Nef is secreted in exosomes from Nef.GFP-expressing and HIV-1-infected human astrocytes

HIV-1 infection of the central nervous system causes HIV-associated neurocognitive disorders, even in aviremic patients. Although astrocyte malfunction was associated to these disorders, their implication is overshadowed by contributions of microglia and macrophages. Astrocytes are infected with HIV-1 in vivo and express a relevant amount of viral protein Nef. Nef was shown to stimulate its own release in exosomes from diverse cell types, which in turn have damaging effects on neighboring cells. Using immunoblotting and electron microscopy, we showed that human astrocytes expressing Nef.GFP similarly release Nef in exosomes. Importantly, Nef.GFP expression increases the secretion of exosomes from human astrocytes up to 5.5-fold, as determined by total protein content and nanoparticle tracking analysis. Protein analysis of exosomes and viruses separated on iodixanol gradient further showed that native or pseudotyped HIV-1-infected human astrocytes release exosomes, which contain Nef. Our results provide the basis for future studies of the damaging role of Nef-exosomes produced by HIV-infected astrocytes on the central nervous system.     

HIV-1 Nef alters the expression of betaII and epsilon isoforms of protein kinase C and the activation of the long terminal repeat promoter in human astrocytoma cells.

In the human immunodeficiency virus type 1 (HIV-1)-infected brain, the virus does not replicate in astrocytes, but a synthesis of viral regulatory proteins occurs in these cells, leading to accumulation of Nef. As an approach to understand the effects of Nef on astrocyte functional activity, we analyzed whether intracellular Nef interferes with the expression and activation of the enzyme protein kinase C (PKC), which is an important regulator of astroglial functions and HIV-1 replication. Astrocytoma clones (U251 MG) not expressing Nef (Neo), or expressing wild-type Nef (Bru) or nonmyristoylated Nef (TH) were used to monitor the expression and activation of 10 PKC isoforms. The same clones were used to evaluate the effect of Nef on the viral long terminal repeat (LTR) promoter after activation of PKC with the phorbol ester 12-myristate 13-acetate (PMA). PKC intracellular distribution and activation were evaluated by Western blot analysis of cytosolic and membrane fractions of control and Nef-expressing clones. PMA-induced LTR activation was analyzed in clones transfected with a plasmid encoding for the CAT reporter gene controlled by the LTR promoter, by using an enzyme-linked immunosorbent assay to measure CAT expression. Nef selectively downregulated the expression and activation of betaII and epsilon PKC isoforms in astrocytoma cells. Such downregulation correlated with an inhibition of LTR activation after PMA stimulation. The myristoylation of Nef and its membrane localization were essential for these effects. These results suggest that Nef may alter astrocytic functions by interfering with PKC expression and activation and contribute to the restriction of HIV-1 replication in astrocytes.     

Gene transfer in astrocytes: comparison between different delivering methods and expression of the HIV-1 protein Nef

To identify a good system to introduce foreign genes into normal and tumoral astrocytes, we studied the efficiency of two chemical methods, calcium phosphate precipitation and lipofection, and of a viral-mediated transfer by a vector derived from the highly attenuated modified vaccinia virus Ankara (MVA). Using the beta-galactosidase (beta-gal) gene (lacZ) as reporter, we searched for optimal experimental conditions to obtain an efficient gene transfer into human embryonic and neonatal rat astrocytes and into a human astrocytoma cell line (U373 MG). The beta-gal protein production was evaluated by cytochemical staining and enzymatic activity assay. Among chemical methods, lipofection was the most efficient system to transfect astrocytes in providing up to 60% of beta-gal-positive cells in all the cell types analyzed. MVA infection also proved to be an efficient system to introduce heterologous genes into human embryonic astrocytes that appeared 80-100% positive 48-96 hr after an infection at a multiplicity of 1-10. In contrast, only a limited infection was observed with rat astrocytes, human astrocytoma cells, and human leptomeningeal cells. A recombinant MVA vector expressing the human immunodeficiency virus-1 (HIV-1) regulatory protein Nef was used to transfect human embryonic astrocytes, and the resulting Nef expression was compared with that detected after lipofection in the same cells. By Western blot analysis, Nef expression was observed in human astrocytes 24-96 hr after infection and was similar to that present in stably HIV-1-infected astrocytoma cells. Lipofection resulted in lower Nef expression. In spite of these promising results, the negative effects of MVA infection on cell viability and the possibility that a productive infection occurs in human embryonic astrocytes limit the use of this vector for gene delivery in developmentally immature human glial cells.     

