Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Magn Reson Med. Author manuscript; available in PMC 2009 February 17.
Published in final edited form as:
PMCID: PMC2643975

Quantification of Choline- and Ethanolamine-Containing Metabolites in Human Prostate Tissues Using 1H HR-MAS Total Correlation Spectroscopy


A fast and quantitative 2D high-resolution magic angle spinning (HR-MAS) total correlation spectroscopy (TOCSY) experiment was developed to resolve and quantify the choline- and ethanolamine-containing metabolites in human prostate tissues in ≈1 hr prior to pathologic analysis. At a 40-ms mixing time, magnetization transfer efficiency constants were empirically determined in solution and used to calculate metabolite concentrations in tissue. Phosphocholine (PC) was observed in 11/15 (73%) cancer tissues but only 6/32 (19%) benign tissues. PC was significantly higher (0.39 ± 0.40 mmol/kg vs. 0.02 ± 0.07 mmol/kg, z = 3.5), while ethanolamine (Eth) was significantly lower in cancer versus benign prostate tissues (1.0 ± 0.8 mmol/kg vs. 2.3 ± 1.9 mmol/kg, z = 3.3). Glycerophosphocholine (GPC) (0.57± 0.87 mmol/kg vs. 0.29± 0.26 mmol/kg, z = 1.2), phosphoethanolamine (PE) (4.4± 2.2 mmol/kg vs. 3.4± 2.6 mmol/kg, z = 1.4), and glycerophosphoethanolamine (GPE) (0.54± 0.82 mmol/kg vs. 0.15± 0.15 mmol/kg, z = 1.8) were higher in cancer versus benign prostate tissues. The ratios of PC/GPC (3.5± 4.5 vs. 0.32± 1.4, z = 2.6), PC/PE (0.08± 0.08 vs. 0.01± 0.03, z = 3.5), PE/Eth (16± 22 vs. 2.2± 2.0, z = 2.4), and GPE/Eth (0.41± 0.51 vs. 0.06± 0.06, z = 2.6) were also significantly higher in cancer versus benign tissues. All samples were pathologically interpretable following HR-MAS analysis; however, degradation experiments showed that PC, GPC, PE, and GPE decreased 7.7± 2.2%, while Cho+mI and Eth increased 18% in 1 hr at 1°C and a 2250 Hz spin rate.

Keywords: phospholipid metabolism, phosphocholine, rotor synchronization, adiabatic mixing scheme

The choline- (Cho) and ethanolamine (Eth)-containing metabolites are major cytosolic precursors and degradation products of phospholipid membrane assembly and catabolism (1). There is currently much interest in correlating the individual Cho- and Eth-containing metabolites with cancer presence, aggressiveness, and therapeutic response because of the information they contain about cellular proliferation, apoptosis, and the activity of key enzymes (2-7). Historically, 31P spectroscopy has been used both in vitro (8-12) and in vivo (13-15) to resolve the phosphomonoesters (phosphocholine [PC] and phosphoethanolamine [PE]) and phosphodiesters (glycerophosphocholine [GPC] and glycerophosphoethanolamine [GPE]). 1H spectroscopy has mainly focused on detection of the choline head group singlets of free Cho, PC, and GPC (16-18), which give rise to a composite choline peak at ≈3.2 ppm in vivo (19). Although Cho, PC, and GPC can be resolved from each other in extracts, they cannot typically be resolved from each other in intact prostate tissue at 11.7T, and free ethanolamine (Eth), PE, GPE, glucose, taurine, and myo-inositol (mI) also co-resonate in this highly overlapping region (17).

Total correlation spectroscopy (TOCSY) (20) is a powerful 2D NMR experiment which has previously been used in HR-MAS studies of cultured melanoma cells (21) and intact prostate (22), liver (23,24), brain (25), and kidney (26) tissues. In TOCSY, an isotropic mixing pulse is used to transfer magnetization between protons within a spin system such that, depending on the duration of the mixing time, correlations can be observed between protons that are several bonds apart. One advantage of TOCSY over conventional correlation spectroscopy (COSY) is that the cross-peak multiplets have an absorption phase rather than dispersive phase character. Consequently, TOCSY can be used quantitatively provided the magnetization transfer (MT) efficiency is known for the specific metabolites under investigation at the mixing time being used (27). Recent studies have demonstrated that TOCSY can be used to resolve and quantify the sidechain CH2-CH2 cross-peaks of several Cho- and Eth-containing metabolites, including PC, GPC, Eth, PE, and GPE (21,22). For the Cho and Eth metabolites, the sidechain methylene protons give rise to a single cross-peak on each side of the diagonal, which simplifies their quantification compared to more complicated spin systems (e.g., myo-inositol).

