Transcription analysis of TIMP-1 and NM23-h1 genes in glioma cell invasion

Análise transcricional dos genes TIMP-1 e NM23-H1 na invasão celular em astrocitoma difuso e glioblastoma multiforme

Transcription analysis of TIMP-1 and NM23-h1 genes in glioma cell invasion

Análise transcricional dos genes TIMP-1 e NM23-H1 na invasão celular em astrocitoma difuso e glioblastoma multiforme

José Augusto Nasser^I; Asdrúbal Falavigna^II; Fernando Ferraz^I; Gregory Duigou^III; Jeffrey Bruce^III

^IEscola Paulista de Medicina, Universidade Federal de São Paulo, SP - Brazil, Universidade Estácio de Sá, Rio de Janeiro RJ - Brazil

^IIEscola Paulista de Medicina, Universidade Federal de São Paulo, SP Brazil, Universidade Caxias do Sul, RS - Brazil

^IIIBartolli Brain Tumor Center of College of Physicians and Surgeons of Columbia University (Department of Neurosurgery) New York, NY, USA

ABSTRACT

PURPOSE: To evaluate using transcription analysis the presence and importance of two genes: NM23-H1 and TIMP-1 on control of tumor cell invasion in diffuse astrocytomas (WHO II) and glioblastoma multiforme (WHO IV).

METHOD: Northern blot analysis of NM23-H1 and TIMP-1 was performed. Eight diffuse astrocytomas and 19 glioblastomas (WHO IV) were analyzed to determine if TIMP-1 and NM23-H1 were candidates to inhibition of tumor cell invasion quantitated RNA levels. The samples were collected directly from operating room. Total cellular RNA was extracted from frozen tissue samples using guanidinium-isothiocyanate and cesium chloride gradients. Total RNA (10 mg per sample) from tumor tissue were size fractionated through 1% agarose-formaldehyde gel and transferred to nylon filters and then hybridized to ³²P-labeled DNA probes and placed for autoradiography. Levels of specific RNAs were determined by computer-assisted laser densitometry. Blot filters were sequentially hybridized to nm23 and TIMP-1 probes in addition to GAPDH, as a control. Statistical analyses were carried out according to t-test for equality of means.

RESULTS: NM23-H1 were detected in each sample, however it did not correlate with malignancy and invasiveness. On the other side TIMP-1 gene expression showed a clear correlation between low expression and invasiveness.

CONCLUSION: The data suggest that TIMP-1 is an inhibitor of high grade gliomas invasion. NM23-H1 was present in the entire gliomas sample, but it did not vary in diffuse astrocytomas and glioblastomas.

Key words: tumor cell invasion, TIMP-1, NM23-H1, RNA levels, gliomas.

RESUMO

OBJETIVO: Comparar através da análise da expressão dos níveis de RNA, a presença e a relevância dos genes NM23-H1 e TIMP-1 no controle da invasão celular tumoral dentro do tecido cerebral normal em: astrocitoma difuso (OMS II) e glioblastoma multiforme (OMS:IV).

MÉTODO: Análise em "Northern blot" dos genes NM23-H1 e TIMP-1. Oito astrocitomas fibrilares difusos (OMS II) e 19 glioblastomas multiformes foram analisados para determinar se TIMP-1 e NM23-H1 estavam relacinados à inibição da invasão tumoral nas neoplasias do sistema nervoso central, quantificando os níveis de RNA dos respectivos genes extraídos diretamente dos tumores. 10 mg por amostra de RNA total foram fracionados de gel de formaldeído e transferidos para os filmes de hibridação. Níveis específicos de RNAs foram determinados na espectrofotometria. Valores das razões entre NM23-H1/GAPDH e TIMP-1/GAPDH foram submetidos à análise de variabilidade das médias.

RESULTADOS: A análise da expressão do gene TIMP-1 mostrou supressão em tumores gliais malignos.

CONCLUSÃO: Os resultados indicam que existe relação direta entre níveis baixos de TIMP-1 e malignidade dos gliomas. O gene NM23-H1 foi detectado em todas as amostras, mas não foi possível relacionar sua subexpressão ou superexpressão com algum fenótipo de invasividade.

Palavras-chave: invasão celular tumoral, TIMP-1, NM23-H1, níveis de RNA, gliomas.

Tumor cell invasion into normal brain is a major component of brain tumor malignancy. In contrast to other tumors, brain tumors are often invasive, but rarely metastasize to distant sites. An understanding of the genes which control tumor cell invasion will allow the development of therapeutic regimens designed to counteract this complex property. Although there are reports of genes that may modulate the ability of a brain tumor cell to invade adjacent normal brain, the molecular control of brain tumor invasiveness is still not well-defined. Specific secreted proteases including heparinases plus serine-,thiol- and metal-dependent proteinases have been implicated in tumor cell invasion¹.

