Drug exposure and the risk of multiple sclerosis: A systematic review.

BACKGROUND: Several environmental and lifestyle factors have been associated with multiple sclerosis (MS) risk, including some pharmacological treatments. We systematically reviewed the literature on prescription drug exposure and MS risk. METHODS: Six databases were searched for original observational studies reporting drug exposure and MS risk published before 2017. RESULTS: Thirteen articles fulfilled inclusion criteria. Exposure to neither amiloride nor valproic acid was associated with MS (adjusted hazard ratio (adj.HR = 1.34;95% CI:0.81-2.20; adj.HR = 1.30;95%CI:0.44-3.80, respectively). Four studies explored oral contraceptive exposure and reported no association with MS; while a single study found an increased risk (odds ratio [adj.OR] = 1.52;95%CI:1.21-1.91). While penicillin exposure was associated with reduced risk of developing MS (adj.OR = 0.5;95%CI:0.3-0.9), a later study observed an elevated risk for penicillin (adj.OR = 1.21;95%CI:1.10-1.27) and all antibiotics (adj.OR = 1.41;95%CI:1.29-1.53), which was potentially attributed to underlying infection. Anti-tumor necrosis factor-alpha (TNFα) was not associated with MS risk in persons with inflammatory bowel disease (standard morbidity ratio = 4.2;95%CI:0.1-23.0) and arthritis (standardized incidence ratio = 1.38;95%CI:0.69-2.77); however, men exposed to anti-TNFα who also had arthritis and individuals with ankylosing spondylitis were at an increased risk (standardized incidence ratios = 3.91;95%CI:1.47-10.42 and 3.48;95%CI:1.45-8.37, respectively). A reduced risk of MS was observed with exposure to the beta2-adrenergic agonist fenoterol (adj.OR = 0.58;95%CI:0.45-0.76), and the sedating histamine 1-receptor antagonists (adj.OR = 0.2;95%CI:0.1-0.8), but not the non-sedating equivalent (adj.OR = 0.8;95%CI:0.4-1.6). CONCLUSIONS: The suggestion that some drugs may prevent MS is intriguing and warrants further study. In addition, further pharmacovigilance is needed to assess the safety of anti-TNFα drugs in specific populations in the context of MS risk.

Long-term behavioral impairment following acute embryonic ethanol exposure in zebrafish.

BACKGROUND: Developmental exposure to ethanol has long been known to cause persisting neurobehavioral impairment. However, the neural and behavioral mechanisms underlying these deficits and the importance of exposure timing are not well-characterized. Given the importance of timing and sequence in neurodevelopment it would be expected that alcohol intoxication at different developmental periods would result in distinct neurobehavioral consequences. METHODS: Zebrafish embryos were exposed to ethanol (0%, 1%, 3%) at either 8-10 or 24-27 h post-fertilization (hpf) then reared to adolescence and evaluated on several behavioral endpoints. Habituation to a repeated environmental stimulus and overall sensorimotor function were assessed using a tap startle test; measurements of anxiety and exploration behavior were made following introduction to a novel tank; and spatial discrimination learning was assessed using aversive control in a three-chambered apparatus. Overt signs of dysmorphogenesis were also scored (i.e. craniofacial malformations, including eye diameter and midbrain-hindbrain boundary morphology). RESULTS: Ethanol treated fish were more active both at baseline and following a tap stimulus compared to the control fish and were hyperactive when placed in a novel tank. These effects were more prominent following exposure at 24-27 hpf than with the earlier exposure window, for both dose groups. Increases in physical malformation were only present in the 3% ethanol group; all malformed fish were excluded from behavioral testing. DISCUSSION: These results suggest specific domains of behavior are affected following ethanol exposure, with some but not all of the tests revealing significant impairment. The behavioral phenotypes following distinct exposure windows described here can be used to help link cellular and molecular mechanisms of developmental ethanol exposure to functional neurobehavioral effects.

Molecular polymorphism of the intermediate filament protein transitin.

