The use of dynamic nuclear polarization ¹³C-pyruvate MRS in cancer

Abstract

In recent years there has been an immense development of new targeted anti-cancer drugs. For practicing precision medicine, a sensitive method imaging for non-invasive, assessment of early treatment response and for assisting in developing new drugs is warranted. Magnetic Resonance Spectroscopy (MRS) is a potent technique for non-invasive in vivo investigation of tissue chemistry and cellular metabolism. Hyperpolarization by Dynamic Nuclear Polarization (DNP) is capable of creating solutions of molecules with polarized nuclear spins in a range of biological molecules and has enabled the real-time investigation of in vivo metabolism. The development of this new method has been demonstrated to enhance the nuclear polarization more than 10,000-fold, thereby significantly increasing the sensitivity of the MRS with a spatial resolution to the millimeters and a temporal resolution at the subsecond range. Furthermore, the method enables measuring kinetics of conversion of substrates into cell metabolites and can be integrated with anatomical proton magnetic resonance imaging (MRI). Many nuclei and substrates have been hyperpolarized using the DNP method. Currently, the most widely used compound is ¹³C-pyruvate due to favoring technicalities. Intravenous injection of the hyperpolarized ¹³C-pyruvate results in appearance of ¹³C-lactate, ¹³C-alanine and ¹³C-bicarbonate resonance peaks depending on the tissue, disease and the metabolic state probed. In cancer, the lactate level is increased due to increased glycolysis. The use of DNP enhanced ¹³C-pyruvate has in preclinical studies shown to be a sensitive method for detecting cancer and for assessment of early treatment response in a variety of cancers. Recently, a first-in-man 31-patient study was conducted with the primary objective to assess the safety of hyperpolarized ¹³C-pyruvate in healthy subjects and prostate cancer patients. The study showed an elevated ¹³C-lactate/¹³C-pyruvate ratio in regions of biopsy-proven prostate cancer compared to noncancerous tissue. However, more studies are needed in order to establish use of hyperpolarized ¹³C MRS imaging of cancer.

Introduction

Increasingly, therapies are tailored either through targeting specific abnormalities, e.g. genetic mutations, or through early response monitoring. Sensitive methods for non-invasive in vivo imaging assessment of early treatment response as well as assisting in developing new drugs are warranted. Positron Emission Tomography (PET) imaging with fluorodeoxyglucose (FDG) has been established as a clinical tool in molecular imaging. FDG is a glucose analog labeled with the positron-emitting isotope ¹⁸F. It is transported across the cell membrane by glucose transporters (GLUTs), phosphorylated by hexokinase, trapped in the cell and accumulates in proportion to local metabolism. Increased ¹⁸F-FDG uptake is then detected using a PET scanner [1]. ¹⁸F-FDG uptake can be measured using standardized uptake value (SUV), which is a quantitative measure of radioactivity concentration in PET images normalized to injected dose and body weight, ¹⁸F-FDG SUV thus measures tumor metabolic activity. However, this standard has certain limitations and can be affected by factors such as blood glucose level in the patient [2]. Conventional computed tomography (CT) is a sensitive anatomical scanning technique in tumor imaging, however it lacks physiological information. Introduction of new CT techniques (dual energy CT and CT perfusion imaging with iodine contrast agents) together with recent advances in hardware has improved spatial and temporal resolution, and physiological information is now obtainable from CT datasets. Nevertheless, when evaluating treatment response, consecutive acquisitions are usually performed yielding a substantial radiation burden; therefore non-ionizing methods should be considered as an alternative. Diffusion Weighted Magnetic Resonance Imaging (DW-MRI) is non-ionizing and measures the movement of proton nuclei, primarily from water, and thereby demonstrates tissue cellularity and integrity of cell membranes. Diffusion based MRI has proven to be useful for staging and predicting prognosis of cancer patients. However, diffusion based MRI is only sensitive to structural information and lacks metabolic information [3].

