Structural bases for central nervous system malfunction in the quaking mouse: dysmyelination in a potential model of schizophrenia.

The dysmyelinating mouse mutant quaking (qk) is thought to be a model of schizophrenia based on diminution of CNS myelin (Andreone et al., 2007) and downregulation of the Qk gene (Haroutunian et al., 2006) in the brains of schizophrenic patients. The purpose of this study was to identify specific structural defects in the qk mouse CNS that could compromise physiologic function and that in humans might account for some of the cognitive defects characteristic of schizophrenia. Ultrastructural analysis of qk mouse CNS myelinated fibers shows abnormalities in nodal, internodal, and paranodal regions, including marked variation in myelin thickness among neighboring fibers, spotty disruption of paranodal junctions, abnormal distribution of nodal and paranodal ion channel complexes, generalized thinning and incompactness of myelin, and on many axonal profiles complete absence of myelin. These structural defects are likely to cause abnormalities in conduction velocity, synchrony of activation, temporal ordering of signals, and other physiological parameters. We conclude that the structural abnormalities described are likely to be responsible for significant functional impairment both in the qk mouse CNS and in the human CNS with comparable myelin pathology.

Spinal cord dysmyelination caused by an antiproteolipid protein IgM antibody: implications for the mechanism of central nervous system myelin formation.

Antiglycolipid IgM antibodies are known to induce formation of "wide spaced" or "expanded" myelin, a distinctive form of dysmyelination characterized by a repeat period approximately two or three times normal, which is seen also in diseases, including multiple sclerosis. To determine whether an antibody directed against a myelin protein would cause equivalent pathology, we implanted O10 hybridoma cells into the spinal cord of adult or juvenile rats. O10 produces an IgM directed against PLP, the major protein of CNS myelin. Subsequent examination of the cords showed focal demyelination and remyelination. In addition, however, some juvenile cords, but none of the adult cords, displayed wide-spaced myelin with lamellae separated by an extracellular material comprising elements consistent with IgM molecules in appearance. Wide spacing tended to involve the outer layers of the sheath and in some cases alternated with normally spaced lamellae. A feature not seen previously consists of multiple expanded myelin lamellae in one sector of a sheath continuous with normally spaced lamellae in another, resulting in variation in sheath thickness around the axonal circumference. This uneven distribution of wide-spaced lamellae is most simply explained based on incorporation of IgM molecules into immature sheaths during myelin formation and implies a model of CNS myelinogenesis more complex than simple spiraling. The periaxonal space never displays widening of this kind, but the interface with adjacent myelin sheaths or oligodendrocytes may. Thus, wide spacing appears to require that IgM molecules bridge between two PLP-containing membranes and does not reflect the mere presence of immunoglobulin within the extracellular space.

Mice with conditional inactivation of fibroblast growth factor receptor-2 signaling in oligodendrocytes have normal myelin but display dramatic hyperactivity when combined with Cnp1 inactivation.

Fibroblast growth factor receptors (Fgfr) comprise a widely expressed family of developmental regulators implicated in oligodendrocyte (OL) maturation of the CNS. Fgfr2 is expressed by OLs in myelinated fiber tracks. In vitro, Fgfr2 is highly upregulated during OL terminal differentiation, and its activation leads to enhanced growth of OL processes and the formation of myelin-like membranes. To investigate the in vivo function of Fgfr2 signaling by myelinating glial cells, we inactivated the floxed Fgfr2 gene in mice that coexpress Cre recombinase (cre) as a knock-in gene into the OL-specific 2',3'-cyclic nucleotide phosphodiesterase (Cnp1) locus. Surprisingly, no obvious defects were detected in brain development of these conditional mutants, including the number of OLs, the onset and extent of myelination, the ultrastructure of myelin, and the expression level of myelin proteins. However, unexpectedly, a subset of these conditional Fgfr2 knock-out mice that are homozygous for cre and therefore are also Cnp1 null, displayed a dramatic hyperactive behavior starting at approximately 2 weeks of age. This hyperactivity was abolished by treatment with dopamine receptor antagonists or catecholamine biosynthesis inhibitors, suggesting that the symptoms involve a dysregulation of the dopaminergic system. Although the molecular mechanisms are presently unknown, this novel mouse model of hyperactivity demonstrates the potential involvement of OLs in neuropsychiatric disorders, as well as the nonpredictable role of genetic interactions in the behavioral phenotype of mice.

Sulfatide is essential for the maintenance of CNS myelin and axon structure.

