《Nature Neuroscience》：与多发性硬化症有关的古老病毒

多发性硬化症（MS）是一种中枢神经系统疾病，全世界患者约有100万人，迄今这一疾病还无法治愈。新研究发现一种能够侵染5000万年前灵长类祖先的古老病毒可能与造成多发性硬化症的神经系统损伤有关。这种病毒DNA的一个片断（现在已经整合进人类基因组）能够刺激患类似MS的小鼠大脑中破坏性化合物的产生。这些研究的报告发表在十月份的Nature Neuroscience上。
对于多发性硬化症的病因有两种假说：病毒感染假说和自家免疫反应假说。但是自家免疫反应假说缺乏证据。有研究者指出已经发现有许多种感染性物质都与多发性硬化症有关。
新的研究表明造成这种疾病的罪魁祸首可能已经藏进了人类DNA中。在人类的基因组中，大约有8％是由在进化过程中整合进基因组中的病毒片断构成。但是大部分的这种古病毒不再有什么功能了。但是HERVs-W（人类内源性反转录病毒W）则是个例外，这种病毒携带制造syncytin蛋白的指令，而这种蛋白是胎盘形成的关键成分。但是这种病毒可能也有副作用。
研究过程中，卡尔加里大学的Christopher Power领导的研究组发现16个患MS的尸体的大脑的synctin水平比对照高三倍，这意味着这种蛋白可能是影响这种疾病的一个因素。研究人员用培养的人类大脑细胞进行实验时发现syncytin的过量生产能够导致自由基的释放。并且，研究人员能够通过给MS小鼠喂食抗过氧化剂来防止这种损伤的发生。但是，研究人员仍然不能确定造成syncytin过量生产的根本原因。
这项研究为研究外源DNA如何影响人类疾病开辟了道路。而且，这项研究将促进用抗生素治疗MS的新疗法的研究。

Nature Neuroscience 7, 1088 - 1095 (2004)
Published online: 26 September 2004; | doi:10.1038/nn1319

Human endogenous retrovirus glycoprotein–mediated induction of redox reactants causes oligodendrocyte death and demyelination

Joseph M Antony1, Guido van Marle1, Wycliffe Opii2, D Allan Butterfield2, François Mallet3, Voon Wee Yong1, John L Wallace4, Robert M Deacon5, Kenneth Warren6 & Christopher Power1
1 Department of Clinical Neurosciences, University of Calgary, Calgary, Alberta T2N 4N1, Canada.

2 Department of Chemistry and Center of Membrane Sciences, University of Kentucky, Lexington, Kentucky 40506-0055, USA.

3 UMR CNRS-bioMerieux, IFR128 BioSciences Lyon-Gerland, Ecole Normale Superieure de Lyon, Lyon 69364, France.

4 Department of Pharmacology & Therapeutics, University of Calgary, Calgary, Alberta T2N 4N1, Canada.

5 Department of Experimental Psychology, University of Oxford, Oxford OX1 3UD, UK.

6 Department of Medicine, University of Alberta, Edmonton, Alberta T6G 2B7, Canada.

Correspondence should be addressed to Christopher Power power@ucalgary.ca


Human endogenous retroviruses (HERVs) constitute 8% of the human genome and have been implicated in both health and disease. Increased HERV gene activity occurs in immunologically activated glia, although the consequences of HERV expression in the nervous system remain uncertain. Here, we report that the HERV-W encoded glycoprotein syncytin is upregulated in glial cells within acute demyelinating lesions of multiple sclerosis patients. Syncytin expression in astrocytes induced the release of redox reactants, which were cytotoxic to oligodendrocytes. Syncytin-mediated neuroinflammation and death of oligodendrocytes, with the ensuing neurobehavioral deficits, were prevented by the antioxidant ferulic acid in a mouse model of multiple sclerosis. Thus, syncytin's proinflammatory properties in the nervous system demonstrate a novel role for an endogenous retrovirus protein, which may be a target for therapeutic intervention.


As much as 8% of the human genome is derived from retrovirus-like elements, which are presumably remnants of retroviral infections that occurred during primate evolution1, 2, 3. Many human endogenous retroviruses (HERVs) have retained functional promoter, enhancer and polyadenylation signals, and these regulatory sequences have the potential to modify the expression of adjacent genes4, 5. Despite the fact that most HERVs are unable to replicate because of mutations in structural retrovirus genes, specific open reading frames corresponding to HERV genes encode detectable proteins6.

An increase in the expression of HERV genes may be important in modulating host innate and adaptive immune responses with ensuing effects on the development of disease, although definitive proof of specific pathogenic effects linked to HERVs is lacking. Several human and animal exogenous retrovirus proteins, particularly the envelope proteins encoded by env, show a proclivity for causing neuropathogenic effects7. The precise functions of endogenous retrovirus proteins in the nervous system remain uncertain, despite the abundant expression of these proteins in many species including rodents, cats and non-human primates4, 5, 7.

Studies have suggested that HERV expression in human brain is augmented in conditions of neuroinflammation8. The prototypic neuroinflammatory disease multiple sclerosis is characterized by infiltration of inflammatory cells, damage to and death of oligodendrocytes and demyelination, which result in physical and cognitive disabilities. Indeed, cytokines, arachidonic acid metabolites and redox reactants including nitric oxide are key determinants of pathogenicity in multiple sclerosis9, 10, 11, which is also influenced by the genetic susceptibility of an individual. The role of both exogenous and endogenous infectious pathogens in the pathogenesis of multiple sclerosis is unknown, but several viruses and bacteria have been implicated through specific mechanisms including transactivation of aberrant immune responses and molecular mimicry.

Here we have examined the effects of HERV expression on neural cell function and survival, focusing in particular on the HERV-W envelope glycoprotein syncytin and its pathogenic effects.

Results
HERV-W env is upregulated in multiple sclerosis lesions
To investigate the expression of phylogenetically related HERV genes (Fig. 1a) in neuroinflammatory diseases, we examined the abundance of different HERV env mRNAs in brains from individuals diagnosed with multiple sclerosis or with other neurological diseases as controls. HERV-W env mRNA expression was selectively upregulated in brain tissue from individuals with multiple sclerosis as compared with controls, whereas other HERV env mRNAs were not (Fig. 1b). Sequencing of the resulting polymerase chain reaction (PCR) products confirmed that HERV-W env was overexpressed in the brain samples on the basis of comparisons with the HERV-W env sequence and other HERV env sequences in GenBan ( http://www.ncbi.nlm.nih.gov/genbank). Western blotting indicated that syncytin (75 kDa), the protein encoded by HERV-W env, was also present in brain tissue from individuals with multiple sclerosis but showed limited expression in controls (Fig. 1c). Comparison of syncytin immunoreactivity in brain showed a roughly 3.0-fold increase in the multiple sclerosis group relative to controls (Fig. 1d).