The most well-studied of these enzymes are the metalloproteinases. Metalloproteinases are a family of protein-degrading enzymes that digest components of the extracellular matrix (ECM), the supporting material that functions to cement cells in place within tissues. These proteins have been implicated in wound healing, uterine, involution, embryogenesis, tissue remodeling, angiogenesis, parasite and bacterial invasion, rheumatoid arthritis and tumor cell invasion. The metalloproteinases are sub grouped into three categories: interstitial collagenases, type IV collagenases and stromelysins (transin). These enzymes are secreted as inactive proenzyme forms and must be activated by removal of an amino terminal fragment. Left unchecked, the activated metalloproteinases would destroy the structural integrity of tissues. Cells have involved mechanisms to protect themselves and their surroundings by excreting proteins that inactivate the metalloproteinases. Important representatives of this group of proteins are the tissue inhibitors of metalloproteinases (TIMP). TIMP-1, TIMP-2, TIMP-3 and TIMP-4 genes have been isolated and their role in tumor and normal cell invasion is being investigated^2-4. The production of TIMP in a small sample of cultured brain tumor explants and cell lines has been examined in tumor tissues, normal brain, low passage number tumor-derived cell lines and cultured normal astrocytes⁵. Another gene that could function to inhibit tumor cell invasion is the NM23-H1. This gene is expressed at higher levels in tumor tissues derived from primary site melanomas and breast carcinomas compared to distant metastases^6-8. Although a direct functional role for this gene in the establishment of metastases has not yet been experimentally determined, it appears that reduced levels of gene product could aid metastatic spread. It is unclear if the gene is involved in tumor cell invasion or in seeding of invaded tumor at distant sites. The cloned human NM23-H1 gene encodes nucleoside diphospate (NDP) kinase A⁹ and has a nucleic acid sequence homology to NDP-kinase from Drosophila melanogaster. NDP kinase functions in the synthesis of nucleotide triphosphates others than ATP.

Since NM23-H1 could be involved in tumor cell invasion inhibition, we have examined the levels of NM23-H1 in diffuse astrocytomas and glioblastoma multiforme (GMB).

METHOD

Tumor samples – 27 tumor tissues were removed from patients undergoing neurosurgical procedures at the Neurological Institute of Columbia University. A portion of the tumor was sent for histopathological diagnosis by a neuropathologist. Detailed reports of the tumor specimen were the basis for tumor classification. Another portion was snap-frozen in liquid nitrogen and stored at –70°C for subsequent molecular characterization. A subset of tumors was divided into a third portion and these tissues were established as tumor-derived cell lines. Cells for culture were dispersed in trypsin: EDTA and grown in plastic tissue culture dishes in Dulbecco's minimal essential medium containing 20% fetal bovine serum.

Extraction of RNA – Total celular RNA and DNA were simultaneously extracted from frozen tissue samples using guanidinium-isothiocyanate and cesium chloride gradients

Northern blot – Total RNA (10 mg per sample) from tumor tissues were size-fractionated through 1% agarose-formaldehyde gels and transferred to Amersham nylon filters. RNA immobilized on the filters was hybridized to 32P-labeled DNA probes and placed next to X-ray film for autoradiography. Levels of specific RNAs were determined by computer-assisted laser densitometry (Molecular Dynamics, Inc.) of the autoradiographs. Blot filters were sequentially hybridized to radiolabeled NM23-H1 and TIMP-1 probes in addition to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe to determine normal continually expressed gene RNA levels. The GADPDH control RNA levels were important to accurately compare the various samples to one another since they can reflect differences in RNA loading onto gels and spectrophotometric quantitation errors. The GAPDH levels were compared to ethidium bromide-stained total RNAs to insure the accuracy of using GAPDH RNA level standardization among samples.

Statistical analysis was carried out according to t-test for equality of means.

RESULTS

The bands from the autoradiographs are depicted in Figures 1A-4A and graphs comparing relative gene expression, based on computer-assisted laser densitometry of the autoradiographic signals, are presented in Figures 1B-4B. The relative levels of RNAs were different depending on the probe used in the blot hybridization. TIMP-1 RNA levels were more variable from tumor to tumor compared to levels of NM23-H1 RNA. GAPDH RNA levels were used to accurately determine the physical amount of sample loaded into each gel. Total RNAs were visualized after gel electrophoresis and documented by photography. The 28S and 18S ribosomal RNAs served as size markers to determine the size of the mRNAs that were detected as bands after Northern blot analysis. The ribosomal RNAs also served to document differences in the physical amounts of RNAs loaded per sample. There were variations in the amounts of RNA loaded per lane and these variations were in 100% correspondence with the variations detected after laser densitometry of bands generated using a GAPDH probe. For each RNA sample, absorbance values of TIMP-1-specific bands or NM23-H1-specific bands were divided by the absorbencies of GAPDH bands to generate relative gene expression values.

