Blocking chemokine responsive to gamma-2/interferon (IFN)-gamma inducible protein and monokine induced by IFN-gamma activity in vivo reduces the pathogenetic but not the antiviral potential of hepatitis B virus-specific cytotoxic T lymphocytes.

Using transgenic mice that replicate hepatitis B virus (HBV) at high levels in the liver as recipients of HBV-specific cytotoxic T lymphocytes (CTLs), we showed that the chemokines responsive to gamma-2/IFN-gamma inducible protein ([Crg2]IP-10) and monokine induced by interferon-gamma (Mig) are rapidly and strongly induced in the liver after CTL transfer. The transferred CTLs produce neither chemokine; rather, they activate (via the secretion of IFN-gamma) hepatocytes and nonparenchymal cells of the liver to produce (Crg2)IP-10 and Mig. Importantly, blocking these chemokines in vivo reduces the recruitment of host-derived lymphomononuclear cells into the liver and the severity of the liver disease without affecting the IFN-gamma-dependent antiviral potential of the CTLs. The finding that neutralization of these chemokines is associated with maintenance of antiviral effects but diminished tissue damage may be significant for the development of immunotherapeutic approaches for the treatment of chronic HBV infection.

Review: The chemokine receptor CXCR3 and its ligands CXCL9, CXCL10 and CXCL11 in neuroimmunity--a tale of conflict and conundrum.

The chemokine receptor CXCR3 and its ligands CXCL9, CXCL10 and CXCL11 in neuroimmunity - a tale of conflict and conundrum The chemokines CXCL9, CXCL10 and CXCL11 (also known as monokine induced by interferon-gamma, interferon-inducible protein-10 and interferon-inducible T cell alpha-chemoattractant, respectively) are structurally and functionally related molecules within the non-ELR CXC chemokine subgroup. These chemokines are generally not detectable in most non-lymphoid tissues under physiological conditions but are strongly induced by cytokines, particularly interferon-gamma, during infection, injury or immunoinflammatory responses. CXCL9, CXCL10 and CXCL11 each bind to a common primary receptor, CXCR3, and possibly to additional receptors. They are best known for their role in leucocyte trafficking, principally acting on activated CD4+ Th1 cells, CD8+ T cells and NK cells. An abundance of data demonstrates that CXCL9, CXCL10 and CXCL11 are produced in many diverse pathologic conditions of the central nervous system. More recent attention has focussed on the function of these chemokines in the central nervous system inflammation. The results of these studies have proven to be sometimes surprising and other times contradictory. Here we discuss the likely more subtle and perhaps divergent roles for these chemokines in the pathogenesis of neuroinflammatory diseases.

Systemic interferon-alpha regulates interferon-stimulated genes in the central nervous system.

The prime anti-viral cytokine interferon-alpha (IFN-alpha) has been implicated in several central nervous system (CNS) disorders in addition to its beneficial effects. Systemic IFN-alpha treatment causes severe neuropsychiatric complications in humans, including depression, anxiety and cognitive impairments. While numerous neuromodulatory effects by IFN-alpha have been described, it remains unresolved whether or not systemic IFN-alpha acts directly on the brain to execute its CNS actions. In the present study, we have analyzed the genes directly regulated in post-IFN-alpha receptor signaling and found that intraperitoneal administration of mouse IFN-alpha, but not human IFN-alpha, activated expression of several prototypic IFN-stimulated genes (ISGs), in particular signal transducers and activators of transcription (STAT1), IFN-induced 15 kDa protein (ISG15), ubiquitin-specific proteinase 18 (USP18) and guanylate-binding protein 3 (GBP3) in the brain. A similar temporal profile for the regulated expression of these IFN-alpha-activated ISG genes was observed in the brain compared with the peripheral organs. Dual labeling in situ hybridization combined with immunocytochemical staining demonstrated a wide distribution of the key IFN-regulated gene STAT1 transcripts in the different parenchyma cells of the brain, particularly neurons. The overall response to IFN-alpha challenge was abolished in STAT1 knockout mice. Together, our results indicate a direct, STAT1-dependent action of systemic IFN-alpha in the CNS, which may provide the basis for a mechanism in humans for neurological/neuropsychiatric illnesses associated with IFN-alpha therapy.

