| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Published online before print June 5, 2008
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Abteilung für Neuropathologie,* and the Institut für Medizinische Mikrobiologie, Immunologie und Hygiene, Universität zu Köln, Köln, Germany; and the Institut für Medizinische Mikrobiologie, Otto-von-Guericke Universität Magdeburg, Magdeburg, Germany
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
and interleukin-17 production, regulated the number of tumor necrosis factor-positive and inducible nitric oxide synthase-positive macrophages in the diseased sciatic nerve, and mediated axonal damage and subsequent neuronal apoptosis and neurogenic muscular atrophy. In contrast, the demyelination of peripheral nerves was only slightly ameliorated in CD4 T cell-depleted mice. In conclusion, P0106–125-induced experimental autoimmune neuritis is a CD4 T cell-mediated autoimmune disease that affects both the peripheral and central nervous systems.
The hallmark of rat experimental autoimmune neuritis (EAN) is the infiltration of the peripheral nervous system by lymphocytes and macrophages. However, it is still controversial whether CD4 T cells and macrophages are sufficient to induce EAN or whether the integrated attack of CD4 and CD8 T cells, B cells, and macrophages is crucial for immunopathology of EAN.4-7
Traditionally, most studies of EAN were performed in rat and disease was induced by either adoptive transfer of autoimmune T cells8,9 or by immunization with neuritogenic epitopes of the P2 or the P0 protein,10-13 respectively. Recently, we have identified the P0106–125 peptide as a neuritogenic epitope in mice.14 Immunization of C57BL/6 mice with P0106–125 induced a well reproducible and robust EAN with axonal neuritis and mild demyelination of peripheral nerves.14 Clinically, P0106–125-induced EAN was more severe and homogenous compared to murine EAN induced by immunization with the traditionally used P0180–199 peptide.14-16 Therefore, the improved P0106–125-model of EAN should allow a clear dissection of pathogenetic mechanisms.
The aim of the present study was to selectively study the role of CD4 and CD8 T cells in P0106–125-induced murine EAN and to determine their impact on the autoimmune-disease manifestation in the peripheral nerve, and also in its target organ, ie, the gastrocnemius muscle, as well as in its origin, ie, the spinal cord. EAN of P0106–125-immunized mice was characterized by a combined axonal and demyelinating inflammatory neuritis resulting in a severe dying back axonopathy with apoptosis of a large fraction of spinal cord motor neurons and a rapidly progressive neurogenic muscular atrophy. T cell depletion experiments identified CD4 T cells as major pathogenetically relevant T cell population with a predominant impact on axonal damage and, subsequently, motor neuron apoptosis and muscular atrophy.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
C57BL/6 mice were bred in the specific pathogen-free animal facility of the University of Cologne (Germany). All animals were housed under specific pathogen-free conditions throughout the experiments. The experiments were carried out according to local governmental regulations. Six-to-eight-week-old male mice were immunized with 200 µg of P0106–125 peptide (Leids Universitair Medisch Centrum, Leiden, The Netherlands; amino acid sequence: I-V-G-K-T-S-Q-V-T-L-Y-V-F-E-K-V-P-T-R-Y) dissolved in dimethyl sulfoxide (Sigma, Taufkirchen, Germany) and 0.5 mg Mycobacterium tuberculosis (strain H 37 RA; Difco; Detroit, MI) emulsified in 25 µl PBS and 25 µl Freund’s incomplete adjuvant (Difco) into their hind foot pads on day (d) 0 and into the back on day 7, respectively. In addition, 500 ng pertussis toxin per mouse (Sigma) was injected i.p. at day 0 and day 2. Control mice were pretreated with PBS instead of P0106–125-peptide.
T Cell Depletion
For depletion of CD4 and CD8 T cells, mice were treated with either rat anti-mouse CD4 (clone GK1.5; concentration 5 mg/ml) and rat anti-mouse CD8 (clone 2.43; concentration 2.5 mg/ml) antibodies, respectively, as described before.17
Control mice were treated with rat IgG (Sigma, Deisenhofen, Germany; concentration 1.25 mg/ml). T cells were depleted by i.p. injection of 200 µl of the respective antibody at days –2, –1, 2, 5, and 7. Successful depletion was determined in the spleen and in the popliteal lymph node (PLN) at days 8, 15, 22, and 40 post-immunization (p.i.) by flow cytometric analysis (depletion efficiency always 
97%).