Figure 1. In vivo expression of HERV env.

(a) Phylogenetic tree showing the evolutionary relationship between different endogenous (ERV-FRD, MuLV, ZFERV, HERV-W, -E, -H, -K) and exogenous (HIV, HTLV-1) retrovirus env genes in mice, fish, cats, non-human primates and humans including those studied here. (b) Real-time RT-PCR analysis, showing a significant (threefold) increase in the mRNA of HERV-W env in individuals with multiple sclerosis (MS) relative to other HERVs and to control individuals affected with other neurological diseases. (c) Representative western blot of brain tissue lysates from individuals with MS, showing an increase in syncytin immunoreactivity as compared with controls. (d) Quantification of western blots, showing an increase in syncytin immuno-reactivity in MS brain as compared with controls. (e–j) Active lesions from MS brain show an increase in syncytin expression. (e,f) Active demyelinating (D) lesion from frontal lobe sections of a MS brain, showing myelin debris in macrophages (e) as compared with normal myelin (f). Sections were stained with Luxol fast blue and hematoxylin and eosin. (g,h) Serial sections from the same active lesion, showing an increase in syncytin expression in an area of active demyelination (g) that is absent in control sections (h). iNOS immunoreactivity (dark blue) is colocalized with syncytin (brown) in glia (g, inset). (i) Double-label immunohistochemical assessment of acute lesions in MS brain, showing that syncytin expression (blue) in activated astrocytes colocalizes (arrows) with GFAP immunoreactivity (brown). Inset, GFAP-immunoreactive astrocyte. (j) As i, but showing that syncytin expression (brown) in microglia and macrophages colocalizes (arrowheads) with Iba-1 immunoreactivity (blue). Inset, Iba-1–immunoreactive microglia. Original magnification, 50 (e–h); 400 (i,j); 1,000 (g,i,j, insets). Values are the mean s.e.m. *P < 0.05, **P < 0.01.

Full Figure and legend (49K)


In the multiple sclerosis group, acute lesions containing active demyelination with numerous lipid- or myelin-filled macrophages and hypertrophied astrocytes (Fig. 1e) showed syncytin immunoreactivity (Fig. 1g) in cells resembling activated glia, which also contained inducible nitric oxide synthase (iNOS) immunoreactivity (Fig. 1g, inset). Lipid vacuole–filled syncytin-immunopositive cells resembling phagocytic macrophages were evident at the margins and the cores (Supplementary Fig. 1 online) of both acute and chronic demyelinating lesions (Table 1). Syncytin-positive cells were not detected in control brain sections (Fig. 1h), however, which showed normal myelination (Fig. 1f).


Table 1. Clinical and neuropathological features of individuals with multiple sclerosis showing syncytin immunoreactivity in demyelinating lesions



Full Table


Because astrocytes12 and microglia13 are important modulators of neuroinflammation, we determined whether syncytin expression was selectively upregulated in these immunologically active cells. Double-label immunohistochemical assessment detected enhanced expression of syncytin in the astrocytes (Fig. 1i) and microglia (Fig. 1j) of brain sections (frontal white matter) from individuals with multiple sclerosis, but not in other neural cells including neurons and myelin-forming oligodendrocytes (data not shown).

To examine the relative expression of different HERV env mRNAs in specific types of cell that have been implicated in neuroinflammation, we studied human cell lines with and without cellular activation. In monocytoid (U937) cells treated with phorbol-12-myristate-13-acetate (PMA), peak mean HERV expression was detected at 4 h, along with a significant increase in HERV-W env mRNA as compared with other HERV mRNAs (Supplementary Fig. 2 online). In PMA-treated astrocytic (U373) cells, the most highly expressed HERV was HERV-H env mRNA at 24 h, but there was also a significant increase in HERV-W env. By contrast, PMA stimulation did not induce HERV env expression in peripheral blood lymphocytes (PBLs), although expression of interleukin-1 (IL-1) was significantly increased (Supplementary Fig. 2 online), similar to our findings in monocytoid and astrocytic cells (data not shown). In fact, we observed suppression of HERV env mRNA when PBLs were stimulated with PMA.

Taken together, these studies indicate that in individuals affected with multiple sclerosis, syncytin is upregulated in active demyelinating lesions and shows selective expression in cells that mediate neuroinflammation.

Syncytin activates proinflammatory molecules in glia
Because syncytin was abundantly expressed in vivo in brain tissue from individuals with multiple sclerosis (Fig. 1b–d,g,i,j), we constructed a SINrep5-based vector that efficiently expressed the HERV-W env open reading frame encoding syncytin (Fig. 2a) and that also mediated syncytia formation in baby hamster kidney (BHK) cells (Fig. 2b). The SINrep5-syncytin virus infected human fetal astrocytes (HFAs; Fig. 2f) which expressed glial fibrillary acidic protein (GFAP; Fig. 2c–e) and macrophages, but not oligodendrocytes (data not shown). We examined the syncytin-mediated induction of genes related to neuroinflammation 24 h after infection, which showed that the proinflammatory cytokine IL-1 was significantly increased in both HFAs (Fig. 2g) and monocyte-derived macrophages (MDMs; Fig. 2h) infected with SINrep5-syncytin, as compared with control cells including cells infected with SINrep5 expressing enhanced green fluorescent protein (EGFP) and mock-infected cells. In addition, the mean expression of iNOS was enhanced in HFAs (Fig. 2g), but not in MDMs (Fig. 2h), after infection with SINrep5-syncytin. The anti-inflammatory cytokine IL-10, however, was not induced in either cell type, suggesting that syncytin expression selectively induces proinflammatory responses.


Figure 2. Syncytin induces proinflammatory molecules in glial cells.

