TI - Results . AB - Mice positive for the IL-10 transgene as determined by PCR analysis developed hind limb paralysis ( Fig 1A ) . Paralysis was restricted to hind limbs of transgene positive ( Tg+ ) mice , as paralysis of the tail was not observed . Transgene negative ( Tg - ) littermates did not develop hind limb paralysis ( Fig 1A ) . Disease onset varied from 2-4 months of age , and Tg+ mice <2 months of age showed no overt loss of motor function . As stated above , VMD2 was originally identified as being localized to RPE cells . However , subsequent data has shown that human VMD2 is expressed in multiple sites within the human body [7] . GeneAtlas analysis reveals that in humans , VMD2 is highly expressed in the nervous system , including the spinal cord and multiple areas of the brain . For mice , murine VMD2 appears to be primarily expressed in the eye and testes [8] . However , the full expression pattern of human VMD2 within murine tissues is unknown . Since human VMD2 was driving murine IL-10 expression in the Tg+ mice , we examined gene expression of murine IL-10 in multiple tissues from Tg+ sick mice compared to normal Tg - control littermates , as Tg - tissues should express baseline levels of IL-10 . We found increased expression of IL-10 in the eye , sciatic nerve , spinal cord , testes and brain of Tg+ sick mice relative to IL-10 levels in the same tissues from Tg - mice ( Fig 2A ) . Similar expression patterns of IL-10 were observed in Tg+ healthy mice relative to Tg - mice ( data not shown ) . Based on studies demonstrating high VMD2 expression in the testes , and since IL-10 gene expression was 55,000 fold higher in the testes of Tg+ mice , we sought to determine if increased IL-10 gene expression in the testes of male Tg+ mice may exacerbate development of disease compared to female Tg+ mice . Male Tg+ mice developed a slightly accelerated paralysis compared to female Tg+ mice ( Fig 1B ) . However , these differences in progression of disease were not significant and ultimately both male and female Tg+ mice developed paralysis . To determine if IL-10 protein levels were upregulated systemically or localized to the neural tissue micromilieu , plasma and cerebrospinal fluid ( CSF ) of male and female Tg+ healthy mice , Tg+ sick mice , and Tg - littermates were harvested and examined by IL-10 ELISA . We found a significant increase of IL-10 protein levels in the CSF of Tg+ healthy mice compared to Tg - littermates ( p lt 0001 ) , with significantly ( p = 0002 ) increasing levels that correlated with development of disease ( Fig 2B ) . Plasma levels of IL-10 protein were not elevated in Tg+ mice , indicating that systemic levels of IL-10 do not change and suggestingG that localized IL-10 expression was driving development of disease . Additionally , we examined local expression of IL-10 protein by immunohistochemistry in organs directly affected by disease . We observed increased expression of IL-10 on the sciatic nerve of Tg+ healthy and sick mice ( Fig S1B and C ) . IL-10 staining was not observed on Tg - sciatic nerve sections ( Fig S1A ) . Spinal cord expression of IL-10 appeared to be localized to the dorsal roots of Tg+ sick mice ( Fig S1F ) , and was not apparent on Tg - or Tg+ healthy spinal cord tissue sections ( Fig S1D and E ) . In order to examine the pathology of disease in VMD2-IL-10 Tg mice , histological examination of various tissues from Tg - , Tg+ healthy , and Tg+ sick mice was performed . H&E staining of the sciatic nerve from Tg - and Tg+ healthy mice was normal ( Fig 3A and B ) , however , Tg+ sick early ( defined as having a disease score of tissue sections revealed a dramatic infiltration of mononuclear cells (Fig 3C) . Tg+ sick late ( disease grade gt = 3 ) mice continued to have a dramatic infiltration of mononuclear cells (Fig 3D) , and also exhibited a significant loss of sciatic nerve tissue , as evidenced by a decrease in tissue diameter . Luxol fast blue staining of sciatic nerve tissue revealed no loss of myelin in Tg - and Tg+ healthy mice ( Fig 3E and F ) . Conversely , Tg+ sick tissues revealed