Essential role for interferon-gamma and interleukin-6 in autoimmune insulin-dependent diabetes in NOD/Wehi mice.

Experimental studies in vitro suggest that cytokines are important mediators in the pathogenesis of autoimmune insulin-dependent diabetes mellitus (IDDM). However, there is little evidence for the role of cytokines in vivo, either in humans or in the spontaneous animal models of IDDM such as the NOD mouse or BB rat. To address this question, we used the model of cyclophosphamide (CYP)-induced autoimmune diabetes in the NOD/Wehi mouse to examine for (a) the production of IFN-gamma and IL-6 from isolated islets, and (b) the effect of anti IFN-gamma or anti IL-6 monoclonal antibodies on the development of diabetes. After cyclophosphamide, the majority of these mice develop of mononuclear cell infiltrate (insulitis) which by 10-14 d is associated with beta cell destruction. IFN-gamma activity at low levels (2.7 +/- 0.3 U/ml) could be detected only in culture supernatants from islets isolated at day 7 post-cyclophosphamide. In contrast, IL-6 activity progressively increased from 457 +/- 44 U/ml at day 0 to 6,020 +/- 777 U/ml at day 10. Culture of islets with anti-CD3 monoclonal antibody resulted in a significant increase in IFN-gamma activity from 41 +/- 7 U/ml at day 0 to 812 +/- 156 U/ml at day 10. Mice given either anti-IFN-gamma or anti-IL-6 antibody had a significantly reduced (P less than 0.001) incidence of diabetes and especially with IFN-gamma, decreased severity of insulitis. We conclude that IFN-gamma and IL-6 have essential roles in the pathogenesis of pancreatic islet beta cell destruction in this model.

Macrophage/microglial-mediated primary demyelination and motor disease induced by the central nervous system production of interleukin-3 in transgenic mice.

Activated macrophage/microglia may mediate tissue injury in a variety of CNS disorders. To examine this, transgenic mice were developed in which the expression of a macrophage/microglia activation cytokine, interleukin-3 (IL-3), was targeted to astrocytes using a murine glial fibrillary acidic protein fusion gene. Transgenic mice with low levels of IL-3 expression developed from 5 mo of age, a progressive motor disorder characterized at onset by impaired rota-rod performance. In symptomatic transgenic mice, multi-focal, plaque-like white matter lesions were present in cerebellum and brain stem. Lesions showed extensive primary demyelination and remyelination in association with the accumulation of large numbers of proliferating and activated foamy macrophage/microglial cells. Many of these cells also contained intracisternal crystalline pole-like inclusions similar to those seen in human patients with multiple sclerosis. Mast cells were also identified while lymphocytes were rarely, if at all present. Thus, chronic CNS production of low levels of IL-3 promotes the recruitment, proliferation and activation of macrophage/microglial cells in white matter regions with consequent primary demyelination and motor disease. This transgenic model exhibits many of the features of human inflammatory demyelinating diseases including multiple sclerosis and HIV leukoencephalopathy.

Cytokines in viral diseases.

As in many other areas of cytokine biology, recent studies of the role of cytokines in viral disease reveal numerous complex interactions that in many instances may contribute directly to the development of pathology. For example, data from the rapidly evolving field of human retrovirology has shown that these viruses, as well as inducing the expression of many cellular cytokine genes, can be activated from latency and driven into replication/expression by the very same cytokines. The continuing rapid expansion of knowledge in the cytokine area augers well for eventual development of novel antiviral therapeutic strategies based on manipulation of the cytokine network.

Chemokines in the CNS: plurifunctional mediators in diverse states.