Assessment of Clinical Score
Before the first depletion and afterward, at days 0, 2, 5, 7, 8, 9, 12, 15, 18, 22, and 40 p.i., mice were clinically examined in a blinded fashion to determine disease activity according to a previously published scale14 : 0 = normal; 1 = floppy tail, weak grip, ruffled fur and slightly reduced motility; 2 = mild paraparesis with impaired motility; 3 = severe paraparesis with significantly reduced motility; 4 = tetraparesis with complete immobilization; 5 = death due to autoimmune neuritis.
Flow Cytometry
At days 8, 15, 22, and 40 p.i., mice were intracardially perfused with 0.9% NaCl in deep CO2 anesthesia. Splenic and PLN leukocytes were isolated and analyzed by triple immunofluorescence staining followed by flow cytometry as described.17 Triple stainings were performed as follows: (1) rat anti-mouse CD25-fluorescein isothiocyanate (FITC; clone 7D4), rat anti-mouse CD4-phosphatidylethanolamine (clone GK1.5), and rat anti-mouse CD8a-peridinin-chlorophyll-protein complex (clone 53-6.7); (2) rat anti-mouse CD4-FITC (clone GK1.5), rat anti-mouse CD45R/B220-phosphatidylethanolamine (clone RA3-6B2) and rat anti-mouse CD8a-peridinin-chlorophyll-protein complex; (3) rat anti-mouse F4/80-FITC (clone BM8), mouse anti-mouse major histocompatibility complex (MHC) class II-antigen-phosphatidylethanolamine (clone AF6-120.1), and rat anti-mouse CD45-peridinin-chlorophyll-protein complex (clone 30-F114); rat anti-mouse CD44-FITC (clone IM7), rat anti-mouse CD62L-phosphatidylethanolamine (clone MEL-14), and rat anti-mouse CD3-peridinin-chlorophyll-protein complex (clone 145-2C11). All antibodies were purchased from BD Biosciences (Heidelberg, Germany), except for anti-F4/80, which was obtained from Caltag (Hamburg, Germany). Flow cytometry was performed on a BD FACSCalibur (BD Biosciences) and analyzed with the CellQuest Pro software (BD Biosciences).
Enzyme-Linked Immunospot Assay
The frequency of P0106–125-specific lymphocytes in spleen and PLN was determined by use of an interferon (IFN)-- and interleukin (IL)-17-specific enzyme-linked immunospot assay at days 8, 15, 22, and 40 p.i. as described previously.17
Triplicates of freshly isolated leukocytes at densities of 2 x 103/well, 2 x 104/well, and 2 x 105/well, respectively, were placed in an enzyme-linked immunospot assay plate (Nunc, Wiesbaden, Germany) coated with rat anti-mouse IFN- (clone RMMG-1 (10 µg/ml; Biosource, Camarillo, CA) or rat anti-mouse IL-17A (clone eBio17CK15A5, 10 µg/ml, eBiosciences). Cells were cocultured with syngeneic C57BL/6 spleen cells from non-immunized mice (4 x 105/well) preloaded with P0106–125 peptides at a concentration of 1 x 10–6 M overnight. Plates were developed with biotin-labeled rat anti-mouse IFN- (BD Pharmingen) or biotin-conjugated rat anti-mouse IL-17A (clone eBio17B7, eBiosciences) followed by peroxidase-conjugated streptavidin (Dianova, Hamburg, Germany), and amino-ethylcarbazole dye solution (Sigma). Spots were counted microscopically. The number of the corresponding epitope-specific lymphocytes was expressed as the number of spots per leukocytes seeded. Isolated spleen and lymph node cells from immunized or PBS-injected mice incubated with non-stimulated spleen cells (negative control) and spleen and lymph node cells from PBS-injected mice incubated with P0106–125-stimulated spleen cells served as controls.
Histopathology and Immunohistochemistry
For histopathology and immunohistochemistry, animals were intracardially perfused with 0.9% NaCl in deep CO2 anesthesia at days 8, 15, 22, and 40 p.i.