Galactocerebroside (GalC) and sulfatide are abundant myelin lipids. In mice incapable of synthesizing these lipids, myelin is thin and regionally unstable and exhibits several subtle structural abnormalities. Although galactolipid-null mice have been beneficial in the analysis of galactolipid function, it has not been possible to differentiate between the functions of GalC and sulfatide with these mice alone. In the present work, we have analyzed a murine model that forms normal levels of GalC but is incapable of synthesizing sulfatide. By comparing a plethora of morphological features between the galactolipid-null and the sulfatide-null mice, we have begun to differentiate between the specific functions of these closely related lipids. The most striking difference between these two mutants is the reduction of myelin developmental abnormalities (e.g., redundant and uncompacted myelin sheaths) in young adult sulfatide-null mice as compared with the galactolipid-null animals. Although sulfatide appears to play a limited role in myelin development, this lipid is essential for myelin maintenance, as the prevalence of redundant, uncompacted, and degenerating myelin sheaths as well as deteriorating nodal/paranodal structure is increased significantly in aged sulfatide-null mice as compared with littermate wildtype mice. Finally, we show that the role played by sulfatide in CNS maintenance is not limited to the myelin sheath, as axonal caliber and circularity are normal in young adult mutant mice but are significantly altered in aged sulfatide-null animals.

Axon-glia interactions and the domain organization of myelinated axons requires neurexin IV/Caspr/Paranodin.

Myelinated fibers are organized into distinct domains that are necessary for saltatory conduction. These domains include the nodes of Ranvier and the flanking paranodal regions where glial cells closely appose and form specialized septate-like junctions with axons. These junctions contain a Drosophila Neurexin IV-related protein, Caspr/Paranodin (NCP1). Mice that lack NCP1 exhibit tremor, ataxia, and significant motor paresis. In the absence of NCP1, normal paranodal junctions fail to form, and the organization of the paranodal loops is disrupted. Contactin is undetectable in the paranodes, and K(+) channels are displaced from the juxtaparanodal into the paranodal domains. Loss of NCP1 also results in a severe decrease in peripheral nerve conduction velocity. These results show a critical role for NCP1 in the delineation of specific axonal domains and the axon-glia interactions required for normal saltatory conduction.

No obvious abnormality in mice deficient in receptor protein tyrosine phosphatase beta.

The development of neurons and glia is governed by a multitude of extracellular signals that control protein tyrosine phosphorylation, a process regulated by the action of protein tyrosine kinases and protein tyrosine phosphatases (PTPs). Receptor PTPbeta (RPTPbeta; also known as PTPzeta) is expressed predominantly in the nervous system and exhibits structural features common to cell adhesion proteins, suggesting that this phosphatase participates in cell-cell communication. It has been proposed that the three isoforms of RPTPbeta play a role in regulation of neuronal migration, neurite outgrowth, and gliogenesis. To investigate the biological functions of this PTP, we have generated mice deficient in RPTPbeta. RPTPbeta-deficient mice are viable, are fertile, and showed no gross anatomical alterations in the nervous system or other organs. In contrast to results of in vitro experiments, our study demonstrates that RPTPbeta is not essential for neurite outgrowth and node formation in mice. The ultrastructure of nerves of the central nervous system in RPTPbeta-deficient mice suggests a fragility of myelin. However, conduction velocity was not altered in RPTPbeta-deficient mice. The normal development of neurons and glia in RPTPbeta-deficient mice demonstrates that RPTPbeta function is not necessary for these processes in vivo or that loss of RPTPbeta can be compensated for by other PTPs expressed in the nervous system.

A brief history of myelinated nerve fibers: one hundred and fifty years of controversy.

The early controversies over myelinated nerve fibers focused on whether nerves are hollow or not, whether the fatty "marrow" (myelin) is inside the nerve fiber or around it, whether myelin is secreted by the axon or formed by another cell, whether nerve fibers are discrete or part of a syncytial network, whether nodes of Ranvier are present in central myelin or only in peripheral myelin. Since Geren's seminal discovery that peripheral myelin is formed by the Schwann cell plasma membrane wrapped around the axon, the focus has shifted. Myelin is clearly a living cell appendage, and the myelin sheath is dependent upon intercellular interactions not only during its formation, but throughout its lifetime and during pathological processes affecting either the axon or the myelin-forming cell. The myelinated fiber is a functional unit, an exquisite symbiosis, whose ability to perform optimally, in some cases whose very survival, depends on the effects the respective cells exert on one another. How are these interactions mediated? Which structures and functions depend on such interaction and which are independent of it? How do cells of the size and shape of myelin-forming cells cope with their metabolic demands and support their most distal components? What are the mechanisms and mutual consequences of demyelination or axonopathy? Relevant studies have burgeoned with the development of molecular biological and genetic engineering methods, and with improvements in microscopy, in vitro culture and specific immunostaining methods. This introductory essay provides an overview of the structural background and continuing controversies relevant to the articles that follow, which represent a sampling of current work and present new information on the molecular structure, function and pathology of myelin and axoglial interactions.

