Chinese-German Journal of Clinical Oncology

February 2008, Vol. 7, No. 2, P103–P106

DOI 10.1007/s10330-007-0162-9

Preliminary study in vitro brain tumor with high resolution magic angle spinning proton magnetic resonance spectroscopy Wei Tan1, Yaojun Ke1, Guangyao Wu2 1

2

Radiology Department of the Affiliated Hospital of Wuhan Science and Technology University, Wuhan 430064, China Department of MRI, Zhongnan Hospital, Wuhan University, Wuhan 430071, China Received: 8 October 2007 / Revised: 9 November 2007 / Accepted: 29 December 2007 Abstract Objective: To explore water soluble metabolite features of brain tumor specimens with HRMAS-1HMRS and its potential clinical value. Methods: There were thirty cases of pathologically proven brain tumor, including 6 I–II grade astrocytomas, 7 III grade anaplastic astrocytomas, 10 IV grade glioblastomas and 7 meningiomas. Used Varian Company 600 MHz spectrometer with the Nano-probe for acquisition HRMAS-1HMRS, which was postprocessed with jMRUI 3.2 version software. These metabolic probability and their ratios to Cr were summed. Results: (1) HRMAS-1HMRS could resolve NAA, PCr/Cr, GPC + PCho + Cho, Glu/Gln, Gly, Tau, Ala, Lac, mI and so on. All samples showed Lac, 6 samples showed unknown single peak at 3.72 ppm or 3.90 ppm. (2) The mean Cho/Cr of 6 I–II grade astrocytomas was 2.42 ± 1.01 (P = 0.003, compared with glioblastoma). The mean Cho/Cr of 7 anaplastic astrocytomas was 3.48 ± 0.59 (P = 0.01, compared with glioblastoma). The Cho/Cr of 10 glioblastomas broadly ranged from 0.9 to 11.3 (mean 5.40 ± 1.23). From I–II grade astrocytoma to glioblastoma, Ala/Cr, Tau/Cr and Gly/Cr trends were increased; the mean Ala/Cr of glioma was 0.31 ± 0.13. (3) Meningiomas showed higher Ala and Cho. Their Cr was lower than that of gliomas. 4/7 cases had no NAA, 3/7 patients had lower NAA. Mean Cho/Cr was 3.56 ± 1.01, Ala/Cr was 0.53 ± 0.28 (P = 0.006, compared with glioma). Conclusion: HRMAS-1HMRS can show further details in vivo MRS, resolve in vivo spectroscopic metabolite of Cho compound and differentiate the extent of benign and malignant glioma. With the increase in the malignant degree of gliomas, Cho, mI, Ala, Tau and Gly will increase. HRMAS-1HMRS is the only method of isotropic spectroscopy for pathological specimens. Key words magic angle spinning; high resolution proton magnetic resonance spectroscopy; in vitro; brain tumor Abbreviations high-resolution magic angle spinning proton magnetic resonance spectroscopy (HRMAS-1HMRS); N-acetyl aspartate (NAA); phosphocreatine/creatine (PCr/Cr); Glycerophosphocholine (GPC); PhosphoChoine (PCho); Choine (Cho); Glutamate/glutamine (Glu/Gln); Glycine (Gly); Taurine (Tau); Alanine (Ala); myo Inositol (mI); Lactate (Lac); phosphate monoester (PME); phosphate ester (PDE); gamma aminobutyrate (GABA); lipid (Lip)

Articles about in vivo brain tumor 1HMRS have been widely reported at home and abroad, reflected some of the characteristics of brain tumors [1]. But in vivo 1HMRS by low field strength, partial volume effect and other constraints has limited value on the metabolism of information and diagnosis of diseases. In vitro specimens can be applied to the higher field strength spectrometer and 1 HMRS can detect some compounds for in vivo spectroscopy difficult detection. But former extraction method of HR-1HMRS destructed histological characteristics of the sample, separated the spectral results and disease state, its application in clinic remains limited. However, recently developed and perfected a new technique that HRMAS1 HMRS did not suffer greatly destruction in the sample Correspondence to: Guangyao Wu. Email: [email protected]

and improved resolution, maintaining a state of the disease and the results of spectral consistency. This study explored water soluble metabolite features of brain tumor (meningiomas and gliomas) specimens with HRMAS1 HMRS and its potential clinical value.

