Clonal analysis of tumor-infiltrating lymphocytes from human primary and metastatic liver tumors

Phenotypic and functional characteristics of tumor-infiltrating lymphocytes (TIL) obtained from human primary and metastatic liver tumors were studied. Lymphocytes isolated from 18 tumors and autologous (A) peripheral blood (6 cases) were phenotyped by 2-color flow cytometry and cloned in a limiting dilution system, which allows virtually all normal T lymphocytes to proliferate; 70-80% of fresh TIL were T cells (i.e., CD3+), and the ratio of CD4+/CD8+ cells was 1.2 in both primary and metastatic liver tumors. TIL contained significantly more CD56+ (NKHI+) cells, half of which were CD3+CD56+, CD3+CD25+ cells and CD3+HLA-DR+ cells, than A-PBL. The frequencies of proliferating T-cell precursors (PTL-p) and cytolytic T-lymphocyte precursors (CTL-p) reactive with K562, allogeneic tumor cells and autologous tumor cells, were determined. Mean PTL-p frequencies for TIL from hepatocellular carcinomas, cholangiocarcinomas and metastatic liver tumors were 0.52 (0.22-0.83), 0.10 (0.05-0.16) and 0.16 (0.01-0.30), respectively. The frequency of CTL-p with natural-killer-like activity was lower in TIL than in A-PBL. The frequency of CTL-p for autologous tumor cells in fresh TIL isolated from primary liver tumors was 0.02-0.13 and 12/81 clones were reactive against autologous tumor. In contrast, only 1/66 TIL clones obtained from colon carcinomas metastatic to liver showed autotumor reactivity. No clones reactive with autologous tumor were obtained from peripheral blood of patients with liver cancer. These data indicate that substantial differences in anti-tumor functions of TIL between primary and metastatic liver tumors exist, which can be detected at a clonal level.     

p53 mutations in primary human lung tumors and their metastases

In a total of 26 primary human lung tumors and 60 metastases derived from them, exons 5–8 of the p53 tumor suppressor gene were analyzed by single-strand conformation polymorphism and subsequent direct DNA sequencing of amplified DNA. Mutational inactivation of the p53 gene was identified in four of five squamous cell carcinomas, three of nine adenocarcinomas, and two of nine small-cell carcinomas, the overall incidence being 35%. Point mutations occurred at a similar incidence in exons 5–8, with a preference for G←T transversions. In seven of nine cases (78%), mutations were identical in the primary tumor and all of its metastases, indicating that in lung tumors, p53 mutations usually precede metastasis and that hematogenic and lymphogenic dissemination of tumor cells to other tissues is not associated with a selection against p53 inactivation. In one case, a kidney metastasis had the same mutation as the primary squamous cell carcinoma, whereas a liver metastasis had no mutation, indicating heterogeneity of the primary lung neoplasm and selective metastasis of mutated and nonmutated tumor cells to kidney and liver, respectively. Only in one liver metastasis was a mutation identified that was neither present in the primary lung tumor nor in a kidney metastasis, suggesting that occasionally p53 mutations occur after metastatic spread. © 1994 Wiley-Liss, Inc.     

Candidate mutator genes in mismatch repair-deficient thymic lymphomas: no evidence of mutations in the DNA polymerase delta gene

DNA mismatch repair (MMR) proteins recognize nucleotides that are incorrectly paired. Deficiencies in MMR lead to increased genomic instability reflected in an increased mutation frequency and predisposition to tumorigenesis. Mice lacking the MMR gene, Msh2, develop thymic lymphomas that exhibit much higher mutational frequencies than other Msh2(-/-) tumours and Msh2(-/-) normal thymic tissue, suggesting that an additional mutator may have been acquired in a tissue-specific manner. Clustered mutations observed exclusively in the thymic lymphomas suggests that a gene(s) associated with the replication machinery might have become altered during tumorigenesis. Based on mutation studies in Saccharomyces cerevisiae lacking Msh2 and DNA polymerase delta (DNA pol delta), we hypothesized that the acquisition of mutations in DNA pol delta could contribute to the hypermutator phenotype and tumorigenesis in Msh2(-/-) thymic tissue. Furthermore, previous reports have suggested that genes containing mononucleotide repeats are non-random mutational targets in the absence of MMR. Therefore, we sequenced all 26 exons of the DNA pol delta catalytic subunit, including the six exons containing mononucleotide repeats of >5 bp, from nine Msh2(-/-) thymic lymphomas and two wild-type controls. No DNA pol delta pathogenic mutations were found in the thymic lymphomas, although several DNA base differences compared with published DNA pol delta sequences were observed. We conclude, therefore, that inactivating mutations in DNA pol delta are not a contributing factor in the development of the hypermutator phenotype in MMR-deficient murine thymic lymphomas.     