Simian immunodeficiency virus mac251 infection of astrocytes

Human immunodeficiency virus type 1 (HIV-1) infection of astrocytes has been demonstrated in the brains of patients with AIDS dementia complex (ADC) and may play an important role in neuropathological pathways of HIV-related encephalopathy. SIVmac-infected monkeys develop an acquired immunodeficiency syndrome (AIDS) with CNS involvement which is quite similar to that seen in human AIDS. We investigated the in vitro infection of primary astrocytes derived from adult macaques with SIVmac251 or an isogenic virus that expresses a non-functional Nef protein (SIVmac251-DeltaNef). In both cases we observed that viral expression was mostly limited to early regulatory genes after a transient phase of late viral gene expression (i.e. env and gag), as reported for HIV-1-infected astrocytes in vivo. Late viral gene expression could be reactivated by TNF-alpha, GM-CSF and IFN-gamma treatment of SIVmac251-infected astrocytes but not by similarly treated SIVmac251-DeltaNef-infected cells. Our findings suggest that Nef is not involved in the restricted expression of SIV in astrocytes, but may be important for astrocytes to function as a viral reservoir in the CNS. In additional experiments, we demonstrated Rev and Nef expression in 17 of 27 primary astrocyte cultures derived from macaques infected by SIVmac251. Nef was located in the cytoplasm of astrocytes infected by SIVmac251 in vivo, but displayed perinuclear localisation after infection in vitro. Attempts to activate late viral gene expression by astrocytes infected in vivo using cytokines or by coculture with human cord blood mononuclear cells were unsuccessful.     

Modulation of astrocyte proliferation by HIV-1: differential effects in productively infected, uninfected, and Nef-expressing cells

Although quiescent in normal brain, reactive astrocytes can proliferate in various disorders. We examined the impact of HIV-1 on astrocyte proliferation in cultures exposed to VSVg env-pseudotyped HIV-1 which yields high levels of infection. HIV-1, while increasing the proliferation of uninfected (p24-) astrocytes, strongly inhibited proliferation of productively infected (p24+) cells. The cell cycle arrest was G1/S rather than G2/M, a type commonly attributed to Vpr. No clear role of Vpr or Nef could be identified. Adenovirus-mediated expression of Nef (a model of "restricted" infection) induced M-phase arrest of astrocytes. We speculate that HIV-1 is a significant modulator of astrocyte proliferation in vivo.     

Replication of different clones of human immunodeficiency virus type 1 in primary fetal human astrocytes: enhancement of viral gene expression by Nef.

Dementia is a common complication of AIDS which is associated with human immunodeficiency virus type 1 (HIV-1) infection of brain macrophages and microglia. Recent studies have shown that astrocytes are also infected in the brain but HIV-1 replication in these cells is restricted. To determine virus specificity of this restriction we tested the expression of 15 HIV-1 molecular clones in primary human fetal astrocytes by infection and DNA transfection. Infection with cell-free viruses was poorly productive and revealed no clone-specific differences. In contrast, transfected cells produced transiently high levels of HIV-1 p24 core antigen, up to 50 nanograms per ml culture supernatant, and nanogram levels of p24 were detected 3-4 weeks after transfection of some viral clones. The average peak expression of HIV-1 in astrocytes varied as a function of viral clone used by a factor of 15 but the differences and the subsequent virus spread did not correlate with the tropism of the viral clones to T cells or macrophages. Functional vif, vpu, and vpr genes were dispensable for virus replication from transfected DNA, but intact nef provided a detectable enhancement of early viral gene expression and promoted maintenance of HIV-1 infection. We conclude that primary astrocytes present no fundamental barriers to moderate expression of different strains of HIV-1 and that the presence of functional Nef is advantageous to virus infection in these cells.     

