High-Resolution Magic Angle Spinning Proton Magnetic Resonance Spectroscopy in Prostate Cancer

Abstract

Prostate cancer (PCa) is the most frequently diagnosed malignancy in men worldwide, due in large part to the increased use of the annual serum prostate specific antigen (PSA) screening test for detection. PSA screening has saved lives, but it has also resulted in the overtreatment of many PCa patients because of a limited ability to accurately localize and characterize PCa lesions through imaging. High resolution magic angle spinning (HRMAS) proton magnetic resonance spectroscopy (¹H MRS) has proven to be a strong potential clinical tool for PCa diagnosis and prognosis. The HRMAS technique allows studies to obtain valuable metabolic information from ex vivo intact tissue samples and also perform histopathology on the same tissue specimens. Studies have found that the quantifications of individual metabolite levels and metabolite ratios, as well as metabolomic profiles, show strong potential to improve accuracy in PCa detection, diagnosis and monitoring. Ex vivo HRMAS has also been a valuable tool for interpreting in vivo results, including the localization of tumors, and thus has the potential to improve in vivo diagnostic tests used in the clinic. Here we primarily review publications of HRMAS ¹H MRS and its use in studying intact human prostate tissue.

Introduction

Prostate cancer (PCa) is the most frequently diagnosed malignancy in men worldwide, and the second leading cause of cancer death for men in the United States (1). Currently, most PCa cases are detected by the annual serum prostate specific antigen (PSA) screening test, the use of which rose rapidly in the early 1990s. Serum PSA testing has proven a great asset to the PCa clinic as it greatly increased the number of PCa diagnoses and shifted the diagnosed population toward earlier stages (2,3). While PSA screening has saved lives, it also created controversies for PCa management due to a limited ability to accurately localize and characterize PCa lesions through imaging.

With the confirmation of elevated PSA levels in patients, biopsies are taken from the prostate and evaluated with histopathology. However, as radiological imaging is currently unable to visualize suspicious areas for targeted biopsy in general, theses randomly conducted biopsies often result in false negatives for early-stage PCa patients as the method is ineffective in detecting small, heterogeneously distributed lesions (4). In addition, even with the presence of cancer glands in a biopsy core, current histology is often unable to characterize PCa and distinguish aggressive from indolent disease. Both of these issues present complications in the attempt to balance the benefits of therapies with the morbidities of aggressive treatments in the selection of the most appropriate treatment option for PCa patients. Possible treatments for biopsy proven PCa range from active surveillance without immediate intervention to a number of radical procedures, including prostatectomy, radiotherapy, and chemotherapy, with the procedures often resulting in serious side effects to the patient. At present, without the ability to appropriately characterize PCa, these radical procedures are often elected in cases of indolent PCa in order to ensure the few aggressive cases are not missed, and result in unnecessary overtreatments that cause adverse effects and impair quality of life for many (5). These dilemmas could be greatly minimized if new PCa biomarkers and their imaging applications could differentiate PCa from benign tissue, predict tumor stage and location, and estimate malignant potentials before the implementation of radical procedures.

High resolution magic angle spinning (HRMAS) proton magnetic resonance spectroscopy (¹H MRS), introduced to intact biological tissue analysis in 1996 (6,7), has proven to be a strong candidate to rectify these challenges seen in PCa clinic. This method applies mechanical rotation of the sample at a 54.7 degree angle to the magnetic field, and produces high resolution spectra allowing for the observation of individual metabolites, without sacrificing tissue architectures. Thus, the HRMAS method allows for subsequent histopathology analysis, the current gold standard for PCa diagnosis and monitoring, of the same tissue specimen. Since the initial application of the HRMAS method, studies have found that the quantifications of individual metabolite levels and metabolite ratios, as well as metabolomic profiles, i.e. collective evaluations of the entire measurable metabolome, show strong potential to improve accuracy in PCa detection, diagnosis and monitoring. Here, we review research studies on PCa using HRMAS ¹HMRS, as summarized in Table 1; the use of HRMAS in the studies of other cancers, including lung, breast, brain, colorectal, and cervical cancer, has been reviewed recently (8).