(a,b) Analysis of BHK cells transfected with SINrep5-syncytin. Syncytin immunoreactivity by western blotting (a) and syncytia formation (b) were detected in cells transfected with SINrep5-syncytin, but not in mock-infected cells or cells transfected with SINrep5-EGFP. (c–f) Confocal microscopy of uninfected HFAs immunostained for GFAP (c) and syncytin (d), and astrocytes infected with SINrep5-syncytin (MOI 1.0) immunostained for GFAP (e) and syncytin (f). (g) HFAs infected with SINrep5-syncytin showed a significant increase in IL-1 and iNOS mRNA expression and an increase in protein carbonyl levels relative to cells infected with SINrep5-EGFP and mock-infected cells, but showed similar expression of IL-10 mRNA. (h) Expression of IL-1 mRNA was increased in MDMs relative to controls after SINrep5-syncytin infection, whereas that of IL-10 and iNOS mRNA and levels of protein carbonyls did not differ. Original magnification, 1,000 (b); 600 (c–f); **P < 0.01, ***P < 0.001.

Full Figure and legend (33K)


As compared with controls, a marked increase in mean protein carbonyl levels was also observed in the conditioned medium from HFAs infected with SINrep5-syncytin (Fig. 2g), but not in that derived from MDMs (Fig. 2h). Conversely, there was no significant difference in the mean quantities of malondialdehyde (a product of lipid peroxidation)14 in conditioned medium from SINrep5-syncytin infected HFAs or MDMs as compared with controls (Supplementary Fig. 3 online). These observations indicated that syncytin induces a proinflammatory molecular profile in astrocytes that includes an increase in the oxidation of cellular proteins.

Syncytin causes oligodendrocyte damage and death
Because oligodendrocytes are the principal cell type that is susceptible to injury associated with neuroinflammation and demyelination, we examined their morphology and survival after treatment with conditioned medium from HFAs and MDMs infected with SINrep5-syncytin. Conditioned medium from HFAs infected with SINrep5-syncytin was highly cytotoxic to human oligodendrocytes as compared with that derived from control HFAs that were either infected with SINrep5-EGFP or mock infected (Fig. 3a). Notably, conditioned medium from HFAs infected with SINrep5-syncytin also induced a higher mean amount of oligodendrocyte death than did that from MDMs infected with SINrep5-syncytin (Fig. 3b). In addition, human oligodendrocytes treated with medium from HFAs infected with SINrep5-syncytin also showed a significant retraction of cellular processes as compared with controls (P < 0.001; Fig. 3c).


Figure 3. Syncytin causes oligodendrocyte damage and death.

(a) Human oligodendrocytes (hOLs) treated with conditioned medium (CM) from HFAs infected with SINrep5-syncytin showed retracted cellular processes (arrowhead) with cell loss, whereas SINrep5-EGFP–infected and mock-infected controls showed abundant cell numbers and intact processes (arrowhead). (b) hOLs treated for 24 h with CM from HFAs infected with SINrep5-syncytin showed significantly higher cytotoxicity than did controls. CM from MDMs infected with SINrep5-syncytin was not toxic to hOLs. (c) hOLs treated for 24 h with CM from HFAs infected with SINrep5-syncytin showed significantly more cells with retracted processes than did controls. (d) CM from HFAs infected with SINrep5-syncytin, but not with a SINrep5-expressed HIV envelope protein, induced significant cytotoxicity in rat OLs. (e) Syncytin (S)-induced OL cytotoxicity and protein carbonyl abundance in the CM of HFAs were reduced when HFAs were treated with the antioxidants ferulic acid (FA; 50 M), NCX-2216 (6 M), L-NIL (0.5 M) and L-NAME (5.0 M). Original magnification, 400. *P < 0.05; ***P < 0.001.

Full Figure and legend (44K)


We confirmed these results in rat oligodendrocytes using the same cytotoxicity protocol, which showed that conditioned medium from HFAs infected with SINrep5-syncytin similarly caused a significantly higher mean amount of cell death than did that from MDMs infected with SINrep5-syncytin (Fig. 3d). In addition to cell loss, rat oligodendrocytes treated with conditioned medium from HFAs infected with SINrep5-syncytin showed marked process retraction as compared with controls (Supplementary Fig. 4 online).

As a control, the envelope protein from another neurotropic retrovirus HIV-JRFL, obtained from an individual with HIV-associated dementia, was expressed in HFAs using the same vector (SINrep5), and was found not to cause oligodendrocyte cytotoxicity (Fig. 3d). Conditioned medium from HFAs and MDMs infected with SINrep5-syncytin was not cytotoxic to human neurons under similar conditions, whereas the envelope protein from HIV-JRFL was highly cytotoxic to neurons (Supplementary Fig. 5 online). Thus, soluble factors released from astrocytes that are selectively induced by syncytin cause cellular injury and death in oligodendrocytes.

Antioxidants prevent syncytin-induced oligodendrocyte injury
Because protein carbonyl formation is mediated by redox reactants such as nitric oxide and its metabolite peroxynitrite14, we considered that compounds that scavenge redox reactants might reduce oligodendrocyte death. HFAs infected with SINrep5-syncytin were treated with a polyphenolic antioxidant, ferulic acid15; a non-steroidal anti-inflammatory-based antioxidant, NCX-2216; and two iNOS inhibitors, N6-(1-iminoethyl)-lysine, hydrochloride (L-NIL; 0.5 M) and N-nitro-L-arginine methyl ester (L-NAME; 5.0 M). Oligodendrocytes treated with the conditioned medium from these astrocytes showed a marked reduction in both mean oligodendrocyte cytotoxicity and protein carbonyl levels, as compared with oligodendrocytes not treated with either drug (Fig. 3e).

Indeed, protection of oligodendrocytes by ferulic acid and NCX-2216 against syncytin-mediated toxicity was found to be dependent on dose (Supplementary Fig. 6 online). Treatment with ferulic acid did not affect infection or expression by SINrep5-syncytin (data not shown). Conversely, treatment with other established neuroprotectants including MK801, an N-methyl-D-aspartate (***A) receptor antagonist, NBQX, an -amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptor antagonist, interferon-(IFN-) and glatiramer acetate did not reduce syncytin- related oligodendrocyte death (Supplementary Fig. 7 online). Thus, syncytin-induced oligodendrocyte cytotoxicity is probably mediated by the pathogenic effects of redox reactants in this in vitro system.