MR spectroscopy

Magnetic Resonance Spectroscopy (MRS) imaging is like MRI based on the principle of nuclear magnetic resonance (NMR). MRS is a potent technique for non-invasive in vivo investigation of tissue chemistry. MRS has been widely applied in biochemistry research for the analysis of basic organic compounds and provides information about e.g. tissue metabolites. Conventional proton MRI uses magnetic gradients to create controlled perturbations of the magnetic field, thus creating anatomical images. MRS likewise codes the spatial dimensions, but in addition, utilizes the miniscule differences in resonance frequency of the various nuclei, depending on the molecular structure, creating an accurate illustration of metabolic chemistry in situ. In theory, many biochemically relevant nuclides, such as hydrogen (¹H), carbon (¹³C), nitrogen (¹⁵N), fluorine (¹⁹F), sodium (²³Na) and phosphor (³¹P) possess a magnetic moment and can be used for MRS. However, only the nuclei of phosphor (³¹P) and hydrogen (¹H) exist in high enough concentrations to be used for routine clinical evaluation, see Table 1. Depending on the chemical structure of the molecule, the nuclei responsible for the NMR signal is exposed to a slightly different magnetic environment, hence resonating at a slightly different frequency. This way various metabolites can be quantified. The method is primarily used in brain and prostate cancer and spectra associated with tumor tissue are distinguished from those associated with normal tissue by the characteristic increase in the resonance peak of choline corresponding to increased cell membrane density [4,5]. There are several limitations to this technique, with consequential decrease of temporal resolution to the minute scale and of spatial resolution to 1 cm³ depending on the nucleus detected and of the magnetic field strength [6,7]. Moreover, spectroscopic images of tissue metabolites lack dynamic information about metabolic fluxes [8].

Table 1.

Table with MR biologically interesting nuclides and demonstrating, that only the nuclei of phosphor (³¹P) and hydrogen (¹H) exist in high enough concentrations to be used for routine clinical evaluation [70]

Nuclide	Abundance in tissue	MR-detect. isotope	Abundance of isotope	Abundance in tissue
H	63%	1H	99.99%	63%
C	12%	13C	1.1%	0.13%
N	0.58 %	15N	0.37%	0.0021%
O	24%	17O	0.038%	0.0091%
P	0.22%	31P	100%	0.22%

Dynamic nuclear hyperpolarization

A more sensitive method that allows real time imaging of the dynamics of the metabolism is therefore necessary. Hyperpolarization in the context of NMR refers to the process of relocating the nuclei temporarily into a redistribution of energy levels [9]. The hyperpolarization process can be accomplished by a number of techniques; 1) optical pumping, which is primarily in use for noble gases as ³He and ¹²⁹Xe in lung perfusion imaging [10], 2) PHIP (para hydrogen-induced polarization) which suffers from the limited number of metabolic substrates that can be hyperpolarized [11,12] or 3) dynamic nuclear hyperpolarization (DNP). A more extensive description of the two former methods can be found elsewhere and is beyond the scope of this paper.

Ardenkjaer-Larsen et al. first described hyperpolarization with DNP and was established on DNP in the solid state. The method is based on the development of a dissolution technique that preserves the nuclear polarization during the transition from solid to liquid state. The method is capable of creating solutions of molecules with polarized nuclear spins in various molecules and has enabled the real-time investigation of in vivo metabolism [13,14]. The procedure consists of polarization transfer from electrons to nuclei in solids by microwave irradiation. The solution to be polarized typically contains ¹³C-enriched substrates; a free stable radical is dissolved in a glass-forming solvent and then positioned into a high magnetic field (e.g. 3 T). The free radical provides unpaired electrons and the glassing material optimize the efficiency of microwave energy transfer. The compound is brought to very low temperature (≈1 K) and irradiated. After the polarization transfer, the microwaves are turned off, and the metabolite is warmed rapidly to room temperature with substantial retention of the polarization that was present in the frozen solid state. It is then quickly transferred and injected into the specimen or animal for MRI acquisition. The nuclear spin polarization is proportional to gamma (γ) and B₀/T where γ is the nucleus specific gyromagnetic ratio constant, B₀ is the applied magnetic field and T is temperature. With this method, a liquid state polarization of up to 64% has been achieved [15]. The time for the polarization to build up to maximum level is about one hour.