Purification of astrocytes from adult human optic nerve heads by immunopanning

Most in vitro studies in the CNS require pure cultures of astrocytes. Astrocytes from the human optic nerve head (ONH, type 1B) represent a specialized population of astrocytes. Primary cells grown from human optic nerve head explants were cultured for 3-4 weeks. To select astrocytes by immunopanning, cell suspensions were placed on a P100 panning dish coated with C5 anti-neuroepithelial antibody and allowed to attach for 30 min. Nonadherent cells were plated on a second dish coated with anti-Thy1.1 antibody to deplete microglia and meningeal cells. Finally, remnant nonadherent cells were plated on a noncoated dish. Purified cells were immunostained with astrocyte markers: GFAP, vimentin, Pax2, A2B5, nestin and NCAM. Other cell types were characterized by HLA-DR for microglia and smooth muscle actin for vascular smooth muscle. The proportion of GFAP+ astrocytes in the cultures was determined by flow cytometry. About 95% of the cells that adhered to the C5 dish were GFAP+ astrocytes. GFAP+ astrocytes expressed vimentin, Pax2, nestin and NCAM, but not A2B5. From the Thy1.1 dish, 60-75% cells were GFAP+ astrocytes and the remainder cells were GFAP- cells. Using cloning rings, we eliminated fibroblast-like cells, smooth muscle and meningeal cells from astrocyte cultures. Smooth muscle cells and fibroblasts grew on the noncoated dish. In conclusion, immunopanning is an efficient method to get high yields of viable type 1B astrocytes from adult human ONH. The current described culture system may provide a valuable tool in studying human optic nerve head biology and disease.     

Proteomic analysis of astroglial connexin43 silencing uncovers a cytoskeletal platform involved in process formation and migration

Connexin43 (Cx43) is the most abundant gap junction protein of the brain, where it is predominantly expressed in astrocytes. Recent studies imply a role of Cx43 in the regulation of important cellular processes, including migration, proliferation, and shape formation. These processes are assumed to be reflected by the proteome of the Cx43 expressing cells. To analyze the influence of Cx43 on the astrocytic proteome, we used RNA interference to downregulate the expression of this connexin in cultures of mouse astrocytes. We applied difference gel electrophoresis (DIGE) to compare silenced astrocytes with control cells. The differential proteome analysis revealed 15 significantly regulated proteins (between 1.2‐ and 1.6‐fold), of which six are known to belong to a group of cytoskeletal proteins involved in cortical platform formation. Astrocytes treated with Cx43 small interfering (si)RNA showed an increased expression of the cytoskeletal proteins: actin, tropomyosin, microtubule‐associated protein RP/EB1, transgelin, and GFAP, and a decreased expression of cofilin‐1. Quantitative immunocytochemistry and Western blotting revealed similar results showing an upregulation of actin, tubulin, tropomyosin, EB1, transgelin and GFAP, and a downregulation of Ser‐3‐phosphorylated cofilin. Furthermore, Cx43 silencing led to phenotypical changes in cell morphology, migratory activity, and cell adhesion. Our results provide mechanistic clues for an understanding of Cx43 interaction with cellular motor activities such as migration and process formation in astrocytes. © 2009 Wiley‐Liss, Inc.     