Ferulic acid inhibits syncytin-induced neurological deficits
As syncytin caused a significant increase in oligodendrocyte cytotoxicity, we examined the in vivo effects of syncytin after stereotaxic implantation of the SINrep5-syncytin virus into the corpus callosum of CD-1 mice. This brain region was selected because it is abundant in myelin, oligodendrocytes and astrocytes and is specifically injured in demyelinating diseases such as multiple sclerosis, resulting in motor and cognitive abnormalities (Fig. 4).


Figure 4. Syncytin induces neuroinflammation and neurobehavioral abnormalities in mice.

(a–c) Syncytin immunoreactivity was detected in the corpus callosum for up to 14 d in mice implanted with SINrep5-syncytin (c) but not in mock-implanted (a) or SINrep5-EGFP–implanted (b) mice. (d–f) Astrogliosis indicated by intense GFAP immunoreactivity was observed in the corpus callosum of mice implanted with SINrep5-syncytin (f), whereas normal astrocytes were observed in controls (d,e). (g–i) White matter of mice implanted with SINrep5-syncytin showed a vacuolar appearance (i), whereas healthy myelin was seen in controls (g,h). GSTYp-immunopositive oligodendrocytes were decreased in the white matter of mice implanted with SINrep5-syncytin (l) as compared with controls (j,k). (m–o) Ferulic acid (FA) abrogates syncytin-induced neurobehavioral changes in mice. Mice implanted with SINrep5-syncytin showed a significantly reduced ability to grasp a horizontal rod (m), a significantly diminished ability to grasp and to escape from an inverted screen (n) and delays in the time taken to cross a cantilevered beam, as compared with SINrep5-EGFP–implanted and mock-implanted controls (n = 6 mice per group in each test). Treatment of mice implanted with SINrep5-syncytin with FA (20 mg/kg administered daily by oral gavage) increased ability to grasp the bar (m), improved ability to grasp and to escape from the inverted screen (n) and reduced the time taken to cross the cantilevered beam (o). Original magnification, 400 (a–c); 1,000 (c, inset); 200 (d–f); 200 (g–i); 600 (j–l). Values are the mean s.e.m.; *P < 0.05, **P < 0.005, ***P < 0.0001.

Full Figure and legend (48K)


After the implantation of SINrep5-syncytin, syncytin was detected near the implantation site (Fig. 4c), and syncytin-expressing astrocytes were visible in the corpus callosum (Fig. 4c, inset). Serial tissue sections showed increased numbers of hypertrophied astrocytes (Fig. 4f,m) and microglia (Supplementary Fig. 8 online) in the corpus callosum of mice implanted with SINrep5-syncytin as compared with mice implanted with SINrep5-EGFP (Fig. 4e) or control conditioned medium (Fig. 4d). In addition, myelin in the corpus callosum of mice implanted with SINrep5-syncytin showed a vacuolar appearance after immunostaining for myelin basic protein (MBP; Fig. 4i) as compared with controls (Fig. 4g,h). There were substantially fewer immunoreactive oligodendrocytes (GSTYp positive) in mice that received the syncytin-expressing virus (Fig. 4l) than in controls (Fig. 4j,k).

To verify the latter observation, we carried out stereological cell counts of immunoreactive oligodendrocytes, which showed a significant reduction in mean cell numbers in mice implanted with SINrep5-syncytin as compared with controls (Table 2). By contrast, astrocyte counts in the same mice indicated that there was a significant increase in activated astrocytes. The decrease in oligodendrocyte numbers and increase in hypertrophied astrocytes caused by syncytin expression were reversed by daily treatment with ferulic acid (Table 2).


Table 2. Cell counts of immunoreactive astrocytes and oligodendrocytes in corpus callosuma



Full Table


We confirmed the above neuropathological findings by neurobehavioral testing of mice receiving SINrep5-EGFP, mock-implanted control conditioned medium and SINrep5-syncytin with and without concurrent ferulic acid treatments. At days 3 and 7, there were no differences in performance among the different groups (Supplementary Fig. 9 online). At days 10 and 14 after implantation, however, mice implanted with SINrep5-syncytin grasped the horizontal rod for significantly less time than did the control mice (P < 0.05; Fig. 4m). In addition, although mice implanted with SINrep5-syncytin retained some ability to hold on to the screen, they were slower to reach the screen edge (Fig. 4n), whereas mice implanted with SINrep5-EGFP or control conditioned medium were more curious and reached the edges of the inverted screen more quickly (P < 0.0001).

Mice implanted with SINrep5-syncytin showed mean delays in the time taken to cross a cantilevered beam as compared with mice implanted with SINrep5-EGFP or control conditioned medium (P < 0.005; Fig. 4o), suggesting that the SINrep5-syncytin implanted mice showed diminished motor activity and exploratory behavior. When mice implanted with SINrep5-syncytin were treated with ferulic acid for 14 d, the neurobehavioral outcomes in the horizontal bar test (Fig. 4m), modified screen test (Fig. 4n) and beam test (Fig. 4o) were significantly improved. Striatal implantation of SINrep5-syncytin in mice did not result in neuronal damage, as assessed by neurobehavioral studies16 (Supplementary Fig. 5 online).

In accordance with our in vitro observations, these in vivo studies indicate that syncytin induces neuroinflammation, oligoden-drocyte and myelin damage, and neurobehavioral abnormalities that are abrogated by the antioxidant ferulic acid.

Top



Discussion
We have shown that expression of both the mRNA and protein (syncytin) encoded by the HERV-W env gene is increased in the brains of individuals with multiple sclerosis and in specific cell types involved in neuroinflammation in demyelinating and demyelinated lesions. In addition, we have shown that in vitro syncytin mediates the production of proinflammatory molecules such as iNOS, IL-1 and redox reactants that at high levels are damaging to the brain9, 14, 17. An explicit function for syncytin is unknown in the brain in contrast to the placenta, where it seems to be important for placental development18, 19, 20, 21. However, our findings, including increased levels of iNOS and redox reactants, are consistent with the neuroinflammatory profile of astrocytes associated with multiple sclerosis22, 23, 24. In addition to the proinflammatory effects of syncytin in glial cells, soluble factors derived from syncytin-expressing astrocytes are toxic to oligodendrocytes but not to neurons, implying that syncytin uses a select mechanism for killing and/or targeting individual cell types—a possibility that is congruent with studies showing that astrocytes may influence oligodendrocyte survival25, 26, 27.