The use of this new method has enhanced the nuclear polarization more than 10,000-fold. It has resulted in a significantly increased sensitivity of the MRS and increased the spatial resolution to the millimeter scale and the temporal resolution to the sub second range. Furthermore, these advances makes it possible to measure the kinetics of the conversion of the metabolite into other cell metabolites and can be integrated with anatomic proton MRI [14]. One limitation, however, is a relatively short half-life (10-30 s in vivo) of the hyperpolarized metabolites. Therefore, the solution requires rapid transfer into the patient and the imaging sequence must be efficient and accomplished within 2-3 minutes and the metabolites must be metabolized and transferred across the cell membrane very rapidly [14].

Many nuclei have been hyperpolarized using the DNP method. Currently, the most widely used metabolite is ¹³C-pyruvate due to the optimal gyromagnetic ratio, the relatively long relaxation time and finally the possibility of incorporating ¹³C into biological relevant metabolites [16,17]. Intravenous injection of the hyperpolarized ¹³C-pyruvate results in an increase of ¹³C-lactate, ¹³C-alanine and ¹³C-bicarbonate resonance peaks depending on the tissue, disease and the metabolic state probed. In cancer, ¹³C-lactate is increased due to increased glycolysis. Numerous studies have shown that glycolysis and production of lactate and alanine is increased in tumor cells compared to normal cells even when oxygen is sufficient a phenomenon known as Warburg effect [18], as shown in Figure 1.

Figure 1.

A key metabolic fingerprint in cancer cells is the switch to glycolysis with production of lactate even in the presence of sufficient oxygen (reproduced with permission from ref [68]).

Use of hyperpolarized ¹³C-pyruvate in cancer

Detecting tumor response to treatment

Currently treatment response is assessed by imaging measurement to monitor anatomical tumor reductions in accordance with the latest guidelines of the “Response Evaluation Criteria In Solid Tumors” (RECIST). Here, partial response to treatment is defined as “at least a 30% decrease in the sum of the longest diameter of target lesions, taking as reference the baseline sum longest diameter” [19]. However these anatomic changes defining partial response may take weeks to develop in tumors and in some cases not develop at all or even demonstrate growth due to tumor stasis after treatment [20,21]. In these situations ineffective treatment or even harmful treatment can be given. Consequently, ¹⁸F-FDG-PET imaging has or is under investigation for therapy response assessment in many tumors. Quantitative estimation of ¹⁸F-FDG uptake by SUV in tumor has demonstrated a distinct reduction of glucose to predict response of treatment, despite the fact that tumor increased in size [22]. However, ¹⁸F-FDG-PET lacks sensitivity to certain tumor types, due to low ¹⁸F-FDG uptake (e.g. prostate cancer) [23], high uptake in surrounding tissue (brain cancers) [24] or specificity due to ¹⁸F-FDG uptake in infected or inflamed tissue [25]. In addition, high ¹⁸F-FDG uptake of infiltrating immune cells after treatment can mask the decreased uptake by the dying cancer cells and correlates with the number of metabolically active viable cells, regardless of whether they are tumoral or non-tumoral in nature. Sustained high ¹⁸F-FDG uptake can occur in a tumor mass after administration of chemotherapy due to infiltrating macrophages [26]. Lastly, PET exposes the patient to significant radiation burden and can only be repeated a limited number of times.