Transitin is an avian intermediate filament protein whose transient expression in the progenitor cells of the muscle and nerve tissues is similar to that of mammalian nestin. Both proteins contain an alpha-helical core domain flanked by a short N-terminal head and a long C-terminal extremity. However, the tail region of transitin is significantly different from that of nestin in that it harbors a unique motif containing more than 50 leucine zipper-like heptad repeats which is not found in any other intermediate filament protein. Despite the absence of introns in this region of the transitin gene, it was reported that different isoforms of the protein were produced by exclusion or inclusion of a number of repeats generated by an unusual splicing mechanism recognizing consensus 5' and 3' splice sites contained within the coding sequence of the heptad repeat domain [Napier et al. (1999) J Mol Neurosci 12:11-22]. Two monoclonal antibodies (mAbs) reacting with repeated epitopes of this motif were used to monitor transitin expression during in vitro myogenesis of the quail myogenic cell line QM7. Confocal microscopy revealed that the subcellular domains decorated with mAbs A2B11 and VAP-5 were mutually exclusive: the intermediate filament network visualized with mAb VAP-5 appeared to abut on a submembranous domain defined by mAb A2B11. When QM7 cells were induced to differentiate by switching to medium containing low serum components, an early effect was the local loss of A2B11 cortical staining at the points of cell-cell contacts. The A2B11 signal also disappeared before that of VAP-5 in newly formed myotubes. Unexpectedly, the mutually exclusive staining pattern of the mAbs could not be explained by alternative splicing since both epitopes mapped to a repeated element preceding the consensus 5' splice sites of the heptad repeat domain. An alternative explanation would be that the central repeat domain of transitin is a polymorphic structure from which different conformations exist depending on the local context. This hypothesis is strengthened by the observation that in cultured neural crest cells, the A2B11 antigen is preferentially expressed by freely migrating crest cells whose intracellular pH and calcium concentrations are different from those of non-migrating cells.

Avian transitin expression mirrors glial cell fate restrictions during neural crest development.

During development, trunk neural crest cells give rise to three primary classes of derivatives: glial cells, melanocytes, and neurons. As part of an effort to learn how neural crest diversification is regulated, we have produced monoclonal antibodies (MAbs) that recognize antigens expressed by neural crest cells early in development. One of these, MAb 7B3 (7B3), was found to recognize an avian transitin-like protein by co-immunostaining with a series of transitin-specific monoclonal antibodies and by Western blot analysis. In neural crest cell cultures, we found that 7B3 initially recognizes the majority of neural crest cells as they emerge from the neural tube. Subsequently, 7B3-immunoreactivity (IR) is progressively restricted to a smaller subpopulation of cells. In fully differentiated trunk neural crest cell cultures, 7B3-IR is expressed only by cells that do not express neuronal markers and lack melanin granules. During development in vivo, 7B3-IR is evident in neural crest cells on the medial, but not the lateral migration pathway, suggesting that it is not expressed by melanocyte precursors. Later, the antigen is detected in non-neuronal, presumptive glial cells in dorsal root ganglia (DRG) and sympathetic ganglia, as well as along ventral roots. Cultures of E5 DRG confirm that 7B3-IR is restricted to non-neuronal cells of ganglia, many of which closely associate with neuronal processes. Therefore, of the three major classes of differentiated trunk neural crest derivatives, 7B3 exclusively recognizes glial cells, including both satellite glia and Schwann cells. Since the pattern of 7B3 expression in vitro mirrors the pattern of glial cell fate-restrictions in the trunk neural crest lineage, and is expressed by neural crest-derived glia in vivo, we conclude that 7B3 is an early pan-glial marker for neural crest-derived glial cells and their precursors.

Localization of transitin mRNA, a nestin-like intermediate filament family member, in chicken radial glia processes.

