Evolution of VHL tumourigenesis in nerve root tissue.

Haemangioblastomas are the key central nervous system manifestation of von Hippel-Lindau (VHL) disease, which is caused by germline mutation of the VHL gene. We have recently shown that 'tumour-free' spinal cord from patients with VHL disease contains microscopic, poorly differentiated cellular aggregates in nerve root tissue, which we descriptively designated 'mesenchymal tumourlets'. Here we have investigated spinal cord tissue affected by multiple tumours. We show that a small subset of mesenchymal tumourlets extends beyond the nerve root to form proliferative VHL-deficient mesenchyme and frank haemangioblastoma. We thus demonstrate that tumourlets present potential, but true precursor material for haemangioblastoma. We further show that intraradicular tumourlets consist of scattered VHL-deficient cells with activation of HIF-2alpha and HIF-dependent target proteins including CAIX and VEGF, and are associated with an extensive angiogenic response. In contrast, activation of HIF-1alpha was only observed in the later stages of tumour progression. In addition, ultrastructural examination reveals gradual transition from poorly differentiated VHL-deficient cells into vacuolated cells with a 'stromal' cell phenotype. The evolution of frank haemangioblastoma seems to involve multiple steps from a large pool of precursor lesions.

Epididymal cystadenomas and epithelial tumourlets: effects of VHL deficiency on the human epididymis.

Although epididymal cystadenomas (ECAs) are among the most frequent VHL disease-associated tumours, fundamental questions about their pathogenesis have remained unanswered. Classification of ECAs is controversial, and the cell of origin is unknown. It is also unknown whether ECAs-like other VHL disease-associated tumours-arise as a result of VHL gene inactivation, and whether ECAs exhibit subsequent activation of hypoxia-inducible factor HIF. Moreover, the morphological spectrum of earliest ECA formation is unknown. In a detailed molecular pathological analysis of a series of epididymides collected from VHL patients at autopsy, we found that ECAs are true neoplasms that arise secondary to inactivation of the wild-type copy of the VHL gene, followed by early and simultaneous activation of HIF1 and HIF2 associated with up-regulation of downstream targets, including CAIX and GLUT-1. The observations also indicate that ECA formation evolves from a variety of microscopic epithelial tumourlets, and that these tumourlets are confined to the efferent ductular system. Although genetic and immunohistochemical analysis of precursor structures consistently revealed VHL gene inactivation and activation of HIF in the precursor lesions, only a small subset appears to progress into frank cystadenoma. Thus, ECA tumorigenesis in VHL disease shares fundamental principles with tumorigenesis in other affected organ systems.

Site-specific immune response to implanted gliomas.

OBJECT: Immunotherapy for glioblastoma has been uniformly ineffective. The immunological environment of the brain, with its low expression of major histocompatibility complex (MHC) molecules and limited access for inflammatory cells and humoral immune effectors due to the blood-brain barrier (BBB), may contribute to the failure of immunotherapy. The authors hypothesize that brain tumors are protected from immune surveillance by an intact BBB at early stages of development. To investigate the immunological characteristics of early tumor growth, the authors compared the host response to a glioma implanted into the brain and into subcutaneous tissue. METHODS: Samples of tumors growing in the brain or subcutaneously in rats were obtained for 7 consecutive days and were examined immunohistochemically for MHC Class I & II molecules, and for CD4 and CD8 lymphocyte markers. Additionally, B7-1 costimulatory molecule expression and lymphocyte-specific apoptosis were examined. CONCLUSIONS: On Days 3 and 4 after implantation, brain tumors displayed significantly lower MHC Class II expression and lymphocytic infiltration (p < 0.05). After Day 5, however, no differences were detected. The MHC Class II expressing cells within the brain tumors appeared to be infiltrating microglia. Minimal B7-1 expression combined with lymphocyte-specific apoptosis were detected in both brain and subcutaneous tumors. Low MHC Class II expression and low lymphocytic infiltration at early time points indicate the importance of the immunologically privileged status of the brain during early tumor growth. These characteristics disappeared at later time points, possibly because the increasing perturbation of the BBB alters the specific immunological environment of the brain. The lack of B7-1 expression combined with lymphocyte apoptosis indicates clonal anergy of glioma-infiltrating lymphocytes regardless of implantation site.

Heterogeneity of subcellular localization of p53 protein in human glioblastomas.