Previous studies have also demonstrated that rotor synchronization, in which the duration of the isotropic mixing pulses are matched to the reciprocal of the spin rate, is crucial for TOCSY experiments under HR-MAS conditions (22,28). Because of their insensitivity to RF (B1) imperfections, adiabatic mixing pulses (e.g., WURST) have also been shown to improve the signal-to-noise ratio (S/N) of TOCSY cross-peaks compared to conventional hard pulse mixing schemes (e.g., MLEV-17) (22,29). In tissue studies the time typically allowed for 2D frequency encoding (≥3 to 5 hr) can severely degrade the metabolic and pathologic integrity of the tissue and compromise subsequent genetic and/or proteomic analyses (17). In this study, a rotor synchronized adiabatic TOCSY experiment was optimized for the quantitative detection of Cho and Eth metabolites in ≈1 hr and then applied to benign and malignant postsurgical prostate tissues prior to pathologic analysis of the same tissues.


Phantom Studies

Choline chloride (Cho), phosphocholine chloride calcium salt tetrahydrate (PC), L-α-glycerophosphorylcholine cadmium chloride complex (GPC), ethanolamine hydrochloride (Eth), O-phosphorylethanolamine (PE), deuterium oxide (D2O), and 1.0 mL vials of deuterium oxide (99.9% atom D) containing 0.75% 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid (D2O+TSP) were purchased from Sigma-Aldrich (St. Louis, MO). Glycerophosphoethanolamine hydrate (GPE) was a kind gift from Euticals S.p.A. (Lodi, Italy). Stock solutions of Cho, PC, GPC, Eth, PE, and GPE were prepared in 1× phosphate-buffered saline and used to optimize the data acquisition parameters and determine MT efficiency constants (KMT) in order to calculate metabolite concentrations in tissue. Samples were analyzed individually and as a mixture of all six metabolites. For solution studies, 3.0 μL of D2O+TSP was pipetted into the bottom of a tared rotor and weighed to 0.01 mg. The balance was tared again and then the solution samples were pipetted into the rotor and weighed.

Tissue Samples

Forty-seven postsurgical prostate tissues were harvested from 22 patients (mean age: 60 ± 13 years), placed in cryogenic vials, and stored at -80°C until use. Samples were analyzed using custom-designed 20 and 30 μL leak-proof zirconium rotors, which contained an oblate spheroid-shaped sample chamber to facilitate shimming and an airtight screw top plug to prevent leakage (17). 3.0 μL of D2O+TSP was pipetted into the bottom of a tared rotor and weighed to 0.01 mg. Tissue cores were then excised from the surgical specimens using a 3.5 mm biopsy punch (Sklar Instruments, West Chester, PA). The balance was tared again and the tissues were weighed to 0.01 mg (mean = 17.36 ± 3.71 mg), then slowly inserted into the rotor. Finally, the top screw was carefully inserted to avoid air bubbles, tightened, and then the rotor was placed in the spectrometer for analysis.

HR-MAS Spectroscopy

Data were acquired at 11.7T (500 MHz for 1H), 1°C, and a 2250 Hz spin rate using a Varian INOVA spectrometer equipped with a 4-mm gHX nanoprobe (Varian, Palo Alto, CA). Prior to beginning each 2D experiment a quantitative 1D “presat” spectrum was acquired with a 2-sec presaturation delay, 2-sec acquisition time (TR = 4 sec for tissue, 6 sec for solutions), and 128 transients (17). TOCSY data were acquired using a rotor synchronized WURST-8 adiabatic mixing scheme flanked by two sets of ≈1.8 ms trim pulses (22). Both the adiabatic and trim pulses were rotor synchronized by setting the pulse lengths equal to the reciprocal of the spin rate (0.444 ms). Solution data were acquired with a repetition time (TR) = 2.04 sec (presaturation delay = 1.5 sec, mixing time [tm] = 40 ms [39.96 ms exactly], acquisition time [AT] = 0.50 sec); steady-state pulses (SS) = 4 (first increment only); number of transients per increment (NT) = 24; number of points (NP) = 4096 (complex) and spectral width = 20,000 Hz in the direct dimension (F2); number of increments (Ni) = 64 (complex) and spectral width = 6000 Hz in the indirect dimension (F1); total time = 1.77 hr.

The experiment time was then optimized for tissue studies by determining the minimum TR and NT needed to achieve adequate S/N, and the minimum Ni needed to resolve the metabolites of interest. For TOCSY, the total experiment time = TR × NT × Ni, where TR is the sum of the combined presaturation and relaxation delay, the average evolution time, tm, and AT. The TR was minimized by determining the power level and duration of presaturation needed to adequately suppress the water signal to the level of the other metabolites, the evolution time was fixed, the optimal tm was previously determined to be 40 ms (39.96 ms exactly) (22), and AT was minimized based on the loss of signal in the free induction decay (FID). For a sample mass ≥10 mg, the following parameters provided optimal water presaturation, S/N, and resolution in a total experiment time of 1.08 hr: TR = 1.24 sec (presaturation delay = 1.0 sec, tm = 40 ms, AT = 0.20 sec); SS = 4 (first increment only); NT = 24; NP = 4096 (complex) and spectral width = 20,000 Hz in F2; Ni = 64 (complex) and spectral width = 6000 Hz in F1. Finally, data were acquired in interleaved mode and phase sensitive using the States et al. method (30). For the 47 tissue samples the total time for sample preparation, shimming, and 1D and 2D data acquisition was <1.5 hr per sample. To investigate the degradation of Cho- and Eth-containing metabolites over time, serial TOCSY experiments were performed on three additional samples for 8 hr at 1°C using the same parameters as for the ≈1-hr experiment.