Proton based MRS has the potential to provide routine assessment of cancer treatment. However, it is still a relatively novel technology and more work is required to validate the biomarkers that MRS can detect [27]. In an orthotopic animal model of glioma, it was investigated whether changes in ¹H MRS metabolites could render reliable biomarkers for an early response to suberoylanilide hydroxamic acid (SAHA) treatment which blocks histone de-acetylation activity and results in the accumulation of acetylated histones and proteins, such as transcription factors essential for expression of genes necessary to induce cell differentiation [28]. Wei et al. demonstrated a decrease in the concentration of lactate in tumors of treated rats. These results indicate that SAHA inhibits tumor growth and induces normal brain tissue-like metabolism in brain tumors [29]. One of the earliest studies using hyperpolarized ¹³C-pyruvate in treatment response in cancer used lymphoma lesion in mice treated with topoisomirase inhibitor (Etoposide). Etoposide induces loss of NAD(H) and decrease in LDH activity and consequently apoptosis with demonstration of decreased flux in hyperpolarized ¹³C-pyruvate and ¹³C-lactate [30], see Figure 2. This flux is inhibited within 24 hours of treatment. Witney et al. compared ¹⁸F-FDG-PET and hyperpolarized ¹³C-pyruvate for detecting tumor response to treatment of murine lymphoma and demonstrated that the timing of response varied. The decreased ¹⁸F-FDG uptake was preceded by the decrease in flux of ¹³C metabolite between pyruvate and lactate, however with equal relative reductions in both ¹⁸F-FDG and hyperpolarized ¹³C-pyruvate/¹³C-lactate. This suggests that the most important advantage of hyperpolarized MRI may be in response assessment in cancers where ¹⁸F-FDG-PET is insensitive [31]. Decreased hyperpolarized ¹³C-lactate was also shown in a cell line of human breast adenocarcinoma treated with another topoisomirase inhibitor (doxorubicin) where the level of hyperpolarized ¹³C-lactate production was decreased before any detectable change in tumor size reflecting DNA [32]. The phosphatidylinositol 3-kinase (PI3K) pathway has been found to have key regulatory roles in many cellular processes, including cell survival, proliferation, glucose metabolism and cell differentiation. Several anti cancer therapeutic drugs targeting the pathway are in clinical trials [33]. Traditional imaging methods such as CT and MRI are inadequate in monitoring early treatment response due to tumor stasis rather than shrinkage associated with PI3K inhibition [34]. Hyperpolarized ¹³C-pyruvate was used in mice with subcutaneously implanted glioblastoma xenografts detecting the effect of PI3K inhibition and produced a decrease of hyperpolarized ¹³C-lactate. This finding was then confirmed in mice with breast cancer cells and associated with reduction in LDH activity of expression and hypoxia-inducible factor-1α proteins (HIF-1α) [35]. This was also shown in a more clinically relevant model involving invasive orthotopic (intracranial) glioblastoma xenografts in rats treated with Everolimus, an mTOR inhibitor that is part of the PI3K pathway [36]. The same mouse model was also treated with the most frequently used chemotherapeutic in treating glioblastoma, Temozlomide, an alkylating agent that works by damaging DNA. It was demonstrated that the ratio of hyperpolarized ¹³C-pyruvate to ¹³C-lactate was increased in the treatment group [37]. Cell signalling through the PI3K pathway can be inhibited by various receptor tyrosine kinase inhibitors (RTK) (Imatinib) and was associated with receptors for the stem cell factor (c-KIT) and platelet-derived growth factor (PDGFR) [38]. In a prostate cancer model inhibition of PDGFR with Imatinib reduced the flux of hyperpolarized ¹³C-lactate by lowering the expression of LDH itself, which was mediated, by reduced expression of their transcriptional factors HIF-1α and c-MYC [39]. Vascular disrupting agents are chemotherapeutics that selectively targets endothelial cell and leads to tumor cell necrosis and is not triggering tumor shrinkage [40]. Vascular function has been measured using dynamic contrast agent-enhanced MRI and ¹⁵O-H₂O-PET, however the necessity of a wide variety of kinetic models and an onsite cyclotron and has limited its use in clinical practice. In a murine lymphoma model, the rate of flux of hyperpolarized ¹³C between pyruvate and lactate was decreased by 34% within 6 and 24 hours of treatment with the vascular disrupting agent combrestatin-A4-phosphate (CA4P). A simultaneously change on dynamic contrast agent-enhanced MRI after 6 hours was demonstrated, which was normal after 24 hours. DWI MR showed no change on the ADC (apparent diffusion coefficient) of tumor after 6 hours of treatment, although a decrease after 24 hours was observed and therefore it was concluded that hyperpolarized ¹³C-pyruvate might provide a more sensitive indicator in vascular disrupting drugs than the traditionally used MR modalities [41]. The same group also demonstrated that hyperpolarized ¹³C MRS can detect the early effects of anti-angiogenic therapy (bevacizumab, Avastin) in 2 colorectal cancer xenograft models (LoVo and HT29), which are known to display differential response to VEGF (vascular endothelial growth factor) blockade [42].

