Demyelination is a hallmark of several human diseases, including multiple sclerosis. To understand better the process of demyelination and remyelination, we explored the use of an in vitro organotypic cerebellar slice culture system. Parasagittal slices of postnatal Day 10 (P10) rat cerebella cultured in vitro demonstrated significant myelination after 1 week in culture. Treatment of the cultures at 7 days in vitro (DIV) with the bioactive lipid lysolecithin (lysophosphatidylcholine) for 15-17 hr in vitro produced marked demyelination. This demyelination was observed by immunostaining for the myelin components myelin basic protein (MBP), myelin oligodendrocyte glycoprotein (MOG), and 2', 3'-cyclic nucleotide 3'-phosphodiesterase (CNPase). After a transient demyelinating insult with lysolecithin in vitro, the cultures recovered with oligodendrocyte differentiation recapitulating a normal time course; there was initially re-expression of CNPase and MBP during this recovery, and this was followed by MOG. In addition, there seemed to be some limited remyelination during the recovery phase. Lysolecithin thus induces demyelination in an in vitro organotypic cerebellar slice culture system, providing a model system for studying myelination, demyelination, and remyelination in vitro.

Lysophospholipids have long been recognized as membrane phospholipid metabolites, but only recently has their role as intercellular signaling molecules been appreciated. Two of the best-studied lysophospholipids, LPA and S1P, signal through cognate G-protein-coupled receptors to activate many well-known intracellular signaling pathways, leading to a variety of biologically important cell responses. Lysophospholipids and their receptors have been found in a wide range of tissues and cell types, indicating their importance in many physiological processes, including reproduction, vascular development, cancer and nervous system function. This article will focus on the most recent findings regarding the biological functions of lysophospholipids in mammalian systems, specifically as they relate to health and disease.

Abstract. We are interested in the relationship between the cytoskeleton and the organization of polarized cell morphology. We show here that the growth cones of hippocampal neurons in culture are specifically stained by a monoclonal antibody called 13H9. In other systems, the antigen recognized by 13H9 is associated with marginal bands of chicken erythrocytes and shows properties of both microtubuleand microfilament-associated proteins (Birgbauer, E., and F. Solomon. 1989. J. Cell Biol. 109:1609-1620. This dual nature is manifest in hippocampal neurons as well. At early stages after plating, the antibody stains the circumferential lamellipodia that mediate initial cell spreading. As processes emerge, 13H9 staining is heavily concentrated in the distal regions of growth cones, particularly in lamellipodial fans. In these cells, the 13H9 staining is complementary to the localization of assembled microtubules. It colocalizes partially, but not entirely, with phalloidin staining of assembled actin. Incubation with nocodazole rapidly induces microtubule depolymerization, which proceeds in the distal-to-proximal direction in the processes. At the same time, a rapid and dramatic redistribution of the 13H9 staining occurs; it delocalizes along the axon shaft, becoming clearly distinct from the phalloidin staining and always remaining distal to the receding front of assembled microtubules. After longer times without assembled microtubules, no staining of 13H9 can be detected. Removal of the nocodazole allows the microtubules to reform, in an ordered proximal-todistal fashion. The 13H9 immunoreactivity also reappears, but only in the growth cones, not in any intermediate positions along the axon, and only after the reformation of microtubules is complete. The results indicate that the antigen recognized by 13H9 is highly concentrated in growth cones, closely associated with polymerized actin, and that its proper localization depends upon intact microtubules.T HE ability of neurons to develop stereotyped morphologies, suitable for the functions of each individual cell, depends upon the activities of their growth cones. All of the motility of nerve cells is confined to these structures at the tips of growing axons and dendrites. Although in migratory cells the position of the leading edge can change frequently, the position of the growth cones persists once they are specified. There have been many analyses of the behavior and properties of growth cones, and of the cues that guide growth cones to their appropriate destinations (for reviews, see Landis, 1983;Lockerbie, 1987;Bray and Hollenbeck, 1988). However, rather less is known about the endogenous determinants of growth cone formation and organization. How does the cell specify the number and position of growth cones? How is growth cone motility coupled to the function of cytoskeletal elements within the growing fiber?Arguably, the unique properties of growth cones might arise not only from the presence of unique constituents but Dr. Goslin's and Dr. Banker's present add...