Data Analysis

Data were processed offline using ACD/Labs 1D and 2D NMR processor v. 9 (ACD/Labs, Toronto). One-dimensional spectra were processed and quantified using Lorentzian-Gaussian peak fitting as previously described (17). TOCSY data were processed using 4×N linear prediction in F1, zerofilled to 4K (F2) and 1K (F1) complex points, apodized using a Gaussian function and Fourier transformed in both dimensions. TOCSY spectra were then phased by displaying slices containing cross-peaks near the beginning, middle, and end of the spectrum in F2 and then F1. Zero- and first-order phase corrections were applied in each dimension and then to the entire 2D matrix. Finally, the spectra were frequency-referenced to TSP at 0.0 ppm.

The cross-peaks above the diagonal were used for quantitation because they contained less environmental or “t1” noise than the cross-peaks below the diagonal (21). Specifically, artifacts arising from the Cho, PC, and GPC singlet diagonal peaks (3.20-3.24 ppm) overlap the Eth cross-peaks below the diagonal, but have no effect on the cross-peaks above the diagonal. Metabolic changes were evaluated by comparing ratios of metabolites to one another and by estimating metabolite concentrations using the diagonal peak of TSP as an internal quantitation reference. Metabolite cross-peaks and the TSP diagonal peak were volume integrated by setting a threshold at 0.05% of each cross-peak's maximum intensity, below which the data were considered to be in the noise. The cross-peak volumes above the threshold were then numerically integrated and exported to an Excel spreadsheet. Metabolite cross-peak to TSP diagonal peak ratios were then calculated and corrected for MT efficiency based on the phantom data as follows. In the 1D spectra of the phantom solutions (n = 3), the downfield metabolite multiplets of Cho (4.07 ppm), PC (4.21 ppm), GPC (4.33 ppm), Eth (3.81 ppm), PE (4.02 ppm), GPE (4.12 ppm), and the TSP singlet (0.00 ppm) were quantified as previously described (17). Empirical MT efficiency constants (KMT) were then calculated by equating the 2D metabolite cross-peak volume (Vmet) to TSP diagonal peak volume (VTSP) ratios to the corresponding 1D metabolite peak area (Amet) to TSP (ATSP) peak area ratios:


KMT was calculated for each metabolite and averaged across the three phantom datasets. Note that normalization factors for the respective numbers of protons are not needed since they would appear on both sides of the equation and thus cancel out. For tissue spectra, metabolite concentrations (mmol/kg wet weight) were then calculated based on the masses of tissue and D2O+TSP in the rotor, and the TSP and metabolite cross-peak volumes using the following equation:

(VmetVTSP)×KMT×Mass D2O+TSP×0.75%(mg)Molecular weight TSP(172.23gmol)×1000gkgMass tissue(mg)=Metabolite Concentration(molkg)

where 0.75% was treated as a constant.

Pathologic Analysis

Following HR-MAS, samples were immediately frozen in optimal cutting temperature (OCT) compound and were then sectioned and stained with hematoxylin and eosin (H&E) as previously described (17). Slides were reviewed by two pathologists who determined the percentage of benign glandular, benign stroma, inflammation, and the percentage and grade of prostate cancer. Pathology readings were then recorded into a database and averaged for both readers. For statistical comparisons, prostate cancer tissues were compared to benign tissues, regardless of relative glandular and stromal composition.

Statistical Analysis

Statistical analyses were performed using SAS statistical analysis software (SAS Institute, Cary, NC). Metabolite concentrations (mmol/kg) were statistically compared between benign and malignant prostate tissues using Z statistics for paired data:


where xh and xc are the mean metabolite values for benign and cancer tissues, σh and σc are the standard deviations of benign and cancer metabolite concentrations, respectively, ρhc is the correlation between the paired values, and nh, nc, and nhc represent the numbers of benign, cancer, and paired samples, respectively. The Z score of greater than 1.96 was considered significant to 5% and 2.58 was considered significant to 1%.


Figure 1 illustrates the problem of spectral overlap in one-dimensional 1H HR-MAS spectra of prostate tissue. Figure 1a shows a 1H HR-MAS spectrum of a phantom solution containing 1.8 mM Cho, PC, GPC, Eth, PE, and GPE, which has been scaled up to emphasize the sidechain CH2-CH2 resonances. Figure 1b shows a 1H HR-MAS spectrum of benign predominantly glandular prostate tissue, and Fig. 1c shows a 1H HR-MAS spectrum of Gleason 3+4 prostate cancer tissue. As shown in the phantom spectrum, the Cho and Eth sidechain methylene resonances are well separated in frequency such that they can be resolved by 2D spectroscopy. Also note that GPC and GPE contain additional resonances due to the glyceryl (CH2-CH-CH2) protons in the regions from 3.59-3.71 and 3.85-3.98 ppm. In tissue (Fig. 1b,c), the multiple resonances of lactate, myo-inositol, taurine, and polyamines dominate those of Cho (4.07 and 3.55 ppm), GPE (4.12 and 3.30 ppm), and Eth (3.15 ppm). In the cancer spectrum shown in Fig. 1c, there is a visually apparent increase in the Cho (3.21 ppm), PC (3.23 ppm), and GPC (3.24 ppm) head group singlets; however, these metabolites cannot be resolved from each other in tissue and PE (3.22 ppm) co-resonates in this highly overlapping region. Furthermore, the remaining Cho and Eth methylene signals (e.g., GPC, PC, PE, and Eth) are difficult to distinguish from the baseline due to low signal intensity and spectral broadness (Fig. 1b,c).

