Imaging of Extracellular pH Using Hyperpolarized Molecules

Quo Vadis Hyperpolarized 13C MRI?

Over the last two decades, hyperpolarized 13C MRI has gained significance in both preclinical and clinical studies, hereby relying on technologies like PHIP-SAH (ParaHydrogen-Induced Polarization-Side Arm Hydrogenation), SABRE (Signal Amplification by Reversible Exchange), and dDNP (dissolution Dynamic Nuclear Polarization), with dDNP being applied in humans. A clinical dDNP polarizer has enabled studies across 24 sites, despite challenges like high cost and slow polarization. Parahydrogen-based techniques like SABRE and PHIP offer faster, more cost-efficient alternatives but require molecule-specific optimization. The focus has been on imaging metabolism of hyperpolarized probes, which requires long T1, high polarization and rapid contrast generation. Efforts to establish novel probes, improve acquisition techniques and enhance data analysis methods including artificial intelligence are ongoing. Potential clinical value of hyperpolarized 13C MRI was demonstrated primarily for treatment response assessment in oncology, but also in cardiology, nephrology, hepatology and CNS characterization. In this review on biomedical hyperpolarized 13C MRI, we summarize important and recent advances in polarization techniques, probe development, acquisition and analysis methods as well as clinical trials. Starting from those we try to sketch a trajectory where the field of biomedical hyperpolarized 13C MRI might go.

Clinically Translatable Hyperpolarized 13C Bicarbonate pH Imaging Method for Use in Prostate Cancer

Solid tumors such as prostate cancer (PCa) commonly develop an acidic microenvironment with pH 6.5-7.2, owing to heterogeneous perfusion, high metabolic activity, and rapid cell proliferation. In preclinical prostate cancer models, disease progression is associated with a decrease in tumor extracellular pH, suggesting that pH imaging may reflect an imaging biomarker to detect aggressive and high-risk disease. Therefore, we developed a hyperpolarized carbon-13 MRI method to image the tumor extracellular pH (pHe) and prepared it for clinical translation for detection and risk stratification of PCa. This method relies on the rapid breakdown of hyperpolarized (HP) 1,2-glycerol carbonate (carbonyl-13C) via base-catalyzed hydrolysis to produce HP 13CO32-, which is neutralized and converted to HP H13CO3-. After injection, HP H13CO3- equilibrates with HP 13CO2 in vivo and enables the imaging of pHe. Using insights gleaned from mechanistic studies performed in the hyperpolarized state, we solved issues of polarization loss during preparation in a clinical polarizer system. We successfully customized a reaction apparatus suitable for clinical application, developed clinical standard operating procedures, and validated the radiofrequency pulse sequence and imaging data acquisition with a wide range of animal models. The results demonstrated that we can routinely produce a highly polarized and safe HP H13CO3- contrast agent suitable for human injection. Preclinical imaging studies validated the reliability and accuracy of measuring acidification in healthy kidney and prostate tumor tissue. These methods were used to support an Investigational New Drug application to the U.S. Food and Drug Administration. This methodology is now ready to be implemented in human trials, with the ultimate goal of improving the management of PCa.

Simultaneous magnetic resonance imaging of pH, perfusion and renal filtration using hyperpolarized 13C-labelled Z-OMPD

pH alterations are a hallmark of many pathologies including cancer and kidney disease. Here, we introduce [1,5-13C2]Z-OMPD as a hyperpolarized extracellular pH and perfusion sensor for MRI which allows to generate a multiparametric fingerprint of renal disease status and to detect local tumor acidification. Exceptional long T1 of two minutes at 1 T, high pH sensitivity of up to 1.9 ppm per pH unit and suitability of using the C1-label as internal frequency reference enables pH imaging in vivo of three pH compartments in healthy rat kidneys. Spectrally selective targeting of both 13C-resonances enables simultaneous imaging of perfusion and filtration in 3D and pH in 2D within one minute to quantify renal blood flow, glomerular filtration rates and renal pH in healthy and hydronephrotic kidneys with superior sensitivity compared to clinical routine methods. Imaging multiple biomarkers within a single session renders [1,5-13C2]Z-OMPD a promising new hyperpolarized agent for oncology and nephrology.

Prediction of multiple pH compartments by deep learning in magnetic resonance spectroscopy with hyperpolarized 13C-labelled zymonic acid

BACKGROUND

Hyperpolarization enhances the sensitivity of nuclear magnetic resonance experiments by between four and five orders of magnitude. Several hyperpolarized sensor molecules have been introduced that enable high sensitivity detection of metabolism and physiological parameters. However, hyperpolarized magnetic resonance spectroscopy imaging (MRSI) often suffers from poor signal-to-noise ratio and spectral analysis is complicated by peak overlap. Here, we study measurements of extracellular pH (pHe) by hyperpolarized zymonic acid, where multiple pHe compartments, such as those observed in healthy kidney or other heterogeneous tissue, result in a cluster of spectrally overlapping peaks, which is hard to resolve with conventional spectroscopy analysis routines.