Figures 1A and ^B, ^4A and ^B depict the bands detected after Northern blot analyses of RNA extracted from 27 primary tumor tissues removed at time of surgery. Total cellular RNA was used in these analyses.

Figures 1A and ^B depicts RNAs from seven GBM samples.

Regarding to TIMP-1 gene expression: RNA sample (694) revealed a significant TIMP-1 increase relative to GAPDH. Lower levels were obtained in four samples according to laser densitometry and RNA sample (721) was significantly decreased relative to GAPDH.

Regarding to NM23-H1 gene expression: Higher RNA level was detected in one sample (437) while three samples (694, 721 and 727) were less than half of the mean densitometry ratios.

Figures 2A and ^B depicts RNAs from three diffuse astrocytomas (541,632,182) and four GBMs (555,657, 728,756).

TIMP-1 RNA levels in diffuse astrocytomas were extremely high compared to levels in all other samples analyzed in this blot. The GBMs samples either severely down-regulated expression of TIMP-1 or they shut off its expression completely.

In contrast to TIMP-1, NM23-H1 RNA levels were easily detected and there were no major differences in the levels of expression in tumors from the astrocytic lineage.

Figures 3A and ^B contains eight GBM RNA samples.

Two tumors contained high levels of TIMP-1 RNA (735,782). While six other tumors had low levels, less than half of the mean densitometry ratios.

NM23-H1 RNA levels in these samples were not as variable. Only one sample (788) contained high level of NM23-H1 compared to the mean. The other tumors did not appear too significantly up-regulate or down-regulate NM23-H1 gene expression at RNA levels.

^{Figures 4A} and ^B depicts RNAs from five diffuse astrocytomas.

There was variable TIMP-1 gene expression in this set of tumors. Sample (64) contained significantly higher amounts of TIMP-1 RNA compared to the other tumors in the series. Tumor sample (4,70) contained an intermediate amount of this RNA.

Statistical analyses were carried out according to t-test for equality of means (Tables 1, 2, 3).

TIMP-1 and diffuse astrocytoma showed that the mean ratio TIMP-1/GAPDH was in 2.17 and standard deviation (SD) 1.99. The confidence interval with level of significance of 5% presented with values from 0.5 and 3.83. TIMP-1 and GBM the mean ratio TIMP1/ GAPDH was 0.45 and SD 0.67. The confidence interval with level of significance of 5% presented limits of 0.13 and 0.77. Analysis of equality of variances (0.05 significance level was considered for all analysis) found to TIMP-1 gene the rejection of hypothesis of equality of variances, then it was used t test to equality of means taking in account unequal variances. The results showed that the means of these two groups were different (p<0.05).

Statistical analysis of NM23-H1 and diffuse astrocytoma showed that the mean ratio NM23-H1/GAPDH was 0.16 and SD 0.02. The confidence interval with level of significance of 5% presented limits of 0.11 and 0.22. NM23-H1 and GBM the mean ratio NM23-H1/GAPDH was 0.30 and SD 0.20. The confidence interval with level of significance of 5% presented limits of 0.20 and 0.40. Analysis of homogeneity of variances (level of confidence: 5%) found to NM23-H1 was not rejected the hypothesis of equality of variances, then it was used t test to equality of means taking in account same variances. The results showed that the means of two groups were the same (p> 0.05).

DISCUSSION

Invasion of normal brain tissue by a glial tumor cell has been a goal of many studies¹⁰. TIMP-1 interacts with metalloproteinases to inhibit its collagenase activity. The balance between collagenase and collagenase inhibitor appears to determine whether a tumor cell can invade adjacent tissue. Most tumors and normal cells produce a number of collagenases¹¹. Our aim was to determine it there was a differential invasiveness into adjacent normal brain as part of their malignant phenotype. Our initial approach to the study of TIMP-1 correlated with protein levels and anticollagenase activity in other tumor systems. There was a clear difference in the amounts of TIMP-1 RNA in histopathologically defined well-differentiated astrocytomas compared to malignant astrocytomas, we have quantitated the levels in 27 different samples from different patients. In the malignant astrocytomas, there was a significant trend toward the down-regulation of TIMP-1 gene expression as evidenced by the lower amounts of accumulated TIMP-1 RNA. This expression trend correlated with the invasive capacity of the tumor. Apodaca examined collagenases and TIMP protein excretes into culture medium of five malignant glioma cell lines and from cultured fetal astrocytes¹. In agreement with our RNA data, they detected collagenase and TIMP-1 activity in conditioned medium from all of these cells.