Immunohistochemical analysis of the p53 protein in human glioblastomas with known genetic profiles of p53 mutations and allele losses on chromosome 17p demonstrated a heterogeneous pattern of subcellular compartmentalization of the p53 protein. Tumors with a single wild type copy of the p53 gene but with allelic deletions on chromosome 17p exhibit nuclear and/or cytoplasmic accumulation of p53, whereas tumors with both copies of the wild type gene and no allele losses on chromosome 17 do not accumulate p53. Glioblastomas with one normal and one mutated copy of the p53 gene and allelic deletions on 17p distal to p53, on the other hand, show predominantly cytoplasmic staining, probably originating from the wild type p53 protein. Furthermore, tumors with mutations in the same codon of p53 display quite different intracellular distribution suggesting that, in addition to the genotype of p53, the intracellular microenvironment of a particular tumor is important in determining the subcellular localization of the p53 protein.

Allelic deletions on chromosome-17 and mutations in the p53 gene in tumors metastatic to brain.

Metastasis associated with tumor progression denotes more aggressive tumor behavior, more malignant histology, and worsening patient prognosis. Brain is one of the common sites to which solid tumors metastasize. Since mutations in the tumor suppressor gene p53 are associated with tumor progression to malignancy in various cancers, we examined the molecular genetic profile of the p53 gene and also analyzed allelic losses of various genes on the short arm of chromosome 17 (17p) in 10 metastatic brain tumors (4 breast, 3 renal, 1 lung, 1 esophageal, and 1 squamous cell carcinoma) and in two of the primary tumors (I breast, 1 renal) corresponding to two of the 10 metastases. Six of the 10 metastatic tumors (4/4 breast, 1/1 esophageal and 1/1 squamous cell carcinoma) contained allelic loss and/or mutations of the p53 gene. The p53 gene profile was identical in both of the primary tumors and their corresponding metastases that were examined. If borne out by a larger series of analysis on tumors, especially from breast, as well as from other organs, detection of chromosome 17/p53 alterations may be of substantial clinical significance in predicting the metastatic potential of primary tumors.

Evidence for the involvement of a potential second tumor suppressor gene on chromosome 17 distinct from p53 in malignant astrocytomas.

Molecular analysis of malignant astrocytomas demonstrated three distinct groups of tumors with chromosome 17p abnormalities, which include (a) deletion of the p53 locus (17p13.1) and mutations in the remaining allele, (b) deletion of the p53 locus but no detectable mutations in the remaining allele, and (c) deletions not including the p53 locus but mutations in one of the alleles. Furthermore, deletion mapping analysis demonstrated allelic loss of genes distal to D17S28/D17S5 markers (17p13.3) in group C tumors. The loss of heterozygosity of genes on chromosome 17 without detectable mutation (group B) or deletion (group C) in the p53 gene implies the presence of a second tumor suppressor gene in the telomeric region of 17p, the homozygous functional inactivation of which may play a role, either alone or in conjunction with p53, in the initiation and/or progression of astrocytic neoplasms.

Implantation of dispersed cells into primate brain.

Although several experimental therapies such as dopaminergic cell implantation in parkinsonian models and intratumoral placement of lymphokine-activated killer cells require intracerebral deposition of dispersed cell suspensions, a successful technique of needle implantation of cells into primate brain has not been demonstrated. The authors have sought to establish a stereotaxic technique to predictably deposit dispersed cells in primate brain. Human lymphocytes were cultured in recombinant interleukin-2, labeled with sodium 51 chromate (51Cr), and stereotaxically injected into the frontal white matter of six anesthetized rhesus monkeys. A 10-microliters aliquot of cell suspension (2 X 10(7) cells/ml) was deposited 16 mm deep to the dura at 5 microliters/min via Hamilton No. 22s or 26s needles. Five control aliquots were counted for each injection. Reflux out of the needle track was absorbed on gauze, and the recovered cells were counted. The animals were sacrificed 1 hour after implantation and the brain was removed and sectioned such that the cortex and white matter along the needle track were separate. The tissue sections were then counted. Recovery was expressed as the percentage of total injected radioactivity (cpm) that was present in each brain section. Two additional injected hemispheres were processed for autoradiography and histological study. Cell recovery in the brain (mean +/- standard deviation) was 87.2% +/- 13.9% (3.3% +/- 4.9% in cortex and 83.9% +/- 15.9% in white matter). The autoradiograms and histological examination showed a dense accumulation of radioactivity (cells) at the target site and minimal radioactivity (cells) in the needle track. Accurate intracerebral deposition of dispersed cells in primates was achieved with the technique described. This knowledge permits reliable stereotaxic implantation of cells into the brains of nonhuman primates and humans for investigation and therapy.