We have examined the gene expression of two radial glia intermediate filament proteins, transitin and vimentin, in the developing chick CNS. Despite global similarities in their mRNA distributions, marked regional differences are observed. Most notably, we show that transitin mRNA is localized along radial glial processes and is localized to radial glia endfeet, whereas vimentin mRNA is not localized in radial glia. Localization of transitin mRNA is best shown in the diencephalic radial glia, as well as cerebellar Bergmann glia. In addition, in the early embryonic optic tectum, telencephalon, and retina, transitin mRNA is highly localized to radial glia endfeet, which is suggestive of its transport in these cells. These in vivo demonstrations of transitin mRNA localization are confirmed by in situ hybridization analysis of cultured chick brain radial glia, which demonstrates the presence of granular staining for transitin mRNA in glial processes. Transitin mRNA distribution in developing muscle also shows a highly regulated expression pattern, especially along the Z-lines of myofibrils. As further support for the transport and localization of transitin mRNA in radial glia and muscle, we have identified a consensus RNA transport signal in transitin mRNA that is absent from vimentin. These data suggest that the local regulation of transitin protein synthesis may contribute to its function as an intermediate filament protein in radial glia.

Agrin binds to beta-amyloid (Abeta), accelerates abeta fibril formation, and is localized to Abeta deposits in Alzheimer's disease brain.

Agrin is an extracellular matrix heparan sulfate proteoglycan (HSPG) well known for its role in modulation of the neuromuscular junction during development. Although agrin is one of the major HSPGs of the brain, its function there remains elusive. Here we provide evidence suggesting a possible function for agrin in Alzheimer's disease brain. Agrin protein binds the amyloidogenic peptide Abeta (1-40) in its fibrillar state via a mechanism that involves the heparan sulfate glycosaminoglycan chains of agrin. Furthermore, agrin is able to accelerate Abeta fibril formation and protect Abeta (1-40) from proteolysis, in vitro. Supporting a biological significance for these in vitro data, immunocytochemical studies demonstrate agrin's presence within senile plaques and cerebrovascular amyloid deposits, and agrin immunostained capillaries exhibit pathological alterations in AD brain. These data therefore suggest that agrin may be an important factor in the progression of Abeta peptide aggregation and/or its persistence in Alzheimer's disease brain.

Composition, synthesis, and assembly of the embryonic chick retinal basal lamina.

To study the biology of basal laminae in the developing nervous system the protein composition of the embryonic retinal basal lamina was investigated, the site of synthesis of its proteins in the eye was determined, and basal lamina assembly was studied in vivo in two assay systems. Laminin, nidogen, agrin, collagen IV, and XVIII are major constituents of the retinal basal lamina. However, only agrin is synthesized by the retina, whereas the other matrix constituents originate from cells of the ciliary body, the lens, or the optic disc. The synthesis from extraretinal tissues infers that the retinal basal lamina proteins must be shed from their tissues of origin into the vitreous body and from there bind to receptor proteins provided by the retinal neuroepithelium. The fact that all proteins typical for the retinal basal lamina are abundant in the vitreous body and a new basal lamina is only formed when the vitreous body was directly adjacent to the retina is consistent with the contention of the vitreous body having a function in retinal basal lamina formation. Basal lamina assembly was also studied after disrupting the retinal basal lamina by intraocular injection of collagenase. The basal lamina regenerated after chasing the collagenase with Matrigel, which served as a collagenase inhibitor. The basal lamina was reconstituted within 6 h. However, the regenerated basal lamina was located deeper in the retina than normal by reconstituting along the retracted neuroepithelial endfeet demonstrating that these endfeet are the preferred site of basal lamina assembly.

Development of the cardiac conduction system involves recruitment within a multipotent cardiomyogenic lineage.