Scatter factor expression and regulation in human glial tumors

Scatter factor (SF) (also known as hepatocyte growth factor [HGF]) is a cytokine that induces cell motility in vitro and angiogenesis in vivo. SF appears to be a determinant of the malignant phenotype in certain systemic cancers. We detected SF in extracts prepared from human gliomas, with the highest levels found in malignant tumors. Human glioblastoma cells expressed both SF and its receptor (c-met protein) in vivo, as demonstrated by immunohistochemistry. Consistent with these observations, we found moderate to high levels of production of immunoreactive and biologically active SF by cultured human glioblastoma cells (3 of 8 lines) and by neural microvascular endothelial cells (NMVEC) (3 of 3 lines). SF stimulated the proliferation of glioblastoma and NMVEC cell lines by paracrine or autocrine mechanisms. Conditioned medium (CM) from both glioblastoma and NMVEC cells contained SF-inducing factor (SF-IF) activity, defined by its ability to stimulate SF production in an indicator cell line (MRC5 human fibroblasts). This activity consisted of a high-molecular-weight (>30 kDa), heat-sensitive component and a low-molecular weight (<30 kDa), heat-stable component. Furthermore, glioblastoma CM stimulated NMVEC SF production, and NMVEC CM stimulated glioblastoma cell SF production, by 3- to 6-fold in each case. Our findings demonstrate that SF-dependent interactions between glioma cells, and between glioma cells and endothelium, can contribute to the heterogeneous proliferative and angiogenic phenotypes of malignant gliomas in vivo. © 1996 Wiley-Liss, Inc.     

Vimentin and glial fibrillary acidic protein in human brain tumors

Using the peroxidase-antiperoxidase (PAP) technique, we examined 35 primary brain tumors for expression of vimentin and GFAP. Both low-grade and high-grade astrocytomas contained vimentin-positive and GFAP-positive cells. Ependymomas also stained for both markers. In gliosarcomas, the glioblastomatous portions stained like astrocytomas, while only vimentin stain was seen in the fibrosarcomatous portions. Medulloblastomas and oligodendrogliomas were negative for both vimentin and GFAP, while meningiomas contained scattered areas of vimentin-positive cells. These results suggest that the expression of vimentin and GFAP is mostly confined to glial-derived tumors, and that vimentin can potentially be a useful marker for distinguishing undifferentiated GFAP-negative glial tumors and mesenchymal tumors from primitive neuroectodermal tumors.     

p16 involvement in primary bladder tumors: analysis of deletions and mutations

Different studies have proposed that genetic alterations leading to inactivation of a tumor suppressor gene on chromosome 9 is an important early event in bladder tumorigenesis. Recent reports have described the p16 gene as the main target. In order to better define its role, we studied 9p21 deletions by microsatellite analysis and its coding sequence. Forty-eight percent of the 44 samples we studied showed LOH surrounding p16. Three of these 44 samples displayed point mutations in p16 and three others were suspected of homozygous deletion. These results suggest that simultaneous loss of both p16 alleles, by point mutation or homozygous deletion, seems to be infrequent in bladder tumors.     

Immunolocalization of granulocyte-colony-stimulating factor in human glial and primitive neuroectodermal tumors