Effects of transforming growth factor-β1 on the extracellular matrix and cytoskeleton of cultured astrocytes

The present study was performed on primary cultures and subcultures of cerebellar astrocytes in order to investigate the effects of transforming growth factor-β1 (TGFβ1) on proliferation, extracellular matrix (ECM) components, and cytoskeletal structures in relation to morphological changes. The expression and cellular distribution of the ECM components laminin and fibronectin and the cytoskeletal proteins glial fibrillary acidic protein (GFAP) and actin were investigated by immunoblotting, immunocytochemistry, and phalloidin staining. The proliferation of primary cultures was strongly inhibited by TGFβ1. Treated cells became enlarged and spread onto the substratum. TGFβ1 promoted the appearance of actin stress fibers and increased the cell actin content. It elicited a slight increase in GFAP expression and induced dispersion of thin filaments of GFAP. TGFβ1 also stimulated the production of laminin and fibronectin and their incorporation into the ECM of primary cultures grown in medium with or without serum Astrocytes grown in serum-containing medium for 1 day after subculturing responded strongly to TGFβ1. Changes promoted by TGFβ1 in cell shape, cytoskeleton, and ECM production of cultured astrocytes may have relevance for understanding the mechanisms of action of TGFβ1 during brain development.     

Analysis of glial fibrillary acidic protein,neurofilament protein,actin and heat shock proteins in human fetal brain during the second trimester

Glial fibrillary acidic protein (GFAP), the 66 kDa neurofilament protein (NF-66), actin, the 27 kDa heat shock protein (HSP27) and the 70 kDa constitutive heat shock protein (HSC70) were analyzed in human fetal brain during the second trimester, from 10 to 24 gestational weeks (GW). By immunohistochemistry, cell-type specific localization of GFAP and NF-66 in astrocytes and neurons, respectively, was confirmed. HSP27 was expressed mostly in the nuclear region of neurons and non-neuronal cells, and HSC70 was widely distributed throughout the tissue. By quantitative immunoblotting, GFAP was not detectable in gray matter of prefrontal cortex prior to 16 GW. Between 16 and 21 GW, the content of GFAP rose slowly. Thereafter, GFAP accumulated rapidly. The content of GFAP in different brain regions (prefrontal, parietal, and occipital cortices) differed significantly at 22 GW. In contrast, NF-66 was already highly expressed at 10 GW, slowly rose to maximal values by 18 GW, and thereafter remained stationary. In contrast to GFAP, the content of NF-66 was similar in different brain regions at 22 GW. Although actin was abundant throughout the second trimester, a sharp drop in its content in the prefrontal cortex was detected at 17 GW. To explain such a decrease, two heat shock proteins were analyzed. HSP27, known to modulate actin polymerization, was found to increase sharply at 16–17 GW. In contrast, HSC70 remained constant during the second trimester and was highly expressed in the fetal brain, at a level comparable to that in the adult brain.     

Knockdown of MLC1 in primary astrocytes causes cell vacuolation: a MLC disease cell model

Megalencephalic leukoencephalopathy with subcortical cysts (MLC) is a rare type of leukodystrophy, in the majority of cases caused by mutations in the MLC1 gene. MRI from MLC patients shows diffuse cerebral white matter signal abnormality and swelling, with evidence of increased water content. Histopathology in a MLC patient shows vacuolation of myelin, which causes the cerebral white matter swelling. MLC1 protein is expressed in astrocytic processes that are part of blood- and cerebrospinal fluid-brain barriers. We aimed to create an astrocyte cell model of MLC disease. The characterization of rat astrocyte cultures revealed MLC1 localization in cell-cell contacts, which contains other proteins described typically in tight and adherent junctions. MLC1 localization in these contacts was demonstrated to depend on the actin cytoskeleton; it was not altered when disrupting the microtubule or the GFAP networks. In human tissues, MLC1 and the protein Zonula Occludens 1 (ZO-1), which is linked to the actin cytoskeleton, co-localized by EM immunostaining and were specifically co-immunoprecipitated. To create an MLC cell model, knockdown of MLC1 in primary astrocytes was performed. Reduction of MLC1 expression resulted in the appearance of intracellular vacuoles. This vacuolation was reversed by the co-expression of human MLC1. Re-examination of a human brain biopsy from an MLC patient revealed that vacuoles were also consistently present in astrocytic processes. Thus, vacuolation of astrocytes is also a hallmark of MLC disease.     