FIG. 1
One-dimensional 1H HR-MAS spectra of (a) a phantom solution containing 1.8 mM Cho, PC, GPC, Eth, PE, and GPE, (b) benign predominantly glandular prostate tissue, and (c) Gleason 3+4 prostate cancer tissue. The sidechain methylene signals which overlap ...

Figure 2 shows the upper diagonal CH2-CH2 region of the 2D TOCSY spectra obtained on the same (Fig. 2a) phantom solution, (Fig. 2b) benign predominantly glandular, and (Fig. 2c) Gleason 3+4 prostate cancer tissue samples shown in Fig. 1. The Cho, PC, GPC, Eth, PE, and GPE cross-peaks are identified and their relative concentrations are reflected by the number of contour levels as in a topographic map. The corresponding 1D spectral regions are shown above and to the left of the 2D plots and representative histopathologic sections (H&E, 10× magnification) are shown in insets in Fig. 2b,c. Based on the phantom data acquired at tm = 40 ms, the MT efficiency constants (KMT) for each metabolite were: Cho (71 ± 1%), PC (54 ± 1%), GPC (46 ± 1%), Eth (79 ± 0%), PE (61 ± 1%), and GPE (50 ± 0%). Consistent with previously reported chemical shift values (21), the Cho and Eth cross-peaks were observed in tissue at: Cho (3.55×4.07 ppm); PC (3.62×4.17 ppm); GPC (3.71×4.33 ppm); Eth (3.14×3.82 ppm); PE (3.23×3.99 ppm); and GPE (3.30×4.12 ppm); however, note that the frequencies observed in tissue differ slightly from those observed in solution. The Cho cross-peak co-resonates with myo-inositol and is therefore labeled as Cho+mI in tissue. On visual inspection, PC was observed in 11/15 (73%) cancer tissues but only 6/32 (19%) benign tissues. Ethanolamine and PE were observed in all benign tissues and 14/15 (93%) cancer tissues, but the spectra in which Eth and PE were not observed were from different patients. GPE was observed in 28/32 (87%) benign tissues and 12/15 (80%) cancer tissues. GPC was observed in all cancer tissues and 30/32 (94%) benign tissues. In two benign tissues from different patients, PC, GPC, and GPE were not observed.

FIG. 2
Upper diagonal CH2-CH2 sidechain region of the two-dimensional TOCSY spectra acquired from (a) the phantom solution, (b) benign predominantly glandular, and (c) Gleason 3+4 prostate cancer tissue shown in Figure 1. Representative histopathologic sections ...

Table 1A summarizes the individual Cho and Eth metabolite concentrations observed in benign and malignant prostate tissues. The PC concentration was significantly higher in malignant versus benign prostate tissues, while the Eth concentration was significantly lower in malignant versus benign prostate tissues. GPC, PE, and GPE concentrations were all higher in cancer versus benign prostate tissues, but the differences were not statistically significant. Note that PE and Eth had the highest concentrations of all metabolites, GPC and GPE had similar concentrations, and PC concentrations were the lowest in both benign and cancer tissues. Several metabolite ratios were also calculated and are summarized in Table 1B. Ratios were calculated for each spectrum and then averaged across all cases; data were excluded when the value in the denominator was zero. PC/GPC, PC/PE, PE/Eth, and GPE/Eth ratios were all significantly higher in malignant versus benign prostate tissues. Figure 3 shows the overlap of Cho and Eth metabolite ratios in benign versus prostate cancer tissues and dashed lines indicate the metabolite ratios above which all samples were malignant, i.e., specificity = 100%. Positive cutoff ratios, sensitivities, and specificities that yielded minimum false positives and false negatives and represented the best compromise between sensitivity and specificity were: PC/GPC >0, sensitivity = 73%, specificity = 81%; PC/PE >0, sensitivity = 73%, specificity = 81%; PE/Eth >4.0, sensitivity = 50%, specificity = 77%; and GPE/Eth >0.1, sensitivity = 64%, specificity = 81%.