Morphology of interleukin-2-stimulated human peripheral blood mononuclear effector cells killing glioma-derived tumor cells in vitro.

This is the first morphological study of interleukin-2-stimulated human peripheral blood mononuclear (PBM) cells resulting in lymphokine-activated killer (LAK) cell activity against human glioma-derived tumor cells in vitro, in which high-resolution differential interference video light microscopy, scanning electron microscopy, and transmission electron microscopy were used. A subset of cells within the LAK cell population are the effector cells and have an asymmetric cellular architecture characteristic of cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells. Upon binding to target cells, the LAK effector cell nucleus is positioned away from the target cell, whereas the granules, Golgi apparatus, and microtubules orient toward the target cell. These LAK-glioma cell conjugates form very tight plasma membrane bonds with numerous interdigitations, and vesicles were found in the small extracellular spaces between the cells. This morphology was not observed in unstimulated PBM-glioma cell co-cultures. Glioma-derived cells react to LAK effector cells by blebbing, becoming round, and rapidly detaching from the substrate. The injured glioma-derived cells had a highly condensed cytoplasm and chromatin, lobular nucleus, and severe plasma membrane blebs, which are consistent with an apoptotic rather than an osmotic lysis mechanism of cell death. This study provides morphological evidence that supports a common cytotoxic mechanism for CTLs, NK cells, and LAK effector cells. The cytotoxic mechanism is based on the local exocytosis of vesicles by the effector cell into the small extracellular space between the effector-target cell conjugate. Granules found in CTLs, NK cells, and LAK cells contain a pore-forming protein that inserts holes in the target cell's plasma membrane through which a lethal substance(s) not yet identified is thought to enter the cell.

Lymphokine activated killer (LAK) cell mediated killing of human glioma: effect of pretreating glioma with various membrane modifying agents.

The killing of human glioma by lymphokine activated killer (LAK) cells was studied. LAK cells generated by culturing recombinant interleukin-2 (IL-2) with human peripheral blood lymphocytes (PBL) obtained from normal volunteers markedly lysed allogeneic glioma grown in tissue culture. Susceptibility of glioma to lysis by LAK cells was abrogated by pretreating the glioma cells with trypsin or chymotrypsin, but was unaffected by pretreatment with hydrocortisone, neuraminidase, glycosidases or sodium periodate. These results suggest that the cell surface determinant on human glioma cells responsible for its tumor selective lysis by LAK is a protein sensitive to trypsin and chymotrypsin.

Human oncogene-transfected tumor cells display differential susceptibility to lysis by lymphokine-activated killer cells (LAK) and natural killer cells.

NIH 3T3 tertiary transfectants containing the N-ras or c-Ha-ras oncogenes derived from human tumors were tested for susceptibility to lymphokine-activated killer (LAK) cell and natural killer (NK) cell lysis. N-ras tertiary transfectants contained a human acute lymphocytic leukemia-derived N-ras oncogene. C-Ha-ras transfectants contained either the position 61-activated form of the oncogene (45.342, 45.322, and 45.3B2) or the position 12-activated form (144-162). In 4 hr 51Cr release assays, seven of seven in vivo grown human oncogene transfected NIH 3T3 fibroblasts were lysed by murine LAK effectors, whereas six of seven were lysed by human LAK effectors. There was no difference in susceptibility to lysis between cells transfected with the N-ras oncogene, the position 61 activated c-Ha-ras oncogene, or the position 12 activated c-Ha-ras oncogene. Cultured NIH 3T3 fibroblasts, as well as in vitro and in vivo grown NIH 3T3 tertiary transfectants were resistant to lysis by murine NK effectors and were relatively resistant (4/6 were not lysed) to lysis by human NK effectors. We conclude that human oncogene-transfected tumors are susceptible to lysis by both murine and human LAK cells while being relatively resistant to lysis by murine and human NK cells. Different oncogenes or the same oncogene activated by different point mutations do not specifically determine susceptibility to lysis by LAK or NK. Also the presence of an activated oncogene does not appear to be sufficient for inducing susceptibility to these cytotoxic lymphocyte populations.