Materials and methods Collected surgical resection 30 cases of brain tumor by pathologically confirmed, according to the WHO criteria, 6 cases were I–II grade astrocytomas, 7 cases were III grade anaplastic astrocytomas, 10 cases were IV glioblastomas and 7 patients were meningiomas. Thirteen cases were males and seventeen cases were females, aged 13 to 70 years, with an average age of 48 years old. Normal brain tissue HRMAS-1HMRS referred to the No. 2 refer-

104

ence [2]. Specimen handling: Each specimen was divided into two, one for the study of pathology, the other for spectral analysis. Pathology specimens were fixed with 10% formalin, then sliced and stained. Each about 20 to 60 mg specimen for spectral analysis was rapidly frozen at –170 ć in liquid nitrogen tanks after resection. During measuring, specimens were washed in 0.9% D2O saline, then were added to trace probe, measured with the Varian Company 600 MHz spectrometer at room temperature. HRMAS-1HMRS method: Carr-Purcell-Gill pulse acquisition. The repeat time was 3.0 s, and the spectral width 8 kHz, 20 ppm. Sampling data points 20 K, postprocessing software with jMRUI 3.2 version. The metabolic probability and their ratios to Cr were summed. In SPSS 11.5 software for statistical analysis, research findings presented mean ± standard deviation (ǒ ± s), and P < 0.05 was significant.

Results HRMAS-1HMRS could resolve NAA, PCr/Cr, GPC + PCho + Cho, Glu/Gln, Gly, Tau, Ala, Lac, mI and others cerebral metabolites. All samples showed Lac, 6 samples showed unknown single peak at 3.72 ppm or 3.90 ppm which might be related with mannitol therapy for decompression. The mean Cho/Cr of 6 I–II grade astrocytomas was 2.42 ± 1.01 (P = 0.003, compared with glioblastoma). The mean Cho/Cr of 7 anaplastic astrocytomas was 3.48 ± 0.59 (P = 0.01, compared with glioblastoma). The Cho/Cr of 10 glioblastomas broadly ranged from 0.9 to 11.3 (mean 5.40 ± 1.23). From I–II grade astrocytoma to glioblastoma, Ala/ Cr, Tau/Cr and Gly/Cr trends were increased; the mean Ala/Cr of glioma was 0.31 ± 0.13. HRMAS-1HMRS could resolve Cho compound details consisting of GPC, PCho and Cho, which contributed significantly different to Cho compound of different grade gliomas. For high grade gliomas, Cho compound was composed of the GPC, PCho and Cho balanced, or PCho oriented. While for low-grade gliomas, Cho compound was mainly from the GPC with a small peak Cho and PCho. Meningiomas showed higher Ala and Cho. Cr was lower than that of gliomas. 4/7 cases had no NAA, 3/7 patients had lower NAA. Mean Cho/Cr was 3.56 ± 1.01, Ala/ Cr was 0.53 ± 0.28 (P = 0.006, compared with glioma).

Discussion In vivo magnetic resonance spectroscopy is due to the magnetic field strength, partial volume effect and patients with such constraints, its spectral resolution remains limited. However, in vitro specimens can choose high resolution spectrometer detection which can resolve in vivo