Antibody-mediated CNS demyelination: focal spinal cord lesions induced by implantation of an IgM anti-galactocerebroside-secreting hybridoma.

O1 hybridoma cells, which secrete an IgM antigalactocerebroside, were implanted into the spinal cord of cyclosporine-treated juvenile or adult rats, and the animals were sacrificed approximately 2-3 wk later. About half the recipient animals developed myelin lesions. In some, sharply circumscribed foci of demyelination formed within the dorsal columns. Cellular reaction consisted of macrophages containing refractile globules in the parenchyma and within enlarged perivascular spaces as well as thickened endothelial cells. "Shadow plaques" also developed, i.e. regions in which axons were surrounded by thin myelin sheaths, compatible with remyelination. In addition, we found damaged axons, some of which were swollen with organelles, comparable to the enlarged axon profiles seen at sites of constriction or interruption. Compromise of the blood-brain barrier at sites of hybridoma growth was demonstrated by extravasation of Evans blue dye. Discontinuation of cyclosporine was followed by an anti-hybridoma, complement-fixing antibody response within 2-3 d. This model of focal CNS demyelination and remyelination, with evidence of some axon damage, is mediated by a defined IgM antiglycolipid monoclonal antibody secreted within the spinal cord parenchyma. The lesions, which are similar to those of multiple sclerosis, probably result from the interaction between the intrathecally secreted IgM antibody and complement entering from the circulation at foci of compromised blood-brain barrier plus activation of endogenous or hematogenous macrophages via their complement receptors.

Xenotransplantation of transgenic oligodendrocyte-lineage cells into spinal cord-injured adult rats.

Spinal cord trauma is associated not only with loss of nerve cells and fibers but also with damage to oligodendrocytes and demyelination. In order to assess the potential of transplanted oligodendrocyte-lineage cells to repair the demyelination that follows spinal cord injury, we have used donor glia derived from a transgenic mouse line containing the LacZ transgene under control of the myelin basic protein promoter. Glia derived from fetal or neonatal transgenic mice were injected into the spinal cords of immunosuppressed adult rats at the site of an experimental traumatic lesion 1-16 days after injury. Cells expressing LacZ were identified 15-18 days later in cryosections rostral and caudal to the transplant site, most conspicuously within white matter defects. Some of these cells within the dorsal columns gave rise to approximately 30- to 60-microns processes, consistent with myelin segments, which are oriented parallel to the fiber tract. Glial transplantation may thus be a feasible means of replacing damaged host oligodendrocytes with donor oligodendrocyte-lineage cells capable of reforming myelin and potentially restoring functional lost as a result of demyelination associated with spinal cord injury.

Spinal cord dysmyelination induced in vivo by IgM antibodies to three different myelin glycolipids.

It was shown previously (Rosenbluth et al.: J. Neurosci. 16:2635-2641, 1996) that implantation of hybridoma cells that produce an IgM antigalactocerebroside into the spinal cord of young rats results in the development of myelin sheaths with a repeat period approximately 2-3x normal, similar to the abnormal peripheral myelin sheaths seen in human IgM gammopathies. We now present evidence that this effect can be reproduced in the spinal cord by implanting either of two other hybridomas, O4 and A2B5, that secrete, respectively, antisulfatide and antiganglioside IgM antibodies. The formation of expanded CNS myelin thus does not depend on antibodies to galactocerebroside specifically but can be mediated by IgM antibodies that react with other myelin glycolipids as well.

Myelin structure in proteolipid protein (PLP)-null mouse spinal cord.