Figure 2.

Color maps representing ¹³C-lactate and ¹³C-pyruvate peak intensities obtained from ¹³C chemical-shift images in the same mouse before and after treatment with etoposide. The images were acquired between 20 and 25 s after intravenous injection of 0.2 ml 75 mM hyperpolarized ¹³C-pyruvate and 20 h after injection of the mice with 67 mg/kg etoposide. Tumor margins are indicated by the white lines. Reprinted with kind permission from Nature Medicine [30].

Other categories of treatment against cancer have also been examined. In the TRAMP model, anti-androgenic therapy demonstrated that hyperpolarized ¹³C-pyruvate correlated with disease progression and increased in untreated animals [43]. In addition, the pre-treatment ¹³C-lactate to ¹³C-pyruvate ratio was a predictor of treatment outcome [2]. Day et al. investigated a rat model of gliobastoma in treatment of radiotherapy and ¹³C hyperpolarization. It was demonstrated that a response could be detected after radiotherapy through decreased ¹³C labeling of tumor lactate. Response could be detected even though an increase in tumor volume was demonstrated [44]. This study shows that it may be possible to distinguish between true progression and pseudo progression due to tumor stasis and could be used to guide subsequent treatment using hyperpolarized ¹³C-pyruvate. An overview of treatment response studies is given in Table 2.

Table 2.

Overview of studies investigating treatment of response of cancer with hyperpolarized ¹³C-pyruvate

Therapy	Host	Cancer type	Cancer cell line	Tumor inoculation	13C-pyruvate/lactate response	Other imaging modalities	Reference
Etoposide	Mice	Lymphoma	EL-4 (moue)	S.c.	Decreased 13C-lactate		[30]
Etoposide	Mice	Lymphoma	EL-4 (mouse)	S.c.	Decreased 13C-lactate	18F-FDG-PET: Decreased SUV.	[31]
Doxyrubicin	Cell media	Breast adenocarcinoma	MDA-MB-231 (Human)		Decreased 13C-lactate		[32]
LY294002	Cell media	Glioblastoma	GS-2 (human)		Decreased 13C-lactate		[35]
Everolimus	Cell media	Breast adenocarcinoma	MDA-MB-231 (human)		Decreased 13C-lactate		[35]
Everolimus	Mice	Glioblastoma	GS-2 (human)	S.c.	Decreased 13C-lactate		[35]
Everolimus	Mice	Glioblastoma	GS-2 (human)	I.c.	Decreased 13C-lactate	MRI (T2, T1GD): No response.	[36]
Imatinib	Mice	Prostate adenocarcinoma	PC-3MM2 (human)	Tibia	Decreased 13C-lactate	MRI (DCE): Significant decease in vascular permeability.	[39]
Imatinib/ Paclitaxel	Mice	Prostate adenocarcinoma	PC-3MM2 (human)	Tibia	Decreased 13C-lactate	MRI (DCE): Significant decease in vascular permeability.	[39]
Combretastatin-A4-phosphate	Mice	Lymphoma	EL-4 (mouse)	S.c.	Decreased 13C-lactate	MRI (DCE): Significant suppression of uptake at 6 h after treatment, returning to above pretreatment values at 24 h. DWI: 6 h after treatment no change, 24 h after treatment increase in ADC.	[41]
Bevacizumab (Avastin)	Mice	Colorectal adenocarcinoma	LoVo (human)	S.c.	Decreased 13C-lactate	MRI (DCE): Antivascular effects were observed at 24 h after the first dose of, but began to recover after the second dose (72 hours. DWI: No change in ADC).	[42]
Bevacizumab (Avastin)	Mice	Colorectal adenocarcinoma	HT29 (human)	S.c.	Increased 13C-lactate	MRI (DCE): No antivascular effects. DWI: Decrease in ADC after 72 h of treatment.	[42]
Androgen deprivation	Mice	Prostate Adenocarcinoma	TRAMP	(orthotopic)	Decreased 13C-lactate		[43]
Radiotherapy	Rats	Glioma	C6 (rat)	I.c.	Decreased 13C-lactate		[44]