Table 1.

Summary of all ex vivo intact human prostate tissue studies using HRMAS MRS.

Year	Results	Samples	Ref
1998	High quality MAS spectra reveal more intense lipid signals in PCa vs. BPH	12	(9)
2001	Correlations between concentrations of spermine and citrate, and volume percentage of benign epithelia	16	(10)
2003	Slow-spinning preserves tissue pathological structures and the DANTE pulse sequence suppressed spinning sidebands	22	(41)
2003	HRMAS identified distinctive metabolic patterns and helped combined MRI/3D-MRSI accurately identify and locate PCa	54	(14)
2003	Storage conditions affect metabolite intensities for absolute concentrations but not relative intensities	12	(39)
2005	Metabolites can be quantified when spinning sidebands are suppressed with a simple minimum function, Min(A,B)	31	(42)
2005	Metabolomic profiles can differentiate PCa from benign samples, correlate with PSA, and delineate a subset of less aggressive cancer	199	(23)
2005	Rotor-synchronized adiabatic TOCSY allowed for the identification of choline- and ethanolaminecontaining metabolites	10	(34)
2006	PCa has higher concentrations of lactate and alanine, and lower concentrations of citrate and polyamines	60	(15)
2007	Storage-induced metabolites changes are not significant for HRMAS tissue analysis	15	(40)
2008	Technique can be used to quantify choline- and ethanolamine-containing metabolites	47	(16)
2008	Significant increases in lactate and alanine concentrations in PCa tissue	130	(17)
2008	Significant metabolic differences between cancerous and benign tissue, and correlations with tumor Gleason score	48	(18)
2009	It is possible to detect PUFAs in prostate tissue with HRMAS NMR	81	(21)
2009	Histopathology and genetic analysis can be performed on prostate tissue samples following HRMAS	40	(46)
2009	Enabled quantification of metabolic products of [3-13C] pyruvate in PCa cells	5	(37)
2009	ERETIC method provides improved quantification accuracy for HRMAS spectroscopy of prostate tissue	60	(36)
2010	New quantification method, HR-QUEST reliably quantified 16 metabolite and reference signals	11	(35)
2010	Metabolomic profiles can differentiate cancer from benign and locate malignancy to direct biopsy	199	(24)
2010	Metabolomic profiles can predict PCa recurrence with an accuracy of 78%	79	(25)
2010	Correlations between PSA velocity, and ODC1 and OAZ1 mRNA expression levels	18	(49)
2011	High-grade PCa shows greater cellular proliferation and higher concentrations of choline- and ethanolamine-containing metabolites	49	(19)
2011	Potential to determine if PCa may be present near benign prostate tissue biopsies	149	(20)
2011	RNA extracted after HRMAS was still intact with high integrity, allowing for comparisons with metabolomic profiles	53	(48)
2012	RNA quality was high after HRMAS and revealed mechanisms underlying low citrate and high choline levels in PCa	133	(47)
2012	Significant correlations between PSA velocity, density, and percent-free PSA, and citrate concentrations in benign epithelia	27	(13)

Evaluations of major metabolites in normal prostate and PCa

Perhaps the earliest study employing HRMAS MRS for the analysis of intact human prostate tissue was reported by Tomlins et al. in 1998, shortly after the proposal of the HRMAS MRS concept (9). Whole pieces of human prostate tissue were obtained from patients with benign prostatic hyperplasia (BPH) and PCa, and HRMAS spectra were compared with spectra acquired from solutions obtained from a conventional tissue extraction procedure (Figure 1). The results showed that the HRMAS method produced high resolution spectra able to reveal important biochemical information regarding prostate biochemistry while also preserving tissue components. In addition, they found more intense lipid signals in PCa tissue when compared with BPH tissue (9). This early investigation showed the utility of HRMAS in the study of PCa, and led to a focus on individual metabolites or metabolite ratios that could be used as specific biomarkers of PCa.

Figure 1.