The past decade has witnessed the remarkable ascendance of chemokines as pivotal regulatory molecules in cellular communication and trafficking. Evidence increasingly implicates chemokines and chemokine receptors as plurifunctional molecules that have a significant impact on the CNS. Initially, these molecules were found to be involved in the pathogenesis of many important neuroinflammatory diseases that range from multiple sclerosis and stroke to HIV encephalopathy. However, more-recent studies have fuelled the realization that, in addition to their role in pathological states, chemokines and their receptors have an important role in cellular communication in the developing and the normal adult CNS. For example, stromal-cell-derived factor 1, which is synthesized constitutively in the developing brain, has an obligate role in neurone migration during the formation of the granule-cell layer of the cerebellum. Many chemokines are capable of directly regulating signal-transduction pathways that are involved in a variety of cellular functions, which range from synaptic transmission to growth. Clearly, the potential use of chemokines and their receptors as targets for therapeutic intervention in CNS disease might now have to be considered in the context of the broader physiological functions of these molecules.

Immunoinflammatory diseases of the central nervous system - the tale of two cytokines.

Cytokines are potent mediators of cellular communication that have crucial roles in the regulation of innate and adaptive immunoinflammatory responses. Clear evidence has emerged in recent years that the dysregulated production of cytokines may in itself be causative in the pathogenesis of certain immunoinflammatory disorders. Here we review current evidence for the involvement of two different cytokines, IFN-α and IL-6, as principal mediators of specific immunoinflammatory disorders of the CNS. IFN-α belongs to the type I IFN family and is causally linked to the development of inflammatory encephalopathy exemplified by the genetic disorder, Aicardi-Goutières syndrome. IL-6 belongs to the gp130 family of cytokines and is causally linked to a number of immunoinflammatory disorders of the CNS including neuromyelitis optica, idiopathic transverse myelitis and genetically linked autoinflammatory neurological disease. In addition to clinical evidence, experimental studies, particularly in genetically engineered mouse models with astrocyte-targeted, CNS-restricted production of IFN-α or IL-6 replicate many of the cardinal neuropathological features of these human cytokine-linked immunoinflammatory neurological disorders giving crucial evidence for a direct causative role of these cytokines and providing further rationale for the therapeutic targeting of these cytokines in neurological diseases where indicated.

Binding of cytoplasmic islet cell antibodies is blocked by human pancreatic glycolipid extracts.

With biochemical and enzymatic treatment of frozen sections of pancreas, we have previously shown that cytoplasmic islet cell antibodies (ICAs) react with carbohydrate determinants of islet cell glycoconjugates. As a first step toward purifying these glycoconjugates, human pancreas tissue was extracted in a mixture of chloroform and methanol, and the glycolipids were obtained by effecting a Folch partition. The protein pellet, lipid fraction, and glycolipid fraction so obtained were assessed for their ability to block the binding of ICAs to frozen sections of human pancreas, the effect being quantitated with a photometer. Only the glycolipid extract could block ICA binding, and blocking was dose dependent. Subfractionation of the glycolipid extract by hydrophobic interaction on C18 cartridges demonstrated that blocking activity resided in the fraction bound and eluted with methanol, consistent with the autoantigen being a glycolipid. Furthermore, the binding of an anti-islet cell ganglioside monoclonal antibody, 3G5, could be blocked with these extracts, whereas the binding of an anti-islet cell protein monoclonal antibody, 4F2, was unaffected. The major gangliosides of the pancreas were seen to be GM3 and GD3 by thin-layer chromatography (TLC). Fractions scraped and eluted from TLC plates were tested for their ability to block ICA binding to pancreatic sections. Neither GM3- nor GD3-containing fractions could block ICA binding; however, a fraction containing minor pancreatic gangliosides (including GM2) of monosialoganglioside mobility was a potent inhibitor of ICA binding to pancreas sections. TLC of a chloroform-methanol extract of human islets demonstrated that islets differentially express monosialogangliosides (especially GM2).

Reovirus infection enhances expression of class I MHC proteins on human beta-cell and rat RINm5F cell.