significant loss of myelin staining ( Fig 3G and H ) . This correlation of cell infiltrate with loss of myelin in the sciatic nerve suggested that disease progression in Tg+ mice was due to an autoimmune attack of the peripheral nerves . The central nervous system , including the brain ( data not shown ) and the spinal cord itself , seemed resistant to the effects of increased IL-10 , as H&E staining of the spinal cord tissue revealed no evidence of cell infiltrate in Tg - , Tg+ healthy , nor Tg+ sick mice ( Fig 3I-L ) . The dorsal roots of Tg+ sick mice do reveal a slight "moth-eaten" appearance , and some loss of myelin staining was observed in these dorsal roots ( Fig 3O and P ) . No loss of myelin was observed in the spinal cords of Tg - or Tg+ healthy mice ( Fig 3M and N ) . In addition to the sciatic nerve , other peripheral nerves were examined for cellular infiltrate . In nerves from the brachial plexus , significant infiltrate was observed in Tg+ sick mice ( Fig S2C ) , but not in Tg - or Tg+ healthy mice ( Fig S2A and B ) . This is interesting as no paralysis was observed in the forelimbs during the course of disease . This suggests that disease in Tg+ mice may be ascending to the forelimbs , as often seen in human CIDP , yet mice perish prior to forelimb paralysis occurs . No cellular infiltrate was observed in the femoral nerve , regardless of disease state ( Fig S2D and F ) . In the optic nerve , no cellular infiltration was observed in Tg+ sick mice ( Fig S2G ) , despite the increased expression of VMD2 and IL-10 in the eyes of these animals . A summary of tissues examined , cellular infiltration , and phenotype is displayed in Table S1 . To identify the cellular infiltrate observed in the sciatic nerve of Tg+ sick mice , tissues from Tg - and Tg+ sick mice were harvested and flow cytometry was used to identify various immune cells , including macrophages ( F4/80 ) , monocytes ( Cd11b ) , T cells ( CD3 ) and neutrophils ( Gr1 ) . Spleens from Tg+ sick mice revealed decreased macrophages ( p lt 0001 ) and increased T cell populations ( p = 0008 ) compared to Tg - littermates ( Fig 4A ) . The sciatic nerve of Tg+ sick mice revealed a significant infiltration of macrophages ( p = 0001 ) and monocytes ( p = 0002 ) compared to Tg - mice , but no difference in neutrophils ( p = 0469 ) was observed ( Fig 4B ) . These results suggest that there is recruitment of macrophages from the spleen to these tissues . Although not evident in H&E staining , spinal cord tissue from Tg+ sick mice had a significant infiltration of macrophages ( p = 0002 ) as well ( Fig 4C ) , which likely corresponds to cellular infiltration into the dorsal roots and not actual cord tissue . Additionally , sciatic nerve tissue sections were stained with APC-conjugated anti-F4/80 antibody and examined by confocal microscopy . Tg - and Tg+ healthy mice exhibited no infiltration of macrophages into the sciatic nerve ( Fig 4D and E ) . Tg+ sick early mice exhibited abundant macrophages in the sciatic nerve , virtually infiltrating the entire tissue (Fig 4F) . Sciatic nerve tissue from Tg+ sick late mice demonstrated infiltrating macrophages (Fig 4G) , although to a much lesser degree than Tg+ sick early tissues . Again , tissue loss was evident in Tg+ sick late mice . This , combined with the above observation of macrophages infiltrating the sciatic nerve at the time of disease suggested that macrophage-mediated demyelination of the sciatic nerve and subsequent loss of tissue resulted in clinical disease . To determine the gene expression cytokine profile of various disease-affected tissues , spleen , sciatic nerve , and spinal cord tissue from Tg+ healthy and sick mice were harvested and examined for various pro - and anti-inflammatory cytokines by quantitative real-time PCR . Using Tg+ healthy tissues as a baseline comparison , Tg+ sick early and sick late spleens exhibited very little change in cytokine gene expression (Fig 5A) . Sciatic nerves , however , exhibited a primarily pro-inflammatory cytokine profile , with cytokine expression levels increasing with disease