Metabolic characterization of human soft tissue sarcomas in vivo and in vitro using proton-decoupled phosphorus magnetic resonance spectroscopy.

We applied 1H-decoupling and nuclear Overhauser enhancement to obtain well-resolved 31P magnetic resonance spectra accurately localized to 20 soft tissue sarcomas in vivo, using three-dimensional chemical shift imaging. Fifteen spectra had large phosphomonoester signals (21% of total phosphorus) that contained high amounts of phosphoethanolamine (compared to those of phosphocholine) but no signals from glycerophosphoethanolamine, and glycerophosphocholine was detected in only four cases. Prominent nucleoside triphosphates (52% of phosphorus) and low inorganic phosphate (10% of phosphorus) indicated that a large fraction of these 15 sarcomas contained viable cells, and this impression was confirmed histologically in 13 of the sarcomas. High-resolution in vitro 31P spectra of extracts of surgical specimens of four of the sarcomas studied in vivo and six additional sarcomas confirmed the in vivo assignments of metabolites and revealed considerable inter- and intratumoral variations of metabolite concentrations associated with histological variations in the relative amounts of cells and of matrix materials or spontaneous necrosis. Seven sarcomas, all high grade with pleomorphic or round cells rather than spindle cells, contained an unidentified phosphodiester signal in vivo; its absence in the extract spectra indicates that it may be from an abnormally mobile membrane component. We have documented a means to obtain new information about in vivo metabolism in human sarcomas and to provide a basis on which to examine the uses of 31P magnetic resonance spectroscopy in the clinical management of sarcomas.

31P MRS of human tumor cells: effects of culture media and conditions on phospholipid metabolite concentrations.

31P magnetic resonance spectroscopy has been used to determine phosphate metabolite profiles in five human tumor cell lines in culture and as solid tumor xenografts in nude mice. Significant differences between cell lines, in particular in their phospholipid metabolite levels, were observed. The largest differences between metabolite profiles in vivo and in culture were observed for cell lines which exhibit low phosphoethanolamine levels in culture. One of these lines, the colon carcinoma CX-1, was studied in more detail in both incubated and perfused DMEM cultures with variation of the concentrations of glucose, choline and ethanolamine. Highly significant alterations of phospholipid metabolite concentrations and UDP-hexoses (primarily UDP-GlcNAc and UDP-GalNAc) were observed as a function of the precursor concentrations, culture time or perfusion time. A strong interaction between phospholipid metabolic pathways and UDP-hexose pathways could be demonstrated.

Metabolism of phosphonium choline by rat-2 fibroblasts: effects of mitogenic stimulation studied using 31P NMR spectroscopy.

Phospholipid turnover increases with both mitogenic stimulation and oncogenic transformation (1-9). Recent 31P nuclear magnetic resonance (NMR) spectroscopy studies of human tumors, animal tumor models and cell systems have reported elevated phosphomonoesters with growth and oncogenic transformation, as well as changes in these levels associated with treatment (10). In order to gain insights into the mechanisms underlying these changes, we used a phosphonium analog of choline and 31P NMR spectroscopy to study choline metabolism in quiescent and mitogenically stimulated Rat-2 fibroblasts. Cell growth status of these cells has a significant effect on choline metabolism. While overall uptake of the analog was similar in both quiescent and growing cells, distribution among metabolite pools differed. Quiescent cells accumulate label in the phosphodiester pool, with little or none in the phosphomonoester pool. On the other hand, mitogenic stimulation resulted in a significant fraction of the label in the phosphomonoester pool.

Mobile lipid accumulation in necrotic tissue of high grade astrocytomas.

Multiple samples from 42 astrocytomas were investigated ex vivo by 1H MR spectroscopy followed by histological assessment. MR visible lipids were detected in 27 of 32 grade 4 astrocytomas. These lipids were heterogeneously distributed within the tumours. Their amount correlated positively with the amount of histologically detected necrosis. Mobile lipids were also observed in grade 4 astrocytoma samples without necrosis, as well as in one of three grade 3, two of three grade 2 and two of four grade 1 astrocytomas. The clinical significance of MR visible lipids, their cellular location, and their possible biological bases are discussed.

A 1H-NMR method for determining temperature in cell culture perfusion systems.

This report describes a noninvasive 1H-NMR method for measuring absolute temperatures (+/- 0.2 degrees C) in biological samples and, in particular, in cell culture perfusion systems, utilizing the linear temperature dependence of the water chemical shift relative to the temperature-independent shift of one of the components of the biological medium, e.g., pyruvate, acetate or lactate. The effects of flow on temperature can be monitored and appropriate adjustment of the temperature controller can be made.

1H MRS of high grade astrocytomas: mobile lipid accumulation in necrotic tissue.

