Key words

1 Introduction

While renal MRI has traditionally been limited largely to morphologic depiction of vascular disease and neoplasia, hyperpolarized (HP) 13C MRI supplies new metabolic and/or functional insight that could be valuable for a range of kidney diseases. The high renal delivery of intravenously injected small molecules, which provides extraordinary sensitivity for renal imaging of HP 13C MRI probes, suggests great potential for application of this nascent medical imaging modality to kidney disease.

Commonly used clinical markers for kidney disease have significant limitations, especially in the context of early disease. Serum creatinine is frequently utilized as a functional marker, but has a wide reference interval, and the mathematical corrections applied for the estimation of glomerular filtration rate (GFR) are inexact. For example, GFR is unpredictably overestimated in glomerulopathic patients as a result of increased creatinine secretion [1]. Even accurately measured GFR, such as by the “gold standard” inulin clearance, is insensitive to early disease [2]. Blood urea nitrogen (BUN) fluctuates due to factors unrelated to kidney function, such as hydration status and diet. Proteinuria at baseline is a significant risk factor for renal disease, but has poor negative predictive value [3]. In general, clinical markers do not exhibit definitive changes until a significant fraction of kidney function is already lost.

Medical imaging has clear potential to address these limitations by providing localized functional data, but this has yet to be translated on a large scale. Nuclear medicine plays a significant role, with (mercaptoacetyltriglycine) MAG3 scintigraphy estimated to account for >400,000 renograms per year [4], commonly for evaluation of renal function (often pre- or post-transplant), split function, collecting duct/urinary tract obstruction, renovascular hypertension, and renal artery stenosis. However, nuclear scans typically have relatively poor spatial resolution and lack the rich tissue contrast of MRI. Furthermore, these studies carry the risks of exposure to ionizing radiation. Ultrasound is also commonly used for assessing kidney disease [5], but is generally limited to evaluating advanced damage reflected in alterations in kidney size and shape, and to some extent blood flow. For patients with impaired renal function, iodinated Computer Tomography (CT) contrast media carry an increased risks of acute kidney injury [6]. Gadolinium based MRI contrast media are associated with nephrogenic systemic fibrosis [7].

HP 13C MRI has unique potential for improved, safer clinical imaging studies of renal structure and function, and perhaps most notably offers an unprecedented avenue toward assessing renal metabolism in vivo. Indeed, results of numerous preclinical studies conducted to date, as summarized below, show the clear potential of this new approach to fill pressing unmet clinical needs for improved markers of various diseases affecting the kidneys, both malignant and nonmalignant. More preclinical work with realistic renal disease models is, however, needed in order to bridge effectively into clinical studies.

This introduction chapter is complemented by two separate chapters describing the experimental procedure and data analysis, which are part of this book.

This chapter is part of the book Pohlmann A, Niendorf T (eds) (2020) Preclinical MRI of the Kidney—Methods and Protocols. Springer, New York.

2 Measurement Concept

2.1 Basic Concept of HP 13C MRI

HP 13C MRI is based on a transient enhancement of the nuclear magnetism of 13C-enriched liquids, by up to five orders of magnitude over states readily attainable in a clinical MRI magnet, thereby offsetting the limitation of poor sensitivity that previously hampered in vivo 13C MRI. Although multiple alternate approaches exist for producing HP material, the method of dissolution dynamic nuclear polarization (DNP) [8] provides an especially robust avenue for hyperpolarizing a wide variety of 13C-labeled substrates, using commercially available instruments. In this approach, the 13C substrate is first prepolarized in a separate magnet at low temperature (see Note 1), primarily through microwave irradiation applied near the electron paramagnetic resonance (EPR) frequency in the presence of a stable organic free radical (aka electron paramagnetic agent or EPA, usually a trityl radical) (see Note 2). Following a period of polarization buildup, the solid sample is rapidly dissolved into a liquid state where it temporarily maintains its hyperpolarization, and is quickly transferred to the MRI scanner for intravenous injection and imaging (see Note 3). The overall process is illustrated in Fig. 1. The transient hyperpolarized state persists for a period of time determined by the substrate’s T1 relaxation time, an exponential time constant governing the rate of decay of polarization (see Note 4). Inverse scaling of T1 with molecular size limits the scope of applicable probes to small molecules, with labels preferably located distal to bonded hydrogen atoms, which can destroy the 13C magnetization via 13C-1H dipolar coupling (see Note 5). During the decay period, 13C MRI data is collected, capitalizing on the sensitivity enhancement afforded by hyperpolarization. Specific methods for data acquisition and processing of HP 13C MRI data are described in the relevant sections to follow.