We did not detect syncytin in supernatants from astrocytes infected with SINrep5-syncytin, precluding the possibility that a direct interaction between syncytin and target cells leads to cytotoxicity. Nonetheless, syncytin expression in astrocytes resulted in cellular stress, manifested by the induction of IL-1, iNOS and redox reactants. Notably, syncytin expression was evident in glia at both the margin and the core of acute demyelination (Fig. 1f), emphasizing its role in cellular stress. Of interest, polymorphisms in retrovirus-encoded envelope proteins from HIV28 and murine leukemia virus29 that alter intracellular envelope expression in glial cells have been associated with pathogenic effects in the nervous system. These effects may occur through misfolding of the envelope protein or protein accumulation in the endoplasmic reticulum, resulting in a stress response by the cell and the subsequent release of neurotoxic molecules including redox reactants30.

Given that iNOS was induced in astrocytes, probable redox stress products include nitric oxide, reactive nitrogen-oxygen species, peroxynitrite and superoxide anions, which are capable of damaging target tissues, particularly the brain14, 22, and are also toxic to oligodendrocytes31. By contrast, we did not observe an increase in the in vitro generation of lipid peroxidation products, in agreement with studies on the pathogenesis of multiple sclerosis32. Although the precise pathway by which oxidation of a protein released by astrocytes mediates oligodendrocyte toxicity remains to be determined, a potential mechanism might involve oxidation of a released matrix metalloproteinase33.

We have also shown that by inducing proinflammatory molecules and redox reactants, ultimately resulting in oligodendrocyte death, syncytin modulates neurobehavioral changes in a mouse model that shows pathology similar to other animal models of multiple sclerosis. The present neurobehavioral changes are also reminiscent of those seen in individuals with multiple sclerosis including weakness, gait unsteadiness and altered executive functions34, 35. In addition, our studies have shown that ferulic acid, the plant-derived phenolcarboxylic acid that acts as an antioxidant15, 36, ameliorates oligodendrocyte death in vitro and in vivo and significantly improves neurobehavioral outcomes.

In summary, our study indicates that syncytin maybe involved in the pathogenesis of active demyelination, principally by evoking redox reactant–mediated cellular damage in the brain. Alternatively, given its persistent expression in lesion cores, the syncytin-mediated glial stress reaction might antagonize remyelination. In this regard, ferulic acid and equivalent compounds might be considered in trials designed to evaluate their efficacy for reducing demyelination or enhancing remyelination, similar to the protective effects of ferulic acid in neurons15. In addition, because HERVs represent a substantial proportion of the human genome and often express proteins, the potential pathogenic (or reparative) effects of other HERV proteins expressed in the nervous system warrant further investigation.

Top



Methods
Cell cultures.
Human and rat oligodendrocytes37, HFAs38, human LAN-2 neuronal cells16, human MDMs8 and PBLs39, U373 astrocytoma40 and U937 monocytoid cells8 (American Type Culture Collection) were cultured as described. To assess the effect of astrocyte and monocyte stimulation on HERV expression, we treated U373, PBL and U937 cells with 50 ng/ml of PMA (Sigma) for 4–72 h, collected them and isolated RNA as described8.

Real-time PCR with reverse transcription (RT-PCR).
Total RNA and cDNA were prepared from brain tissue or cells as described40. Primers for GAPDH, IL-1 and iNOS have been described40. We also used primers for IL-10 (5'-CCTCTCACCGTCTTGCTTTC-3' and 5'-GCAGAGTTGCTTGTTCTCC-3'), HERV-W env (5'-TGCCCCATCGTATAGAGTCT-3' and 5' CATGTACCCGGGTGAGTTGG-3'), HERV-H env (5'-TTACCCCATCATCAGTCCCCATTAC-3' and 5'-GAGCTCTTCGGTCCCATTTG-3'), HERV-E env (5'-TCGCCAAAGCCAGAGTT-3' and 5'-AAGGGGGAAATGAGGA-3') and HERV-K (II) env(5'-TTTATGGGGCCGAGACTTGTTA-3' and 5'-AGCGGCTACTGATTTACCATAC-3'). Semiquantitative real-time RT-PCR analysis was done by monitoring in real-time the increase of fluorescence of SYBR-green dye on a Bio-Rad i-Cycler. Data were normalized to the GAPDH mRNA level and are expressed as the relative fold change (RFC) in mRNA as described40.

Human brain tissue samples.
Brain tissue (frontal white matter) was collected with informed consent at autopsy, as described8. Control subjects included 18 individuals (mean age 56 16.4 yr), who were diagnosed with Alzheimer's disease (n = 6), HIV infection (encephalitis, n = 4; gliosis, n = 4), cerebral arteriosclerosis (n = 2), anoxic encephalopathy (n = 1) or normal brain pathology (n = 1). The multiple sclerosis group included 16 individuals (age 63.3 13.4 yr) who had been classified with primary progressive (n = 4), secondary progressive (n = 10) or relapsing-remitting (n = 2) multiple sclerosis and had an estimated disability status scale score of 7–10 before death. Frozen brain tissue from individuals with multiple sclerosis was obtained from the Multiple Sclerosis Patient Care and Research Clinic, Edmonton, Alberta, Canada. Brain tissue sections from chronic demyelinated plaques from individuals with multiple sclerosis were used for histochemical and immunohistochemical studies.

Immunohistochemistry and histochemistry.
Paraffin-embedded sections were immunostained with antibodies to Iba-1 (1.0 g/ml; ref. 40) for macrophage and microglia detection, GFAP (diluted 1:2,000; DAKO) for astrocyte detection, iNOS (1:500; Santa Cruz Biotech), GSTYp (1:600; Biotrin) and adenomatosis polyposis coli (APC, 1:500; Ab-7, Oncogene Research Products) for oligodendrocyte detection, and syncytin (1:1,000; 6A2B2)19 as reported16. To double label both microglia and astrocytes with cell-type-specific markers and syncytin, sections were probed with antibodies to Iba-1 and GFAP respectively, followed by incubation with biotinylated goat anti–rabbit IgG (H+L) (1:500; Vector Laboratories) and subsequently with Elite ABC reagent (Vector Laboratories) or peroxidase-conjugated Affinipure goat anti–rabbit IgG (H+L) (1:500; Jackson ImmunoResearch), and then developed by diaminobenzidine (DAB) treatment for peroxidase or bromochloroindolyl phosphate (BCIP) and nitro blue tetrazolium (NBT) for alkaline phosphatase activity.