Antitumor reactivity in vitro and in vivo of lymphocytes from normal donors and cancer patients propagated in culture with T cell growth factor (TCGF).

Peripheral blood lymphocytes (PBL) from 38 normal donors and from 27 cancer patients were propagated in bulk cultures for 3-6 weeks using T cell growth factor (TCGF). In addition, cultures derived from lymphocyte preparations enriched for or depleted of natural killer (NK) cells and several clones of cultured cells were studied. The following main observations were made: (a) PBL of both patients and healthy donors could be expanded to large numbers (up to 2500-fold); (b) CLC derived from unfractionated PBL exhibited intermediate levels of cytotoxic activity against autologous and allogeneic fresh lung tumor cells and strong cytotoxicity toward several cultured adherent tumor cells; (c) whereas cultures originated from populations enriched for NK cells were highly cytotoxic against both adherent tumor target cells and against an NK-sensitive leukemic cell line (K562), cultures derived from populations depleted of NK cells were preferentially cytotoxic to adherent target cells; (d) clones of CLC were also strongly cytotoxic, but 2 out of 3 clones tested showed a narrower spectrum of target cytotoxicity than that of uncloned CLC; (e) CLC, when mixed with two carcinoma cell lines, were able to inhibit tumor growth in nude mice.

A rapid technique for isolation of viable tumor cells from solid tumors: use of the tumor cells for induction and measurement of cell-mediated cytotoxic responses.

A rapid and simple technique for the isolation of viable tumor cells from human and mouse solid neoplasms is described. It consists of a 5 to 10-min treatment with trypsin-collagenase-DNase mixture, followed by mechanical disaggregation of the tumor tissue and subsequently by a brief centrifugation on a discontinuous Percoll gradient. With the tumors employed, this procedure usually requires less than 1 hr and results in preparations comprising greater than 80% tumor cells with viability of 80-90%. Cell-mediated cytotoxic response was measured with: (a) unsensitized lymphocytes freshly obtained from tumor-bearing hosts; (b) lymphocytes propagated in culture with T cell growth factor; and (c) lymphocytes stimulated in cocultures with autologous or syngeneic tumor cells. The cytotoxic activity was assessed in a modified [51Cr]-release assay adapted for solid tumor cells, allowing a long incubation period (24 hr) and the use of a low number (200-1000) of highly labeled target cells (2-10 counts/min/cell).

Natural cell-mediated cytotoxicity in vitro and inhibition of tumor growth in vivo by murine lymphoid cells cultured with T cell growth factor (TCGF).

Lymphoid cells obtained from the spleen, thymus, bone marrow, peripheral blood, and peritoneal exudate of normal mice (BALB/c, BALB/c nude, C57BL/6, C3H) and from spleens of mice bearing a transplantable lung carcinoma or primary mammary carcinoma were expanded in culture for 1-9 months, with an increase in cell number of 10(5)- to 10(6)-fold per month, in crude or lectin-depleted medium containing T cell growth factor (TCGF). All these cultured lymphoid cell (CLC) lines exhibited strong cytotoxic activity in vitro (assessed by 51Cr-release assays) toward a variety of freshly harvested and cultured syngeneic, allogeneic, and xenogeneic tumor target cells, both lymphoid and solid (including metastatic growths) in origin. Extensive killing was observed against tumor targets that were resistant to lysis by natural killer (NK) cells as well as to NK-sensitive tumor lines. Low levels of cytotoxic reactivity were also demonstrated against fresh and cultured normal lymphoid cells. The CLC had some characteristics of NK cells but also expressed some typical T cell markers. In local Winn-type neutralization assays, CLC delayed or completely inhibited the growth of lymphomas and carcinomas in syngeneic and allogeneic recipients. In mice with metastatic growth of a second-generation transplant of mammary carcinoma, CLC were shown to have some therapeutic effect when administered IV 1 day after cyclophosphamide. No significant beneficial action of IV administered CLC was observed in the absence of chemotherapy in mice implanted with a lung carcinoma. The possibilities of employing TCGF-propagated cytotoxic effector cells in adoptive immunotherapy of human malignancies are discussed.