S.c.: subcutaneous. I.c.: intracranial. DCE: dynamic contrast enhancement.

Application in brain tumor imaging

¹⁸F-FDG-PET studies have shown increased rate of glucose uptake, increased expression of GLUT transporters in numerous cancers and a correlation of glucose uptake and accumulation of lactate with the aggressiveness and prognosis [45]. The application of ¹⁸F-FDG-PET in clinical neuro-oncology is limited by the high background uptake of normal gray matter and low specificity [24]. Therefore other investigations are warranted in brain tumor imaging. Several studies have demonstrated that increased lactate production is found in human glioma [46] and rat glioma models [47]. Proton MRS (¹H MRS) has been one tool to characterize lactate production of brain tumors and elevated lactate has been associated with poor survival in patients with glioblastoma multiforme [48]. However the interpretation of MRS is difficult in this highly heterogeneous tumor, since increased lactate level is dependent upon increased glycolysis and rate of lactate production and clearance. Furthermore, because of the complexity in separating the lactate peaks from cystic and necrotic regions. Consequently an alternative technique for assessing the tumor metabolism is necessary. Park et al. compared the in-vivo metabolism in malignant glioma and normal rat brain with two different human glioma xenografts (U-251 MG and U-87) and demonstrated the feasibility of hyperpolarized ¹³C-pyruvate in characterizing brain tumors from normal brain and between the different glioma cell lines [37,49].

Application in prostate cancer imaging

Prostate cancer is the second leading cause of cancer deaths in men. Most men diagnosed at a very early stage will die with prostate cancer but not from the disease [50]. Distinguishing lethal from clinically indolent disease is a paramount interest in prostate cancer management. So far prostate cancer is diagnosed by histopathologic samples from the gland by random transrectal biopsies. Gleason grading on histopathological examination is the best prognostic indicator in prostate cancer to date; though, grading on biopsies may not correlate with the prostatectomy specimen because of sampling errors, interobserver variation, and morphologically identical prostate cancer can behave differently [51]. The transgenic adenocarcinoma of mouse prostate (TRAMP) model mimics that of human prostate cancer. The TRAMP model develops progressive, heterogeneous and multifocal disease and spontaneous prostatic neoplasia which progresses to carcinoma and metastasis in lymph nodes, lungs and less frequently liver, adrenal glands and bones [52]. The first study with hyperpolarized ¹³C-pyruvate was applied in the TRAMP model by Chen et al. and presented the feasibility of detecting rapid cellular conversion of ¹³C-pyruvate to ¹³C-lactate in tumor and allowed the acquisition of 3D MRS in approximately 10 s with a spatial resolution of 0.135 cm³ [53]. Serial data from Albers et al. demonstrated that it was possible to differentiate and measure the histological grade on the basis of ¹³C-lactate levels in normal prostate, low-grade tumors and high-grade tumors. ¹³C-lactate levels and histological grade had a correlation coefficient of 0.95. Additionally, lymph node metastasis had significantly elevated levels of ¹³C-lactate as well [54]. Lupo et al. confirmed this and demonstrated that variations in ¹³C-lactate signal intensities were apparent for different stages of disease development, with correlation between elevated ¹³C-lactate levels and advanced disease. The intensity of the ¹³C-lactate level varied spatially within the tumor consistent with the known heterogeneity of the disease [55]. To evaluate the use of hyperpolarized ¹³C-pyruvate on a larger scale, Nelson et al. demonstrated the feasibility in healthy dogs because of the resemblance in anatomy and size of the prostate gland to that of the human. Coil geometry, data acquisition parameters and radiofrequency pulses were developed and provided a realistic test of the methodology. The observed ¹³C-lactate to ¹³C-pyruvate ratio was significantly lower than previously seen in kidneys and tumors of smaller animals but was still increased in the healthy dog prostate [56].