FIG. 3
Overlap of choline and ethanolamine metabolite ratios in benign versus prostate cancer tissues. Dashed lines indicate the metabolite ratio above which all samples were malignant.
Table 1
Choline and Ethanolamine Metabolite Concentrations (A) and Ratios (B) for Benign and Malignant Prostate Tissues

At pathologic evaluation, 32 samples were benign and 15 contained prostate cancer (Gleason score: 3+3 [n = 9]; 3+4 [n = 3]; 4+4 [n = 1]; and 4+5 [n = 2]) with a surface area of at least 20%. Benign samples contained an average of 40 ± 16% glandular tissue (range: 10-70%) and 60 ± 16% stroma (range: 30-90%). Malignant samples contained an average of 53 ± 26% cancer (range: 20-100%), 16 ± 13% glandular tissue (range: 0-50%), and 31 ± 21% stroma (range: 0-65%). In every case, both pathologists could interpret the slides, indicating that the duration of time in the spectrometer (<1.5 hr) did not compromise the pathologic integrity of the samples. Figure 4 shows the mean percent change in Cho and Eth metabolites over 8 hr at 1°C and a 2250 Hz spin rate. For this experiment, samples were chosen that contained high levels of PC+GPC in the 1D spectrum. During the course of a 1-hr TOCSY experiment, PC and GPC decreased by an average of 8.9% and 8.6%, respectively, while Cho+mI increased by 12%; PE and GPE decreased by 8.9% and 4.5%, respectively, while Eth increased by 24%. However, over the course of 8 hr, PC, GPC, PE, and GPE levels decreased even more dramatically, while the combined Cho+mI and Eth cross-peaks increased by nearly 150 and 300%, respectively. This indicates that HR-MAS spectroscopy data should be acquired as quickly and efficiently as possible to preserve the integrity of the tissue for subsequent analysis.

FIG. 4
Average percent change in choline and ethanolamine metabolites in prostate tissue over 8 hr at 1°C and 2250 spin rate. Cross-peak volumes were normalized to TSP.


The purpose of this study was to develop a fast and quantitative 2D HR-MAS TOCSY experiment to resolve and quantify the individual Cho- and Eth-containing metabolites in postsurgical prostate tissues prior to pathologic analysis of the same tissues. A nominal ≈ 1-hr acquisition time was chosen to preserve the metabolic and pathologic integrity of the tissue and to make the experiment feasible for routine use in a high-throughput environment. To achieve this goal, a rotor synchronized WURST-8 adiabatic mixing scheme was optimized to obtain the necessary S/N and resolution in the shortest possible time and this combination was previously shown to provide up to 10 times greater S/N in tissue compared to conventional isotropic mixing pulses (MLEV-17) (22). Magnetization transfer efficiency constants (KMT) were determined empirically using phantom solutions containing Cho, PC, GPC, Eth, PE, and GPE and then used to calculate the corresponding metabolite concentrations in prostate tissue based on the cross-peak to TSP diagonal peak volumes, and the masses of tissue and TSP in the rotor. At a 40 ms mixing time the calculated magnetization transfer efficiencies for each metabolite decreased from Cho (71 ± 1%) to PC (54 ± 1%) to GPC (46 ± 1%), and from Eth (79 ± 0%) to PE (61 ± 1%) to GPE (50 ± 0%). These data indicate that MT has a substantial influence on the cross-peak volumes of different metabolites, and this information should be taken into account when quantifying the data.

Consistent with previous in vivo (14) 31P MRS studies of prostate cancer patients, and in vitro (8,10) 31P MRS studies of prostate cancer tissue and xenograft extracts, higher concentrations of phosphomonoesters (PC and PE) were observed in prostate cancer versus benign prostate tissues. Note, however, that the increase in PE was not significant in prostate cancer samples in this study. The current results are also consistent with previous in vivo 1H MRSI (31) and ex vivo HR-MAS spectroscopy (16,17) studies that demonstrated an overall increase in Cho-containing metabolites in prostate cancer versus benign prostate tissues, and suggest that the Eth-containing metabolites, particularly PE, are major contributors to the in vivo “choline” region. Based on the data in Table 1, and assuming that free Cho concentrations are similar to PC and GPC concentrations in prostate cancer tissue, the total area of Eth-containing metabolites could be almost as much as the total area of Cho-containing metabolites, after correcting for the respective number of protons (2 vs. 9). However, it must be noted that other metabolites including taurine, myo-inositol, glucose, creatine, and polyamines also contribute to the in vivo Cho region, and that the intensity of the Eth metabolites observed in vivo is reduced due to J-modulation and T2 effects. TOCSY could potentially be used to resolve the Cho and Eth metabolites in vivo, and a localized 1D TOCSY sequence has recently been applied to the brain (32).

The dramatic increase in PC concentration in prostate cancer observed in this study is also consistent with many previous studies involving prostate (18), breast (5,33), and other cancers (1,2,13). Specifically, PC was observed in 11/15 (73%) cancer tissues but only 6/32 (19%) benign tissues and PC increased from 0.02 ± 0.07 to 0.39 ± 0.40 mmol/kg or ≈19 fold between benign and cancer tissues. GPC concentrations also increased between benign and cancer tissues; however, the difference was not as dramatic (≈2-fold) as that observed for PC and did not reach statistical significance. Nonetheless, the PC/GPC ratio was over 10 times greater in cancer than benign prostate tissues. These findings are in agreement with those of Ackerstaff et al. (18), who reported that both PC and GPC are elevated in human prostatic epithelial cells after malignant transformation. PC/PE, PE/Eth, and GPE/Eth ratios were also significantly higher in prostate cancer versus benign tissues. As shown in Fig. 3, there was considerable overlap of individual metabolite ratios; however, good sensitivity and specificity values were obtained with a minimum number of false-positives and false-negatives. Ongoing studies involve determining whether there are grade-dependent changes in the concentrations and metabolite ratios of phospholipid intermediates in a larger patient cohort.