Granulocyte-colony-stimulating factor (G-CSF) is a hematopoietic cytokine that regulates the differentiation of myeloid progenitors and the function of mature neutrophils. It is produced in vitro by monocytes/macrophages, mesothelial cells, fibroblasts and endothelial cells after appropriate induction by inflammatory mediators like IL-I and TNF. Normal as well as tumorous glial cells can also be induced to produce CSFs in vitro. However, little is yet known about the in vivo expression of G-CSF as a mediator in inflammation and malignancy withiin the human central nervous system. The aim of the present study was to investigate by immunostaining the expression of the G-CSF protein within non-tumorous and tumorous glial tissues, and primitive neuroectodermal tumors. Using the murine monoclonal anti-G-CSF TM 82/60 antibody, we found high G-CSF expression in astrocytoma WHO grades I and II and reactive brain tissue, low expression in astrocytoma WHO grade III, and none in glioblastoma, oligodendroglioma WHO grades II and III, none medulloblastoma. In consecutive sections of the tissue samples, G-CSF protein was localized in GFAP-positive glial cells, but not in macrophages/microglial cells, which expressed HLA-DR, detected by the antibody CR3/43. Computer-assisted microdensitometric evaluation of the intensity of immunostaining for G-CSF and statistic analysis of the data revealed significant differences between the diagnostic entities studied (p < 0.0001). We conclude that in vivo expression of G-CSF is a characteristic of reactive as well as tumorous astrocytes, with the latter losing this feature at higher degrees of dedifferentiation. © 1994 Wiley-Liss, Inc.     

High-resolution genome-wide mapping of genetic alterations in human glial brain tumors

High-resolution genome-wide mapping of exact boundaries of chromosomal alterations should facilitate the localization and identification of genes involved in gliomagenesis and may characterize genetic subgroups of glial brain tumors. We have done such mapping using cDNA microarray-based comparative genomic hybridization technology to profile copy number alterations across 42,000 mapped human cDNA clones, in a series of 54 gliomas of varying histogenesis and tumor grade. This gene-by-gene approach permitted the precise sizing of critical amplicons and deletions and the detection of multiple new genetic aberrations. It has also revealed recurrent patterns of occurrence of distinct chromosomal aberrations as well as their interrelationships and showed that gliomas can be clustered into distinct genetic subgroups. A subset of detected alterations was shown predominantly associated with either astrocytic or oligodendrocytic tumor phenotype. Finally, five novel minimally deleted regions were identified in a subset of tumors, containing putative candidate tumor suppressor genes (TOPORS, FANCG, RAD51, TP53BP1, and BIK) that could have a role in gliomagenesis.     

p53 mutations in nonastrocytic human brain tumors

Genomic DNA from 51 primary human brain tumors was screened for the presence of mutations in the tumor suppressor gene, p53, using the polymerase chain reaction and single strand conformation polymorphism analysis, followed by direct DNA sequencing. Mutations leading to an amino acid change were found in 2 of 17 (12%) oligodendrogliomas and 2 of 19 (11%) medulloblastomas but none of 15 ependymomas. Sites of mutations were in exon 5 (codon 141), exon 6 (codon 193 and 213), and exon 7 (codon 246). In addition, there were silent mutations in exon 6 (codon 213) in one oligodendroglioma and in one ependymoma. This study points to the possible role of the p53 tumor suppressor gene in some central nervous system neoplasms of divergent histogenesis.     

p53 mutations in human lung tumors.

Mutation of one p53 allele and loss of the normal p53 allele [loss of heterozygosity (LOH)] occur in many tumors including lung cancers. These alterations apparently contribute to development of cancer by interfering with the tumor suppressor activity of p53. We directly sequenced amplified DNA in the mutational hot spots (exons 4-8) of p53 in DNA samples from 40 lung cancers. Most (31 of 40) samples were preselected for LOH in the region of p53. We detected 23 p53 mutations within these exons in 22 lung cancers; no p53 mutations were found in normal tissue of the patients. One-half of the mutations were G to T transversions on the nontranscribed strand, consistent with mutagenesis by tobacco smoke. Mutations of C to A on the nontranscribed strand, which would result from G to T mutations on the transcribed strand, were detected only in one sample. Three of 23 mutations were nonsense mutations; to date, nonsense mutations of p53 have not been reported in lung cancer. Mutation of this p53-coding region was detected in 20 of 27 small cell lung cancer samples, representing a 70% occurrence. Mutation of the p53 gene is apparently very frequent in small cell lung cancers. When LOH in the p53 region could be determined, complete concordance occurred between a sample having both a p53 mutation and LOH in the region of p53 (18 of 18 samples). Twelve samples of lung cancer had LOH in the region of p53, but the samples had no detectable p53 mutations, suggesting either alterations outside the known mutational hot spots of p53 or alterations of another unidentified tumor suppressor gene in the region of p53.     