Sixty-four samples from six grade 4 astrocytomas were investigated ex vivo by 1H MRS at 360 MHz and subsequently by histopathology to obtain percentages of viable and necrotic tumour and grey and white matter. MR-visible lipids were detected in 87% of tumour samples. Necrotic foci were < 3 x 3 x 6 mm3. The means of the intensities/unit weight tissue of the lipid resonances at 5.33, 2.80, 1.29 and 0.89 ppm were significantly higher (p < 0.05) for three sets of comparisons: samples with 85-100% vs 50-75%; with 50-75% vs 10-40% and with 10-40% vs 0-5% necrosis. For the lipid resonance at 2.04 ppm the difference in the means was significant only for samples with 50-75% compared to those with 85-100% necrosis, because for samples with < 50% necrosis resonances from glutamine and possibly small amounts of glutamate, gamma-aminobutyrate and N-acetylaspartate anions contribute significantly to the spectral area at 2.0 ppm. We conclude that necrotic foci below MRI resolution yield the resonances at 1.3 and 0.9 ppm, and contribute to the intense resonance at 2.0 ppm observed in in vivo 1H spectra of some high grade astrocytomas.

1H MR visible lipids in colon tissue from normal and carcinogen-treated rats.

Metabolic characteristics of colon mucosa, submucosa, muscularis and tumour specimens from four control (n = 105) and nine carcinogen (azoxymethane)-treated (n = 91) Sprague-Dawley rats were investigated by ex vivo 1H MRS. Ninety-seven per cent of pure mucosa samples (n = 59) yielded spectra with narrow lipid resonances (chemical shift delta of -(CH2)n-, 1.3 ppm; linewidth at half-height v1/2, 30-50 Hz). Eighty-two per cent of control mucosa samples with histologically proven submucosa contamination (n = 11) and 46% of control cross-sections (containing mucosa, submucosa and muscularis; n = 57) yielded spectra with broad lipid resonances (delta-(CH2)n-, 1.5 ppm; v1/2, 80-100 Hz) identical to those of adipose tissue surrounding rat colon. Thirty per cent of tumour samples (n = 10) yielded spectra with narrow lipid resonances while 70% contained no significant amount of MR visible lipids. We conclude that (i) lipids giving rise to broad resonances are in the heterogeneously distributed adipocytes of submucosa, (ii) lipids giving rise to narrow resonances are within the mucosa in an unknown structural environment, and (iii) the type and distribution of lipids in human and rat colon are similar. Tumours contained significantly more taurine than pure control mucosa (n = 15; p < 0.004) and pure mucosa containing aberrant crypt foci (putative preneoplasm, n = 36; p < 0.002). Our results suggest that the rat colon is a good model for 1H MR investigations of human colon carcinogenesis.

Lipoprotein(a) and CA125 levels in the plasma of patients with benign and malignant ovarian disease.

Elevated plasma levels of apolipoprotein (a) have been reported earlier in cancer patients. In order to investigate the potential of apolipoprotein(a) as an ovarian tumour marker, plasma apolipoprotein(a) and CA125 levels were measured in healthy women and women with benign or malignant pelvic masses. Among women younger than 49 years, 80% of healthy controls and ovarian cancer patients had apolipoprotein(a) levels below 350 U/l. Among women aged 49 years or older, 46% of healthy controls but 73% of ovarian cancer patients and 77% of women with successfully treated ovarian cancer, had low apolipoprotein(a) levels. For both age groups, apolipoprotein(a) is not a suitable marker for ovarian cancer. No correlation was found between apolipoprotein(a), triglyceride or cholesterol in plasma. Healthy women younger than 49 years had significantly higher CA125 levels than women 49 years or older (20 +/- 14 U/ml vs. 13 +/- 12 U/ml, p less than 0.005). Levels of CA125 above 35 U/ml were found in 12% of the younger and 4% of the older healthy women, 73% of the younger and 61% of the older patients with untreated or residual tumours, and in 33% of the younger and 31% of older patients with no evidence of disease, as well as in 58% of women of both age groups with benign pelvic masses. The sensitivity and specificity of CA125 levels for the detection of cancer were 73% and 74% respectively for women younger than 49 years, and 62% and 78% respectively for women 49 years or older.

Mobile lipids and metabolic heterogeneity of brain tumours as detectable by ex vivo 1H MR spectroscopy.

To examine metabolic heterogeneity in primary brain tumour, multiple biopsies (one to five per tumour) from four benign meningiomas, five acoustic schwannomas, eleven glial tumours and five medulloblastomas were each subdivided into one to five samples (total = 194) and investigated using ex vivo 360 MHz 1H MR spectroscopy with histopathological correlation. Low amounts of mobile lipids were detected in meningothelial meningiomas and medulloblastomas, high amounts in some or all samples of all other tumours. The differences in non-lipid metabolites between different tumour types were less marked, with the exception of the medulloblastomas, which had high amounts of N-acetyl aspartate and cholines. The amount of mobile lipid correlated with Antoni type B areas in schwannomas and necrosis in a high grade giant cell tumour. The amount of lipid varied significantly from sample to sample in a grade 2 astrocytoma and grade 4 astrocytoma samples with 0-5% necrosis but did not correlate with the site in the tumour or histopathological characteristics. The metabolic heterogeneity occurred on a spatial scale at least as small as the size of the MR samples, i.e., < or = 60 mg. These results suggest that the detection of metabolic heterogeneity and mobile lipids may have importance for the design of clinical studies.