A phase 1 “Ascending-dose study to assess the safety and tolerability and imaging potential of hyperpolarized pyruvate (¹³C) injection in subjects with prostate cancer” was undertaken at the University of California, San Francisco. This first-in-man 31-patients study was conducted with the primary objective to assess the safety of hyperpolarized ¹³C-pyruvate in healthy men and prostate cancer patients, and confirmed the safety of the agent. In addition it showed elevated ¹³C-lactate/¹³C-pyruvate ratio in regions of biopsy-proven prostate cancer compared to non-cancerous tissue of the prostate [57].

Application in breast cancer

The metabolite flux depends upon delivery of pyruvate via the vasculature, on the rate of pyruvate transport across the plasma cell membrane and on the concentration of lactate dehydrogenase (LDH) and its substrates. Consequently, understanding the changes in lactate levels will require interpreting the relative significance of these steps in establishing the kinetics of metabolite exchange [14]. In human breast cancer cells, it was shown that the flux of hyperpolarized ¹³C-pyruvate assuming simple Michaelis-Menten kinetics was transferred to the cell as determined by the monocarboxylic acid transporter-1 (MCT). MCT is a member of a family of trans-membrane proteins, which introduces pyruvate into the glycolytic pathway. Conversely the addition of a MCT1 inhibitor (Quercetin) markedly reduced the apparent rate of ¹³C-pyruvate to ¹³C-lactate conversion and stressed out that the rate of metabolism of hyperpolarized pyruvate does not reflect the activity of LDH, but rather the activity of the MCT transporters in a breast cancer population [58]. However as shown by Witney et al. in a murine lymphoma cell line, the control of metabolite flux is probably shared between the membrane transporter and LDH [14].

Application in liver cancer

The incidence of liver cancer is continuously increasing and is usually diagnosed at a late stage [59]. Thus a sensitive non-invasive marker of disease with regard to prognosis and treatment of response is warranted. Yen et al. developed a single voxel based MRS sequence that was able to measure T₂ relaxation times of hyperpolarized ¹³C-metabolites in vivo. They demonstrated that hyperpolarized ¹³C-alanine and ¹³C-lactate level were elevated in a rat model of hepatocellular carcinoma compared to normal liver tissue [60]. This was confirmed by Darpolor et al. who demonstrated that the conversion of hyperpolarized ¹³C-pyruvate to ¹³C-alanine and ¹³C-lactate was associated with elevation of the enzymes of LDH and alanine transaminase (ALT) [61]. This suggests that the elevated levels of hyperpolarized ¹³C-lactate and ¹³C-alanine in hepatocellelular carcimona can be dependable upon the up-regulation of LDH and ALT. Hu et al. used a switchable transgenic mouse model with overexpression of the proto-oncogene MYC in the liver, giving rise to tumors that have features of human hepatocellular carcinoma and hepatoblastoma. Interestingly, conversion of ¹³Cpyruvate to ¹³C-alanine was predominant in very early stages of pre-tumor tissue, prior to any observable morphologic or histological changes. The increase of hyperpolarized ¹³C-alanine level highlights the potential of ¹³C-alanine as an early biomarker in liver cancer [62]. Continuously, a pulse sequence methodology was implemented to improve spatial resolution (to as high as 0.034 cm³) and achieve up to a factor of 7.53 in acceleration for 3D spectroscopic imaging in the same transgenic mouse model of liver cancer. In clinical applications, this new pulse sequence could be used to cover a larger field of view [63] and therefore MRS could visualize larger volumes of tumors in one sequence. Menzel et al. demonstrated that hyperpolarized ¹³C-pyruvate correlated with ¹⁸F-FDG-PET in a rat model with hepatocellular carcinoma in most of the animals. Intratumoral heterogeneities due to necrosis could be identified in the ¹³C metabolic and ¹⁸F-FDG images [64].