METHODS

We investigate whether deep learning methods can yield improved pHe prediction in hyperpolarized zymonic acid spectra of multiple pHe compartments compared to conventional line fitting. As hyperpolarized 13C-MRSI data sets are often small, a convolutional neural network (CNN) and a multilayer perceptron (MLP) were trained with either a synthetic or a mixed (synthetic and augmented) data set of acquisitions from the kidneys of healthy mice.

RESULTS

Comparing the networks' performances compartment-wise on a synthetic test data set and eight real kidney data shows superior performance of CNN compared to MLP and equal or superior performance compared to conventional line fitting. For correct prediction of real kidney pHe values, training with a mixed data set containing only 0.5% real data shows a large improvement compared to training with synthetic data only. Using a manual segmentation approach, pH maps of kidney compartments can be improved by neural network predictions for voxels including three pH compartments.

CONCLUSION

The results of this study indicate that CNNs offer a reliable, accurate, fast and non-interactive method for analysis of hyperpolarized 13C MRS and MRSI data, where low amounts of acquired data can be complemented to achieve suitable network training.

Frequency-Selective Manipulations of Spins allow Effective and Robust Transfer of Spin Order from Parahydrogen to Heteronuclei in Weakly-Coupled Spin Systems

We present a selectively pulsed (SP) generation of sequences to transfer the spin order of parahydrogen (pH2 ) to heteronuclei in weakly coupled spin systems. We analyze and discuss the mechanism and efficiency of SP spin order transfer (SOT) and derive sequence parameters. These new sequences are most promising for the hyperpolarization of molecules at high magnetic fields. SP-SOT is effective and robust despite the symmetry of the 1 H-13 C J-couplings even when precursor molecules are not completely labeled with deuterium. As only one broadband 1 H pulse is needed per sequence, which can be replaced for instance by a frequency-modulated pulse, lower radiofrequency (RF) power is required. This development will be useful to hyperpolarize (new) agents and to perform the hyperpolarization within the bore of an MRI system, where the limited RF power has been a persistent problem.

In vitro singlet state and zero-quantum encoded magnetic resonance spectroscopy: Illustration with N-acetyl-aspartate

Magnetic resonance spectroscopy (MRS) allows the analysis of biochemical processes non-invasively and in vivo. Still, its application in clinical diagnostics is rare. Routine MRS is limited to spatial, chemical and temporal resolutions of cubic centimetres, mM and minutes. In fact, the signal of many metabolites is strong enough for detection, but the resonances significantly overlap, exacerbating identification and quantification. Besides, the signals of water and lipids are much stronger and dominate the entire spectrum. To suppress the background and isolate selected signals, usually, relaxation times, J-coupling and chemical shifts are used. Here, we propose methods to isolate the signals of selected molecular groups within endogenous metabolites by using long-lived spin states (LLS). We exemplify the method by preparing the LLSs of coupled protons in the endogenous molecules N-acetyl-L-aspartic acid (NAA). First, we store polarization in long-lived, double spin states, followed by saturation pulses before the spin order is converted back to observable magnetization or double quantum filters to suppress background signals. We show that LLS and zero-quantum coherences can be used to selectively prepare and measure the signals of chosen metabolites or drugs in the presence of water, inhomogeneous field and highly concentrated fatty solutions. The strong suppression of unwanted signals achieved allowed us to measure pH as a function of chemical shift difference.

Litmus-Body: A Molecularly Targeted Sensor for Cell-Surface pH Measurements

Precise pH measurements in the immediate environment of receptors is essential for elucidating the mechanisms through which local pH changes associated with diseased phenotypes manifest into aberrant receptor function. However, current pH sensors lack the ability to localize and target specific receptor molecules required to make these measurements. Herein we present the Litmus-body, our recombinant protein-based pH sensor, which through fusion to an anti-IgG nanobody is capable of piggybacking on IgG antibodies for molecular targeting to specific proteins on the cell surface. By normalizing a pH-dependent green fluorescent protein to a long Stokes shift red fluorophore or fluorescent protein, we readily report pH independent of sensor concentration using a single 488 nm excitation. Our Litmus-body showed excellent responsiveness in solution, with a greater than 50-fold change across the regime of physiological pH. The sensor was further validated for use on live cells and shown to be specific to the protein of interest. In complex with our Litmus-body, cetuximab therapeutic antibody retained its functionality in binding and inhibiting ligand interaction of its target epidermal growth factor receptor (EGFR), triggering receptor-mediated endocytosis that allowed tracking of local pH from the cell surface through the endocytic pathway.