Viruses are implicated in the pathogenesis of beta-cell destruction in type I (insulin-dependent) diabetes. The aim of our study was to investigate whether reovirus 1 or reovirus 3, which are known to infect beta-cells and induce autoimmunity in susceptible mice, could alter the expression of the major histocompatibility complex (MHC) proteins by human beta-cells and rat insulinoma RINm5F cells. Forty-eight hours after infection of either human beta-cells or RINm5F cells with reovirus 1 or reovirus 3, cytopathic effects were noted. By flow-cytofluorometric analysis, infected RINm5F cells exhibited a seven- to eightfold increase in the surface expression of class I MHC proteins. Upregulation of class I MHC proteins on reovirus 3-infected RINm5F cells was inhibited by 80% after preexposure of the virus to reovirus 3 antiserum. When analyzed by double-indirect immunofluorescence microscopy, human beta-cells infected with reoviruses 1 or 3 also exhibited markedly increased levels of class I MHC proteins. Reovirus infection of human beta-cells or RINm5F cells was not accompanied by the induction of class II MHC proteins. These findings suggest that 1) in addition to direct cytopathic effects, reovirus infection may contribute to beta-cell destruction by increasing expression of class I MHC proteins and therefore reactivity with cytotoxic T-lymphocytes; and 2) some viruses may increase MHC protein expression independent of and before the action of cytokines (e.g., interferon-gamma and tumor necrosis factor) released by immunoinflammatory cells.

64,000-Mr autoantigen in type I diabetes. Evidence against its surface location on human islets.

The sera of type I (insulin-dependent) diabetic subjects are reported to contain autoantibodies against a 64,000-Mr protein identified in [35S]methionine biosynthetically labeled pancreatic islet cells. We have attempted to localize this autoantigen to the surface of the beta-cell and to define its properties. Sera from 10 newly diagnosed type I diabetic subjects, including five of the index sera originally used to identify the autoantigen, were shown to specifically precipitate a reduced protein of 67,000 Mr from Triton-solubilized, surface 125I-labeled cultured adult human islet and rat insulinoma (RINm5F) cells but not from fresh rat spleen cells. Further characterization revealed that this protein was bovine serum albumin (BSA) adsorbed to cells from fetal calf serum (FCS)-supplemented culture medium and precipitated by BSA antibodies present in many diabetic sera. No labeled proteins were specifically precipitated when surface 125I-labeled and solubilized human islet or RINm5F cells were precleared with anti-BSA immunoglobulins or when cells were first cultured in human serum. In contrast, a 64,000-Mr protein, clearly not BSA, was precipitated by diabetic globulins from human islets but not from RINm5F cells labeled with [35S]methionine. In addition, a protein of the same size as well as proteins of approximately 35,000, 43,000, 140,000, and 200,000 Mr were specifically precipitated by diabetic globulins from freshly isolated human islets solubilized in Triton X-100 and then labeled with 125I. These findings suggest that the 64,000-Mr antigen is not expressed on the surface of human islet cells, at least in culture, and therefore question its relevance as a target for islet cell surface antibodies in initiating beta-cell damage.(ABSTRACT TRUNCATED AT 250 WORDS)

MHC molecules and beta-cell destruction. Immune and nonimmune mechanisms.

Hyperexpression of major histocompatibility complex (MHC) molecules by islet cells is a prominent, early feature of islet pathology in insulin-dependent diabetes mellitus and concomitant with beta-cell failure after exposure of islets to specific cytokines or viruses. The transgenic expression of a class I MHC gene (H-2Kb) in the beta-cells of either syngeneic or allogeneic mice leads to beta-cell failure by a nonimmune mechanism. Several class II MHC transgenes, with one exception, have the same effect, but the expression of other transgenes that have products that are membrane proteins is not necessarily detrimental. Class I MHC molecules have been shown to interact directly with other membrane proteins. The inappropriate expression of MHC molecules could therefore interfere with key cellular functions. We postulate that the hyperexpression of MHC molecules in the beta-cell, e.g. in response to viruses, is a primary, nonimmune mechanism of beta-cell failure that precedes a secondary autoimmune response.

Interferon-gamma enhances the expression of the major histocompatibility class I antigens on mouse pancreatic beta cells.