Fixed preparations of proteolipid protein (PLP)-null mouse spinal cord show myelin sheaths which in some regions consist of typical alternating major dense lines (MDLs) and intermediate lines (ILs) with a repeat period of 10.3 nm. More commonly, the lamellar structure consists of what appears to be a single population of dense lines, having a repeat period of 5.2 nm. These apparently equivalent lines are, however, sometimes distinguishable as MDLs or ILs based on continuity with cytoplasmic or extracellular regions. Focal separations of lamellae at the intermediate line are common. MDLs too may be replaced focally by cytoplasmic pockets, sometimes in the same quadrant over several lamellae, resembling Schmidt-Lanterman clefts. Occasional densities reminiscent of the "radial component" can be seen. Otherwise, this structure, which is prominent in wild-type myelin, is conspicuously absent. Redundant folding of some lamellae but not others may occur in the same sheath. These observations conform to those made previously on the isolated myelin segments that occur in the myelin-deficient rat central nervous system (CNS), which also lacks PLP. Thus, a compact lamellar structure can be seen in fixed PLP-null myelin, but defects in the apposition of both the extracellular and the cytoplasmic surfaces of the myelin membranes are common. The abnormalities seen suggest a lack of firm intermembrane bonding, resulting in structural instability. PLP-null myelin may therefore be more susceptible than normal myelin to disruption by mechanical or osmotic stresses. Although PLP is not essential for the formation of either major dense lines or intermediate lines, it may play a role in stabilizing the compact structure.

Expanded CNS myelin sheaths formed in situ in the presence of an IgM antigalactocerebroside-producing hybridoma.

When O1 hybridoma cells, which produce an IgM antigalacto-cerebroside, are implanted into the dorsal columns of 4-8 d rat spinal cord, some of the myelin that subsequently develops in the immediate vicinity displays an abnormal periodicity. The spacings that are seen cluster at approximately 19 nm and 31 nm, roughly two and three times the normal 11 nm spacing. In the expanded sheaths, major dense lines are separated by broad extracellular spaces containing a dense material in which single or double rows of approximately 10 nm circular profiles can be identified, consistent with the "central rings" of IgM molecules. Because IgM is multivalent, it may serve to link adjacent lamellae together in place of intrinsic myelin molecules that normally interact at close range. Extensive direct contact between myelin components of successive myelin lamellae is thus not essential to signal the growth of the oligodendrocyte membrane or the spiral wrapping of that membrane around axons during myelinogenesis, or to stabilize the myelin spiral that forms.

Paranodal structural abnormalities in rat CNS myelin developing in vivo in the presence of implanted O1 hybridoma cells.

O1 hybridoma cells, which produce a monoclonal IgM antigalactocerebroside, were implanted into the spinal cords of immature and mature rats and the cords examined 5-24 days later. Study of the younger group, in which myelin was developing at the time of implantation, revealed examples of abnormal myelin sheaths in which the repeat period was markedly increased. The paranodal regions of these abnormal sheaths were superficially normal in configuration; i.e. myelin lamellae terminated one by one as 'terminal loops' that indented the axolemma and formed normal axoglial junctions displaying periodic 'transverse bands'. Neighbouring terminal loops are normally joined by tight junctions that block passage of tracers from the paranodal periaxonal space into the compact myelin, as seen after implantation of a control hybridoma. In the abnormal sheaths that developed after O1 implantation, in contrast, terminal loops were usually widely separated from each other. As a result, multiple pathways from the paranodal periaxonal space into the myelin sheath remained patent, forming potential routes for shunting nodal action currents. This subtle abnormality could thus compromise conduction, even though the sheaths might appear to be normally myelinated at the histological level. Equivalent abnormalities in human neurological diseases, including multiple sclerosis and paraproteinemic neuropathies, could underlie functional loss in the absence of frank demyelination.

Distribution of myelin lipid antigens in adult and developing rat spinal cord.

We examined the distribution of myelin antigens recognized by monoclonal antibodies (mAbs) 01 and 04 in the developing ventral white matter of the cervical spinal cord of the rat using immunogold-labeled ultrathin cryosections. From the beginning of myelination after birth to multilamellar myelin in adult animals, we observed colocalization of 04 and 01 label in myelin. In the oligodendrocyte soma, immunolabel was found primarily over Golgi cisternae. In the oligodendrocyte processes, immunolabeling was also found in the cytoplasm and along the plasmalemma. More cytoplasmic 04 and 01 label was found in the external loop of myelin than in the internal loop. The amount of 01 and 04 label increased over compact myelin in proportion to the number of lamellae, but the label density per unit length of membrane remained approximately the same in compact myelin as in oligodendrocyte plasmalemma. We did not see a concentration gradient for either 04 or 01 label across, or along multilamellar myelin sheaths.

Inhibition of CNS myelin development in vivo by implantation of anti-GalC hybridoma cells.