Fig. 1
figure 1

Illustration of the process of HP 13C MRI via dissolution dynamic nuclear polarization. The agent (e.g., [1-13C]pyruvate) is mixed with the EPA and cooled to 0.8 K in a magnetic field of 5 T, then irradiated at 140GHz to transfer polarization from unpaired electrons in the EPA to the 13C nuclei. The sample is then rapidly dissolved and transferred to the MRI scanner for intravenous injection and MR imaging

Full size image

2.2 Renal Functional and Metabolic Investigations Using HP 13C MRI

HP 13C MRI offers many new possibilities for renal functional and metabolic investigations. Dynamic images of the real-time delivery of metabolically inert tracers such as 13C urea from arterial vasculature to the kidneys can be used to quantify renal function. Real-time in vivo processing of metabolically active HP probes by the kidneys can be tracked by spectroscopic imaging, which yields separate images of injected tracer(s) and their downstream metabolic product(s) based on chemical shift differences. Notably, multiple probes can also be simultaneously copolarized to yield multiparametric data sets [9], for example spanning both renal functional and metabolic parameters. In this section, we review the most promising HP 13C MRI probes of renal function/metabolism proposed to date, and the relevant associated principles. These probes and their metabolic pathways are illustrated in Fig. 2.

Fig. 2
figure 2

HP 13C MRI probes of interest for monitoring renal disease, and the associated metabolic conversions (if any). Positions of the 13C labels are indicated by stars

Full size image

2.2.1 [13C]Urea

The key osmolyte and metabolic end product urea was one of the first 13C-labeled molecules to be hyperpolarized (T1 = 45 s at 3 T) [8] and imaged in vivo [10], and high quality dynamic images of the kidneys and the feeding arterial vasculature can readily be generated [11,12,13]. Quantitative estimates of renal function require dynamic measurements from both the renal parenchyma and arterial vasculature [11, 14]. Urea is largely reabsorbed by urea transporters (UTs) expressed at the inner medullary collecting ducts (IMCD), and accordingly renal imaging shows distinct modulation by hydration state [13, 15], based on differential action of vasopressin. Despite this significant reabsorption, it has been suggested that urea can also be used to estimate glomerular filtration rate (GFR) [14], based on the first-order transfer coefficient between arterial vasculature and renal parenchyma. Secondary labeling of [13C]urea with 15N2 is helpful in terms of prolonging both T1 (at low magnetic fields only) [16] and T2 relaxation times [12] of urea, with the latter factor allowing imaging rat kidneys at spatial resolutions approaching 1 mm isotropic (using refocused image acquisition) [13]. Furthermore, relaxation mapping shows great promise for improved contrast between the individual kidney compartments and various functional states [12, 13, 17]. Urea is distinguished by an exceptionally good safety profile among medical imaging contrast agents, even in patients with reduced kidney function, and therefore has potential for clinical translation for imaging patients with kidney diseases.

2.2.2 [1-13C]Pyruvate

Pyruvate, the end product of glycolysis and key metabolic intermediate, is the most widely studied 13C probe, in C1-labeled form (T1 = 60s at 3 T) [18]. In the kidneys, its interconversions with lactate (via lactate dehydrogenase, LDH) and alanine (via alanine transaminase, ALT), as well as its decarboxylation yielding [13C]bicarbonate (via pyruvate dehydrogenase or PDH) can readily be tracked by spectroscopic imaging. The extent of metabolic conversion observed reflects a complex series of biophysical processes, including vascular delivery, cellular uptake (via monocarboxylate transporters), and finally enzymatic conversion, which in turn may be influenced by multiple factors. For the conversion via LDH, the lactate pool size is especially important [19, 20]. For quantitative comparison among subjects, all of these factors are usually grouped into simple parameters such as metabolite area-under-the-curve (AUC) ratios or related apparent first-order conversion rates (e.g., kpyruvate-to-lactate or kpl) [21, 22], derived from a series of dynamic spectroscopic imaging data. From a metabolic perspective, while the conversion to bicarbonate is clearly a net flux, the conversions to lactate and alanine via the respective bidirectional, highly active enzymes appear to represent largely isotopic exchange fluxes into the respective product metabolite pools [19, 20, 23]. Because of the high bidirectional activity of LDH (which is restricted to cytoplasm), the ratio of lactate to pyruvate has frequently been interpreted as a marker for the cytosolic free NADH/NAD+ ratio [24]. This suggests that the conversion of HP pyruvate to lactate is driven by this NAD(H) redox state, a parameter that is fundamentally inter-connected with numerous related biochemical reactions.

2.2.3 [1,4-13C2]Fumarate

Cellular uptake of injected fumarate, a tricarboxylic acid (TCA) cycle intermediate, is ordinarily highly restricted on the timescale of HP experiments, due to its dicarboxylate structure. Compromise of the cellular membrane (e.g., due to necrosis) allows access of injected HP [1,4-13C2]fumarate (a singlet due to molecular symmetry, with T1 = 58 s at 3 T) [25] to cellular fumarase, and consequently detectable conversion of product HP [1,4-13C2]malate. Thus, HP fumarate has been investigated as a specific marker of cellular necrosis with potentially significant applicability to detection of renal tubular necrosis [26].

2.2.4 [1-13C]Dehydroascorbate

HP [1-13C]dehydroascorbate (T1 = 56 s at 3 T) [27, 28] can be used to probe oxidative stress in a direct manner. Vitamin C, a key physiologic antioxidant, exists in a NADP(H)-mediated equilibrium with its oxidized form, dehydroascorbate (DHA). Conversion of injected HP [1-13C]DHA, which is rapidly taken up by glucose transporters, to [1-13C]vitamin C is detectable in vivo. The observed extent of conversion of HP DHA to vitamin C, a reaction whose reducing power is likely mostly derived from reduced glutathione (GSH), is attenuated by oxidative stress.