A series of 500 MHz MAS and high resolution 1H NMR spectra of BPH tissue samples (A, B and C) and cancer tissue samples (D, E and F), comparing the differences between 1D presat MAS spectra (A and D), the CPMG spectra (B and E) of intact tissue samples and the corresponding normal presat spectra from an acetonitrile–H2O extraction (C and F). The specific assignments for lipid protons are shown in spectrum D. Metabolite key: (1) isoleucine (2) leucine (3) valine (4) lysine (5) glutamate (6) glutamine (7) acetate (8) alanine (9) lactate (10) threonine (11) N-acetylglycoproteins (12) arginine (13) citrate (14) aspartate (15) asparagine (16) creatine (17) choline (18) taurine (19) tyrosine (20) histidine (21) myo-inositol (22) glycine (23) fatty acid side chains of mobile triglycerides (24) phosphatidylcholine (25) hypoxanthine (26) formate (27) phenylalanine (28) a-glucose (29) guanidine diphosphate. (Reproduced from Tomlins et al. 1998, Figure 1)

In 2001, a study used HRMAS on a 9.4 Tesla (400 MHz) MR spectrometer and quantitative histopathology to evaluate tissue specimens from removed prostates of PCa patients (10). For the first time, positive linear correlations of the volume percentage of benign prostatic epithelial cells with concentrations of spermine, a polyamine abundant in the prostate and a proposed endogenous inhibitor of PCa growth (11), and citrate, a critical prostate metabolite related to zinc depletion in malignant prostate tissue (12), were reported. These results are critical for an improved understanding of the biochemical characteristics of prostate tissue. In addition, by successfully showing that metabolite concentrations could be measured accurately in intact tissue with HRMAS, and also that quantitative histopathology could follow HRMAS analysis on the same specimens, future studies began to look into the functions of these PCa metabolites. Most recently, citrate concentrations measured by HRMAS from benign epithelial glands in the peripheral zone of PCa patients were found to be correlated with PSA velocity, PSA density, and serum percent free PSA, such that low citrate concentrations in unit benign epithelial glands represent rapidly increasing PSA values, and therefore, likely fast growing cancer (Figure 2) (13). These results advance previous findings that citrate and polyamine levels are reduced in higher-grade, or more aggressive, cancers (14), by presenting the potential of using citrate to estimate PCa growth rates. These studies contribute not only to our knowledge of tumor biology, but may aid in the avoidance of PCa patient overtreatment by differentiating fast from slow-growing PCa. Thus PCa growth rates may be predicted if prostate biopsies could be evaluated for their benign epithelial citrate concentrations, either by HRMAS or another method.

Figure 2.

Relationships between levels of citrate in histo-benign epithelia PSA velocity (a), density (b), and percent-free PSA (c) respectively. The ratios of the concentrations of citrate and the volume percentages of histology quantified benign epithelia decrease with increases in PSA velocities and densities, and with decreases in the percent of blood free PSA. The points in the plot represent the average value measured for each individual case, where multiple samples were analyzed. The vertical error bars are the standard errors of these multiple measurements. DPSA, PSA densities; VPSA, PSA velocities. (Reproduced from Dittrich et al. 2012, Figure 3)

Other studies by Swanson et al. found a number of differing metabolite levels in cancerous and benign prostate tissue (14-16). Phosphocholine, taurine, myo-inositol, scylloinositol, lactate, and alanine concentrations were significantly higher in cancer compared with benign tissue, while citrate and ethanolamine concentrations were lower in PCa tissue. The use of lactate and alanine as metabolic biomarkers of PCa was further investigated in a 2008 study using HRMAS of prostate “snap-frozen” needle biopsy tissues, where very low concentrations of lactate and alanine were reported in benign prostate biopsy tissues, while their levels were increased significantly in biopsies containing cancer (Figure 3) (17). Although the use of alanine and lactate as ex vivo biomarkers of PCa has been disputed due to issues of metabolic degradation for postsurgical samples (15), biopsy tissues that were frozen within seconds of collection and analyzed using the electronic reference to access in vivo concentrations (ERETIC) method provided a more accurate representation of metabolite levels in vivo (17). van Asten et al. studied metabolite concentrations and metabolite ratios in prostate needle biopsies and found significant differences between cancerous and benign tissues, along with significant correlations between Gleason score (GS) and metabolite ratios involving choline, creatine, and citrate (18). Similar correlations with GS were reported by other studies showing high-grade prostate cancers (GS ≥ 4+3) have higher concentrations of choline- and eth-containing metabolites than lower grade cancers (GS ≤ 3+4); in addition, a number of metabolite ratios correlated with distances of tissue samples to the nearest tumor, the fraction of tumor cells present in the sample, and the amount of cell proliferation as measured by Ki-67 staining (19,20).