Recombinant nef HIV-IIIB protein is toxic to human neurons in culture

The expression of HIV-1 negative factor (nef) has been positively correlated with HIV disease progression [Z. Hanna, D.G. Kay, N. Rebai, A. Guimond, S. Jothy, P. Jocicoeur, Nef harbors a makor determinant of pathogenicity for an AIDS-like disease induced by HIV-1 in transgenic mice. Cell 95 (1998) 163-175]. Nef expression has been detected in HIV infected human brains with neuronal damage [A. Ranki, M. Nyberg, V. Ovod, M. Haltia, I. Elovaara, R. Raininko, H. Haapsalo, K. Krohn, Abundant expression of HIV Nef and Rev proteins in brain astrocytes in associated with dementia, AIDS 9(9) (1995) 1001-1008; Y. Saito, L.R. Sharer, M.G. Epstein, J. Michaels, M. Mintz, M. Londer, K. Golding, B.M. Blumberg, Overexpression of nef as a marker for restricted HIV-1 infection of astrocytes in postmorten paediatric central tissues, Neurology 14 (1994) 474-480]. It is postulated that nef may contribute to the neuronal damage observed in the brain of those with late HIV disease. To test this, the potential toxicity of recombinant nef (from HIV-1 IIIB) was compared to the neurotoxin human tumour necrosis alpha (TNFalpha) on human brain cells in culture. SK-N-SH neuroblastoma, primary human neurons and glial cells were exposed to recombinant nef or TNFalpha protein for 3 days or twice over 6 days. Cell viability was assessed by Trypan Blue, lactate dehydrogenase (LDH) release and MTT assays. Nuclear fragmentation was detected using the Hoechst Blue nuclear dye assay. Both nef and TNFalpha (100 ng/ml) caused a significant 30% reduction of SK-N-SH cell numbers after 3 days exposure (P=0. 001). At this time, exposure to nef caused evident fragmented nuclei in these cultures. Human neuronal cultures had a 32 and 33% decrease in cell number after 6 days exposure to either nef or TNFalpha, respectively (P<0.001). Furthermore, as previously shown [J. He, C.M. DeCastro, G.R. Vandenbark, J. Busciglio, D. Gabuzda, Astrocyte apoptosis induced by HIV-1 transactivation of the c-kit protoonocogene, Proc. Natl. Acad. Sci. 94 (1997) 3954-3959], a 3-day exposure to nef significantly reduced human glial cell number by 25% (P=0.001). Recombinant nef and TNFalpha compromise human neurons in culture. Thus, like other virotoxins, it is shown for the first time that nef may also contribute to neuronal damage that has been reported in dementia in late HIV disease.     

Immune activation, viral gene product expression and neurotoxicity in the HIV-1 transgenic rat