One limitation of the current study is the unknown state of metabolic degradation encountered when working with postsurgical tissues and that caused by the freeze-thaw process. To solve this problem, studies are under way using biopsy tissues, which can be harvested and frozen in seconds and therefore provide a much better metabolic snapshot of in vivo metabolism than surgical tissues. However, the degradation experiments performed in this study indicate that, even at 1°C, PC, GPC, PE, and GPE decompose to form free Cho and Eth, and over several hours, breakdown of phospholipid membranes contributes to the continuing increase in Cho and Eth metabolites. These observations are consistent with a previous 1D HR-MAS study from our group, which showed a temperature-dependent decrease in the combined GPC+PC peak and a concomitant increase in the free choline peak over 12 hr. While the mechanisms for the degradation of the phosphorylated Cho and Eth derivatives into their free bases are not completely understood, several Kennedy cycle enzymes including phospholipases could be biologically active at 1°C or upregulated under ischemic conditions. Additionally, sheer mechanical degradation could occur due to the fast spinning rates used for HR-MAS. Therefore, in order to preserve the integrity of the tissue, the HR-MAS data should be acquired at 1°C and as fast and efficiently as possible. While slower spinning 1D HR-MAS spectroscopy has been shown to preserve prostate tissue histology (34), no slower spinning 2D HR-MAS TOCSY experiments have been reported to date. The development of a quantitative slower spinning TOCSY experiment would be technically challenging but is worthy of future investigation. Under the conditions used in the current study we have shown that the Cho and Eth metabolites can be resolved and quantified in a reasonable amount of time and that the same tissues can be interpreted pathologically following HR-MAS analysis.

A potential source of error in this study arises from the choice of TR. Although the T1s of the Cho and Eth sidechain protons cannot be determined in tissue due to spectral overlap and low signal intensity, they are assumed to be similar to each other and to that of taurine, which has a T1 of ≈600 ms at 1°C and 11.7T (17). As in 1D NMR, a TR ≈1.25 × T1 provides the maximum S/N during signal averaging in 2D NMR and is quantitative when the T1s of the metabolites of interest are similar. Therefore, the repetition time of 1.24 sec used in this study is a reasonable compromise between maximizing S/N and the time needed to presaturate the water signal. Another potential source of error arises because a solution was used to calibrate the magnetization transfer efficiency. For small molecules in solution, T1T2, and the primary difference is due to magnetic field inhomogeneity (T2*). As the correlation time (τc) increases, T1 and T2 both decrease and then diverge from each other. Excluding citrate and polyamines, the T1 relaxation times of prostate metabolites are ≈1.5 to 3 times T2 at 11.7 T and 1°C (17). Since 1/T1 < 1/T ) 1/T2, concentration errors due to differences in T2 between solution and tissue samples are estimated to be 10% or less at a mixing time of 40 ms.

There are other approaches that could be used to resolve and quantify the Cho and Eth metabolites in tissue. Conventional COSY can be applied quantitatively and the magnitude mode cross-peaks are linearly proportional to concentration. However, TOCSY provides better sensitivity than COSY because of the net magnetization transfer and lengthening of T provided by isotropic mixing and because COSY cross-peak multiplets have both absorptive and dispersive components, which partially cancel each other. Therefore, COSY would not be practical in a 1-hr setting (35). Payne et al. (36) recently demonstrated the feasibility of 31P HR-MAS spectroscopy of intact tissues. Although 31P MRS provides complementary and additional information to 1H MRS, 31P sensitivity is critically dependent on coil design and, therefore, a dedicated probe designed for 31P detection may be necessary. Additionally, 31P requires faster spin rates than 1H under HR-MAS conditions to overcome broadening due to increased chemical shift anisotropy. Therefore, 31P may be impractical for delicate samples such as biopsies, which typically weigh less than 10 mg and are ≈1 mm thick. Loening et al. (37) recently described a 31P edited 1H MRS sequence to quantify the concentrations of Cho, PC, and GPC in biological samples and demonstrated its utility in human brain tumor extracts; however, this technique has not yet been applied in tissue.


A fast and quantitative 2D high-resolution magic angle spinning (HR-MAS) total correlation spectroscopy (TOCSY) technique was developed that can be used to resolve and quantify the choline- and ethanolamine-containing metabolites in human prostate tissues in ≈1 hr prior to pathologic analysis of the same tissues. Phosphocholine concentrations were significantly higher, while Eth concentrations were significantly lower in cancer versus benign prostate tissues. PC/GPC, PC/PE, PE/Eth, and GPE/Eth ratios were also significantly higher in cancer versus benign prostate tissues.


The authors thank Drs. Eriks Kupce, Daina Avizonis, and Bao Nguyen of Varian Instruments (Palo Alto, CA) for valuable discussions and continued support of this research.