Figure 3.

Representative 1H HR-MAS spectra and corresponding histopathologic sections of (a) benign predominantly glandular (40% glandular, 60% stroma) and (b) prostate cancer (70% Gleason 3+3) biopsy tissues. The major metabolites, TSP, and ERETIC resonances are shown. Lactate to alanine regions from 144-ms CPMG spectra are magnified above spectra. (Reproduced from Tessem et al. 2008, Figure 1)

HRMAS has also been used to detect the presence of polyunsaturated omega-6-fatty acid species (PUFAs), known promoters of PCa, in malignant prostate tissues using two-dimensional (2D) ¹H-¹³C correlation spectra (21). These results indicate that the analysis of individual metabolites by HRMAS can provide valuable information regarding the metabolic changes in PCa that may be relevant in the clinical setting. A summary of all individual metabolite and metabolite ratio findings from studies of intact human prostate tissue using HRMAS is listed in Table 2.

Table 2.

Metabolite level changes and correlations observed in human prostate cancer using HRMAS MRS.

Metabolites	Observation	Reference
Alanine	Increase	(17); (15)
Citrate	Decrease; correlates with PSA velocity, PSA density and percent-free PSA	(15); (14); (13)
Citrate / Creatine	Decrease; correlates with GS	(18)
Choline / Creatine	Increase; correlates with GS and tumor fraction	(18); (20)
Ethanolamine (Eth)	Decrease	(16)
Glycerophosphocholine (GPC)	Increase, correlates with GS	(14); (16); (19)
Glycerophosphoethanolamine (GPEA)	Increase	(16)
GPEA / ETA	Increase	(16)
(GPC + PC)	Increase	(15)
(GPC + PC) / Creatine	Increase; correlates with tumor fraction, rate of cell proliferation, distance to nearest tumor, and GS	(18); (20)
Lactate	Increase	(17); (15)
Lactate/Alanine	Increase	(18)
Lipids	Increase	(9)
myo-Inositol	Increase	(14)
myo-Inositol / scyllo-Inositol	Correlates with GS, tumor fraction and distance to nearest tumor	(20)
Phosphocholine (PC)	Increase; correlates with GS	(14); (16); (19)
Phosphoethanolamine (PE)	Increase	(16)
PC / GPC	Increase	(16)
PC / PE	Increase	(16)
PE / Eth	Increase	(16)
Polyamines	Decrease; lower levels in more aggressive cancers	(15); (14)
Polyunsaturated omega-6 fatty acids	Accumulate in PCa tissue, related to higher GS	(21)
scyllo-Inositol	Increase	(14)
scyllo-Inositol / Creatine	Correlates with tumor fraction	(20)
Spermine	Correlates with volume percent of normal epithelia	(10)
Taurine	Increase	(14)
Total choline	Increase	(15)
Total choline / Citrate	Increase, correlates with GS	(18)

PCa Metabolomics

Studies focusing on individual metabolites in prostate tissue are valuable in advancing our understanding of PCa; however, metabolomic profiles have shown even greater promise for improving PCa diagnosis and monitoring disease progression. Metabolomics evaluates physiological or pathological conditions by profiling the entire measurable metabolome, instead of focusing only on certain metabolites or isolated metabolic pathways (22). Using PCa metabolomic profiles through different combinations of metabolite spectral intensities obtained from a single set of HRMAS data, different PCa characteristics can be probed. For instance, results show that metabolomic profiles can improve sensitivity in PCa detection, when compared with indications from single metabolites, and in PCa characterization, when compared with histopathological evaluations for predictions of PCa stage and recurrence potential. A 2005 study showed that different metabolomic profiles obtained through HRMAS from 199 samples of 82 PCa patients can differentiate cancerous and benign samples from the same patients with 98.2% accuracy, and correlate with serum PSA levels (23). Furthermore, the metabolomic profiles can delineate a subset of less aggressive tumors and predict tumor perineural invasion within the subset, all achievable using concentration based metabolite intensities measured from a single set of tissue spectra (23). Later, it was shown that metabolomic profiles obtained from relative spectral intensities (normalized by the spectral region of 0.5-4.5 ppm) of the same patient population could also differentiate cancerous from histologically benign samples with 93% accuracy (Figure 4) (24).

