine 5, with enantiomeric excesses greater than 99%. Isotopic enrichments on position C2 of [1-¹³C,2-²H] and [2-¹³C,2-²H]sodium lactate (3 and 8) were 97% and 98%, respectively, with lower enantiomeric excesses (86 and 94%). Moderate chemical yield on [1-¹³C,2-²H] valine 5, 53%, may be due to its lower solubility in D₂O. Enantiomeric excess was 98% for both [1-¹³C,2-²H] and [2-¹³C,2-²H]serine (2 and 7) whereas chemical yields were 78% and 77%, respectively. Their lower isotopic enrichments on position C2 (52 and 90%) may be due to the additional deuterium labelling on their position C3. In a few cases, side reactions were encountered which led to decomposition of the desired product (ESI^†). Taken together with prior reports,^27,35 our data indicate that this is a versatile method for deuterium incorporation in biologically relevant molecules.

Scheme 1:

Regioselective catalytic deuterium labelling via 1H/2H exchange using ruthenium on carbon (RuC).

Table 1:

Structures of ¹³C-enriched molecules after deuterium enrichment. The bracketed number indicates the isotopic enrichment determined by ¹H,¹³C NMR and HRMS (analyses described in the ESI^†). ee: enantiomeric excess.

Molecule	Structure	Chemical yield	ee
13C on position C1 and 2H on position C2
[1-13C,2-2H]alanine 1		99 %	99 %
[1-13C,2,3-2H3]serine 2		78 %	98 %
[1-13C,2-2H]lactate 3		98 %	86 %
[1-13C,2-2H2]glycine 4		79 %	-
[1-13C,2-2H]valine 5		53 %	99 %
13C on position C2 and 2H on position C2
[2-13C,2-2H]alanine 6		89 %	99 %
[2-13C,2,3-2H3]serine 7		77 %	98 %
[2-13C,2-2H]lactate 8		99 %	94 %

In order to determine the impact of deuterium incorporation on T₁, we then prepared the labelled substrates for hyperpolarization. Solutions of 4 to 6 M substrate with 1 to 1.2 equivalents NaOH and 23 to 24 mM free radical (OX063) were prepared for hyperpolarization using DNP.³⁶ Following polarization, T₁ measurements were performed on a 3T preclinical MR scanner. Deuterium substitution at the C2 position yielded significant improvements of the T₁ with ¹³C at the C1 position, ranging from 16–29% (Figure 1). The relatively modest improvement in T₁ yielded larger signal gains at later time points. For example, in the case of [1-¹³C]alanine, deuteration yielded an increase in signal to noise ratio of 60% at 90s after the start of the experiment (Figure S68c, ESI^†). Remarkably, in the case of [2-¹³C,2-²H]alanine 6, [2-¹³C,2,3-²H]serine 7 and [2-¹³C,2-²H]lactate 8, deuterium labelling led to a greater than four-fold increase in T₁. Due to rapid signal decay on [2-¹³C]alanine, [2-¹³C]serine and [2-¹³C]lactate, their T₁ could not be measured using a hyperpolarized method³⁶ and were instead assayed using an inversion recovery-sequence. Part of the reason why the T₁ gains due to deuteration are relatively limited is because of chemical shift anisotropy (CSA) which is likely the dominant relaxation mechanism at 3T.^37,38 Therefore, at 1.5T, there could be further improvements in T₁ prolongation.³⁹

Figure 1:

T₁ relaxation times at 3T for proton and deuterium-labelled ¹³C-substrates (n = 3, ± s.d., p < 0.02). Due to very low polarization for commercial non-deuterium labeled [2–¹³C]alanine, [2–¹³C]serine and [2–¹³C]lactate, T₁ could not be evaluated using hyperpolarized methods, and inversion-recovery was used at 11.7T: T₁ = 4.9 s, 3.6 s and 7.2 s, respectively.