Occurrence of Ki-ras and p53 mutations in primary colorectal tumors

Twenty-four colorectal tumors and their paired mucosa were examined for allele loss on chromosomes 5 and 17, using polymerase chain reaction-amplified DNA, and for mutations in the Ki-ras and p53 genes. In addition, 19 benign polyps were analysed for mutations in the p53 gene. RFLP analysis of the long arm of chromosome 5 indicated an allele loss frequency of 47% for malignant tumors, a value somewhat higher than previously observed. Examination of the short arm of chromosome 17 indicated an allele loss of 60%. Sequence analysis of the p53 gene in colorectal tumors indicated that 64% contained a mutation. All tumors showing allele loss on chromosome 17 were mutant for p53, suggesting that mutation of one p53 allele precedes the hemizygotic loss of the wild-type allele. Sequence analysis of the p53 gene in 19 benign polyps, devoid of severe dysplasia, did not reveal any mutants, suggesting that the mutation of one p53 allele is an event that takes place after polyp formation. Ki-ras mutations were observed in 48% of the tumors examined. All tumors which were mutant for Ki-ras, except for one, were also mutant for p53.     

Gene expression abnormalities in human glial tumors identified by gene array

Novel genes specific for human oligodendroglioma and glioblastoma multiforme (GBM) were detected using the gene array analysis [18,376 genes Gene Discovery Array (GDA) from Incyte Genomics, Inc.]. Eleven genes were chosen based on the highest ratios of differential expression identified by GDA between histologically normal adjacent tissue and brain tumor tissue. The differential expression of those 11 genes was verified by semiquantitative RT-PCR and Northern analysis on 22 samples of glial and other tumors of the brain, as well as of normal embryonic and adult brain tissue. Gene no. 5 (an EST) was more expressed by GDA analysis in histologically normal adjacent brain tissue than in the corresponding oligodendroglioma. By RT-PCR, this gene was expressed in a number of brain tumors but not in normal adult and embryonic brain. By GDA analysis, gene no. 7 (oligophrenin-1) gave the highest ratio compared to other genes in brain tissue adjacent to the GBM vs. GBM. By RT-PCR, oligophrenin-1 was expressed in tumors and tumor-adjacent tissue, whereas meningioma and corpus callosum were negative. Gene no. 11 (an EST) was expressed only in brain tumors but not in normal brain by Northern analysis (message size 1.5 kb) and RT-PCR. GDA analysis successfully identified genes preferentially expressed in brain tumors, which was confirmed by Northern analysis and semiquantitative RT-PCR. The validity of gene arrays for tumor-specific gene discovery is discussed. Study of differential gene expression in glial tumors should help identify the mechanism/s of transformation of normal glial cells to malignant.     

Presence and significance of oxytocin receptors in human neuroblastomas and glial tumors

To determine whether oxytocin (OT) could be added to the list of growth factors acting on neoplastic cells of nervous origin, we investigated the presence of oxytocin receptors (OTR) in human primary neuroblastomas and glioblastomas and related cell lines. OTR were demonstrated both at mRNA level (using a RT-PCR procedure) and at protein level (using immunocytochemical and immunofluorescence procedures). In order to clarify whether OT exerts any biological effect on these tumors through OTR, we also studied cell proliferation in 3 human neuroblastoma cell lines (SK-N-SH, SH-SY5Y, IMR-32) and one human anaplastic astrocytoma cell line (MOG-G-UVW) treated with OT 1 nM to 100 nM for 48 and 96 hr. At these doses, a dose-dependent inhibitory effect on cell proliferation was demonstrated. This inhibition was accompanied by a significant increase in the intracellular concentration of cAMP, which we have reported to be the intracellular mediator of the OT anti-proliferative effect in breast-carcinoma cell lines. Our data indicate that specific OTR are present in human neuroblastomas and glioblastomas. Through these receptors, OT could inhibit cell proliferation and modulate tumor growth. Int. J. Cancer 77:695–700, 1998. © 1998 Wiley-Liss, Inc.     

Hypoxia is important in the biology and aggression of human glial brain tumors.