One sciatic nerve per animal was excised from the lumbar spinal cord to the periphery including parts of the gastrocnemius muscle, which is innervated by the sciatic nerve, mounted on thick filter paper with Tissue Tek OTC compound (Miles Scientific, Naperville, IL), snap-frozen in isopentane (Fluka, Neu-Ulm, Germany) precooled on dry ice, and stored at –80°C until preparation of serial 10 µm frozen sections. The other sciatic nerve and corresponding gastrocnemius muscle were postfixed in 3.9% phosphate buffered glutaraldehyde (Merck, Darmstadt, Germany) for 48 hours, osmicated, and embedded in epoxy resin (Serva, Heidelberg, Germany) for preparation of 1 µm semithin sections. The spinal cord including the meninges was prepared and postfixed with 4% buffered paraformaldehyde before dissection, embedding in paraffin, and preparation of 4-µm paraffin sections.
Sections of the respective tissues were stained with H&E, Masson-Goldner, 0.1% toluidine blue O (Merck) in 2.5% sodium carbonate (Merck), silver impregnation (Bielschowsky), luxol fast blue, cresyl violet, enzyme histochemistry with reactions for myofibrillary ATPase pH at 9.4, and following preincubation at pH 4.6 in acid phosphatase, and reduced nicotinamide adenine dinucleotide- tetrazolium reductase. In addition, immunohistochemistry with the following rat anti-mouse monoclonal antibodies obtained from hybridomas (American Type Culture Collection, Manassas, VA) were performed: CD45 (clone M1/9.3.4.HL.2), F4/80 (clone F4/80), Mac-1 (clone WEHI-274.1), Mac-3 (clone M3/84), MHC class I-antigen haplotypes (clone M 1/42.3.9.8 HLK), and MHC class II-antigen (I-Ab,d,q haplotypes; clone M5/114.15.2). CD4 (clone H129.19), CD8a (clone 53–6.7), CD19 (clone 1D3), and tumor necrosis factor (TNF; clone MP6/XT22) were purchased from BD Biosciences. In addition, the following primary antibodies were used: glial fibrillary acidic protein (GFAP, Dako, Hamburg, Germany), neurofilament (NF-L, Chemicon/Millipore, Schwalbach, Germany), inducible nitric oxide (iNOS, Calbiochem, Darmstadt, Germany), CD95 (FAS, LabVision, Fremont, CA), CD79a (Zytomed Systems, Berlin, Germany), and synaptophysin antibody (clone SP11, DCS, Hamburg, Germany). Axonal and neuronal damage, respectively, was evidenced by application of monoclonal mouse 
-amyloid precursor protein (-APP, clone 22C11, Chemicon) and monoclonal mouse NeuN (clone A60, Chemicon) and detected by use of the Animal Research Kit (Dako) according to the manufacturer’s instructions.
Immunohistochemistry was performed by use of the avidin-biotin complex technique with appropriate biotinylated secondary antibodies (Vectastain Elite Kit, Vector Laboratories, Burlingame, CA). Peroxidase reaction product was visualized using 3,3'-diaminobenzidine (Sigma) as chromogene and H2O2 as cosubstrate.
The terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) kit (Roche, Mannheim, Germany) was applied according to the manufacturer’s instructions.
Statistical Analysis
To test for differences in the clinical disease activity a two-tailed Mann-Whitney U-test was applied. The number of CD4 and CD8 T cells and macrophages was determined by counting immunoreactive cells in high power fields (microscopic magnification x400) of CD4, CD8, or Mac3 immunostained sections, respectively, of the spinal cord and the sciatic nerve in at least four animals per experimental group. The number of apoptotic neurons, identified by eosinophilic cytoplasm, and condensed hyperchromatic nuclei, or by TUNEL staining was counted in high power fields of cresyl violet stained spinal cord paraffin sections of at least two animals per experimental group. Data are presented as mean ± SD. Differences in the number of inflammatory cells and antigen-specific T cells in spleen and PLN, were determined using a two-tailed Student’s t-test. P values <0.05 were considered significant. All experiments were performed at least three times with four animals per experimental group. Thus, overall 840 animals were studied.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
All animals immunized with P0106–125 developed clinical signs of neuritis up to day 15 (Figure 1A)
as published previously.14
Immunized mice having received rat IgG did not differ from animals without application of antibodies. At day 8 p.i., individual mice were diseased and showed a floppy tail, weak grip, ruffled fur, and a slightly reduced motility (Figure 1B
; mean 0.87 ± 0.2). At this time point, single CD4 T cells had homed to the sciatic nerve. Myelin sheaths, axons, and the gastrocnemius muscle were normal.