Acquisition strategies for spatially resolved magnetic resonance detection of hyperpolarized nuclei

Hyperpolarization is an emerging method in magnetic resonance imaging that allows nuclear spin polarization of gases or liquids to be temporarily enhanced by up to five or six orders of magnitude at clinically relevant field strengths and administered at high concentration to a subject at the time of measurement. This transient gain in signal has enabled the non-invasive detection and imaging of gas ventilation and diffusion in the lungs, perfusion in blood vessels and tissues, and metabolic conversion in cells, animals, and patients. The rapid development of this method is based on advances in polarizer technology, the availability of suitable probe isotopes and molecules, improved MRI hardware and pulse sequence development. Acquisition strategies for hyperpolarized nuclei are not yet standardized and are set up individually at most sites depending on the specific requirements of the probe, the object of interest, and the MRI hardware. This review provides a detailed introduction to spatially resolved detection of hyperpolarized nuclei and summarizes novel and previously established acquisition strategies for different key areas of application.

pH Dependence of T1 for 13 C-Labelled Small Molecules Commonly Used for Hyperpolarized Magnetic Resonance Imaging

Hyperpolarization is a method to enhance the nuclear magnetic resonance signal by up to five orders of magnitude. However, the hyperpolarized (HP) state is transient and decays with the spin-lattice relaxation time (T₁ ), which is on the order of a few tens of seconds. Here, we analyzed the pH-dependence of T₁ for commonly used HP ¹³ C-labelled small molecules such as acetate, alanine, fumarate, lactate, pyruvate, urea and zymonic acid. For instance, the T₁ of HP pyruvate is about 2.5 fold smaller at acidic pH (25 s, pH 1.7, B₀ =1 T) compared to pH close to physiological conditions (66 s, pH 7.3, B₀ =1 T). Our data shows that increasing hydronium ion concentrations shorten the T₁ of protonated carboxylic acids of most of the analyzed molecules except lactate. Furthermore it suggests that intermolecular hydrogen bonding at low pH can contribute to this T₁ shortening. In addition, enhanced proton exchange and chemical reactions at the pK_a appear to be detrimental for the HP-state.

Ab Initio Simulation of pH-Sensitive Biomarkers in Magnetic Resonance Imaging

An ab initio simulation scheme is introduced as a theoretical prescreening approach to facilitate and enhance the research for pH-sensitive biomarkers. The proton ¹H and carbon ¹³C nuclear magnetic resonance (NMR) chemical shifts of the recently published marker for extracellular pH, [1,5-¹³C₂]zymonic acid (ZA), and the as yet unpublished ( Z)-4-methyl-2-oxopent-3-enedioic acid (OMPD) were calculated with ab initio methods as a function of the pH. The influence of the aqueous solvent was taken into account either by an implicit solvent model or by explicit water molecules, where the latter improved the accuracy of the calculated chemical shifts considerably. The theoretically predicted chemical shifts allowed for a reliable NMR peak assignment. The p K_a value of the first deprotonation of ZA and OMPD was simulated successfully whereas the parametrization of the implicit solvent model does not allow for an accurate description of the second p K_a. The theoretical models reproduce the pH-induced chemical shift changes and the first p K_a with sufficient accuracy to establish the ab initio prescreening approach as a valuable support to guide the experimental search for pH-sensitive biomarkers.

Hyperpolarized Amino Acid Derivatives as Multivalent Magnetic Resonance pH Sensor Molecules

pH is a tightly regulated physiological parameter that is often altered in diseased states like cancer. The development of biosensors that can be used to non-invasively image pH with hyperpolarized (HP) magnetic resonance spectroscopic imaging has therefore recently gained tremendous interest. However, most of the known HP-sensors have only individually and not comprehensively been analyzed for their biocompatibility, their pH sensitivity under physiological conditions, and the effects of chemical derivatization on their logarithmic acid dissociation constant (pK_a). Proteinogenic amino acids are biocompatible, can be hyperpolarized and have at least two pH sensitive moieties. However, they do not exhibit a pH sensitivity in the physiologically relevant pH range. Here, we developed a systematic approach to tailor the pK_a of molecules using modifications of carbon chain length and derivatization rendering these molecules interesting for pH biosensing. Notably, we identified several derivatives such as [1-¹³C]serine amide and [1-¹³C]-2,3-diaminopropionic acid as novel pH sensors. They bear several spin-1/2 nuclei (¹³C, ¹⁵N, ³¹P) with high sensitivity up to 4.8 ppm/pH and we show that ¹³C spins can be hyperpolarized with dissolution dynamic polarization (DNP). Our findings elucidate the molecular mechanisms of chemical shift pH sensors that might help to design tailored probes for specific pH in vivo imaging applications.