Halaka et al. examined conditioned medium from eight glioma explants cultured up to seven days⁵. They reported varying amounts of TIMP production and a tendency for invasive tumors to release less TIMP-1 into culture medium. Nakagawa et al. confirmed the TIMP-1 involvement with gliomas and meningiomas, they did protein experiments and their results are similar as ours¹². Nakano and his group published a very elegant study showing their conclusions about the relationship of metalloproteinases, TIMP-1 and TIMP-2, and also the interaction with PMA, TNFa, EGF, IL-1a, IL-6 and TGF b1. They did Northern blot analysis in glioma cells and chemoinvasion assay¹³.

Invasion is a complex interaction between metalloproteinases and inhibitors, in addition with the modulation of some growth factors and cytokines¹⁴. TIMP-1 RNA and NM23-H1 RNA in diffuse astrocytomas and GBM samples extracted from tissue culture are mostly the same, because it is not a good model for something which is very active. Other mechanisms are clearly involved and there is likely to be a continual balance between the metalloproteinases and their inhibitors¹⁵.

There are individual tumors that contained significantly higher amounts of TIMP-1 RNA. It is possible believe the existence of chemical mediators of TIMP-1 overexpression. It is supported by the fact of high levels of TIMP-1 RNA are influenced by external stimuli, internal cell regulatory control or a combination of both gene expression control mechanisms.

Since TIMP-1 levels are lower in specific malignant astrocytomas, this gene may be involved in regulating tumor cell invasion. It is possible that therapeutic regimens could be devised to deliver TIMP-1 protein or increase TIMP-1 gene expression to destroy the activity of the TIMP-1 target, collagenase, and ameliorate tumor cell invasion. Increased expression of TIMP-1 mediated by the addition of exogenous DNA can decrease the metastatic potential of the genetically engineered cell^16,17. The efficacy of TIMP-1 gene replacement therapy will be dependent on useful gene expression vectors that can be easily delivered to patients¹⁸.

The role of NM23-H1 in tumor cell invasion was determined by Liotta¹⁹. This gene is overexpressed in primary melanomas compared to their respective metastases and decreased in human primary infiltrating ductal breast carcinomas that exhibit lymph node involvement. RNA expression of this gene in glial tumors has not been evaluated. The levels of NM23-H1 in glial tumors samples were not as varied as those of TIMP-1 RNA in our data. We did not detect a significant trend toward low expression of NM23-H1 in the invasive astrocytomas. However, NM23-H1 was present in all glial tumor sample. Our results did not show any variation in NM23-H1 gene expression from diffuse astrocytomas and glioblastoma cells. Researchers have suggested that the NM23-H1 is responsible by the internal control of the invasion in malignant tumors^19-23.

The data presented in this paper suggest that TIMP-1 is a candidate inhibitor of brain tumor invasion. We presented samples from GBM in which low or undetectable amounts of TIMP-1 RNA suggest that one of the known mechanisms to inhibit tumor cell invasion is impaired in these tumors. Replacement of this gene product may be of therapeutic value to patients with brain tumors, slowing or statically inhibiting their disease^24,25.

NM23-H1 was present in the entire glioma samples, but it did not vary in diffuse astrocytomas and glioblastomas.

Received 21 November 2005, received in final form 11 May 2006. Accepted 25 May 2006.

This paper was supported by a Grant from CAPES (PhD in Science of Universidade Federal de São Paulo - EPM-UNIFESP, 2005).

Dr. José Augusto Nasser - Avenida Ataúlfo de Paiva 135/914 - 22440-032 Rio de Janeiro RJ - Brasil. E-mail: [e-mail]