|
; mean 2.83 ± 0.3), the sciatic nerve harbored prominent inflammatory infiltrates consisting of F4/80+ macrophages, which expressed TNF and iNOS protein (Figure 2A)
, CD4 and CD8 T cells (Figure 3A)
, and single CD19+ B cells. Axons were damaged as demonstrated by a reduction in axonal NF protein expression and APP+ swelling (Figure 2, B and C)
. Semithin sections showed axonal damage evidenced by myelin ovoids within the nerve fibers as an indication of axonal degeneration forming bands of Bungner (Figure 4A)
. Remarkably, damaged axons were associated with macrophages (Figure 4, A and B)
. These alterations caused a severe and progressive neurogenic muscular atrophy with macrophages phagocytosing damaged muscle fibers (Figure 4C)
. However, demyelination of axons was mild (Figure 4, A and B)
.
|
|
|
; mean 2.5 ± 0.5). In the sciatic nerve, CD4 and CD8 T cells, and CD19+ B cells had significantly declined (P < 0.05, Figure 3B
), whereas F4/80+ macrophages were more numerous as compared to day 15 p.i. (P < 0.05, Figure 3B
). Segmental demyelination as well as axonal damage including axonal swelling of the sciatic nerve persisted. Leukocytic infiltrates composed of CD4 and CD8 T cells, macrophages, and single CD19+ B cells were still present in the gastrocnemius muscle, which showed prominent neurogenic muscular atrophy. Phagocytic activity in muscle fibers was ongoing. In addition to these alterations of the sciatic nerve and the gastrocnemius muscle, the spinal cord showed marked pathological alterations. As early as day 8 p.i., coinciding with the clinical onset of disease, in the spinal cord a few Mac-3+ MHC class II-antigen+ microglial cells and some ubiquitously distributed GFAP+ astrocytes were weakly activated. However, inflammatory infiltrates, in particular CD3+ T cells were absent from both the parenchyma and the leptomeninges.
From day 8 to 15 p.i., pathology had progressed. In the lumbar and thoracic spinal cord, APP staining revealed axonal swelling, and a strong reduction of NF protein indicated axonal loss. Although some Mac-3+ MHC class II-antigen+ macrophages were present in the white matter of the spinal cord, demyelination was not obvious (Figure 5A
; Figure 6A
). Remarkably, anterior horn motor neurons were apoptotic as evidenced by cell shrinkage, a condensed nucleus, an eosinophilic cytoplasm as well a positive TUNEL reaction (Figure 5A, 6B
, and inset in 6B). They did not express Fas (CD95, data not shown). TUNEL-positive neurons did not express NeuN and their synaptic density was markedly reduced compared to controls as evidenced by a reduced synaptophysin expression. Although activation of microglial cells and astrocytes had moderately increased from day 8 to 15 p.i. (Figure 6A)
, they did neither form clusters nor were they preferentially associated with neurons. CD3+ T cells were absent from the spinal cord.
|
|
At day 40 p.i., the anterior horn of the spinal cord harbored patchy loss of neurons (Lückenfelder). Remaining neurons were viable, had a triangular shape with a round nucleus, a prominent basophilic nucleolus, without evidence for chromatolysis or nuclear fragmentation, a normal cytoplasm, and a prominent axon (Figures 5C and 6C)
. Neuronal viability was further confirmed by a prominent nuclear and cell surface expression of NeuN (Figure 6C
, inset) and synaptophysin, respectively, indicative of a normal nuclear structure and synaptic density. Negative TUNEL staining lends further support to the absence of neuronal apoptosis (Figure 5A)
. A strong and ubiquitous NF protein expression evidenced a dense axonal network, and axonal damage as well as demyelination was absent. In the spinal cord parenchyma as well as in the leptomeninges neither CD3+ T cells nor other CD45-expressing leukocytes or Mac3+ macrophages were detected (Figure 5B)
.
At day 40 p.i., inflammatory infiltrates were absent from the sciatic nerve (Figure 3C)
. Only single Mac3+ macrophages were present (Figure 3C)
. There was neither axonal damage nor segmental demyelination. Accordingly, neurogenic muscular atrophy did not progress in the gastrocnemius muscle. At this late time point, hypertrophic muscle fibers and fiber type grouping indicated compensatory collateral re-innervation.
CD4 T Cells are Important for Development and Severity of EAN
To precisely assess the role of CD4 and CD8 T cells in murine EAN, CD4 and CD8 T cell depletion experiments were performed.