The cardiac pacemaking and conduction system sets and maintains the rhythmic pumping action of the heart. Previously, we have shown that peripheral cells of the conduction network in chick (periarterial Purkinje fibers) are selected within a cardiomyogenic lineage and that this recruitment occurs as a result of paracrine cues from coronary arteries. At present, the cellular derivation of other elements of this specialized system (e.g. the nodes and bundles of the central conduction system) are controversial, with some proposing that the evidence supports a neurogenic and others a myogenic origin for these tissues. While such ontological questions remain, it is unlikely that progress can be made on the molecular mechanisms governing patterning and induction of the central conduction system. Here, we have undertaken lineage-tracing strategies based on the distinct properties of replication-incompetent adenoviral and retroviral lacZ-expressing constructs. Using these complementary approaches, it is shown that cells constituting both peripheral and central conduction tissues originate from cardiomyogenic progenitors present in the looped, tubular heart with no detectable contribution by migratory neuroectoderm-derived populations. Moreover, clonal analyses of retrovirally infected cells incorporated within any part of the conduction system suggest that such cells share closer lineage relationships with nearby contractive myocytes than with other, more distal elements of the conduction system. Differentiation birthdating by label dilution using [(3)H]thymidine also demonstrates the occurrence of ongoing myocyte conscription to conductive specialization and provides a time course for this active and localized selection process in different parts of the system. Together, these data suggest that the cardiac conduction system does not develop by outgrowth from a prespecified pool of 'primary' myogenic progenitors. Rather, its assembly and elaboration occur via processes that include progressive and localized recruitment of multipotent cardiomyogenic cells to the developing network of specialized cardiac tissues.

Identification of extracellular matrix ligands for the heparan sulfate proteoglycan agrin.

Agrin is a major brain heparan sulfate proteoglycan which is expressed in nearly all basal laminae and in early axonal pathways of the developing central nervous system. To further understand agrin's function during nervous system development, we have examined agrin's ability to interact with several heparin-binding extracellular matrix proteins. Our data show that agrin binds FGF-2 and thrombospondin by a heparan sulfate-dependent mechanism, merosin and laminin by both heparan sulfate-dependent and -independent mechanisms, and tenascin solely via agrin's protein core. Furthermore, agrin's heparan sulfate side chains encode a specificity in interactions with heparin-binding molecules since fibronectin and the cell adhesion molecule L1 do not bind agrin. Surface plasmon resonance studies (BIAcore) reveal a high affinity for agrin's interaction with FGF-2 and merosin (2.5 and 1.8 nM, respectively). Demonstrating a biological significance for these interactions, FGF-2, laminin, and tenascin copurify with immunopurified agrin and immunohistochemistry reveals a partial codistribution of agrin and its ECM ligands in the chick developing visual system. These studies and our previous studies, showing that merosin and NCAM also colocalize with agrin, provide evidence that agrin plays a crucial role in the function of the extracellular matrix and suggest a role for agrin in axon pathway development.

Characterization of the chicken transitin gene reveals a strong relationship to the nestin intermediate filament class.

Our laboratory previously reported that transitin is a radial glial intermediate filament protein sharing the basic structural features common to all intermediate filament (IF) proteins. It contains an alpha-helical core domain flanked by a short nonhelical head and a long COOH-terminal tail. The core sequence of transitin shows the greatest similarity to Xenopus tanabin and to rat and human nestin. We also reported that transitin has multiple splice variants derived from the deletion or inclusion of a leucine-zipper heptad repeat domain in the COOH-terminal tail. In the present study, we provide new evidence to support the classification of nestin and transitin in the same group of IF proteins based on the number and position of its introns. In addition, we suggest that the different isoforms of transitin are produced by a splicing mechanism that recognizes consensus 5' and 3' splice sites contained within the coding sequence of the leucine-zipper heptad repeat domain.

Collagen XVIII is a basement membrane heparan sulfate proteoglycan.