Implantation of hybridoma cells that secrete a monoclonal antigalactocerebroside into the dorsal columns of < or = 9-day-old rat spinal cord results in failure of development of dorsal column myelin in the vicinity of the implant. Clusters of apparently undamaged amyelinated axons remain among the hybridoma cells. Ventral myelin is unaffected. These in vivo results support antibody-mediated inhibition of myelin formation as a potential mechanism underlying failure of remyelination in multiple sclerosis.

Effects of cerebellar lesions on tonic seizures, tremor and lifespan in myelin-deficient rats.

In common with other dysmyelinating mutants, the myelin-deficient rat displays an action tremor and tonic seizures culminating in the death of the animals at approximately 23-26 days. We find that deep lesions of the cerebellar vermis alleviate the manifestations of the myelin deficiency significantly. Such lesions introduced at 20 days or later eliminate both tremor and seizures for periods up to 10 days. Lifespan is prolonged to nearly 30 days, on average, and to 35 days in some cases. Shallow lesions of the vermis or lateral lobe lesions have relatively little effect. Based on these observations we suggest that the cerebellum contributes not only to the action tremor but also to the tonic seizures characteristic of central myelin deficiency. Spontaneous activity originating in myelin-deficient fiber tracts may be carried to the cerebellum and processed there to produce a highly amplified and/or synchronized output to broad areas of the neuraxis. Deep lesions of the vermis presumably interfere with cerebellar output and compromise the cerebellar contribution to the seizures. Tonic seizures and other 'paroxysmal attacks' also occur commonly in human demyelinating diseases including multiple sclerosis [11]. Manipulation of cerebellar output offers a potential approach to the control of such spontaneous activity.

Myelin formation by mouse glia in myelin-deficient rats treated with cyclosporine.

Previous attempts to generate myelin in the myelin-deficient rat spinal cord by transplanting mouse glia were not successful. In order to determine whether this result was due to graft rejection or to interspecies mismatch of cellular or molecular components at the axoglial junction, we have repeated the experiment in cyclosporine-treated rats. Our results show that in the immunosuppressed hosts, foetal glial xenografts form an abundance of myelin within the dorsal columns at or near the injection site about two weeks after the operation. In some cases, myelination extends virtually across the entire width of the dorsal columns. Ultrastructurally, the myelin sheaths are normal in all respects, including the presence of the 'radial component'. The lateral edges of the myelin lamellae form typical paranodal axoglial junctions, some displaying periodic 'transverse bands'. We infer that previous mouse to rat xenograft failures reflect host immune response rather than mismatch of heterologous junctional components. We also compared foetal, early post-natal and adult xenografts. Foetal donor cells, containing an abundance of precursors but virtually no mature oligodendrocytes, are more effective than neonatal donor cells in forming myelin, and after adult grafts, we found no myelin formation. Thus, in xenografts, as in allografts, foetal precursor cells are far more suitable than glia from mature donors in generating significant amounts of myelin.

Structural abnormalities in freeze-fractured sciatic nerve fibres of diabetic mice.

Nodal and paranodal regions of myelinated sciatic nerve fibres from diabetic (db/db) mice were examined in freeze fracture replicas. In some fibres, the axolemma was found to display abnormalities in the paranodal region. These include shallow, undifferentiated junctional indentations, thinning of the indentations with widening of the non-junctional grooves between them, particle clusters within the non-junctional grooves, and patches in which axolemmal E-face particles are distributed randomly rather than in the form of linear strings within grooves. Nodal structure, in contrast, is hardly affected. Nodal E-face and P-face particle densities in db/db axons are not significantly different from those in age-matched controls, although we found a few examples in which the E-face density fell slightly below the normal range. Occasional fibres showing evidence of paranodal or segmental demyelination were also seen. The results support paranodal pathology as a potential basis for reduced nerve conduction velocity in diabetic nerves but provide no evidence for significant changes in nodal structure or in nodal Na channel density in sciatic nerve fibres of the db/db mouse.

Transplantation of labeled fetal spinal cord fragments into juvenile myelin-deficient rat spinal cord.

Minced and triturated fragments from the spinal cord of normal rat fetuses (15-18 days gestation) labeled with the fluorescent dye fast blue (FB) were successfully transplanted into juvenile myelin-deficient rat spinal cord under direct observation. Clusters of myelinated fibers were found subsequently in the recipient spinal cord, and, by fluorescence microscopy, clusters of FB-labeled cells were found at corresponding sites. The results indicate that the surgical approach used is suitable for transplantation of tissue fragments into a defined region of juvenile rat spinal cord, that FB can be used to locate the transplanted cells subsequently, and that FB does not interfere with maturation of the donor glia or with myelin formation.