• 1. Apodaca G, Rutka JT, Bouhana K, et al. Expression of metaloproteinases and metalloproteinases inhibitors by fetal astrocytes and glioma cells. Cancer Res 1990; 50:2322-2329.
• 2. Johnson MD, Housey GM, Kirshmeier PT, Weinstein IB. Molecular cloning of gene sequences regulated by tumor promoters and mitogens through protein kinase C. Mol Cell Biol 1987;7:2821-2829.
• 3. Boone TC, Johnson MJ, DeClerck YA. cDNA cloning and expression of a metalloproteinase inhibitor related to tissue inhibitor of metalloproteinases. Proc Natl Acad Sci USA 1990;87:2800-2804.
• 4. Groft LL, Muzik H, Rewcastlle NB, et al. Differential expression and localization of TIMP-1 and TIMP-4 in human gliomas. Br J Cancer 2001;85:55-63.
• 5. Halaka AN, Bunning RAD, Bird CC, Gibson M, Reynolds JJ. Production of collagenase and inhibitor (TIMP) by intracranial tumors and dura in vitro. J Neurosurg 1983;59:461-466.
• 6. Bevilacqua G, Sobel ME, Liotta LA, Steeg PS. Association of low NM23 RNA levels in human primary infiltrating ductal breast carcinomas with lymph node involvement and other histopathological indicators of high metastatic potential. Cancer Res 1989;49:5185-5190.
• 7. Leone A, Flatow U, King K, et al. Reduced tumor incidence, metastatic potential, and cytokine responsiveness of NM23 transfected melanoma cancer cells. Cell 1991;65:25-35.
• 8. Steeg PS, Bevilacqua G, Kopper L, et al. Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 1988;80: 200-204.
• 9. Gilles AM, Presecan E, Vonica A, Lascu I. Nucleoside diphospate kinase from human erythrocytes: structural characterization of the two olypeptide chains responsible for heterogeneity of the hexameric enzyme. J Biol Chem 1991;266:8784-8789.
• 10. Pasquier B, Pasquier D, N'Golet A, Panh MH, Courdec P. Extraneural metastasis of astrocytomas and glioblastomas. Cancer 1908;45:112-125.
• 11. Kachra Z, Beaulieu E, Delbecchi L, et al. Expression of matrix metalloproteinases and their inhibitors in human brain tumors. Clin Exp Metastasis 1999;17:555-566.
• 12. Nakagawa T, Kubota T, Kabuto M, et al. Production of matrix metalloproteinases and tissue inhibitor of metalloproteinases-1by human brain tumors. J Neurosurg 1994;81:69-77.
• 13. Nakano A, Tani E, Miyazaki K, Yamamoto Y, Furuyama J. Matrix metalloproteinases and tissue inhibitors of metalloproteinases in human gliomas. J Neurosurg 1995;83:298-307.
• 14. Paulus W, Tonn JC. Basement membrane invasion of glioma cells mediated by integrin receptors. J Neurosurg 1994;80:515-519.
• 15. Henessy C, Henry JA, May FEB, Westley BR, Angus B, Lennard TWJ. Expression of the antimetastatic gene NM23 in human breast cancer: an association with good prognosis. J Nat Cancer Inst 1991;83:281-285.
• 16. DeClerck YA, Perez N, Shimada H, Boone TC, Langely KE, Taylor SM. Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res 1992;52:701-708.
• 17. Liotta LA, Stetler-Stevenson WG. Tumor invasion and metastasis: an imbalance of positive and negative regulation. Cancer Res 1991;51:5054-5059.
• 18. Schneider J, Hoffman FM, Apuzzo MLJ. Cytokines and immunoregulatory molecules in malignant glial neoplasms. J Neurosurg 1992;77:265-273.
• 19. Liotta LA. Cell Invasion and metastasis. Scientific American 1992;2: 54-63.
• 20. Steeg PS. Metastasis suppressor genes. J Natl Cancer Inst 2004;96:419.
• 21. Rosengard AM, Krutzsch HCR, Sandeen MA, Margulies IMK, Liotta LA, Steeg PS. Reduced tumor incidence metastatic potential, and cytokine responsiveness of NM23-H1 transfected melanoma cell. Cell 1991;65: 25-35.
• 22. Korabiowska M, Honig JF, Jawien J, et al. Relationship of NM23 expression to proliferation and prognosis in malignant melanomas of the oral cavity. In Vivo 2005;19:1093-1096.
• 23. Li J, Zhou J, Chen G, et al. Inhibition of ovarian cancer metastasis by adeno-associated virus-mediated gene transfer of NM23-H1 in an orthotopic implantation model. Cancer Gene Ther 2006;13:266-272.
• 24. Kebebew E, Peng M, Reiff E, Duh QY, Clark OH, McMillan A. Diagnostic and prognostic value of angiogenesis-modulating genes in malignant thyroid neoplasms. Surgery 2005;138:1102-1109.
• 25. Van Waveren C, Sun Y, Cheung HS, Moraes CT. Oxidative phosphorylation dysfunction modulates expression of extracellular matrix-remodeling genes and invasion. Carcinogenesis 2006;27:409-418.

Publication Dates

• Publication in this collection
  11 Oct 2006
• Date of issue
  Sept 2006

History

• Received
  11 May 2006
• Accepted
  25 May 2006