31P NMR studies of cultured human tumor cells. Influence of pH on phospholipid metabolite levels and the detection of cytidine 5'-diphosphate choline.

202 MHz 31P NMR (11.7 T) was used to study the effects of culture medium pH on the levels of phosphate metabolites in three human tumor cell lines (XP29MAmal, a malignant xeroderma pigmentosum; CX-1, a colon carcinoma; KB, a squamous cell carcinoma of the oral cavity). Cells were cultured in Roux flasks in HAM's F-12 medium, and the pH was varied with the final medium change. After harvesting, 1-5 x 10(8) cells were suspended in Ringer/HEPES buffer at pH 7.4 and 4 degrees C for 31P NMR studies. Cell adhesion and growth rate decreased with decreasing pH, but, down to ca pH 6.1, trypan blue exclusion and the observed levels of nucleoside di- and triphosphates (range: 22-37% of total phosphates detected), phosphocreatine (PCr, 2-5%) and Pi (5-11%) did not vary significantly with pH. For XP29MAmal cells in exponential growth phosphocholine levels decreased from 18-28% at pH 7.0 to ca 5% at pH 6.0, while phosphoethanolamine levels increased from 2-7% to 15%. Glycerophosphocholine (GPC) levels increased from ca 7% at pH 7.2 to 13% at pH 6.3. At pH less than 6.3 cytidine 5'-diphosphate (CDP) choline became detectable (8-16%, delta p:P alpha = -8.13 ppm, P beta = 8.93 ppm, for PCr = 0 ppm). However, confluent cells did not accumulate CDP-choline when the pH was lowered. The cell lines CX-1 and KB also showed the pH effects described herein.

A simple procedure for obtaining high-quality NMR spectra of semiquantitative value from small tissue specimens: cervical biopsies.

The handling of small tissue biopsy samples (< 100 microliters) for NMR investigations poses special problems. Optimal and stable positioning of the samples within the sensitive volume of the radiofrequency coil can be achieved by inserting the sample in a capillary. Methods for quantitation of the spectral information from samples requiring histological evaluation after the NMR experiment are discussed, in particular with respect to cervical biopsies.

Quantitation of resonances in biological 31P NMR spectra via principal component analysis: potential and limitations.

This paper examines the potential and limitations of peak area quantitation of biological NMR spectra using principal component analysis (PCA), including its requirement for prior knowledge. The principles of the method are presented without in-depth mathematical treatment. PCA is illustrated for simulated data, 31P NMR spectra obtained consecutively over 1-2.5 days from perfused Rat-2 cells metabolizing the choline analogue phosphoniumcholine (Chop) and in vivo proton-decoupled, NOE-enhanced, three-dimensional CSI localized 31P NMR spectra of the liver of healthy volunteers. The results show that PCA can be used to quantitate strongly overlapping peaks without prior knowledge of the peak shapes or positions and to reconstruct spectra with significantly reduced noise variance. Two major limitations of PCA are presented: (1) PCA cannot separate peaks whose intensities are well correlated; (2) PCA is sensitive to differences in chemical shift and line-width of peaks between spectra. The discussion focuses on what knowledge of the biological and spectroscopic features of the samples and the principles of PCA is necessary for peak area quantitation via PCA.

Classification of 1H MR spectra of human brain neoplasms: the influence of preprocessing and computerized consensus diagnosis on classification accuracy.

We study how classification accuracy can be improved when both different data preprocessing methods and computerized consensus diagnosis (CCD) are applied to 1H magnetic resonance (MR) spectra of astrocytomas, meningiomas, and epileptic brain tissue. The MR spectra (360 MHz, 37 degrees C) of tissue specimens (biopsies) from subjects with meningiomas (95; 26 cases), astrocytomas (74; 26 cases), and epilepsy (37; 8 cases) were preprocessed by several methods. Each data set was partitioned into training and validation sets. Robust classification was carried out via linear discriminant analysis (LDA), artificial neural nets (NN), and CCD, and the results were compared with histopathological diagnosis of the MR specimens. Normalization of the relevant spectral regions affects classification accuracy significantly. The spectra-based average three-class classification accuracies of LDA and NN increased from 81.7% (unnormalized data sets) to 89.9% (normalized). CCD increased the classification accuracy of the normalized sets to an average of 91.8%. CCD invariably decreases the fraction of unclassifiable spectra. The same trends prevail, with improved results, for case-based classification. Preprocessing the 1H MR spectra is essential for accurate and reliable classification of astrocytomas, meningiomas, and nontumorous epileptic brain tissue. CCD improves classification accuracy, with an attendant decrease in the fraction of unclassifiable spectra or cases.