Cell surface antigens of the major histocompatibility complex (MHC) play a crucial role in the initiation of immune reactions. To investigate whether the expression of MHC antigens on pancreatic islet cells could be altered, we have cultured mouse islets in the presence of interferon-gamma (IFN-gamma) and subsequently examined the levels of MHC antigen by indirect immunofluorescence using monoclonal antibodies. IFN-gamma induced a 10-fold increase in H-2K antigen expression on islet cells, the percentage of cells with detectable H-2K expression increasing from 24% to 98%. The effects of IFN-gamma on H-2D and la antigen expression were less marked, with only a twofold increase in mean fluorescence levels, the percentage of cells with detectable levels of expression increasing from 10% to 48% and 5% to 16%, respectively. Using double-indirect immunofluorescence, it was demonstrated that IFN-gamma enhanced expression of H-2K and H-2D antigens on beta-cells. However la-positive beta-cells were undetectable in the presence or absence of IFN-gamma. The ability of IFN-gamma to induce increased expression of H-2 antigens on beta-cells may represent a mechanism for targeting immune (cytotoxic) reactions to beta-cells, e.g., in autoimmune insulitis or allograft rejection.

Expression of class I MHC proteins on RIN-m5F cells is increased by interferon-gamma and lymphokine-conditioned medium.

To examine whether products of the immune system interact with the pancreatic beta-cell, rat insulinoma cells (RIN-m5F line) were cultured in the presence of conditioned medium from concanavalin A-activated mouse spleen cells (CAS medium). Indirect immunofluorescence and flow cytometry revealed that after culture in CAS medium, RIN-m5F cells had an 8- to 10-fold increase in class I major histocompatibility complex (MHC) proteins, whereas class II MHC proteins remained undetectable, and the level of insulin and/or insulin-like growth factor 1 receptors was unchanged. The stimulation of class I MHC expression on RIN-m5F cells by CAS medium could be mimicked by recombinant interferon-gamma. Analysis of 125I-surface-labeled cells by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography revealed that in the presence of CAS medium, there was a major increase in the expression of proteins of 48,000, 32,000, and 12,000 Mr and a minor increase in proteins of 17,000 and 9,000 Mr. Precipitation with monoclonal antibody identified the 48,000- and 12,000-Mr proteins as the class I MHC protein and beta 2-microglobulin, respectively. The ability of lymphokine-conditioned medium to increase the expression of RIN-m5F cell surface proteins, including the class I MHC proteins, provides a potential mechanism for enhancing the immune-mediated destruction of the beta-cell.

Reactivity to human islets and fetal pig proislets by peripheral blood mononuclear cells from subjects with preclinical and clinical insulin-dependent diabetes.

A simple, direct assay for T-lymphocyte reactivity to islet antigen(s) in human insulin-dependent diabetes mellitus (IDDM) should facilitate preclinical diagnosis and the evaluation of intervention therapy to avert autoimmune-mediated beta-cell destruction. In subjects with preclinical or clinical IDDM, we measured the reactivity of peripheral blood mononuclear cells (PBMCs) incubated over 6 days with either adult human islets or fetal pig proislets, or other fetal pig tissues, and with human insulin. With islets, the stimulation index (SI) of [3H]thymidine uptake by PBMCs exceeded the mean + 2SD of control subjects in 6 of 6 preclinical subjects (SI 8.7 +/- 3.7), 7 of 11 clinical subjects (SI 5.2 +/- 3.4), and 1 of 12 control subjects (SI 2.7 +/- 1.7); with insulin, the responses were less in frequency and magnitude, being 4 of 6 (2.7 +/- 1.6), 3 of 11 (2.2 +/- 1.1), and 0 of 12 (1.20 +/- 0.55), respectively. The mean responses to islets of PBMCs from preclinical and clinical subjects differed significantly from control subjects (P less than 0.02 by 2-tailed Kruskal-Wallis test). Secretion of granulocyte macrophage colony-stimulating factor by PBMCs over 6 days was assayed in the preclinical group and generally paralleled the uptake of [3H]thymidine. PBMC reactivity to islets appeared to be at least as sensitive a marker of preclinical IDDM as autoantibodies to a 64,000-Mr protein, presumably the enzyme glutamic acid decarboxylase, in fetal pig proislets. In conclusion, islet-reactive T lymphocytes in subjects with preclinical and clinical IDDM can be identified in bulk culture of PBMCs.(ABSTRACT TRUNCATED AT 250 WORDS)