Immunofluorescence and confocal laser scanning microscopy.
Paraffin-embedded sections from mice were immunostained with mouse anti-MBP (diluted 1:1,000; Sternberger Monoclonals), rabbit anti–mouse GSTYp (1:600; Biotrin) and anti-APC (1:500; Ab-7, Oncogene Research Products). Slides were examined on an Olympus FV300 confocal laser-scanning microscope. Cultured oligodendrocytes derived from adult human and rat brains were stained with a monoclonal antibody (O1) that recognizes galactocerebroside (Gal-C), a marker for mature oligodendrocytes41. The total number of cells and the number with processes at 24 h were counted per well (10 fields at 3400 magnification) after treatment with conditioned medium from HFAs or MDMs infected with SINrep5-syncytin or SINrep5-EGFP. APC-positive oligodendrocytes and GFAP-positive astrocytes in the vicinity of the implantation site were counted (five fields at 3400; expressed as total number of GSTYp-positive cells in an area of 2,376 m2).

We used three serial sections (5 m) from six mice per treatment group to quantify numbers of GFAP-positive astrocytes and GSTYp-positive oligodendrocytes by a stereological method42. In brief, a systematic random sample of three sections that spanned the corpus callosum was selected for analysis. Sections were selected at equal intervals in the series comprising a known fraction of the section series (ssf). The labeled astrocytes and oligodendrocytes were counted (Q) under a known fraction of the section area (asf). The height of the optical dissectors (h) positioned in the central part of the section thickness (t) was used to determine the ratio h/t (tsf). The total number (N) of astrocytes or oligodendrocytes were estimated as N = Q 1/ssf 1/asf 1/tsf. The coefficient of error was determined by the formula CE (Q) = s.e.m./mean.

Western blot analysis.
Protein extracts were prepared from various brain tissue and cell samples with lysis buffer and concentrations were determined by BCA assay (Pierce). We separated proteins by 10% SDS-PAGE, transferred them to nitrocellulose membranes and probed them with antibodies to syncytin and -actin as reported8. The 75-kDa band was analyzed by densitometry, and the mean results are expressed as the RFC in the density of the band in the multiple sclerosis group as compared with controls.

Construction of SINrep5 virus.
The Sindbis virus–based (SIN) vector system used in this study has been described28. In brief, the 1.832-kbp env open reading frame from phCMVenvpH74 (ref. 19) was cloned into pSINrep5 to obtain pSINrep5-syncytin. Virus stocks were prepared as described28. On average, 106 to 107 infectious virus particles per ml were obtained for SINrep5-syncytin and SINrep5-EGFP, respectively. All infections were done with the same number of virus particles (multiplicity of infection (MOI) 1.0). HFAs and MDMs (5 104 per well) were seeded in 16-well chamber slides and infected with SINrep5-syncytin or SINrep5-EGFP (MOI 1.0 each) or mock-infected with conditioned medium.

Cell survival assay.
To determine whether oligodendrocyte cytotoxicity can be inhibited, HFAs were treated with NBQX (30 M; Sigma), MK-801 (30 M; Sigma), glatiramer acetate (25 g/ml; Teva), IFN- (100 U/ml; Serono), ferulic acid (0.005, 0.5, 5, 50 or 250 M; Sigma), NCX-2216 (60 nM, 600 nM and 6 M; a gift of NicOx S.A.), L-NIL (0.5 M, Sigma) and L-NAME (5.0 M, Sigma) followed by infection with SINrep5-syncytin overnight at 37 °C. Adult rat brain-derived oligodendrocytes were treated subsequently with either conditioned medium (diluted 1:1 with AIM-V medium) from drug-treated or SINrep5-syncytin-infected HFA for 24 h. Conditioned media for toxicity assays were obtained from HFA and MDM infected with SINrep5-syncytin or SINrep5-EGFP. We prepared LAN-2 neurons as described16. All experiments were repeated at least three times. To measure cellular injury, oligodendrocytes were immunostained with antibody to Gal-C and cells with and without processes were counted. To quantify cell death, the trypan-blue exclusion method was used as described28.

Protein carbonyl and 4-hydroxynonenol (HNE) assays.
Protein oxidation was determined14 by an oxidized protein detection kit (Oxyblot, ONCOR). Samples were incubated for 20 min with 12% SDS and 2,4-dinitrophenylhydrazine (DNPH) in 10% trifluoroacetic acid with vortexing every 5 min, and then neutralized with Oxyblot Neutralization solution. We blotted 600 ng of protein onto nitrocellulose paper by the slot blotting technique. Membranes were incubated with blocking buffer for 30 min at 24 °C, incubated with rabbit antibodies to DNPH (diluted 1:150) for 90 min, and then by anti–rabbit IgG coupled to alkaline phosphatase (1:15,000) for 2 h at 24 °C. After being washed and developed with SigmaFast chromogen (Sigma), blots were analyzed by computer-assisted imaging software (Scion Imaging). Samples for HNE detection were similarly analyzed by the slot blotting technique except that a rabbit antibody to HNE (1:4,000; Calbiochem) was used as a primary antibody as described14. Results are expressed as the RFC as compared with values obtained in AIM-V medium.

In vivo implantation of viruses and drug treatment.
SINrep5-EGFP or SINrep5-syncytin virus (0.5 106 particles/ml in 3 l) was stereotactically implanted bilaterally into the corpus callosum of CD-1 mice aged 10 weeks (n = 6 for each treatment). Control mice (n = 6) were implanted with conditioned medium from mock-infected cultures (1-mm anterior, 2.0-mm lateral and 1.5-mm deep relative to bregma). In addition, the right striatum of CD-1 mice aged 10 weeks (n = 4 for each treatment) were stereotactically implanted with SINrep5-EGFP and SINrep5-syncytin (0.5 106 particles/ml) as described28. Similarly, for treatment with ferulic acid, mice (n = 6) were implanted with SINrep5-syncytin and subjected by daily oral gavage with ferulic acid (20 mg per kg (body weight) on a daily basis) for 14 d. Mice were killed on day 14 and intracardially perfused with saline, followed by 4% paraformaldehyde (PFA). All mouse experiments were done in accordance with Canadian Council on Animal Care guidelines.