Figure 4.

Prostate cancer metabolomic profiles of relative metabolic intensities at 14T. To compensate for the lack of an established in vivo concentration reference standard, we reanalyzed tissue metabolomic profiles according to relative metabolite intensities (normalized by the metabolite spectral region of 0.5 to 4.5 ppm) for 42 samples from 13 patients (19). (A) The overall loading factors (combined coefficients from PCA and canonical analysis) for the 36 metabolites and regions included, which provide examples of phosphorylcholine (PCh; 3.22 ppm), spermine (Spm; 3.05 to 3.15 ppm), and creatine (Cr; 3.03 ppm), are labeled. (B) Metabolomic profile expressed as canonical score 1 distinguished cancer (solid dot) from histo-benign (open dot) samples (overall accuracy of 93%, indicated by the ROC curve, not shown here) with statistical significance (P < 0.0001). Median (M) and SD values were calculated for all samples. (Reproduced from Wu et al. 2010, Figure 2)

Metabolomic profiles calculated from HRMAS have also been used to assess PCa biochemical recurrence (BCR) potential, defined as the detection of serum PSA elevations after radical prostatectomy, for which currently there is no biological test to accurately predict the likelihood at the time of surgery. A retrospective study in 2010 analyzed clinically and pathologically stage matched groups with and without BCR as training and testing cohorts (25). By applying the metabolomic profiles obtained for differentiating groups with and without BCR for the training cohort, to the testing cohort, the results revealed the ability of the calculated metabolomic profiles of the testing cohort to predict BCR potential with an overall accuracy of 78% (25), much improved over the 50-50 prediction that can be reached in the current PCa clinic for these matched-cases. If such tests became applicable in the PCa clinic, PCa aggressiveness in terms of cancer recurrence could be determined on a personal basis and help guide the most appropriate course of treatment, for at least a subgroup of PCa patients.

Advances in the methodology

In the study of PCa, ex vivo NMR measurements have evolved from using conventional solution methods starting in the mid 1980’s, applied on either intact tissue samples (26,27) or solutions of tissue extracts (28,29), to the analysis of intact prostate tissue samples using the HRMAS method (30). Prior to the introduction of HRMAS, prostate tissue samples would either be packed in standard NMR tubes, immersed in phosphate buffered solutions, and analyzed on an NMR spectrometer with limited spectral resolution, or analyzed at high spectral resolution with solutions chemically extracted from tissue. While intact tissue studies revealed certain biochemical information, including the presence of spermine in tissue (31), and could be correlated with pathology of the same tissue samples, the spectral resolution was limited and prevented identification of many individual metabolites. On the other hand, although extract solutions could provide high spectral resolution, the extraction process destroyed tissue architecture and prevented subsequent histopathological evaluation of the same tissue. In the study of early-stage PCa, a disease known for its heterogeneously distributed, small lesions, verifying tissue pathology is critical due to the statistical eventuality that the majority of prostate tissue samples obtained from these patients would in fact be histologically benign specimens from cancerous prostates (32). Combining the strengths of these two analytic approaches, the HRMAS method offers a solution to obtain high spectral resolution without destructive tissue chemical extraction procedures, so that tissue specimens can be histologically evaluated after HRMAS analysis. In addition, HRMAS analysis of intact tissue requires less sample volume (~10 mg) and shorter preparation time than solution extraction methods, making it suitable for the analysis of both PCa biopsies and larger specimens (33).