No abstract

Abstract. The marginal band of nucleated erythrocytes is a microtubule organelle under rigorous quantitative and spatial control, with properties quite different from those of the microtubule organelles of cultured cells. Previous results suggest that proteins other than tubulin may participate in organizing the marginal band, and may interact with elements of the erythrocyte cytoskeleton in addition to microtubules. To identify such species, we raised mAbs against the proteins that assemble from chicken brain homogenates with tubulin. One such antibody binds to a single protein in chicken erythrocytes, and produces an immunofluorescence pattern colocalizing with marginal band microtubules. Several properties of this protein are identical to those of ezrin, a protein isolated from brush border and localized to motile elements of cultured cells. A significant proportion of the antigen is not soluble in erythrocytes, as determined by extraction with nonionic detergent. This cytoskeletonassociated fraction is unaffected by treatments that solubilize the marginal band microtubules. The protein has properties of both microtubule-and microfilamentassociated proteins. In the accompanying manuscript (Goslin, K., E. Birgbauer, G. Banker, and E Solomon. 1989. J. Cell Biol. 109:1621-1631), we show that the same antibody recognizes a component of growth cones with a similar dual nature. In early embryonic red blood cells, the antigen is dispersed throughout the cell and does not colocalize with assembled tubulin. Its confinement to the marginal band during development follows rather than precedes that of microtubules. These results, along with previous work, suggest models for the formation of the marginal band.

We are studying the changes in the organization of the cytoskeleton which accompany expression of differentiated neuronal morphology. Of particular interest is the elaboration of growth cones, the motile domains of the neuronal plasma membrane, and the cytoskeletal structures that underlie them. A candidate for a component of the growth cone cytoskeleton of cultured hippocampal neurons is the antigen recognized by the monoclonal antibody, 13H9 (Birgbauer and Solomon, J Cell Biol 109:1609-1620, 1989; Goslin et al., J Cell Biol 109:1621-1631, 1989). That antibody binds strongly to growth cones, but barely stains neurites. The characterization of the antigen, both biochemical and microscopic, suggests that it may interact with microfilaments and microtubules. We have established that 13H9 recognizes a subset of the isoforms of ezrin (unpublished results). Here, we describe the properties and localization of ezrin isoforms in differentiating neuronal cells, using two in vitro systems and developing spinal cord. In embryonal carcinoma cells, both the abundance of ezrin and the proportion of ezrin associated with the cytoskeletal fraction increase upon induction of neuronal differentiation with retinoic acid. In the neuronal cells within such cultures, the 13H9-positive forms of ezrin are enriched in the growth cone, while the bulk of ezrin identified by a polyclonal antibody shows no specific localization. In mouse DRG neurons, 13H9 staining is asymmetrically distributed along the edges of the complex growth cones of these cells. Staining of developing spinal cord in rat embryos also demonstrates that the 13H9-positive forms of ezrin do not colocalize with the majority of ezrin.(ABSTRACT TRUNCATED AT 250 WORDS)

In the development of the nervous system, one of the critical aspects is the proper navigation of axons to their targets, i.e. the problem of axonal guidance. We used the chick visual system as a model to investigate the role of the lysophospholipids lysophosphatidic acid (LPA) and sphingosine-1-phosphate (S1P) as potential axon guidance cues. We showed that both LPA and S1P cause a specific, dose-dependent growth cone collapse of retinal neurons in vitro in the chick model system, with slight differences compared to the mouse but very similar to observations in Xenopus. Because LPA and S1P receptors are G-protein-coupled receptors, we analyzed the intracellular signaling pathways using pharmacological inhibitors in chick retinal neurons. Blocking rho kinase (ROCK) prevented growth cone collapse by LPA and S1P, while blocking PLC or chelating calcium had no effect on growth cone collapse. Inhibition of G_i/o with pertussis toxin resulted in a partial reduction of growth cone collapse, both with LPA and with S1P. Inhibition of p38 blocked growth cone collapse mediated by LPA but not S1P. Thus, in addition to the involvement of the G_12/13-ROCK pathway, LPA- and S1P-induced collapse of chick retinal growth cones has a partial requirement for G_i/o.

One of the major requirements in the development of the visual system is axonal guidance of retinal ganglion cells toward correct targets in the brain. A novel class of extracellular lipid signaling molecules, lysophospholipids, may serve as potential axon guidance cues. They signal through cognate G protein-coupled receptors, at least some of which are expressed in the visual system. Here we show that in the mouse visual system, a lysophospholipid known as lysophosphatidic acid (LPA) is inhibitory to retinal neurites in vitro when delivered extracellularly, causing growth cone collapse and neurite retraction. This inhibitory effect of LPA is both active in the nanomolar range and specific compared to the related lysophospholipid, sphingosine 1-phosphate (S1P). Knockout mice lacking three of the five known LPA receptors, LPA1–3, continue to display retinal growth cone collapse and neurite retraction in response to LPA, demonstrating that these three receptors are not required for these inhibitory effects and indicating the existence of one or more functional LPA receptors expressed on mouse retinal neurites that can mediate neurite retraction.