The severity of EAN were significantly reduced in anti-CD4 (mean 1.08 ± 0.3) and anti-CD4 plus anti-CD8-treated mice (mean 1.4 ± 0.25) as compared to rat IgG-treated control mice (mean 2.83 ± 0.38; Figure 1B
; P < 0.05 at days 8, 15, and 22 p.i.).
At day 8 p.i., when clinical symptoms were still absent, sciatic nerve, gastrocnemius muscle, and spinal cord of anti-CD4 plus anti-CD8-treated animals were neuropathologically normal.
At day 15 p.i., the sciatic nerve harbored only a few F4/80+ macrophages, while T cells were absent. The number of TNF+ and iNOS+ cells was also strongly reduced as compared to rat IgG-treated animals (Figure 2, A and D)
. Moreover, depletion of CD4 & CD8 T cells resulted (i) in remarkably less severe axonal damage of the sciatic nerve as evidenced by only a focal reduction of NF expression (Figure 2E)
, the absence of APP+ axonal swelling (Figure 2F)
, and the absence of myelin ovoids and missing bands of Bungner in semithin sections (Figure 4D)
. In addition, demyelination was slightly reduced (Figure 4E)
, and neurogenic muscular atrophy was markedly reduced as compared to rat IgG-treated mice (Figure 4F)
. Thereafter, inflammatory infiltrates regressed in parallel to the declining disease activity (Figure 1B)
.
At day 15 p.i., in the spinal cord of mice devoid of CD4 and CD8 T cells, demyelination was absent, and axons were only mildly damaged. In addition, the number of apoptotic neurons was significantly decreased as compared to rat IgG-treated mice (P < 0.05; Figures 5A and 6E
). These neurons only weakly expressed NeuN. In addition, synaptophysin was expressed at decreased density as compared to control mice. Microglial activation was weak; only exceptional Mac-3+ macrophages were detected in the spinal cord of these animals (P < 0.05 for CD4 & CD8- versus rat IgG-treated mice, Figure 5B
). These neuropathological alterations persisted unchanged up to day 22 p.i.
At day 40 p.i., mice depleted of CD4 or CD4 and CD8 T cells neither showed axonal damage nor demyelination in their spinal cord as evidenced by a normally dense expression of the NF protein and lack of APP immunoreactivity. The parenchyma and the leptomeninges of the spinal cord did not harbor CD45+ leukocytes including CD3+ T cells and Mac3+ macrophages (Figure 5B)
. Neurons in the anterior horn of the spinal cord were viable, exhibited a triangular shape, a round nucleus with a prominent basophilic nucleolus without signs of chromatolysis or nuclear fragmentation, a normal cytoplasm, and a prominent axon (Figure 6F
, inset in 6F
). Correspondingly, there was no evidence for ongoing apoptosis (Figure 5A)
. However, the absolute number of viable anterior horn neurons was 95.57 ± 5.44 in 5 HPF in CD4+ depleted mice, thus, significantly less decreased as compared to rat IgG-treated animals (52.86 ± 3.34; P < 0.00002; Figure 5C
).
At day 40 p.i., inflammatory infiltrates were absent from the sciatic nerve of CD4-depleted mice (Figure 3C)
. Only single macrophages were detected (Figure 3C)
. In the absence of ongoing axonal damage or segmental demyelination, the gastrocnemius muscle contained a few hypertrophic muscle fibers and incipient fiber type grouping indicative of collateral re-innervation.
Having established that T cells were decisive for the induction of EAN, we next determined whether CD4 or CD8 T cells were more important. Mice having been depleted of CD4 T cells showed a similar mild clinical course and essentially the same morphological alterations as observed for mice having received anti-CD4 plus anti-CD8 treatment (Figure 1, A and B)
. In contrast, mice lacking CD8 T cells exhibited the same severe clinical disease activity as rat IgG-treated animals (Figure 1B
; P > 0.05 at days 8, 15, and 22 p.i.). Accordingly, the morphological characteristics of EAN were similar to the rat IgG-treated group (data not shown). Thus, these data point to an important role of CD4 T cells in murine EAN with a particular impact on axonal damage and spinal cord motor neuron apoptosis.
