PASADENA hyperpolarized ¹³C phospholactate

Abstract

We demonstrate that potassium 1-¹³C-phosphoenolpyruvate becomes hyperpolarized potassium 1-¹³C –phospholactate with ¹³C T₁ = 36 s after molecular hydrogenation by PASADENA (Parahydrogen and Synthesis Allows Dramatically Enhanced Nuclear Alignment). This proof-of-principle study was conducted with a fully protonated molecular precursor. ¹³C was polarized to a level of 1 %, corresponding to nearly 4,000 fold sensitivity enhancement at 3 T. The relevant homo- and heteronuclear spin-spin couplings are reported.

Hyperpolarized magnetic resonance (MR) of ¹³C and ¹⁵N exogenous metabolic contrast agents is a rapidly developing field, which progressed from proof-of-principle studies in the early 2000s^1,2 to the first clinical trial in men.³ The underlying driving force stems in part from the demonstrated capability to detect abnormal metabolic pathways in cancer^4–7 and other diseases with impaired metabolism.⁸ Hyperpolarization of long-lived nuclear spin states increases sensitivity by 4–6 orders of magnitude^2,9 with nuclear spin polarization approaching the order of unity. Hyperpolarized metabolic agents such as 1-¹³C-pyruvate can report on metabolic status of tumors similarly to positron emission tomography (PET) tracers.

One of the main bottlenecks for preclinical and clinical application of hyperpolarized MR is the development and validation of relevant compounds that can probe biochemical pathways in vivo. Dynamic nuclear polarization (DNP)¹⁰ has been most widely used to date with the main drawback of long (~ 1 h) polarization cycles. Parahydrogen induced polarization (PHIP) offers significantly faster preparations with hyperpolarization cycles as short as 1 minute. However, it is limited by the availability of the required unsaturated molecular precursors that are necessary for molecular addition of parahydrogen^11,12, which acts as a source of spin order. An additional requirement for increased relaxation time of the hyperpolarized ¹³C or ¹⁵N site is the absence of directly attached protons allowing for in vivo spin lattice relaxation time T₁ as long as 43 seconds¹³ and 2 minutes¹⁴ respectively. As a result, the smallest PHIP moiety consists of an unsaturated C=C or C≡C bond adjacent to a labeled ¹³C or ¹⁵N site. For ¹³C hyperpolarized compounds, this represents a three carbon limitation successfully exemplified by acrylate moiety by a number of groups.^15,16 However, hydrogenation of acrylate results in propionate, which turns out to be useful only for angiographic applications.^1,15 In contrast, a leading DNP hyperpolarized metabolic agent 1-¹³C-pyruvate is also a three-carbon molecule carrying an extra oxygen atom in addition to the three-carbon skeleton of propionate, Fig. 1. Hyperpolarized 1-¹³C-pyruvate is not amenable by PHIP.

Figure 1.

The molecular diagrams of PHIP and DNP hyperpolarization of 1-¹³C-pyruvate and 1-¹³C-phospholactate (PLAC) and their metabolism for application as metabolic hyperpolarized (HP) agents.

Here, we describe hyperpolarization of 1-¹³C-phospholactate by PHIP. While both 1-¹³C-propionate and 1-¹³C-phospholactate are carboxylic acids with three-carbon backbones, 1-¹³C-phospholactate has an extra phosphate protected oxygen. The latter makes it similar to pyruvate, as it can be potentially metabolized, Fig. 1. This is a proof-of-principle study demonstrating (i) the feasibility of molecular addition of parahydrogen with established and commercially available Rh-based water-soluble catalyst and resulting in ¹³C hyperpolarization, (ii) sufficiently long life time of this potentially new metabolic contrast agent, and (iii) complete characterization of homo- and heteronuclear spin-spin couplings. The latter are necessary for optimization of PHIP polarization efficiency especially for perdeuterated unsaturated precursors, which represent a three spin system consisting of two nascent parahydrogen protons and a ¹³C or ¹⁵N nucleus. In addition, perdeuterated precursors may result in hyperpolarized compounds with longer relaxation time.¹⁷ Deuteration minimizes the number of nearby protons, which have significantly greater magnetic moment compared to that of deuterons. As a result, ¹³C T₁ decrease due to dipolar relaxation mechanism may be somewhat minimized.¹⁷

The double C=C bond is stable with phosphate moiety in the second position, Fig. 1, allowing for molecular hydrogenation and PHIP. The unsaturated precursor is non-toxic 1-¹³C-phosphoenolpyruvate (PEP, Sigma-Aldrich #589,462), which becomes ¹³C hyperpolarized racemic 1-¹³C-phospholacate (PLAC) with D- and L- forms. The protection by phosphate group allows stabilizing the carbon-carbon double bond adjacent to the oxygen atom. While other protecting groups can potentially be used, the phosphate moiety could be quickly removed enzymatically¹⁸ by a suitable choice of phosphatase to yield hyperpolarized 1-¹³C-lactate (LAC), Fig. 1. Moreover, phosphate ions are biologically compatible and therefore do not require additional purification steps other than removal of phosphatase prior to in vivo administration.