Since the advent of HRMAS for intact prostate tissue analysis, many studies have invested in improvements of the methodology to advance its utility in PCa investigations. Much of this research has focused on producing better resolved spectra to optimize metabolite identification and quantification. Two-dimensional (2D) total correlation spectroscopy (TOCSY) experiments provide an enhanced way to identify metabolites that are difficult to resolve in 1D HRMAS spectra. The combination of rotor-synchronized adiabatic pulse sequences and 2D HRMAS TOCSY allowed for the identification of cross-peaks associated with choline- and ethanolamine-containing metabolites (34), and for their quantification (16). Studies have also used different T₁ and T₂ relaxation times to optimize the quantification of major metabolites, as well as developed new semi-parametric time-domain quantification methods (15,35). In addition, the ERETIC method, mentioned previously for the quantification of alanine and lactate, which uses a synthesized radiofrequency (RF) pulse to produce a signal, presents a possible substitute for chemical reference standards in HRMAS MRS (36). Other studies have used HRMAS ¹³CMRS to study prostate cancer cells grown in culture to further characterize PCa metabolism using ¹³C labeled substrates as metabolic probes, and have found indications that citric acid cycle metabolism is responsible for much of the consumption of pyruvate in PCa cells (37). While the use of HRMAS ³¹PMRS in the study of intact human prostate tissue has not been presented, to the best of our knowledge, it has proven to be a useful method for measuring phospholipid metabolites in tumor samples (38).

The quantification of metabolite concentrations also raises concerns regarding the state of prostate tissue before HRMAS analysis, and as to whether storage conditions, especially freezing, affect metabolite concentrations. One study found freezing to affect absolute metabolite concentrations, but not relative intensities (39), and another study found that long-term frozen storage of prostate tissue might only alter\ metabolic concentrations to a degree below HRMAS detection limits (40). A study of prostate needle biopsies found that while the echo gel used for ultrasound guidance could potentially contaminate tissue HRMAS spectra preventing accurate analysis of metabolic concentrations, it could be resolved by avoiding contact of the biopsy needle with the gel (18). With respect to this concern, our lab has established a protocol of transporting fresh, unfrozen prostate biopsy cores for HRMAS analysis on the same day of the biopsy procedure, adoptable in clinic, with which HRMAS may constitute an integrated step in routine PCa biopsy evaluation without interference with histology analysis. In this protocol, the biopsy core is briefly rinsed in deuterium oxide (D₂O) to rid it of contaminants, and then stored on ice in an apparatus specifically designed to maintain hydration of the core while preserving metabolite concentrations (Figure 5). Rinsing of the biopsy cores with D₂O to eliminate contaminants introduced during the biopsy procedure, may result in a certain degree of metabolite loss. However, by rinsing for a standardized minimal time (3~5 seconds) without soaking the core, the potential metabolite losses would be slight and would likely only appear at the superficial surface of the biopsy cores.

Figure 5.

A hydrated apparatus used for transporting fresh prostate needle biopsy cores. A small piece of Kimwipe is placed inside a 1.5 mL microcentrifuge tube and dampened with deuterium oxide (D₂O). A 0.2 mL PCR tube is then placed inside the microcentrifuge tube, resting on the Kimwipe and moisturized by D₂O vapor. During the collection of a prostate biopsy core, the core is first rinsed briefly in a 1.5 mL tube filled with the D₂O and then placed inside the PCR tube before the closure of the microcentrifuge tube; the apparatus is then placed on ice until HRMAS analysis. This design provides a moist environment for the tissue but protects it from the possible loss of metabolites from being completely immersed in liquid.

While the quantification of metabolites with HRMAS is essential, subsequent histopathology is the key to characterizing metabolic profiles according to cancer characteristics, especially given that the majority of analyzed prostate tissues are often eventually defined by histology as benign tissues from PCa patients (23). Therefore, multiple studies have investigated the best procedure for preserving tissue architecture for subsequent pathology evaluation and additional biomolecular analyses. The use of slow spinning rates with HRMAS have shown greater preservation of tissue morphological structures than faster spinning rates (Figure 6) (41). Although beneficial for subsequent pathology, slow spinning rates also present a challenge for metabolite quantification with the presence of spinning sidebands. These spinning sidebands need to be suppressed in order to prevent them from interfering with and confounding spectral regions of interest. Various rotor-synchronized spinning sideband suppression methods have been investigated, and both the rotor-synchronized DANTE pulse sequence and Min(A,B) post-spectral editing scheme have been reported as simple and efficient sideband suppression methods for accurate quantification of metabolic concentrations (Figure 7) (41,42). The ability to perform additional prostate tissue analyses after HRMAS has also been studied, including quantifying tissue pathological features with computer aided image analysis (43).