Cytokines and chemokines as mediators of protection and injury in the central nervous system assessed in transgenic mice.

Cytokines and chemokines are potent biologic response molecules that play a key role in cellular communication in physiologic and pathophysiologic states. An understanding of the actions and roles of these molecules in CNS biology has been greatly facilitated by molecular genetic approaches that permit the targeted manipulation of gene expression in an intact organism. Studies in promoter-driven transgenic mice with CNS production of a number of cytokines or chemokines have demonstrated that these factors can directly induce a spectrum of cellular alterations often resulting in pronounced neurological disease (Table 1). Thus, these factors, in addition to initiating and maintaining immunoinflammatory responses, can be direct mediators of CNS injury. The neuropathological outcomes in the transgenic mice often recapitulate those reported in human neurological disorders such as MS, neurological diseases associated with AIDS and Alzheimer's disease, pointing to the importance of these animal models to our understanding of the role of cytokines and chemokines in these human disorders. Despite problems of timing and tissue specificity as well as some inconsistencies in the findings from different groups, knockout mice have begun to provide insights that are altering our view of the contribution made by individual cytokines to immunoinflammatory responses in the brain. For example, IL-6 and TNF were originally viewed as having minor and major proinflammatory contributions, respectively, in EAE, but now, based on findings from knockout mice, the opposite seems true. Studies in transgenic and knockout mice now offer strong evidence that, in addition to being mediators of damage, cytokines can have beneficial functions, e.g. the antiviral functions of the IFNs or the trophic and/or neuroprotective actions of some cytokines such as IL-6 and TNF. Clearly, studies in mutant mice, as summarized here, will continue to provide important insights into the nature of cytokine and chemokine actions in the CNS and will offer the possibility that we may identify new targets for effective therapeutic intervention in neuroinflammatory disorders.

Regulation of MHC protein expression in pancreatic beta-cells by interferon-gamma and tumor necrosis factor-alpha.

Isolated human and mouse pancreatic islet cells and the rat insulinoma cell line RIN-m5F were used to examine the ability of recombinant interferon-gamma (IFN-gamma) and tumor necrosis factor-alpha (TNF-alpha) to regulate the expression of the class I and class II major histocompatibility (MHC) surface proteins and mRNA in beta-cells. Each cytokine increased significantly the expression of class I MHC proteins as determined by double indirect immunofluorescence microscopy and flow cytofluorimetric analysis. In the RIN-m5F cells, this increase in surface expressed class I MHC proteins was mirrored by an increase in the level of class I MHC mRNA. The order of potency of the cytokines on class I MHC expression was TNF-alpha plus IFN-gamma greater than or equal to IFN-gamma greater than or equal to TNF-alpha. While IFN-gamma or TNF-alpha alone were without effect, in combination they were found to induce class II MHC proteins on 30-40% of human or murine beta-cells. In contrast, IFN-gamma plus TNF-alpha did not induce detectable class II MHC proteins or mRNA in the RIN-m5F cells. These findings indicate that 1) TNF-alpha, in addition to IFN-gamma, upregulates the expression of beta-cell class I MHC proteins and mRNA, and 2) more than one signal is required for the induction of class II MHC proteins on beta-cells. The ability of IFN-gamma plus TNF-alpha to induce class II MHC proteins on only a fraction of the normal beta-cell population and not on RIN-m5F cells suggests that this response is related to the differentiation state of the beta-cell.

Insulin receptor expression in the Burkitt lymphoma cells Daudi and Raji.