Neurobehavioral studies.
Behavioral tests were conducted in mice that had been implanted with the SINrep5-syncytin or SINrep5-EGFP virus or control conditioned medium (mock-implanted) on days 3, 7, 10 and 14 after implantation. The horizontal bar test involved a test of coordination and forelimb strength using a horizontal bar (0.2 cm thick, 38 cm long) held 49 cm above a bench43. The static rod test involved a test of coordination using five rods (60 cm long) of varying thickness (diameters: 35 mm, rod 1; 28 mm, rod 2; 22 mm, rod 3; 15 mm, rod 4; 9 mm, rod 5). These rods were bolted to the edge of a bench so that the rods protruded their full 60-cm length horizontally into space. We placed a mouse at the exposed end of the widest rod and measured the time that it took to orient 180° from the starting position and to travel to the other end43. Muscle strength and seeking behavior were determined by using an inverted screen test. The invertable screen was a 43-cm square wire mesh consisting of 12-mm squares of 1-mm diameter wire surrounded by a 4-cm deep wooden beading43. We modified the inverted screen test by placing the mouse at a point that was equidistant from the edges of the screen. We started the stop-clock once the screen was inverted and noted the time that the mouse took to reach the edge of the inverted screen as a measure of curiosity and seeking behavior. The rotary behavior of the mice with implants in the striatum was analyzed at 3, 7, 10 and 14 d after implantation, as described28.

Statistical analysis.
Because multiple treatments were used in all experiments, a one-way analysis of variance (ANOVA) with Tukey-Kramer multiple comparisons test was used for analysis. Data were analyzed using GraphPad InStat version 3.01 for Windows 95 (GraphPad Software).

Note: Supplementary information is available on the Nature Neuroscience website.

Top



Received 8 April 2004; Accepted 30 August 2004; Published online 26 September 2004.