The HIV-1 transgenic (TG) rat has been shown to be a useful model of nervous system disease that occurs in human HIV-1 infection. Studies were, therefore, performed to examine characteristics of the immune response in the periphery and brain of the animals and expression of factors in the nervous system that might be associated with neurotoxicity. Activated splenocytes from wild-type (WT) and TG rats were stimulated with either CD3/CD28 or with lipopolysaccharide (LPS) and examined for proliferative responses and for proinflammatory cytokine (IFN-γ, TNF-α and IL-1β) secretion. Brain tissue lysates from the rats were also examined for proinflammatory cytokine levels and tissue sections were stained by immunofluorescence for class II MHC+, ED1+ or Iba1+ (for macrophages and microglial cells), and for GFAP+ (for astrocytes) cells and for co-labeling of these cells for TNF-α. Co-labeling was also performed to identify cells expressing HIV-1 gp160, tat, nef and vif. Finally, on Western blots brain tissue lysates were examined for phosphorylation of Erk1/2, p38, JNK-SAPK and Erk5. TG rat splenocyte proliferative responses were higher than for WT with CD3/CD28-stimulation but lower than WT with LPS stimulation. CD3/CD28-stimulated TG rat splenocytes also secreted higher levels of IFN-γ, TNF-α and IL-1β whereas LPS-stimulated TG rat splenocytes secreted higher levels of only TNF-α than cultures from WT rats. Levels of all three cytokines were higher in brain lysates from TG rats than for WT rats. On immunofluorescence staining of corresponding sections of brain, TG rats contained increased numbers of class II MHC+ and ED1+ cells, and there was also increased co-labeling or these cells as well as astrocytes for TNF-α. Iba1+ cells showed positive staining for all of the HIV proteins whereas astrocytes showed significant positive staining for only nef and vif. Phosphorylation of Erk1/2, p38 and JNK/SAPK was detected for both TG and WT rat tissues with higher levels of phosphorylation forms of these proteins detected in the TG rat brain. Phosphorylation of Erk5, a marker that is associated with specifically neuronal repair, was detected only in TG rat brain. These studies suggest that activated nervous system mononuclear phagocytes and astrocytes expressing HIV-1 gene products in specific patterns are associated with neurodegeneration in the HIV-1 TG rat.     

Extracellular matrix proteins expressed by human adult astrocytes in vivo and in vitro: An astrocyte surface protein containing the CS1 domain contributes to binding of lymphoblasts

Primary cultures of human astrocytes, expressing glial fibrillary acidic protein (GFAP), were obtained from postmortem brain tissue samples. These cultured astrocytes produced an extracellular matrix (ECM), containing laminin (Ln) and fibronectin (Fn), as shown with specific antibodies. The perinuclear staining observed in these cells indicated that these proteins were de novo synthesized. Monoclonal antibody (mAb) 90.45, which recognizes the CS1 sequence found in an alternatively spliced form of Fn, also stained cultured astrocytes. Immunohistochemical analysis of normal human brain tissue showed positive staining for the CS1 domain, both on protoplasmic and fibrous astrocytes located in the gray and white matter. In contrast to cultured astrocytes, no immunoreactivity for Ln or Fn was found on astrocytes in normal human brain tissue. These in situ data indicate that the CS1 domain expressed by astrocytes is not part of a splicing variant of Fn. Western blot analysis confirmed that the CS1 domain expressed by cultured human astrocytes is part of an astrocyte protein which is different from human Fn. The CS1 domain is a known ligand for the adhesion receptor α4β1 (VLA-4). We found that the human lymphoma cell lines Jurkat and Ramos, which express α4β1, bound to cultured human astrocytes, and that this interaction could be partly blocked by mAb 90.45 or a synthetic CS1 peptide. Thus, the novel CS1-containing surface protein expressed by astrocytes in vitro and in vivo, contributes to binding of lymphoblasts, and therefore may be a relevant adhesion molecule for the recruitment of α4-integrin expressing leukocytes into the central nervous system (CNS). J. Neurosci. Res. 50:539–548, 1997. © 1997 Wiley-Liss, Inc.     

Human immunodeficiency virus type-1 Nef protein induces blood-brain barrier disruption in the rat: role of matrix metalloproteinase-9

We recently showed that MMP-9 activity was detectable in the cerebrospinal fluid (CSF) of about half of neurologically symptomatic HIV-infected patients. Using an experimental animal model, we detected MMP-9 activity in CSF samples from rats that had been injected intracisternally with recombinant HIV-1 Nef protein, but not after injection of heat-treated Nef, gp120, gp160 or PBS. Nef also induced a breaching of the blood-brain barrier (BBB), which could be inhibited by pretreatment with the matrix metalloproteinase (MMP) inhibitor batimastat. In vitro Nef only slightly induced MMP-9 activity in freshly isolated human peripheral blood mononuclear cells and in the murine macrophage cell line RAW 264.7, but not in endothelial, neuronal or astroglial cell lines. Taken together, our findings indicate that HIV-1 Nef protein can induce BBB disruption in the rat - presumably via MMP induction.     