The specific binding of insulin to either intact or Triton-solubilized Daudi cells (a Burkitt lymphoma cell line) was reduced by over 95% compared to that to control IM-9 lymphocytes due to a decrease in receptor number without a change in affinity. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography revealed that 125I-labeled Daudi cells had reduced amounts (approximately 1/20th) of immunoprecipitable binding (alpha) subunit [mol wt (Mr), 130,000] of the receptor and a relatively abundant 210,000 Mr form not seen in IM-9 cells. The transmembranous (beta) subunit (Mr, 90,000) of the receptor, although not detected by 125I surface labeling, could be phosphorylated and, together with the 210,000 Mr form, exhibited the same 2-fold stimulation of phosphorylation by insulin as that in IM-9 cells. Northern blot hybridization revealed a decrease in Daudi cells of all four major species of insulin receptor mRNA. The Raji cell, another Burkitt lymphoma cell line, also exhibited reduced protein and genetic expression of the insulin receptor, indicating that reduced insulin receptor expression may be representative of other Burkitt lymphoma cell lines.

Evolution of neuropathologic abnormalities associated with blood-brain barrier breakdown in transgenic mice expressing interleukin-6 in astrocytes.

As both astrocytes and cytokines modulate the permeability of cerebral endothelial cells, transgenic animal models which overexpress cytokines, such as interleukin-6 (IL-6), may provide insight into the neuropathological consequences of increased BBB permeability. In this study, a GFAP-IL6 transgenic mouse model and horseradish peroxidase (HRP) were used to investigate BBB permeability and associated neuropathologic changes. In the cerebellum of control mice, the BBB developed between postnatal days 7 and 14. In transgenic mice, the BBB never developed and extensive breakdown was evident in both high- and low-expressor animals by 1 month after birth. Vascular proliferation was apparent from birth in association with development and retention of normal cerebellar architecture until 3 and 6 months in high- and low-expressor animals, respectively. At these times, a leptomeningeal inflammatory infiltrate, vacuolated astrocytic foot processes and endothelial abnormalities were apparent in the cerebellum. At 6 months in high-expressor and 12 months in low-expressor animals, parenchymal inflammation, gliosis, spongiform change, axonal degeneration and macrophage accumulation were evident. The findings suggest that increased production of IL-6 can influence the development and physiologic function of the BBB as well as contribute to parenchymal central nervous system injury.

Distinct expression patterns and levels of enzymatic activity of matrix metalloproteinases and their inhibitors in primary brain tumors.

Matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs) have been implicated in the immense invasive potential and neovascularization of primary brain tumors. We investigated the gene expression profiles of MMPs 1, 2, 3, 7, 9, 12, 13, 14, 16 and of TIMPs 1, 2, 3, and 4 in various primary brain tumors (astrocytoma WHO grade I-III, glioblastoma, PNET, ependymoma III and oligoastrocytoma II) using novel RNase protection assay probe sets. In addition, we determined the level and cellular source of gelatinolytic activity and localized gelatinase B and TIMP-1 RNA. Distinct expression patterns of the MMP and TIMP genes were found in the various brain tumors tested. While the WHO grade I and II tumors had MT1/MT3 ratios below 1, the malignant (grade III and IV) tumors had ratios above 1. Strong expression of TIMP-1 RNA was observed in all malignant tumors and in grade I pilocytic astrocytomas and localized to the walls of neovessels. Quantitative analysis of enzymatic activity in the soluble fraction of protein extracts revealed that in most tumors gelatinases remained in the inactive pro-form. In situ zymography revealed net gelatinolytic activity in neurons of normal brain and in tumor cells and vessel walls of all tumors tested. Immunohistochemistry demonstrated that gelatinase B was localized to vessel walls, to neutrophils in areas of hemorrhage, and in glioblastomas to macrophages. Together these data demonstrate that the different primary brain tumors show distinct regulation of MMP and TIMP genes. The localization of the soluble gelatinase B indicates an association with neovascularization, whereas membrane-bound MMPs may account for the invasive potential of the glial tumor cells.