Grant sponsor: National Institutes of Health; Grant numbers: K01 CA096618, R01 CA102751; Grant sponsor: American Cancer Society; Grant number: RSG-05-241-01-CCE; Grant sponsor: University of California Discovery Grant; Grant number: ITL-BIO04-10148.


1. Podo F. Tumour phospholipid metabolism. NMR Biomed. 1999;12:413–439. [PubMed]
2. Podo F, de Certaines JD. Magnetic resonance spectroscopy in cancer: phospholipid, neutral lipid and lipoprotein metabolism and function. Anticancer Res. 1996;16:1305–1315. [PubMed]
3. Gillies RJ, Morse DL. In vivo magnetic resonance spectroscopy in cancer. Annu Rev Biomed Eng. 2005;7:287–326. [PubMed]
4. Glunde K, Ackerstaff E, Natarajan K, Artemov D, Bhujwalla ZM. Real-time changes in H-1 and P-31 NMR spectra of malignant human mammary epithelial cells during treatment with the anti-inflammatory agent indomethacin. Magn Reson Med. 2002;48:819–825. [PubMed]
5. Glunde K, Jie C, Bhujwalla ZM. Molecular causes of the aberrant choline phospholipid metabolism in breast cancer. Cancer Res. 2004;64:4270–4276. [PubMed]
6. Glunde K, Jie CF, Bhujwalla ZM. Mechanisms of indomethacin-induced alterations in the choline phospholipid metabolism of breast cancer cells. Neoplasia. 2006;8:758–771. [PMC free article] [PubMed]
7. Leach MO, Verrill M, Glaholm J, Smith TAD, Collins DJ, Payne GS, Sharp JC, Ronen SM, McCready VR, Powles TJ, Smith IE. Measurements of human breast cancer using magnetic resonance spectroscopy: a review of clinical measurements and a report of localized P-31 measurements of response to treatment. NMR Biomed. 1998;11:314–340. [PubMed]
8. Kurhanewicz J, Dahiya R, Macdonald JM, Jajodia P, Chang LH, James TL, Narayan P. Phosphorus metabolite characterization of human prostatic adenocarcinoma in a nude-mouse model by P-31 magnetic-resonance spectroscopy and high-pressure liquid-chromatography. NMR Biomed. 1992;5:185–192. [PubMed]
9. Cornel EB, Heerschap A, Smits GAHJ, Oosterhof GON, Debruyne FMJ, Schalken JA. Magnetic-resonance spectroscopy detects metabolic differences between 7 Dunning rat prostate tumor sublines with different biological behavior. Prostate. 1994;25:19–28. [PubMed]
10. Cornel EB, Smits GAHJ, Oosterhof GON, Karthaus HFM, Debruyne FMJ, Schalken JA, Heerschap A. Characterization of human prostate-cancer, benign prostatic hyperplasia and normal prostate by in-vitro H-1 and P-31 magnetic-resonance spectroscopy. J Urol. 1993;150:2019–2024. [PubMed]
11. Cornel EB, Smits GAHJ, Deruijter JE, Oosterhof GON, Heerschap A, Debruyne FMJ, Schalken JA. In-vitro proton magnetic-resonance spectroscopy of 4 human prostate-cancer cell-lines. Prostate. 1995;26:275–280. [PubMed]
12. Macdonald JM, Kurhanewicz J, Dahiya R, Espanol MT, Chang LH, Goldberg B, James TL, Narayan P. Effect of glucose and confluency on phosphorus metabolites of perfused human prostatic adenocarcinoma cells as determined by P-31 MRS. Magn Reson Med. 1993;29:244–248. [PubMed]
13. Negendank W. Studies of human tumors by MRS — a review. NMR Biomed. 1992;5:303–324. [PubMed]
14. Kurhanewicz J, Thomas A, Jajodia P, Weiner MW, James TL, Vigneron DB, Narayan P. P-31 spectroscopy of the human prostate-gland in vivo using a transrectal probe. Magn Reson Med. 1991;22:404–413. [PubMed]
15. Albers MJ, Krieger MD, Gonzalez-Gomez I, Gilles FH, McComb JG, Nelson MD, Bluml S. Proton-decoupled P-31 MRS in untreated pediatric brain tumors. Magn Reson Med. 2005;53:22–29. [PubMed]
16. Swanson MG, Vigneron DB, Tabatabai ZL, Males RG, Schmitt L, Carroll PR, James JK, Hurd RE, Kurhanewicz J. Proton HR-MAS spectroscopy and quantitative pathologic analysis of MRI/3D-MRSI-targeted postsurgical prostate tissues. Magn Reson Med. 2003;50:944–954. [PubMed]
17. Swanson MG, Zektzer AS, Tabatabai ZL, Simko J, Jarso S, Keshari KR, Schmitt L, Carroll PR, Shinohara K, Vigneron DB, Kurhanewicz J. Quantitative analysis of prostate metabolites using H-1 HR-MAS spectroscopy. Magn Reson Med. 2006;55:1257–1264. [PubMed]
18. Ackerstaff E, Pflug BR, Nelson JB, Bhujwalla ZM. Detection of increased choline compounds with proton nuclear magnetic resonance spectroscopy subsequent to malignant transformation of human prostatic epithelial cells. Cancer Res. 2001;61:3599–3603. [PubMed]