The present study shows that collagen XVIII is, next to perlecan and agrin, the third basal lamina heparan sulfate proteoglycan (HSPG) and the first collagen/proteoglycan with heparan sulfate side chains. By using monoclonal antibodies to an unidentified HSPG in chick, 14 cDNA clones were isolated from a chick yolk sac library. All clones had a common nucleotide sequence that was homologous to the mRNA sequences of mouse and human collagen XVIII. The deduced amino acid sequence of the chick fragment shows an 83% overall homology with the human and mouse collagen XVIII. Similar to the human and mouse homologue, the chick collagen XVIII mRNA has a size of 4.5 kilobase pairs. In Western blots, collagen XVIII appeared as a smear with a molecular mass of 300 kDa. After treatment with heparitinase, the protein was reduced in molecular mass by 120 kDa to a protein core of 180 kDa. Collagen XVIII has typical features of a collagen, such as its existence, under non-denaturing conditions, as a non-covalently linked oligomer, and a sensitivity of the core protein to collagenase digestion. It also has characteristics of an HSPG, such as long heparitinase-sensitive carbohydrate chains and a highly negative net charge. Collagen XVIII is abundant in basal laminae of the retina, epidermis, pia, cardiac and striated muscle, kidney, blood vessels, and lung. In situ hybridization showed that the main expression of collagen XVIII HSPG in the chick embryo is in the kidney and the peripheral nervous system. As a substrate, collagen XVIII moderately promoted the adhesion of Schwann cells but had no such activity on peripheral nervous system neurons and axons.

The developmentally regulated avian protein IFAPa-400 is transitin.

Transitin and IFAPa-400 are developmentally regulated high M(r) proteins expressed transiently in early chick embryogenesis. Both are associated with radially oriented fibers in the developing CNS and with various neural and myogenic tissues before their down-regulation at later stages. Previous studies have shown that IFAPa-400 colocalized and copurified with intermediate filament proteins and recent molecular cloning has indicated that transitin is a member of this family of cytoskeletal proteins. Here, we provide evidence that IFAPa-400 and transitin are the same protein. The sequence of a composite cDNA corresponding to more than 700 amino acids of IFAPa-400 carboxy-terminal extremity is identical to that of transitin. Both proteins exhibit identical apparent M(r) and isoelectric point. Immunopurified IFAPa-400 reacts with different antibodies to transitin and vice-versa. The patterns of expression of both proteins show a perfect coincidence at the tissue level. At the subcellular level, most antibodies to IFAPa-400/transitin decorate a typical intermediate filament network. However, monoclonal antibody A2B11, at the origin of transitin identification, exhibits a staining more typical of a cortical component, suggesting that different populations of transitin exist within the cell.

Immunocytochemical localization of a novel radial glial intermediate filament protein.

We have examined by immunocytochemistry the subcellular localization of a chick radial glial protein, named transitin, that by molecular cloning has been shown to be a novel member of the intermediate filament protein superfamily. In astrocytes cultured from E10 chick brain, transitin is localized to the intermediate filament network in accordance with its structural properties. Using confocal microscopy we examined the expression of transitin, vimentin and glial fibrillary acidic protein (GFAP) in cultured astrocytes, and show that transitin co-distributes with these other glial intermediate filament proteins. The expression of transitin, vimentin and GFAP was also compared in embryonic chick spinal cord and brain radial glia, with these studies showing that these intermediate filament proteins display distinct expression patterns during CNS development. Of particular note is the absence of vimentin and GFAP in spinal cord midline radial glia that express transitin protein, and a transient expression of transitin in brain midline radial glia that continue to express vimentin. Our studies presented here therefore indicate that transitin, a novel radial glial intermediate filament protein, may have functions that are unrelated to GFAP or vimentin during CNS development, since transitin is localized to the processes of midline radial glia and is transiently expressed during chick CNS development.

Distribution and substrate properties of agrin, a heparan sulfate proteoglycan of developing axonal pathways.