www. springerlink. com/content/1613-9089

spectroscopy indistinguishable some metabolites. Nevertheless, the former high resolution spectral method is mostly based on extraction method. Isolated specimens must be extracted, centrifugated, and dissolved, then changed liquid samples. The whole process was complex, their structures were damaged. MAS technique was been used for high resolution spectroscopy of solid research in 1982. In 1997, the technique was introduced into the in vitro biological research [3]. The technology was called direct method, only need a small amount of sample (20–60 mg) and avoids partial volume effects, has high sensitivity. Samples can be used directly in the operation without a physical or chemical treatment. The technology is also fully automatic procedure, the observer is not dependent on technology and experience and can quickly analyze results, including the number of quantitative analysis and modular report, reducing the potential costs of pathology. The solid elements in a relatively fixed position, the effect of anisotropy significantly, NMR spectroscopy is much more than the width of liquid. For example, the water molecule of linewidth is about 0.1 Hz, but ice is 105 Hz. Liquid molecules rapid irregular movement averages the anisotropy into zero. Some, however, has yet to find a suitable material solvents, and some in dissolving or melting process may be changed the structure of molecules. The main reasons of solid line broadening were as follows: dipole-dipole interaction, the nuclear tetrapolar interaction, chemical shift anisotropy, and spin coupling anisotropy. The use of rotating machinery, and the samples around the field B0-beta angle rotation, 3cos2ǃ±  ǃ ƒĻLVFDOOHGPDJLFDQJOH[4]. Ideal MAS can cause dipole-dipole interaction and the nuclear tetrapolar interaction into zero and the chemical shifts of anisotropy, spin coupling anisotropy partial into zero as the liquid with the same high resolution. MAS linewidth is strongly dependent on the relationship with spinning speed and the magic angle setting accuracy. When narrowing factor of the spectrum of more than 1000 times, require accurate PDJLFDQJOHWRVHYHUDOGHJUHHVSHUǃGHYLDWHGIURP the more, peaks are more width, the more impact from the greater angle. This is the magic angle own limitations. It is the only available method of solid samples of isotropic, similar liquid high resolution spectroscopy [4]. HRMAS-1HMRS owing to high field strength and uniform magnetic field can distinguish more metabolites than in vivo spectroscopy. Our group HRMAS-1HMRS had better SNR and showed lots of metabolites such as NAA, PCr/Cr, GPC + PCho + Cho, Glu/Gln, Gly, Tau, Ala, Lac, and mI. NAA has two resonance peaks: first, the 2.02 ppm -CH3 single peak is a major contribution to NAA and quantitative statistical peak; second, the 4.39 ppm Į&+2 double peak is lower almost ten times than -CH3 single peak, concealed by water peak. Lac has also two resonance peaks: first, the 1.33 ppm -CH3 double peak;

Chinese-German J Clin Oncol, February 2008, Vol. 7, No. 2

WKHRWKHULVĮ&+TXDUWHWZKLFKFDQQRWEHUHVROYHGEHcause of baseline instability. PCr/Cr has two resonance peaks, the main peak is 3.03 ppm N-CH3 single peak, the other is 3.93 ppm S-CH2 single peak, all response to cellular energy state. Choline compound consisting of three components: GPC, 3.24 ppm N-(CH3)3 single peak; PCho 3.22 ppm N-(CH3)3 single peak; Cho (free choline) 3.20 ppm N-(CH3)3 single peak, and the three components can be indistinguishable with in vivo spectroscopy. PCho is the main ingredient of PME, and is concerned with myelinization, cell proliferation, cell membrane synthesis of glial cell growth and cell nutritional status. GPC is the main ingredient of PDE, and relates to phospholipase decomposition products. With the increase in myelin, myelin update, PDE must increase. In high resolution spectroscopy from 2.03 ppm to 2.51 ppm, because of peaks overlap, more metabolites, and J-coupling phenomenon serious, these peaks are difficult to quantify and calculate [1]. For example, Gln chemical shift is 2.46 ppm, 2.13 SSPDQGSSPPXOWLSOHWIURPDŽ&+2ǃ&+2DQGĮ&+ respectively. Glu chemical shift is 2.35 ppm, 2.09 ppm DQGSSPPXOWLSOHWIURPDŽ&+2ǃ&+2DQGĮ&+UHspectively. Glu as excitatory neurotransmitter is the most abundant amino acid in the brain. Gln is a precursor and storing form of Glu. Structures of Gln and Glu are similar and both are difficult to be resolved in vivo spectroscopy, collectively referred to as Glx. GABA is the brain main inhibitory neurotransmitter, the derivatives of glutamate through decarboxylation. Its chemical shift is 1.91 ppm, SSPDQGSSPPXOWLSOHWDQGWULSOHWIURPDŽ&+2, ǃ&+3DQGĮ&+2 respectively. Our group did not compare these. mI chemical shift is 3.54 ppm, 3.28 ppm and 3.63 ppm multiplet and triplet from H1, H3, H5 and H2. mI relates to intracellular sodium content and glial activity, and is considered as sign of glial cells [5]. Ala chemical shift is SSPDQGSSPGRXEOHSHDNIURPĮ&+DQGǃ&+3 respectively. In vivo spectroscopy, Ala peak easily overlaps with the Lip peak. Ala increased is considered as the characteristic performance of meningioma. Tau chemical shift is 3.42 ppm and 3.27 ppm triplet from N-CH2 and S-CH27DXLVDWUDQVIRUPDWLRQIURPF\VWLQHRQHǃDPLQR acid. Its many biological functions show osmoregulation, neurotransmitter regulation, promote the growth and development body. Gly chemical shift is 3.56 ppm, and overlaps with mI peak [2, 4]. Gly as inhibitory neurotransmitter and antioxidants widely exists in the central nervous system [5]. On in vivo spectroscopy of normal brain tissue, mean Cho/Cr was 0.8 ± 0.2, Cho/Cr from in vitro is about 0.8. In low grade astrocytoma, Cho/Cr was lower. In six cases of I–II grade astrocytoma of our group, average Cho/Cr was 2.42 ± 1.01; in anaplastic astrocytomas, average Cho/Cr was 3.48 ± 0.59; in glioblastomas, Cho/Cr broadly ranged from 0.9 to 11.3, the average was 5.40 ± 1.23. Glioblas-