Experimental polarization procedures and equipment were identical to those described earlier.^{9,19 13}C PHIP polarization of PLAC, Fig. 2a, was conducted in a 0.0475 T polarizer⁹ with 2.5 mM water soluble Rh-based catalyst^19,20 using the Gold-man polarization transfer sequence.¹⁶ 99.8% D₂O (Sigma-Aldrich-Isotec # 617,385) was used as a medium, pH = 11 using 18 mM phosphate buffer, for 4 second long hydrogenation of 1.8 mM PEP for high resolution ¹³C studies using an 11.7 T Bruker NMR spectrometer equipped with a 5 mm X-H RF probe, Fig 2. Following the PHIP procedure, hyperpolarized solution was injected in a 5 mm NMR tube. The NMR tube was inserted into the NMR spectrometer and the solution degassed during an ~ 30 second delay. The degassing step is necessary to eliminate residual parahydrogen-nitrogen bubbles and their associated susceptibility gradients and broad lines. ¹H decoupled ¹³C spectra, Fig. 2b, were acquired approximately 60 seconds or 3×T₁ after PHIP hyperpolarization resulting in hyperpolarization decay by 20 fold. Nevertheless, a 50 fold signal enhancement of PLAC was observed, Fig. 2b, when compared to the ¹³C signal recorded from the same sample at Boltzmann levels of ¹³C polarization, Fig. 2c. This 50-fold enhancement corresponds to a ¹³C polarization of 0.05% at 11.7 T and 298 K. Taking into account 20-fold polarization loss during ~ 60 s delivery and degassing procedure, polarization of 1% was produced in the polarizer. The conversion of PEP (¹³C δ = 171.9 ppm) to PLAC (¹³C δ = 182.1 ppm) was only 25% due to non-optimal catalyst preparation. The Rh-based catalyst was further optimized for low field in situ NMR studies at 0.0475 T by titration of Bis(norbornadiene)rhodium(I) tetrafluoroborate (Strem # 45-0230, Newburyport, MA) and 3, 3′-(1,4-butanediylbis(phenylphosphinidene))bispropanesulfonic, disodium salt (Sigma-Aldrich-Isotec #Q36333-SPEC), Shchepin, R.V., unpublished results.

Figure 2.

¹³C PASADENA hyperpolarization of potassium 1-¹³C-phospholactate; a) the diagram of molecular hydrogenation of potassium 1-¹³C-phosphoenolpyruvate (PEP) yielding 1-¹³C-phospholactate (PLAC), b) proton decoupled single scan ¹³C spectrum of hyperpolarized DL-1-¹³C-PLAC; the insert shows T₁ decay of ¹³C hyperpolarized signal of PLAC measured with 15° excitation RF pulses. All ¹³C spectra are acquired on a 500 MHz Bruker spectrometer equipped with X-H dual tuned RF coil and identical acquisition parameters, c) proton decoupled ¹³C spectrum of Boltzmann polarized mixture of DL-1-¹³C-PLAC and unreacted 1-¹³C-PEP using the same sample as in spectrum 2b except for nuclear spin polarization; note the doublet peak appearance due to J³_PC ~ 6.5 Hz, a signature of carbon-phosphorous spin-spin couplings in PLAC and PEP. The conversion of PEP (¹³C δ = 171.9 ppm) to PLAC (¹³C δ = 182.1 ppm) was only 25% due to non-optimal catalyst preparation.

Low field in situ detection of ¹³C hyperpolarization was conducted with 2.0 mM PEP in aqueous medium with double de-ionized water, pH = 9.8 using 22 mM phosphate buffer resulting in 99+% chemical conversion to PLAC, pH = 8.8. While excellent PEP chemical conversion is achieved, in vivo use of the hyperpolarized contrast agent would require additional pH neutralization and catalyst removal. It should also be noted that D-isomers may not be metabolized in vivo and may be toxic. The signal intensity of a single scan ¹³C hyperpolarized spectrum, 0.007 millimoles of PLAC, Fig. 3b, was compared to that of a ¹³C spectrum acquired using 170 millimoles of sodium 1-¹³C-acetate at Boltzmann polarization at 0.0475 T and 308 K, Fig. 3a.⁹ 1.0 % polarization was achieved corresponding to ¹³C signal enhancement ε = 240,000 fold at 0.0475 T and 308 K. It is in good agreement with our high field studies. ¹³C T₁ was measured by monitoring ¹³C hyperpolarization decay using a time series of ¹³C spectra acquired every 20 seconds with 30° excitation pulses, Fig. 3c. The T₁ of 1-¹³C at 0.0475 T was 36 ± 2 seconds, which is twice as long as the T₁ measured at 11.7 T in D₂O. The T₁ decay was simulated as an exponential decay taking into account the effect of RF pulses on ¹³C magnetization. A likely explanation for longer T₁ at low magnetic field is a reduced chemical shift anisotropy, which is expected to dominate T₁ relaxation at high magnetic fields.²¹ Similar field effect was reported for the ¹³C T₁ in 2-hydroxyethyl-1-¹³C-propionate-2,3,3-d₃.⁹

Figure 3.