The distribution and substrate properties of agrin, an extracellular matrix heparan sulfate proteoglycan (HSPG), was investigated in the developing chick nervous system by immunocytochemistry, Western blotting, and in neurite outgrowth assays. By comparing the distribution of agrin with that of laminin-1, merosin (laminin-2), neurofilament, and neural cell adhesion molecule (NCAM), it was found that throughout development, agrin is a constituent of all basal laminae. From embryonic day (E) 4 onwards, agrin is also abundant in axonal pathways of the central nervous system, such as the optic nerve, the tectobulbar pathway, the white matter of the spinal cord, and the marginal and the molecular layers of the forebrain and the cerebellum. The abundance of agrin in brain decreases from E13 onwards. In the peripheral nervous system, agrin is present throughout development as a constituent of the Schwann cell basal laminae. Western blots confirmed the immunocytochemical data, showing maximum expression of agrin occurs during the early to medium stages of brain development. Western blots also showed that in mouse and human brain, agrin exists as an HSPG. Purified agrin did not support neurite outgrowth, rather it inhibited retinal neurite extension on mixed agrin/merosin substrates. Despite the fact that agrin, when used as a substrate inhibited neurite outgrowth, its temporal and spatial overlap with growing axons suggests that agrin has a supportive role in the development of axonal pathways, possibly as a binding component for growth factors and cell adhesion proteins.

Molecular cloning of a new intermediate filament protein expressed by radial glia and demonstration of alternative splicing in a novel heptad repeat region located in the carboxy-terminal tail domain.

In the present study we describe the molecular cloning of transitin, formerly named EAP-300. We show that transitin is an intermediate filament protein with a core domain most closely resembling nestin and tanabin. Transitin also contains a novel heptad amino acid repeat domain, comprising multiple leucine zipper repeats, located in its tail region. Based on these structural motifs we propose that a novel intermediate filament protein that is transiently expressed by radial glia during CNS development has been identified. We also show the existence of splice variants of transitin with splicing occurring in the novel heptad repeat domain to give rise to transitin isoforms that lack this heptad repeat. By in situ hybridization analysis we show that transitin mRNA is expressed by midline radial glial structures, by several axon commissures, and by Bergmann glia of the developing cerebelium. Based on the structural properties of the transitin protein, and expression of its mRNA, we suggest that transitin is a new member of the intermediate filament gene superfamily that is transiently expressed by radial glia.

An alternatively spliced, 5'-truncated MAP1B isoform is expressed in the developing chick nervous system.

Our laboratory has previously characterized a keratan sulfate proteoglycan, named claustrin, and shown by molecular cloning that claustrin and the mouse MAP1B protein share high homology, with claustrin representing a 5'-truncated fragment of MAP1B. In the present study, we examine further the relationship between claustrin and MAP1B, and also describe the isolation of a cDNA encoding the 3'-region of MAP1B, which shares 3'-untranslated sequence, but not coding sequence, with claustrin. We call this partial cDNA 3'-MAP1B-related clone (3'-MRC), since it is homologous to the 3'-region of the mouse MAP1B sequence. We show by Northern analysis that distinct mRNAs are recognized by the claustrin and 3'-MRC cDNAs, and by RT-PCR that mRNAs encoding these distinct MAP1B-related molecules are present in embryonic chick brain and cardiac and smooth muscle. Our data also suggest a higher level of expression of claustrin mRNA in astrocyte cultures, when compared to 3'-MRC. Our data therefore provide new evidence that alternatively spliced variants of MAP1B are expressed in brain, and that at least one of these variants encodes the claustrin proteoglycan.

Differential effects of diabetes and glomerulonephritis on glomerular basement membrane composition.

The hallmark of renal diseases involving the glomerulus is the presence of proteinuria. While the routes of pathogenesis of proteinuria have not been established, alterations in the barrier function of the glomerular basement membrane (GBM) have been implicated. We evaluated the effect of streptozotocin diabetes and passive Heymann nephritis (PHN) over time on the macromolecular composition of rat GBM to determine if changes in composition correlate with proteinuria. Six to twelve rats from each group (control, diabetic, and PHN) were sacrificed 1, 5, 28, 56, or 84 days after induction of disease. Identical amounts of GBM were subjected to a sequential extraction procedure, and type IV collagen, entactin, laminin, fibronectin, and anionic charge content were quantitated in the extracts. Type IV collagen and entactin content did not change with time or disease. Both laminin and fibronectin contents increased with time in GBM in all groups, but this increase was significantly greater in diabetic GBM. A significant decrease in anionic charge content of GBM coincided with the onset of albuminuria at Day 28 in diabetes, but no change was seen in PHN. In diabetic rats, the increase in laminin content over control preceded the onset of albuminuria, while the increase in fibronectin was not apparent until after albuminuria was present. In PHN, no differences in type IV collagen, entactin, laminin, fibronectin, or anionic charge content of GBM were found compared with control, even though profound albuminuria was evident from Day 5 through 84. Thus, while alterations in laminin and fibronectin content may contribute to the loss of glomerular permselectivity in streptozotocin diabetes, such changes apparently are not involved in PHN.