progression ( Fig 5B ) . These cytokines include FasL , IFN-gamma and IL-12 . The anti-inflammatory marker arginase (Arg1) was also increased in sciatic nerve tissues from sick mice , suggesting a possible attempt to squelch this burst of autoimmune activity . Pro - and anti-inflammatory cytokine gene expression is also evident in spinal cord tissue of Tg+ sick mice ( Fig 5C ) , albeit at drastically reduced levels compared to the sciatic nerve tissue . Pancreas specific IL-10 transgenic mice have been shown to develop spontaneous inflammation and diabetes [9] , [10] . Cellular infiltration into the pancreas correlated with increased expression of adhesion molecules on pancreatic endothelium , including intercellular adhesion molecule 1 ( ICAM-1 ) [10] . To determine if VMD2-IL-10 Tg mice exhibited increased ICAM-1 expression , sciatic nerve and spinal cord tissues sections were stained with PE-conjugated anti-ICAM-1 antibody . Tg - littermates exhibit no expression of ICAM-1 on sciatic nerve tissue sections (Fig 6A) , whereas Tg+ healthy and sick mice demonstrated significant staining of ICAM-1 antibody on the surface of the sciatic nerve tissue ( Fig 6B and C ) . Increased ICAM-1 expression was not evident on spinal cord dorsal roots in Tg+ healthy mice ( Fig 6D ) , but was evident once Tg+ mice develop disease ( Fig 6E ) . No increase in ICAM-1 expression was observed in the cerebellum of Tg - , Tg+ healthy , or Tg+ sick mice ( data not shown ) , reflecting that sites spared of cellular infiltration lack ICAM-1 expression . In our VMD2-IL-10 transgenic mice , it appears that disease progression correlated with infiltration of macrophages into the sciatic nerve and subsequent demyelination . However , paralysis can be initiated by a variety of mechanisms other than antecedent demyelination . In addition , although mice did not clinically demonstrate evidence of weakness or paralysis prior to demyelination , it remained possible there was some subclinical axonal dysfunction prior to initiation of demyelination . This includes inherent neuronal defects resulting in neuromuscular dysfunction . In order to determine if Tg+ mice have underlying axonal dysfunction prior to development of paralysis , we first stained sciatic nerve tissue sections with an antibody specific for normal neurofilament phenotype . The SMI-31 antibody reacts with a phosphorylaTED epitope in extensively PHOSphorylated neurofilament H and , to a lesser extent , with neurofilament M , a sign of normal neurofilament status . SMI 31 reacts broadly with thick and thin axons and some dendrites such as basket cell dendrites , but not Purkinje cell dendrites . Nerve cell bodies are generally unreactive , and other cells and tissues are unreactive except for peripheral axons . Damage to neurofilaments results in dePHOSphorylation of neurofilaments , and thus less reactivity with SMI-31 . Wild-type mice demonstrated significant positive staining with SMI-31 , reflecting a healthy axonal phenotype ( Fig 7A ) . Tg+ mice developed progressively decreased staining with SMI-31 as they developed weakness and paralysis reflecting axonal damage and loss with onset of paralysis ( Fig 7B and C ) . We next examined whether the loss of neurofilament staining observed in Tg+ mice correlated with a loss of activity at times preceding macrophage infiltration and hind limb paralysis . This was done by examining voluntary wheel running ( VWR ) of Tg+ mice and their Tg - age-matched littermates . Three month-old Tg+ mice , prior to developing clinical signs of weakness or motor dysfunction , showed no significant difference in VWR compared to Tg - littermates , as determined by distance , time , and speed on the wheel ( Fig 7D ) . This demonstrates that although mice at 3 months of age demonstrate decreased neurofilament staining (Fig 7B) , loss of function in these animals does not progress until cellular infiltration and demyelination occurs , indicating that macrophages are the primary mediators of disease in Tg+ mice . This also demonstrates that paralysis and loss of activity is not due to an