Immunization with P0106–125 Induces Antigen-Specific, IFN-- and IL-17-Producing CD4 T Cells
In the PLN and spleen of immunized mice, P0106–125-specific IFN--producing cells were induced (Figure 7)
. In the PLN, maximal numbers of P0106–125-specific cells were reached as early as day 8 p.i. declining thereafter (Figure 7)
. In contrast, numbers of splenic antigen-specific IFN--producing cells were maximal at day 15 p.i. (Figure 7B)
.
|
-producing cells was massively reduced in PLN and spleen (Figure 7, A–C)
illustrating that the immunization with P0106–125 induced antigen-specific T cells. To define which further T cell subset was specific for P0106–125, either CD4 or CD8 T cells were selectively depleted. In anti-CD4-treated mice, only very small numbers of P0106–125-specific IFN--producing cells were detectable in PLN and spleen, whereas the depletion of CD8 T cells did not significantly reduce numbers of P0106–125-specific IFN--producing cells compared to rat IgG-treated mice (Figure 7, A–C)
. Thus, CD4 T cells are the major P0106–125-specific T cell population producing the pro-inflammatory cytokine IFN-.
In PLN and spleen of immunized mice, P0106–125-specific IL-17-producing cells were induced. At day 8 p.i., numbers of IL-17-producing T cells were low in PLN and spleen of rat IgG-treated and CD8-depleted mice with a rise up to day 15 and 22 p.i (Figure 8A–C)
. Antigen-specific IL-17-producing cells were completely absent from the PLN and spleen of CD4- or CD4&CD8-depleted mice throughout the study, further lending support to the importance of CD4 T cells in the pathogenesis of EAN.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
EAN of the peripheral nerve was characterized by a predominant and severe axonopathy with a mild accompanying demyelination. Since the P0-protein expressing Schwann cells are MHC class II negative18 (and this study), it is likely that the EAN-inducing CD4 T cells, which were activated on P0106–125 immunization, homed to the sciatic nerve and activated local resident endoneurial macrophages. In fact, the infiltrates in the diseased sciatic nerve contained CD4 T cells and activated macrophages expressing iNOS and TNF protein. Since both iNOS and TNF are toxic for axons,17,19,20 the axonopathy may be caused by these activated macrophages. In fact, nitric oxide can promptly and reversibly block axonal conduction, which is impaired in P0106–125-induced EAN as shown previously.14 nitric oxide affects several axonal channels, including sodium, potassium, and calcium channels and may block them, thereby impairing conduction.19 In this regard, murine EAN is similar to multiple sclerosis, in which the extent of axonal destruction is related to the number of activated macrophages, which are present in close apposition with injured axons.19,21 The rather mild demyelination of the sciatic nerve in murine P0106–125-induced EAN may occur secondarily to axonal damage or, mutually not exclusive, may also be caused by iNOS- and TNF-producing macrophages, and may render axons more sensitive to nitric oxide toxicity, since demyelinated axons are particularly vulnerable to the axonal conduction blocking effects of nitric oxide.22 This scenario is supported by our finding that a depletion of CD4 T cells resulted in a clinically highly attenuated EAN with significantly reduced numbers of iNOS+ and TNF+ macrophages in the sciatic nerve.
The severity of sciatic nerve damage P0106–125-induced EAN is further highlighted by the prominent neurogenic muscular atrophy. Importantly, inflammatory infiltrates, in particular CD4 and CD8 T cells, were absent from affected muscles and macrophages were confined to atrophic muscle fibers in the process of myophagia.
Addressing the mechanisms underlying P0106–125-induced EAN, functional T cell depletion experiments unequivocally identified the important role of CD4 T cells for development and severity of disease. P0106–125-specific CD4 T cells were the major cell population responsible for antigen-specific IFN- production, while CD8 T cells were less important. In addition, CD4 T cells produced IL-17. Since IL-17-producing CD4 T cells play a crucial role in the development of experimental autoimmune encephalomyelitis,23
this cytokine may also be important in our EAN model. In contrast to IL-17, IFN- is dispensable for the induction of experimental autoimmune encephalomyelitis, but may contribute to propagation of the disease.