¹³C NMR spectroscopy conducted at 0.0475 T. a) ¹³C reference spectrum of 170 millimoles (14 g in 50 mL D₂O) of sodium 1-¹³C-acetate, Boltzmann polarization at 308 K, 32 averages, b) ¹³C spectrum of 0.007 millimoles, 1.5 mg in 3.6 mL H₂O at pH = 9.0, hyperpolarized potassium DL-1-¹³C-PLAC, polarization P = 1.0% corresponding to signal enhancement by 240,000 fold measured by comparing ¹³C signal intensity of hyperpolarized 0.007 millimoles of PLAC with that of 170 millimoles of sodium 1-¹³C-acetate at Boltzmann polarization at 0.0475 T and 308 K, Fig. 3a.⁹ c) decay of ¹³C hyperpolarized PLAC signal measured with 30° excitation pulses. The T₁ decay was simulated (red trace) as an exponential decay taking into account the effect of RF pulses on ¹³C magnetization.

The solution used in PHIP low field studies was reduced in volume using a rotary evaporator, dissolved in 99.8% D₂O and used for high-resolution studies of PLAC spin-spin couplings, Fig. 4. The combination of heteronuclear spectra, Fig. 4, over-determines the problem of complex multiplets and allows for accurate measurements of the spin-spin coupling magnitudes shown in Fig. 2a.: J³_HH = 6.90 Hz, J³_HC = 4.07 Hz, J²_HC = 3.80 Hz, J³_PC = 6.52 Hz, J³_HP = 8.60 Hz. The knowledge of spin-spin couplings is critical for optimal performance of PHIP polarization transfer sequences that convert spin order of nascent parahydrogens to detectable ¹³C magnetization.

Figure 4.

Multinuclear high-resolution 11.7 T NMR spectroscopy (blue trace) and simulations (red trace) of PLAC at pH = 8.8: ¹H spectroscopy with ³¹P decoupling, (a) – methine peak and (c) – methyl peak and without ³¹P decoupling, (b) – methine peak and (d) – methyl peak, ¹³C spectra with ¹H decoupling (e) and without ¹H decoupling (f), ³¹P spectra with ¹H decoupling (g) and without ¹H decoupling (h). The resonances marked with * correspond to peaks from natural abundance PLAC without 1-¹³C enrichment.

1-¹³C-phospholactate is suitable for measurement of spin-spin couplings due to ¹³C enrichment. The PHIP product of perdeuterated 1-¹³C-PEP, PLAC-d₂, could lengthen ¹³C T₁ somewhat and yield higher polarization, with the caveat that the extra deuterons in PLAC would pose synthetic challenges and complicate the spectra of ¹H, ¹³C and ³¹P, making spin-spin coupling measurements more arduous. An alternative would be to use natural abundance PEP similarly to the strategy utilized for ¹³C hyperpolarized succinate,¹⁷ but this would require recording ¹³C spectra at natural abundance levels of ¹³C resulting in many hours of acquisition time. If PEP is perdeuterated, the spin system participating in polarization transfer dynamics would be reduced to three spins: ¹³C and two nascent parahydrogens in PLAC. Since Goldman’s polarization transfer¹⁶ is designed for three-spin systems and is not expected to perform well for systems with more than 3 spins, it is likely to result in a significantly greater hyperpolarization using perdeuterated PEP compared to the 5-spin system used here.^17,22

While the metabolic relevance of PLAC itself for biomedical imaging is yet to be seen, the corresponding relevance of LAC as an imaging agent is certainly well-established²³. One possibility to produce LAC from PLAC would be to enzymatically cleave the phosphate moiety.¹⁸ It is also possible that phosphatases of blood could cleave phosphate moiety after intravenous injection of PLAC. Fast hyperpolarization of LAC by PHIP method would present a clear advantage over DNP method and would make this promising contrast agent available to a broader range of biomedical scientists for cancer and cardiovascular research.⁷ Moreover, if successful, this strategy of using phosphate protected −C=C-O- motifs and subsequent cleavage could be extended to hyperpolarization of other metabolic contrast agents including ¹⁵N-choline.