Figure 6.

The effects of HR-MAS stress on tissue morphology. a: A human prostate tissue sample taken directly from a tissue bank without HR-MAS testing. The tissue showed highly organized ductal cellular structures with well-defined epithelial layers. b: Another sample from the same patient but after an HR-MAS experiment, which involved spinning at 600 Hz for 45 min and then 700 Hz for 15 min. The tissue still exhibits a normal ductal structure that cannot be differentiated from the original sample (a). c: A sample, again from the same clinical case, after MAS analysis at higher spinning frequency, 3.0 kHz, for 1 hr. The tissue ductal structures are visibly distorted compared with the natural specimens. (Images presented at the same magnification) (Reproduced from Taylor et al. 2003, Figure 2, color version of figure is available online).

Figure 7.

Human prostate CW water presaturated spectra at spinning rates of (a) 600 Hz (A), and (b) 700 Hz (B). c: A spectrum that was edited, using A and B, with Min(A, B) to be visually compared with d, a spectrum obtained at a spinning rate of 3.0 kHz, plotted with a different vertical scale in order to produce e, a digital analysis presents the differences between spectra d and c. * SSBs from tissue water and rubber standard signals. (Reproduced from Burns et al. 2005, Figure 6)

While most PCa investigations with HRMAS have used human prostate tissues, HRMAS analyses of PCa animal models have also been reported. Specifically, two recent studies have used the HRMAS method as a way to determine the characteristics of mouse and rat PCa models, including their differences and common features with human PCa. In 2008, Teichert et al. compared tissue from the mouse model, Transgenic Adenocarcinoma of the Mouse Prostate (TRAMP), with wild-type mice (44). They found that while citrate trends were similar to those observed in human PCa, changes in choline species did not agree with human PCa as the levels were decreased in TRAMP tumors (44). A more recent study by Stenman et al. found significant metabolic differences between human and rat PCa, where many of the key metabolic markers for human PCa, such as citrate, aspartate, lysine, taurine, glutamate, glutamine, creatine, inositol, and choline compounds, did not translate to the rat model (45). These studies presented HRMAS as a promising method to help improve future interpretations in PCa studies using animal models.

Future directions and in vivo translations

An exciting area of intact prostate tissue study is the combination of genetic analysis with HRMAS evaluation. In an early study, RNA was extracted from tissues after their analysis with HRMAS (46). While the RNA integrity was significantly lower with biopsy samples after HRMAS analyses, as compared to control biopsy samples without HRMAS measurements, the integrity for HRMAS analyzed prostatectomy samples was not significantly different from the control samples, and in fact, the mean integrity values for the control biopsy samples were very similar to all surgical samples (with and without HRMAS). In addition, there were no significant differences in histopathology integrity for surgical and biopsy samples versus control samples (46). Not only did this show that histopathology and genetic analysis could be successfully performed on prostate tissue after HRMAS analysis, but it contributed an expanded understanding of the genetic mechanisms underlying the different concentrations of certain metabolites in cancerous and benign prostate tissue, specifically lower levels of citrate in aggressive PCa cases (47). Furthermore, a standardized method for fresh tissue harvesting from radical prostatectomy specimens has been developed to ensure that the collected fresh frozen tissue would be suitable for HRMAS analysis and subsequent gene expression profiling (48).