We investigated whether increasing levels of tissue hypoxia, measured by the binding of EF5 [2-(2-nitro-1-H-imidazol-1-yl)-N-(2,2,3,3,3-pentafluoropropyl) acetamide] or by Eppendorf needle electrodes, were associated with tumor aggressiveness in patients with previously untreated glial brain tumors. We hypothesized that more extensive and severe hypoxia would be present in tumor cells from patients bearing more clinically aggressive tumors. Hypoxia was measured with the 2-nitroimidazole imaging agent EF5 in 18 patients with supratentorial glial neoplasms. In 12 patients, needle electrode measurements were made intraoperatively. Time to recurrence was used as an indicator of tumor aggression and was analyzed as a function of EF5 binding, electrode values and recursive partitioning analysis (RPA) classification. On the basis of EF5 binding, WHO grade 2 tumors were characterized by modest cellular hypoxia (pO2s approximately 10%) and grade 3 tumors by modest-to-moderate hypoxia (pO2s approximately 10%- 2.5%). Severe hypoxia (approximately 0.1% oxygen) was present in 5 of 12 grade 4 tumors. A correlation between more rapid tumor recurrence and hypoxia was demonstrated with EF5 binding, but this relationship was not predicted by Eppendorf measurements.     

An immunocytochemical comparison of glial fibrillary acidic protein,S-100p and vimentin in human glial tumors

Immunostaining patterns of glial fibrillary acidic protein (GFAP), S-100 protein (S-100p) and vimentin were studied using immunohistochemical techniques on 48 paraffin embedded glial tumors. GFAP was positive in all tumor cases except in two oligodendrogliomas. S-100p was found in most astroglial tumors and in half of the oligodendrogliomas. Vimentin was positive in many astrocytomas but in no oligodendrogliomas. Most astroglial tumors showed similar immunoreactivity for GFAP and S-100p. Fibrillary processes, however, showed stronger and more crisp staining with anti-GFAP than with anti-S-100p, whereas cell nuclei were labeled only for S-100p. Vimentin was localised mainly in juxtanuclear positions.In many astrocytomas with different degrees of malignancy co-expression of GFAP, S-100p and vimentin was found. The presence of GFAP and S-100p was not correlated with the degree of differentiation in astrocytomas. Vimentin was more positive in anaplastic astrocytomas but this finding was not statistically significant. It seems that GFAP is a superior marker to S-100p and vimentin in the identification of human gliomas.     

Overexpression of regenerating islet-derived 1 alpha and 3 alpha genes in human primary liver tumors with beta-catenin mutations

The Wnt/beta-catenin signaling pathway is activated in many human hepatocellular carcinomas (HCC). We tried to identify the genes involved in carcinogenesis and progression of HCC with beta-catenin mutations. We used PCR-based subtractive hybridization to compare gene expression between malignant and benign components of a human HCC occurring in pre-existing adenoma activated for beta-catenin. Two of the genes identified belong to the Regenerating gene (REG) family. They encode the Regenerating islet-derived 3 alpha (REG3A/HIP/PAP/REG-III) and 1 alpha (REG1A) proteins, both involved in liver and pancreatic regeneration and proliferation. Using siRNA directed against beta-catenin, we demonstrated that REG3A is a target of beta-catenin signaling in Huh7 hepatoma cells. The upregulation of REG3A and REG1A expression is significantly correlated to the beta-catenin status in 42 HCC and 28 hepatoblastomas characterized for their beta-catenin status. Thus, we report strong evidence that both genes are downstream targets of the Wnt pathway during liver tumorigenesis.     

Collagen biosynthesis in human neuroblastoma cell lines: evidence for expression of glial cell properties

An analysis was done on the synthesis of collagen, an extracellular matrix protein synthesized by Schwann cells, by several human neuroblastoma lines. Cultured cells were incubated in the presence of L-[2,3-3H]proline, and the collagens synthesized and secreted into the culture medium were analyzed by electrophoresis on acrylamide gels, ion exchange chromatography, and immunoprecipitation. The amount of collagen secreted by 4 cell lines tested represented less than 3% of the total protein synthesis, indicating a low degree of collagen biosynthesis by this tumor type. Analysis of collagen types secreted by all cell lines revealed the presence of a high-molecular-weight collagen precursor (Mr = 165,000) identified as type IV procollagen. In addition, several cell lines synthesized stromal type I and type III collagens. These studies show that neuroblastoma cells produce collagenous proteins including basement membrane collagen (type IV) and stromal collagens (types I and III), indicating that these cells express properties of glial cells such as Schwann cells.