2.2.5 [13C]Acetoacetate

The ketone body acetoacetate, a universal oxidative fuel, rapidly interconverts with its reduced form, beta-hydroxybutyrate, in mitochondria via beta-hydroxybutyrate dehydrogenase. In analogy with lactate/pyruvate, the BOHB–AcAc ratio has frequently been interpreted as a marker of the free mitochondrial NADH–NAD+ ratio, suggesting an analogous interpretation of conversion of HP acetoacetate [29, 30] as being driven by the mitochondrial NAD(H) redox state. Metabolic conversion of injected [1,3-13C2]acetoacetate (58 s at 3 T, C1 position) can be detected in rat kidney in vivo (via the C1 label) [29], suggesting applicability of this new probe to interrogate mitochondrial-driven kidney disease.

3 Overview of Applications

3.1 Kidney Cancer

Multiple studies have reported elevated conversion of HP [1-13C]pyruvate to [1-13C]lactate and rapid export of [1-13C]lactate in renal cell carcinoma (RCC), in both RCC cells ex vivo [31,32,33] and in vivo in mice orthotopically implanted with human RCC cells [34]. These findings suggest that HP [1-13C]pyruvate MRI could have clinical value for improved characterization of kidney cancer in patients. Moreover, a recent study showed that HP [1-13C]pyruvate MRI could be used to predict RCC treatment response to mTOR inhibition, which varies greatly among individual patients, thus potentially informing treatment decisions [35].

3.2 Diabetic Nephropathy

Diabetes results in profound shifts in central carbon metabolism, directly shifting several pathways accessible to HP 13C MRI. Although diabetes is detectable using relatively simple testing, sustained hyperglycemia over time leads to serious end-organ complications including diabetic nephropathy, a significant source of morbidity, the detection/prediction of which is a difficult clinical problem where HP 13C MRI could be valuable. Though cancer has been the main focus of the HP 13C MRI community, several recent studies have reported differences in renal HP 13C signals detected in type 1 and type 2 diabetes models as compared to normal controls, based on multiple HP probes including [1-13C]pyruvate [23, 36, 37], [13C]urea [38, 39], and [1-13C]DHA [40]. Not surprisingly, standard antidiabetic agents including insulin and metformin have been found to induce large renal HP signal changes as well [29, 41, 42]. Renal HP lactate levels are attenuated with hyperbaric [43] and antioxidant treatment [44] in diabetic rats. HP [13C]acetate has so far failed to show any significant metabolic change associated with the diabetic kidney [45]. Further work is needed in animal models of frank diabetic nephropathy, which may display changes additive to the fundamental metabolic shifts observed in diabetes, to address this important clinical problem.

3.3 Ischemic Kidney Injury

Effective detection of acute kidney injury (AKI) is an unresolved clinical problem especially affecting hospitalized patients. Rapid treatment of AKI is critical but is impeded by existing clinical indicators which are poor in diagnosing early AKI. Several studies of models of ischemia reperfusion injury (i.e., unilateral renal artery clamp) have shown clear effects of AKI on renal HP 13C MRI using [13C]urea [46, 47], pyruvate [48], and [1-13C]pyruvate–[1-13C]DHA combination [49], and fumarate [50]. However, these relatively simple models likely do not reflect the etiology of ischemic injuries more frequently encountered in patients, such as those resulting from atherosclerotic disease. Further work using more realistic models of ischemic kidney injury is needed to answer, for example, the important question of predicting treatment response in renal artery stenosis. Finally, HP [1,4-13C2]fumarate has also been investigated as a probe of tissue necrosis in a toxic folic acid model of acute tubular necrosis (ATN) [26]. A nonzero urinary fumarase activity could potentially be used as a clinical indicator for a hyperpolarized fumarate examination to investigate the extent and the origin of the renal damage [50].

4 Notes

  1. 1.

    The state-of-the-art GE SPINLab polarizer operates at 5 T and a temperature of 0.8 K. Microwave irradiation is applied near the EPR frequency corresponding to this magnetic field (~140 GHz), with a typical power of ~20 mW. A buildup time of ~3 h is typically required.

  2. 2.

    Urea and many other agents including fumarate and acetoacetate require the addition of a glassing agent such as glycerol or DMSO to prevent crystallization upon freezing, which impedes the polarization process.

  3. 3.

    The 13C concentration is made as high as possible in the starting material (in the molar range), facilitating efficient polarization, but the concentration typically drops >10-fold on dissolution. In contrast, the concentration of the radical is only ~15 mM in the starting material.

  4. 4.

    The T1 exponential decay constant of HP 13C probe is directly proportional to its “half-life,” by a factor of ln 2 (i.e., t1/2 = 0.69 × T1).

  5. 5.

    T1 values of 30–60 s are typical for the most useful probes, and T1 values are magnetic field- and temperature-dependent.