105

toma metabolite distribution was pleomorphic as their inhomogeneous morphology. Despite Cr peak in glioma variation, but their average level was rising with the malignancy of glioma. Increased Cho was related to cell proliferation and the degree of malignancy. High Cho/Cr was a typical tumor spectroscopy. Furthermore, HRMAS1 HMRS could distinguish Cho-under-metabolites, namely PCho, GPC and Cho. In the high level or low grade gliomas, every ingredients contributed significantly different to Cho compound. For high grade gliomas, Cho compound was composed of the GPC, PCho and Cho balanced, or PCho oriented. While for low grade gliomas, Cho compound was mainly from the GPC with a small peak Cho and PCho. PCho/Cho was linear correlation with malignant glioma cells percentage. These metabolites in different concentrations and the different ratios are malignant glioma marker [3]. This study showed that with the increase of malignancy of gliomas, Gly, Tau, mI and Ala had greater present possibility, Ala/Cr, Tau/Cr and Gly/Cr of low-grade astrocytomas were significantly lower than those of glioblastomas, suggesting that Ala, Tau and Gly as the basis for the identification of different-grade gliomas. Astrocyte cells are known to be degradation Gly site in the brain, which relate to Gly metabolic pathways and the functional modulation, including absorption Gly receptor. In normal brain tissue, Gly is lower level, however, in high grade glioma, astrocyte cells were replaced by tumor cells, and Gly increased. Tau was reported in the literature which may be related to increased proliferation of glial cells, Tau and glial cells had high affinity [5]. Jiang et al [6] found in lacking sugar and hypoxia induced brain injury, inhibitory amino acid (IAA) increased, it is thought that the body launched a self-protection mechanism, but in high-grade gliomas, the inhibitory amino acids such as Gly and Tau increase may be in a similar role [7]. NAA mainly produced by the mitochondria widely exists in neurons and synapses. With an increase of malignant glioma, Cho is a upward trend, NAA showed a downward trend. In literatures, there were reports that NAA from the high-resolution spectra had no significant difference between I–II grade astrocytoma and primitive neuroectodermal tumor. Glioblastoma Cho/NAA ratio increased, is mainly due to Cho increase rather than NAA decline. NAA peak measured is the smaller by HRMAS1 HMRS than by in vivo 1HMRS, probably the in vivo 1 HMRS is because greater voxel from normal brain tissue contamination. Some documents have reported astrocytoma could be detected valine, leucine peak, but no cases of this group showed clearly that these amino acids overlap by the 0.9 ppm Lip. Meningioma is non-neurological mesodermal tumor, with the lack of neurons and glial cells. Typical in vivo spectroscopy is lack of Cr, mI and NAA, and has high Cho