inherent neurological dysfunction in Tg+ mice . Disease progression in VMD2-IL-10 transgenic mice appears to develop in two stages . First , overexpression of IL-10 results in upregulation of ICAM-1 on sciatic nerve tissue . Secondly , macrophages infiltrate sciatic nerve tissue , cause demyelination , neuronal dysfunction , and ultimately paralysis . We wanted to test whether the onset of disease could be delayed or disease severity ameliorated by attempting to remove important components of disease progression : IL-10 and macrophages . We first attempted to block IL-10 protein expression with blocking antibodies to IL-10 . Tg+ mice were treated with intraperitoneal injections of 250 ug anti-IL-10 antibody or rat IgG control antibody i.p. 1-2x weekly . Anti-IL-10 immunotherapy did not successfully delay progression of disease compared to isotype control administration ( Fig S2I ) . Higher doses of anti-IL-10 were not successful and resulted in toxicity ( data not shown ) . The lack of success with i.p.administration of anti-IL-10 was not surprising , as systemic levels of IL-10 were not elevated in transgenic mice ( Fig 2B ) . Moreover , systemically delivered antibodies would not readily cross blood-brain barriers and penetrate nerve tissues . We next attempted to ameliorate disease by depletion of macrophages with clodronate liposomes . Tg+ mice were treated with intraperitoneal injections of clodronate or PBS liposomes ( 200 ul initial injection ; 100 ul each subsequent injection ) twice weekly [11] . Liposome treatment commenced immediately prior to onset of disease ( day 100 ) . Tg+ mice treated with clodronate liposomes developed paralysis at a significantly reduced tempo compared to PBS liposome-treated Tg+ mice ( Fig 8A ) . Clodronate liposome treatment was continued for 2 months , and ceased at day 162 due to toxicity related to repeated liposome injection . At time of cessation , the clinical score of paralysis was 1.5+-0.41 for clodronate treated mice compared to 3.75+-0.96 for PBS-treated mice . PBS liposome-treated Tg+ mice continued to live past day 162 , suggesting that toxicity is directly attributable to the clodronate and not to liposomes . Clodronate-liposome treatment resulted in significantly reduced macrophage infiltrate into the sciatic nerve ( Fig S2H ) . It was previously observed that gene expression of FasL was upregulated in the spleen and sciatic nerve of Tg+ mice ( Fig 4A and B ) . Since membrane bound FasL can regulate cell death [1] , [3] , [12] , we hypothesized that this increase in FasL may be due to increased expression on Tg+ macrophages , and that these macrophages are utilizing FasL to mediate myelin-producing Schwann cell death in the sciatic nerve , leading to demyelination of the sciatic nerve and ultimately paralysis of the animal . To determine if Tg+ macrophages utilize FasL to induce Schwann cell death , a modified macrophage/target cell co-culture experiment was utilized [13] . Macrophages from Tg+ sick mice were isolated and co-cultured with radioactively labeled murine Schwann cells at an effector[?]target ratio of 10[?]1 . Schwann cell death was determined by the release of 3H-thymidine into the cell supernatant . The co culture was performed with anti-FasL blocking antibodies or isotype control antibodies present . Tg+ macrophages induced significant Schwann cell cytotoxicity , and was due to FasL mediated death , as blocking antibodies to FasL significantly inhibited Schwann cell death (Fig 8B) . Therefore , macrophages-mediated cytotoxicity of myelin-producing Schwann cells through FasL -dependent pathways . To determine if FasL was indeed causing demyelination and paralysis in vivo , we bred VMD2-IL-10 Tg+ mice with C57BL/6-Faslpr mice , which harbor a spontaneous mutation in Fas that prevents Fas-FasL interaction and signaling . We found that Tg+ lpr homozygous mice failed to develop progressive disease and paralysis compared to Tg+ mice ( Fig 8C ) , confirming that Fas expression on neural tissues is necessary for macrophage - and FasL -mediated demyelination and paralysis in Tg+ mice .