REFERENCES
Lander, E.S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860–921 (2001). | Article | PubMed | ISI | ChemPort |
Costas, J. Characterization of the intragenomic spread of the human endogenous retrovirus family HERV-W. Mol. Biol. Evol. 19, 526–533 (2002). | PubMed | ISI | ChemPort |
Hughes, J.F. & Coffin, J.M. Evidence for genomic rearrangements mediated by human endogenous retroviruses during primate evolution. Nat. Genet. 29, 487–489 (2001). | Article | PubMed | ISI | ChemPort |
van de Lagemaat, L.N., Landry, J.R., Mager, D.L. & Medstrand, P. Transposable elements in mammals promote regulatory variation and diversification of genes with specialized functions. Trends Genet. 19, 530–536 (2003). | Article | PubMed | ISI | ChemPort |
Sverdlov, E.D. Retroviruses and primate evolution. BioEssays 22, 161–171 (2000). | Article | PubMed | ISI | ChemPort |
Bock, M. & Stoye, J.P. Endogenous retroviruses and the human germline. Curr. Opin. Genet. Dev. 10, 651–655 (2000). | Article | PubMed | ISI | ChemPort |
Power, C. Retroviral diseases of the nervous system: pathogenic host response or viral gene–mediated neurovirulence? Trends Neurosci. 24, 162–169 (2001). | Article | PubMed | ISI | ChemPort |
Johnston, J.B. et al. Monocyte activation and differentiation augment human endogenous retrovirus expression: implications for inflammatory brain diseases. Ann. Neurol. 50, 434–442 (2001). | Article | PubMed | ISI | ChemPort |
Calabrese, V. et al. Nitric oxide synthase is present in the cerebrospinal fluid of patients with active multiple sclerosis and is associated with increases in cerebrospinal fluid protein nitrotyrosine and S-nitrosothiols and with changes in glutathione levels. J. Neurosci. Res. 70, 580–587 (2002). | Article | PubMed | ISI | ChemPort |
Lu, F. et al. Oxidative damage to mitochondrial DNA and activity of mitochondrial enzymes in chronic active lesions of multiple sclerosis. J. Neurol. Sci. 177, 95–103 (2000). | Article | PubMed | ISI | ChemPort |
Bo, L. et al. Induction of nitric oxide synthase in demyelinating regions of multiple sclerosis brains. Ann. Neurol. 36, 778–786 (1994). | PubMed | ISI | ChemPort |
John, G.R. et al. Multiple sclerosis: re-expression of a developmental pathway that restricts oligodendrocyte maturation. Nat. Med. 8, 1115–1121 (2002). | Article | PubMed | ISI | ChemPort |
Carson, M.J. Microglia as liaisons between the immune and central nervous systems: functional implications for multiple sclerosis. Glia 40, 218–231 (2002). | Article | PubMed | ISI |
Butterfield, D.A. & Lauderback, C.M. Lipid peroxidation and protein oxidation in Alzheimer's disease brain: potential causes and consequences involving amyloid -peptide-associated free radical oxidative stress. Free Radic. Biol. Med. 32, 1050–1060 (2002). | Article | PubMed | ISI | ChemPort |
Kanski, J., Aksenova, M., Stoyanova, A. & Butterfield, D.A. Ferulic acid antioxidant protection against hydroxyl and peroxyl radical oxidation in synaptosomal and neuronal cell culture systems in vitro: structure-activity studies. J. Nutr. Biochem. 13, 273–281 (2002). | Article | PubMed | ISI | ChemPort |
Zhang, K. et al. HIV-induced metalloproteinase processing of the chemokine stromal cell derived factor-1 causes neurodegeneration. Nat. Neurosci. 6, 1064–1071 (2003). | Article | PubMed | ISI | ChemPort |
Hua, L.L., Kim, M.O., Brosnan, C.F. & Lee, S.C. Modulation of astrocyte inducible nitric oxide synthase and cytokine expression by interferon is associated with induction and inhibition of interferon gamma-activated sequence binding activity. J. Neurochem. 83, 1120–1128 (2002). | Article | PubMed | ISI | ChemPort |
Blond, J.L. et al. Molecular characterization and placental expression of HERV-W, a new human endogenous retrovirus family. J. Virol. 73, 1175–1185 (1999). | PubMed | ISI | ChemPort |
Blond, J.L. et al. An envelope glycoprotein of the human endogenous retrovirus HERV-W is expressed in the human placenta and fuses cells expressing the type D mammalian retrovirus receptor. J. Virol. 74, 3321–3329 (2000). | Article | PubMed | ISI | ChemPort |
Frendo, J.L. et al. Direct involvement of HERV-W env glycoprotein in human trophoblast cell fusion and differentiation. Mol. Cell. Biol. 23, 3566–3574 (2003). | Article | PubMed | ISI | ChemPort |
Mi, S. et al. Syncytin is a captive retroviral envelope protein involved in human placental morphogenesis. Nature 403, 785–789 (2000). | Article | PubMed | ISI | ChemPort |
Calabrese, V. et al. Disruption of thiol homeostasis and nitrosative stress in the cerebrospinal fluid of patients with active multiple sclerosis: evidence for a protective role of acetylcarnitine. Neurochem. Res. 28, 1321–1328 (2003). | Article | PubMed | ISI | ChemPort |
Smith, K.J. & Lassmann, H. The role of nitric oxide in multiple sclerosis. Lancet Neurol. 1, 232–241 (2002). | Article | PubMed | ISI | ChemPort |
Danilov, A.I. et al. Nitric oxide metabolite determinations reveal continuous inflammation in multiple sclerosis. J. Neuroimmunol. 136, 112–118 (2003). | Article | PubMed | ISI | ChemPort |
Lieberman, A.P., Pitha, P.M., Shin, H.S. & Shin, M.L. Production of tumor necrosis factor and other cytokines by astrocytes stimulated with lipopolysaccharide or a neurotropic virus. Proc. Natl. Acad. Sci. USA 86, 6348–6352 (1989). | PubMed | ChemPort |
Rosen, C.L., Bunge, R.P., Ard, M.D. & Wood, P.M. Type 1 astrocytes inhibit myelination by adult rat oligodendrocytes in vitro. J. Neurosci. 9, 3371–3379 (1989). | PubMed | ISI | ChemPort |
Takahashi, J.L., Giuliani, F., Power, C., Imai, Y. & Yong V.W. Interleukin-1 promotes oligodendrocyte death through glutamate excitotoxicity. Ann. Neurol. 53, 588–595 (2003). | Article | PubMed | ISI | ChemPort |
van Marle, G., Ethier, J., Silva, C., Mac Vicar, B.A. & Power, C. Human immunodeficiency virus type 1 envelope–mediated neuropathogenesis: targeted gene delivery by a Sindbis virus expression vector. Virology 309, 61–74 (2003). | Article | PubMed | ISI | ChemPort |
Lynch, W.P. & Sharpe, A.H. Differential glycosylation of the Cas-Br-E env protein is associated with retrovirus-induced spongiform neurodegeneration. J. Virol. 74, 1558–1565 (2000). | Article | PubMed | ISI | ChemPort |
Wesselingh, S.L. & Thompson, K.A. Immunopathogenesis of HIV-associated dementia. Curr. Opin. Neurol. 14, 375–379 (2001). | Article | PubMed | ISI | ChemPort |
Scott, G.S., Virag, L., Szabo, C. & Hooper, D.C. Peroxynitrite-induced oligodendrocyte toxicity is not dependent on poly(ADP-ribose) polymerase activation. Glia 41, 105–116 (2003). | Article | PubMed | ISI |
Ravikumar, A., Arun, P., Devi, K.V., Augustine, J. & Kurup, P.A. Isoprenoid pathway and free radical generation and damage in neuropsychiatric disorders. Indian J. Exp. Biol. 38, 438–446 (2000). | PubMed | ChemPort |
Gu, Z. et al. S-nitrosylation of matrix metalloproteinases: signaling pathway to neuronal cell death. Science 297, 1186–1190 (2002). | Article | PubMed | ISI | ChemPort |
Paty, D.W., Noseworthy, J.H. & Ebers, G.C. Diagnosis of multiple sclerosis. in Multiple Sclerosis (eds. Paty, D.E. & Ebers, G.C.) 68–75 (F.A. Davis Company, Philadelphia, 1998).
Anlar, O., Tombul, T. & Kisli, M. Peripheral sensory and motor abnormalities in patients with multiple sclerosis. Electromyogr. Clin. Neurophysiol. 43, 349–351 (2003). | PubMed | ChemPort |
Ogiwara, T. et al. Inhibition of NO production by activated macrophages by phenolcarboxylic acid monomers and polymers with radical scavenging activity. Anticancer Res. 23, 1317–1323 (2003). | PubMed | ISI | ChemPort |
Yong, V.W. & Antel, J.P. Culture of glial cells from human brain biopsies. in Protocols for Neural Cell Culture (eds. Fedoroff, S. & Richardson, A.) 81–96 (Humana, Totowa, NJ, 1997).
Power, C. et al. Neuronal death induced by brain-derived human immunodeficiency virus type 1 envelope genes differs between demented and nondemented AIDS patients. J. Virol. 72, 9045–9053 (1998). | PubMed | ISI | ChemPort |
Bar-Or, A. et al. Analyses of all matrix metalloproteinase members in leukocytes emphasize monocytes as major inflammatory mediators in multiple sclerosis. Brain 126, 2738–2749 (2003). | Article | ISI |
Power, C. et al. Intracerebral hemorrhage induces macrophage activation and matrix metalloproteinases. Ann. Neurol. 53, 731–742 (2003). | Article | PubMed | ISI | ChemPort |
Bansal, R., Warrington, A.E., Gard, A.L., Ranscht, B. & Pfeiffer, S.E. Multiple and novel specificities of monoclonal antibodies O1, O4, and R-mAb used in the analysis of oligodendrocyte development. J. Neurosci. Res. 24, 548–557 (1989). | PubMed | ISI | ChemPort |
West, M.J., Ostergaard, K., Andreassen, O.A. & Finsen, B. Estimation of the number of somatostatin neurons in the striatum: an in situ hybridization study using the optical fractionator method. J. Comp. Neurol. 370, 11–22 (1996). | Article | PubMed | ISI | ChemPort |
Guenther, K., Deacon, R.M., Perry, V.H. & Rawlins, J.N. Early behavioural changes in scrapie-affected mice and the influence of dapsone. Eur. J. Neurosci. 14, 401–409 (2001). | Article | PubMed | ISI | ChemPort |