We then evaluated the T₁ of one of our substrates, [1-¹³C,2-²H]alanine 1, in an in vivo experiment in a mouse model and compared its properties with those of [1-¹³C]alanine. MR measurements where performed on a preclinical 3T scanner (Figure S71, ESI^†). 300 μL of 80 mM solutions of hyperpolarized [1-¹³C]alanine and [1-¹³C,2-²H]alanine 1 were injected intravenously immediately followed by dynamic acquisition of ¹³C MRS spectra. As expected, based on the in vitro studies, we found an increase in the apparent in vivo T₁ at 3T, from 32 s, for [1-¹³C]alanine, to 42 s, for [1-¹³C,2-²H]alanine 1.

As a demonstration of the utility of the deuteration method, we next applied the labelled alanine probes in an in vitro enzyme assay using alanine transaminase (ALT). ALT is an abundant enzyme and a biomarker for liver disease, which converts alanine and α-ketoglutarate to pyruvate and glutamate, respectively (Figure 2a). Previous reports have studied this enzyme both in vitro and in vivo using hyperpolarized ¹³C methods.^36,40,41 Therefore, we developed an assay for the detection of ¹³C pyruvate production by incubation of polarized [1-¹³C]alanine or [1-¹³C,2-²H]alanine 1 with α-ketoglutarate, glutamate and ALT based on prior reports.⁴² As expected, ¹³C pyruvate was rapidly formed during the time course of the hyperpolarized experiment (Figures 2b-d). Furthermore, the initial rate of pyruvate signal growth, which approximates the forward conversion rate, was about 2.42-fold lower for the [1-¹³C,2-²H]alanine 1 as compared with the [1-¹³C]alanine (n = 3 each, p < 0.002, neutral pH). This agrees closely with the previously reported kinetic isotope effect of 2.3.⁴² In order to confirm these findings, we fit the dynamic alanine and pyruvate MRS data to a kinetic model accounting for HP signal exchange between protonated and deuterated [1-¹³C]alanine and [1-¹³C]pyruvate pools as well as signal loss due to RF sampling and T₁ loss (Fig. S72, ESI^†).⁴³ We thus obtained pseudo-first order rate constants of (1.87 ± 0.174) x 10⁻³ s⁻¹ (n = 3) and (0.736 ± 0.015) × 10⁻³ s⁻¹ (n = 3) for protonated and deuterated alanine, respectively. This difference in kinetic rates suggested a kinetic isotope effect of 2.53, in close agreement with our previous analysis and with the literature.⁴²

Figure 2:

Influence of deuterium labelling on [1–¹³C]pyruvate formation after conversion from [1–¹³C]alanine and from [1–¹³C,2–²H]alanine 1 in solution in the presence of ALT enzyme. a) Metabolic pathways of hyperpolarized [1–¹³C]alanine and [1–¹³C,2–²H]alanine 1 via ALT. b) and c) Time courses of integrated spectra showing the evolution of HP [1–¹³C]alanine,[1–¹³C,2–²H]alanine 1 and their metabolite [1–¹³C]pyruvate (normalized peak integrations) (n = 3, ± s.d.). For clarity, pyruvate integrals were ten-fold upscaled. Shaded areas denote the experimental error bars. Spectral acquisition started 11s (b) and 10 s (c) after incubation of the HP probe and the enzyme solution in an NMR tube. d) Measurements of [1–¹³C]pyruvate/total ¹³C-labeled signals (tC) ratios (n = 3, ± s.d.).

In summary, these data indicate that the RuC labelling method represents a versatile method for high-yield deuteration of ¹³C labelled substrates, ideal for application to hyperpolarized ¹³C MRI. When the deuterium was incorporated adjacent to a ¹³C-enriched carbonyl, the effect on T₁ prolongation was moderate, ranging from 16–29%. In contrast, when applied to ¹³C nuclei with directly attached protons ([2-¹³C,2-²H]alanine 6, [2-¹³C,2,3-²H]serine 7 and [2-¹³C,2-²H]lactate 8), an approximately 4-fold increase in T₁ was observed. To further study the behavior of doubly-enriched substrates, we applied [1-¹³C]alanine and [1-¹³C,2-²H]alanine 1 to an in vitro enzyme assay with purified ALT enzyme, demonstrating a kinetic isotope effect, in agreement with prior reports. We anticipate that this versatile method will find application to a variety of substrates for hyperpolarized ¹³C MRI.