106

and Ala. The characteristic spectral HRMAS-1HMRS is broad range of Cho/Cr and Ala/Cr, or very high Cho/Cr and Ala/Cr. In our group, 4/7 cases had no NAA, 3/7 patients had lower NAA peaks, may be contaminated with normal brain tissue and adjacent Gln/Glu drift and interference. In meningioma, the mean Cho/Cr was 3.56 ± 1.01, and there was not significant difference with glioma. Shino et al [8] thought that in malignant meningioma, Cho/Cr was significantly higher than that of benign meningiomas, and may be as a basis for identification of benign and malignant meningioma. But this group was only one case of malignant meningioma, Cho/Cr up to 7.90, about Cho/Cr average of two times. Reports showed that schwannoma and hemangiopericytoma had 3.56 ppm high mI, and may be as the identification of meningioma. Although in vivo studies found that Ala peak of meningioma, HRMAS-1HMRS also detected Ala in glioma. However meningiomas had higher Ala, Ala/Cr was 0.53 ± 0.28, and glioma Ala/Cr was 0.31 ± 0.13, the difference was significant (P < 0.01). In meningiomas higher Glu/ Gln peaks also detected, due to J-coupling, we did not have the statistical analysis. Ala peak in the meningiomas was more inclined to the effect of glutamate rather than glycolysis for energy metabolism, and Glu/Gln may be as the energy substrate in meningioma to support Glu/Gln the assumption for meningioma energy metabolism other than brain tumors [9]. In all of the samples HRMAS-1HMRS showed double peak at 1.33 ppm and 4.11 ppm quartet for Lac signal. Because the sampling and frozen specimens in the course of the inevitable happened anaerobic glycolysis made Lac content increased, the detected Lac was higher than that in vivo 1HMRS, the isolated samples of Lac was limited value. Another 6 patients showed unknown a single peak at 3.72 ppm and 3.90 ppm in our group, which may be related to the mannitol treatment for cerebral decompression during a craniotomy [5]. In short, HRMAS-1HMRS can show further details in vivo MRS, resolve in vivo spectroscopic Cho compound and differentiate the extent of benign and malignant glio-

www. springerlink. com/content/1613-9089

mas. With the increase of the degree of malignancy, Cho, mI, Ala, Tau and Gly will increase. HRMAS-1HMRS is the only method of isotropic spectroscopy for pathological specimens. We believe that with the development of hardware, the determinable metabolites will be increased, and HRMAS-1HMRS will increasingly become the powerful tool of studying cellular and molecular pathology and disease diagnosis in imaging, pathology, biology, and other fields.

References 1. Tzika AA, Cheng LL, Goumnerova L, et al. Biochemical characterization of pediatric brain tumors by using in vivo and ex vivo magnetic resonance spectroscopy. J Neurosurg, 2002, 96: 1023–1031. 2. Govindaraju V, Young K, Maudsley AA. Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed, 2000, 13: 129–153. 3. Lehnhardt FG, Bock C, Röhn G, et al. Metabolic differences between primary and recurrent human brain tumors: a 1H NMR spectroscopic investigation. NMR Biomed, 2005, 18: 371–382. 4. Martínez-Bisbal MC, Martí-Bonmatí L, Piquer J, et al. 1H and 13C HR-MAS spectroscopy of intact biopsy samples ex vivo and in vivo 1 HMRS study of human high grade gliomas. NMR Biomed, 2004, 17: 191–205. 5. Tugnoli V, Tosi MR, Bertoluzza A, et al. In vitro magnetic resonance spectroscopy of health and neoplastic brain tissues. J Mol Struc, 1999, 483: 365–369. 6. Jiang QS, Xu SY, Yang JQ, et al. Effects of anoxia and/or aglycaemia in the release of free calcium and amino acids from rat brain synaptosomes. J Pediatr Pharm (Chinese), 2003, 9: 8–10. 7. Barba I, Moreno A, Martinez-Pérez I, et al. Magnetic resonance spectroscopy of brain hemangiopericytomas: high myoinositol concentrations and discrimination from meningiomas. J Neurosurg, 2001, 94: 55–60. 8. Shino A, Nakasu S, Matsuda M, et al. Noninvasive evaluation of the malignant potential of intracranial meningiomas performed using proton magnetic resonance spectroscopy. J Neurosurg, 1999, 91: 928–934. 9. Howe FA, Barton SJ, Cudlip SA, et al. Metabolic profiles of human brain tumors using quantitative in vivo 1H magnetic resonance spectroscopy. Mag Res Med, 2003, 49: 223–232.