Immunocytochemical staining for glial fibrillary acidic protein and the metabolism of cytoskeletal proteins in experimental allergic encephalomyelitis

Spinal cord sections from Lewis rats with acute experimental allergic encephalomyelitis (EAE) showed greatly increased staining of astrocytes when stained immunocytochemically for glial fibrillary acidicc protein (GFAP). Fibrous processes in white matter were heavily stained early in the course of the disease when paralysis was first evident (10–12 days after injection of guinea pig spinal cord myelin), then protoplasmic astrocytes were stained in the gray matter and became more heavily stained at 20 dats post-injection. The stained astrocytes were evenly distributed throughout the tissue, and did not correspond to the sites of the lesions. Spinal cord slices of control and EAE rats were incubated with [³H]amino acids, then cytoskeletal proteins were prepared in an enriched fraction, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and the protein bands counted for radioactivity. In the EAE rat all cytoskeletal proteins, including the neurofilaments, vimentin, microtubules, GFAP and actin, showed increased uptake of radioactive amino acids. Immunoprecipitation of GFAP with specific antiserum showed increased radioactivity in the complex beginning at day 10 when cellular infiltration was beginning in the EAE animals. As the disease became acute, the radioactivity in the immunoprecipitated GFAP increased, in some cases to very high levels, then by day 18 when recovery was underway, the radioactivity had fallen to normal levels. Possible agents causing metabolic activation of protein synthesis in EAE animals include stimulating substances elaborated by infiltrating lymphoid scells, and the generalized edema accompanying the demyelinative condition. The activation of GFAP protein staining and metabolism in EAE might serve as a model for the activated growth of astrocyte processes which cause the severe gliosis seen in multiple sclerosis.     

Heterogeneity of astrocytes in human optic nerve head

Previous studies demonstrated regional differences in the synthesis of extracellular matrix by astrocytes during optic nerve head (ONH) maturation and in glaucomatous optic neuro pathy, suggesting heterogeneity of astrocytes. To characterize different types of glial cells in human fetal and adult ONH, we used a variety of neural cell markers such as HNK-1/N-CAM, A2B5, galactocerebroside (GalC), myelin basic protein (MBP), and glial fibrillary acidic protein (GFAP). Cryostat or paraffin sections were prepared from fetal (16–25 weeks) and mature (8 months to 75 years old) ONH and processed for standard single/double immunocytochemistry. Two subpopulations of type 1 astrocytes were present in the mature prelaminar and laminar regions. Glial celia expressing only GFAP were identified as type 1A astrocytes at the edges of the cribriform plates. Cells forming the glial columns and lining the cribriform plates expressed both GFAP and fINK-1/N-CAM and were identified as type lB astrocytes. In the myelinated nerve, type 1A astrocytes form the glial limiting membrane. Cells labeled with GFAP and A2B5 were identified as type 2 astrocytes, and GFAP-negative cells labeled with GaIC, MBP, and HNK-1/N-CAM were identified as oligodendrocytes. In fetal ONH, all glial cells expressed HNK-1/N-CAM. In older fetal ONH, some glial cells also expressed GFAP. No type 2 astrocytes or oligodendrocytes were present in the fetal ONH. In conclusion, at least two subpopulations of type 1 astrocytes exist in human ONH: Type 1A astrocytes may serve as structural support for a type lB astrocytes, which retain the developmental neural marker HNK-1/N-CAM, may have a more complex function by interfacing between blood vessels and other connective tissue surfaces. These findings demonstrate the heterogeneity of astrocytes in the human ONH and suggest differential regional responses to changes in their microenvironment. © 1995 Wiley-Liss Inc.