NCAM-mediated adhesion of transfected cells to agrin.

The vertebrate neural cell adhesion molecule NCAM mediates heterophilic adhesion to heparan sulfate proteoglycans in embryonic chick brain membranes. In this study, mouse L cells transfected with chicken NCAM were used to identify two of these ligands as agrin and the target of the 6C4 monoclonal antibody. A third heparan sulfate proteoglycan, perlecan, appeared not to support NCAM-mediated adhesion. Enzymatic degradation of chondroitin sulfates decreased adhesion in agrin-containing membrane fractions but increased adhesion if the agrin had previously been removed by immunoprecipitation, suggesting that interactions between heparan sulfate and chondroitin sulfate proteoglycans have important influences on adhesion. Our experiments support the view that NCAM can interact with multiple, but not with all, heparan sulfate and chondroitin sulfate proteoglycans in chick brain membranes in both positive and negative ways to influence cell adhesion.

Agrin: an extracellular matrix heparan sulfate proteoglycan involved in cell interactions and synaptogenesis.

Recent studies have documented important roles for heparan sulfate proteoglycans in the control of nervous system development. Agrin is an extracellular matrix protein identified and named based on its involvement in the aggregation of acetylcholine receptors (AChRs) during synaptogenesis at the neuromuscular junction. Recent studies have demonstrated that agrin is a large extracellular heparan sulfate proteoglycan, with a molecular mass in excess of 500 kDa and a protein core of 220 kDa. Emerging evidence indicates that agrin's function is not limited to its role in AChR aggregation during synaptogenesis, as the majority of agrin expression occurs in the developing central nervous system, especially in developing axonal tracts. This review examines recent studies suggesting a role for agrin in the regulation of cell-cell interactions, most notably by its ability to interact with the neural cell adhesion molecule. In addition, other potential roles for the heparan sulfate chains of agrin during nervous system development are explored.

Identification of a novel alternatively spliced agrin mRNA that is preferentially expressed in non-neuronal cells.

A novel agrin isoform was identified based on the isolation of an agrin cDNA from E9 chick brain that lacked 21 base pairs (bp) in the NH2-terminal encoding region of the agrin mRNA. Reverse transcription-polymerase chain reaction (RT-PCR) of E9 chick brain mRNA confirmed the existence of this agrin isoform in brain, although the novel splice variant represents a minor fraction of agrin mRNA in brain. However, upon analysis of chick brain astrocyte mRNA, smooth muscle mRNA, and cardiac muscle mRNA by RT-PCR, we show that this novel agrin isoform is the predominant agrin isoform in these non-neuronal cell populations. We extended our analyses to examine the expression of this agrin mRNA isoform during chick development and show that the agrin mRNA lacking this 21-bp exon is up-regulated with brain development, consistent with the increase in glial number during brain development, while the agrin isoform that does not undergo splicing and thus contains the 21-bp exon is down-regulated in brain development. Because the 21-bp exon is inserted in the region of chick agrin which encodes the putative signal sequence of agrin, with the signal peptidase site immediately preceding the putative first amino acid of the mature protein being deleted as a result of splicing, these data raise the interesting possibility that the presence or absence of this alternatively spliced exon may differentially regulate processing of the agrin protein in neuronal and non-neuronal cells, respectively.