Applications in other cancers

The first study using hyperpolarized ¹³C-pyruvate in cancer used P22 tumor tissue, a highly malignant rat sarcoma. It was demonstrated that all implanted tumors showed significantly higher lactate content than the normal tissues and indicated that non-invasive quantification of localized Warburg effect may be possible [65]. Renal cell carcinoma is a very heterogeneous cancer with varying aggressiveness and prognosis. Differentiation between malignant and benign small renal tumors is difficult with conventional imaging, like ultrasound, CT and PET [66]. Keshari et al. demonstrated that hyperpolarized MRS of ¹³C-pyruvate to ¹³C-lactate ratio could be used to distinguish normal renal cells from renal cell carcinoma. In addition, they used the technique to show that metastatic renal cell carcinoma had a significantly higher ¹³C-lactate/¹³C-pyruvate ratio compared to the cells derived from localized renal cell carcinoma and that the differences were mediated by the MCT4 transporter [67].

Conclusion and perspective

The use of hyperpolarized ¹³C-pyruvate for characterizing tumor metabolism has been demonstrated pre-clinically in an array of cancers. Recently the application was also shown clinically in prostate cancer. The information obtained is overlapping with that of ¹⁸F-FDG-PET and the question is therefore if there is a role for clinical routine use of ¹³C-pyruvate in cancer. We have demonstrated, for the first time, the feasibility of a new imaging concept combining hyperpolarized ¹³C-pyruvate MRSI and ¹⁸F-FDG-PET imaging [68]. A concept we have named hyperPET [69]. In a canine cancer patient with a biopsy verified sarcoma we found clearly increased ¹³C-lactate production, which also corresponded to high ¹⁸F-FDG uptake on PET. This is in agreement with the fact that glycolysis and production of lactate are increased in tumor cells compared to normal cells [68], see Figures 3 and 4. To the degree that ¹³C-pyruvate and ¹⁸F-FDG-PET leads to similar information there is no doubt that ¹⁸F-FDG is the method of choice: It is less costly, available in most hospitals, and renders whole body evaluation of metabolism. However, in some cancer types, e.g. prostate cancer ¹⁸F-FDG-PET is not useful and here ¹³C-pyruvate may represent an attractive alternative. Also in children, the concern of radiation burden could make ¹³C-pyruvate the method of choice. With the emergence of hybrid PET/MR systems it is possible to conjoin ¹⁸F-FDG-PET and ¹³C-pyruvate MRS, which may be of increased value. With regard to early response monitoring in cancer treatment, the possibility exist that ¹³C-pyruvate can detect such effect earlier and with higher sensitivity than other methods. Also this can only be answered through comparative PET and MR studies.

Figure 3.

View of HyperPET and description of the Warburg effect. Increased ¹⁸F-FDG uptake is a composite quantity of numerous hallmarks of cancer, e.g. metabolic changes, cell proliferation and hypoxia. While measured ¹³C-pyruvate to ¹³C-lactate switch probably expresses the Warburg per se. Modified from [68].

Figure 4.

Showing Hyper PET images of a canine cancer patient with a sarcoma on the back. Demonstrating the agreement of ¹⁸F-FDG and ¹³C-lactate in the tumor region. Unpublished data.