19. Swanson MG, Vigneron DB, Tran TKC, Kurhanewicz J. Magnetic resonance imaging and spectroscopic imaging of prostate cancer. Cancer Invest. 2001;19:510–523. [PubMed]
20. Bax A, Davis DG. MLEV-17-based two-dimensional homonuclear magnetization transfer spectroscopy. J Magn Reson. 1985;65:355–360.
21. Morvan D, Demidem A, Papon J, Madelmont JC. Quantitative HRMAS proton total correlation spectroscopy applied to cultured melanoma cells treated by chloroethyl nitrosourea: demonstration of phospholipid metabolism alterations. Magn Reson Med. 2003;49:241–248. [PubMed]
22. Zektzer AS, Swanson MG, Jarso S, Nelson SJ, Vigneron DB, Kurhanewicz J. Improved signal to noise in high-resolution magic angle spinning total correlation spectroscopy studies of prostate tissues using rotor-synchronized adiabatic pulses. Magn Reson Med. 2005;53:41–48. [PubMed]
23. Duarte IF, Stanley EG, Holmes E, Lindon JC, Gil AM, Tang HR, Ferdinand R, McKee CG, Nicholson JK, Vilca-Melendez H, Heaton N, Murphy GM. Metabolic assessment of human liver transplants from biopsy samples at the donor and recipient stages using high-resolution magic angle spinning H-1 NMR spectroscopy. Anal Chem. 2005;77:5570–5578. [PubMed]
24. Martinez-Granados B, Monleon D, Martinez-Bisbal MC, Rodrigo JM, del Lmo J, Lluch P, Ferrandez A, Marti-Bonmati L, Celda B. Metabolite identification in human liver needle biopsies by high-resolution magic angle spinning H-1 NMR spectroscopy. NMR Biomed. 2006;19:90–100. [PubMed]
25. Martinez-Bisbal MC, Marti-Bonmati L, Piquer J, Revert A, Ferrer P, Llacer JL, Piotto M, Assemat O, Celda B. H-1 and C-13 HR-MAS spectroscopy of intact biopsy samples ex vivo and in vivo H-1 MRS study of human high grade gliomas. NMR Biomed. 2004;17:191–205. [PubMed]
26. Moka D, Vorreuther R, Schicha H, Spraul M, Humpfer E, Lipinski M, Foxall PJD, Nicholson JK, Lindon JC. Biochemical classification of kidney carcinoma biopsy samples using magic-angle-spinning H-1 nuclear magnetic resonance spectroscopy. J Pharmaceut Biomed. 1998;17:125–132. [PubMed]
27. Cavanagh J, Chazin WJ, Rance M. The time-dependence of coherence transfer in homonuclear isotropic mixing experiments. J Magn Reson. 1990;87:110–131.
28. Wieruszeski JM, Montagne G, Chessari G, Rousselot-Pailley P, Lippens G. Rotor synchronization of radiofrequency and gradient pulses in high-resolution magic angle spinning NMR. J Magn Reson. 2001;152:95–102. [PubMed]
29. Kupce E, Keifer PA, Delepierre M. Adiabatic TOCSY MAS in liquids. J Magn Reson. 2001;148:115–120. [PubMed]
30. States DJ, Haberkorn RA, Ruben DJ. A two-dimensional nuclear Overhauser experiment with pure absorption phase in 4 quadrants. J Magn Reson. 1982;48:286–292.
31. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer. J Magn Reson Imaging. 2002;16:451–463. [PMC free article] [PubMed]
32. Marjanska M, Henry PG, Bolan PJ, Vaughan B, Seaquist ER, Gruetter R, Ugurbil K, Garwood M. Uncovering hidden in vivo resonances using editing based on localized TOCSY. Magn Reson Med. 2005;53:783–789. [PMC free article] [PubMed]
33. Aboagye EO, Bhujwalla ZM. Malignant transformation alters membrane choline phospholipid metabolism of human mammary epithelial cells. Cancer Res. 1999;59:80–84. [PubMed]
34. Burns MA, Taylor JL, Wu CL, Zepeda AG, Bielecki A, Cory D, Cheng LL. Reduction of spinning sidebands in proton NMR of human prostate tissue with slow high-resolution magic angle spinning. Magn Reson Med. 2005;54:34–42. [PubMed]
35. Davis DG, Bax A. Assignment of complex H-1-NMR spectra via two-dimensional homonuclear Hartmann-Hahn spectroscopy. J Am Chem Soc. 1985;107:2820–2821.
36. Payne GS, Troy H, Vaidya SJ, Griffiths JR, Leach MO, Chung YL. Evaluation of P-31 high-resolution magic angle spinning of intact tissue samples. NMR Biomed. 2006;19:593–598. [PubMed]
37. Loening NM, Chamberlin AM, Zepeda AG, Gonzalez RG, Cheng LL. Quantification of phosphocholine and glycerophosphocholine with P-31 edited H-1 NMR spectroscopy. NMR Biomed. 2005;18:413–420. [PubMed]