Additional investigations include evaluations of mRNA expression levels of rate-limiting enzymes in the spermine metabolic pathway following the reported function of spermine as an endogenous inhibitor of PCa growth (11). In contrast with the measured concentrations of spermine from HRMAS spectra, expression levels of the spermine anabolic enzymes ornithine decarboxylase (ODC1) and S-adenosylmethionine decarboxylase (AMD1) were reduced in the benign epithelia surrounding cancer glands with increases of PSA velocity, and expression levels of ornithine antizyme increased with increased PSA velocity (49). If PSA velocity is used as a measure for fast growing PCa, and therefore aggressive PCa, then the evaluation of spermine levels with HRMAS and the expression levels of these enzymes could be used as clinical measurements to determine cancer aggressiveness for newly-diagnosed PCa patients.

Possibly the most important research focus for the study of PCa with HRMAS concerns the great potential for translation of PCa metabolomics to in vivo MR imaging and spectroscopy, to aid in the development of non-invasive, clinical protocols for detecting, monitoring and diagnosing PCa patients. A study by Swanson et al. found that in vivo MRSI reveals typical HRMAS patterns(14). The higher resolution achieved with HRMAS allows for the elucidation of spectral patterns associated with different prostate tissue types and cancer grades that may significantly improve the interpretation of in vivo MRSI data (14). Addressing the limitation of current clinical radiology in localizing prostate cancer lesions in vivo, the study also showed the ability of combined MRI/3D-MRSI to guide the selection of PCa tissue for ex vivo HRMAS analysis with a 71% accuracy (Figure 8) (14). Furthermore, to demonstrate the improved ability of PCa metabolomics in differentiating PCa from histologically benign tissues, a study published in 2010 revealed a 93-97% overall accuracy for detecting the presence of cancer using PCa metabolomic profiles, when whole prostates removed by prostatectomy from biopsy-proven PCa patients were evaluated on a 7 Tesla human whole body MR scanner (24). Localized, multi cross sectional, multi voxel spectra were used to generate PCa metabolomic imaging according to intact tissue HRMAS MRS analyses, which was then evaluated against whole-mount prostate histology (24). Therefore, by using HRMAS analysis in tandem with in vivo imaging modalities, we can reveal spectral patterns that can improve clinical interpretations of in vivo MRI/MRSI, and more importantly work towards the development of clinically implementable PCa metabolomic imaging.

Figure 8.

a: T2-weighted MRI image from the prostatic apex of a 56-year-old prostate cancer patient. b: 3D-MRSI spectrum taken from the 0.24 cm3 voxel shown in a. c: 1H HR-MAS spectrum of excised tissue section showing high levels of choline and GPC_PC, and low levels of citrate and polyamines relative to creatine. d: H&E stain of a tissue sample indicating the presence of Gleason 3_4 prostate cancer. (Note: the acetone impurity was introduced during tissue inking.) (Reproduced from Swanson et al. 2003, Figure 2)

Presently, the major concern in the PCa clinic is represented by the urgent need for non-invasive imaging of cancer foci to perform targeted biopsies and to predict PCa aggressiveness prior to prostatectomy. The studies of PCa individual metabolite concentrations and metabolomics with HRMAS have shown reliable potential to remedy the absence of these in vivo diagnostic tools (14). While these analyses are not meant to replace histopathology evaluations, they have the potential to provide additional biological information that may be able to better detect, characterize, and sub-categorize cases according to tumor-biochemical potentials, currently unachievable with existing PCa clinical and pathological criteria. While the utility of HRMAS MRS remains limited to studies of ex vivo prostate tissue samples, metabolic markers thus obtained can ultimately assist in the translation of these findings into imaging paradigms in order to provide a non-invasive technique for the diagnosis and monitoring of PCa.

Abbreviations

AMD1

S-adenosylmethionine decarboxylase

BPH

benign prostatic hyperplasia

DANTE

Delay Alternating with Nutation for Tailored Excitation

ERETIC

electronic reference to access in vivo concentrations

GS

Gleason score

HRMAS

high resolution magic angle spinning

ODC1

ornithine decarboxylase

PUFA

polyunsaturated omega-6 fatty acid

PCa

prostate cancer

PSA

prostate specific antigen

TOCSY

Total Correlation Spectroscopy

TRAMP

Transgenic Adenocarcinoma of the Mouse Prostate

1H

proton

13C

carbon-13