Interestingly, induction of EAN was not completely abrogated in mice devoid of CD4 T cells. Even mice having been depleted of both CD4 and CD8 T cells developed EAN on immunization with the P0106–125 peptide, although with a significantly milder degree. The remaining low disease activity of CD4 T cell-depleted mice is in accordance with studies in CD4–/–, CD8–/–, and CD4–CD8– mutant C57BL/6 mice,7,24
which argues against an induction of EAN in our T cell-depleted animals by the remaining small number of T cells (
3%). Furthermore, neuropathological analysis of serial immunostained sections confirmed the absence of the respective T cell population in the sciatic nerve and the spinal cord. Collectively, these clinical and neuropathological observations point to the involvement of other cell populations in the pathogenesis of EAN. As CD4 and CD4 plus CD8 T cell-depleted mice still mounted a weak residual IFN- response in the spleen and PLN, one may suggest that 
CD8+ T cells and NK cells, which are a cellular source of IFN-,25
also play a role in P0106–125-induced EAN. In fact, these cell populations are present in the cauda equina in EAN of Lewis rats.26
However, so far, their role in murine EAN has not been investigated in detail, and the relative role of these cell populations as well as the importance of IFN- and IL-17 remain to be determined.
The relative impact of CD4 and CD8 T cells on the various manifestations of EAN is still controversially discussed. The significantly reduced clinical disease activity in mice lacking CD4 T cells was attributed to a remarkably less severe axonal neuropathy of the sciatic nerve, whereas the mild demyelination was less affected. The latter may be caused by iNOS- and TNF-producing endoneurial macrophages, which still were present in both CD4 and CD4 & CD8 T cell-depleted mice, albeit in significantly reduced number. These findings point to a particular impact of CD4 T cells on axonal damage and also highlight the importance of axonal damage for clinical disability in EAN. Clinically relevant, patients with the axonal variant of the Guillain-Barré syndrome have an explosive onset and a more severe clinical course as compared to those with the demyelinating variant.27 Moreover, patients with multifocal motor neuropathy present with more pronounced and progressive neurogenic muscular atrophy,28 which is also a prominent feature of P0106–125-induced EAN. In addition, in multiple sclerosis, axonal loss, but not demyelination, is associated with irreversible neurological disability.2,19,29
A comprehensive neuropathological investigation of the sciatic nerve as target organ of the autoimmune response as well as the spinal cord and the gastrocnemius muscle revealed that in the absence of prominent inflammatory infiltrates a significant fraction, ie, up to 44%, of lumbar and thoracic spinal cord motor neurons underwent apoptosis. This remarkable extent of apoptotic neurons may account for the persistence of clinical symptoms up to day 22 p.i. Our observation of a complete remission of neurological symptoms at later time points, ie, 40 days p.i.14 and this study, points to the existence of functional regenerative processes. In this scenario, classical compensation by collateral re-innervation may be the most important factor. In fact, collateral re-innervation was prominent in the gastrocnemius muscle of rat IgG and anti-CD8-treated mice at day 40 p.i. Remarkably, although CD4- and CD4&CD8-depleted mice exhibited only minor symptoms of EAN, these slight alterations were sufficient to stimulate collateral innervation, albeit to a remarkably lesser extent as compared to rat IgG-treated mice. In contrast, the persistence of a reduced number of viable anterior horn neurons in rat IgG- and anti-CD8-treated mice argues against a compensatory effect by the generation of new functionally active neurons in these animals. In addition, the production of neurotrophic factors, eg, brain-derived neurotrophic factor, which can also be produced by autoimmune T cells,30-32 may play a role. However, CD4 & CD8 T cell depleted mice also recovered from their mild EAN indicating that T cells are more important for disease induction than for recovery.
The extensive neuronal damage in the spinal cord of P0106–125-immunized mice is attributed to a severe dying back axonopathy resulting from the severe axonal damage and has, to the best of our knowledge, not yet been reported for EAN, while it is characteristically observed in spinal cord injury, sciatic nerve axotomy, and facial nerve axotomy.33-36
Our histopathological and immunohistochemical observations viable and structurally unimpaired neurons as well as of a normal neurono-axonal system at day 40 p.i., ie, a late time point when clinical symptoms have subsided, lend further support to the concept of a dying back axonopathy underlying the spinal cord damage as opposed to a primary neuronopathy. Interestingly, in multiple sclerosis, neuronal damage has gained increasing attention recently as a key feature of pathology.21,29
In principle, axonal and neuronal injury in the spinal cord as well as other regions of the central nervous system may be caused in a both antigen-specific as well as antigen-nonspecific manner, either by CD8 T cells or macrophages, respectively. T cells aligning along axons and soma of neurons induce substantial neuronal loss in a cell contact-dependent mechanism involving Fas (CD95)-FasL (CD95L) interactions,37,38
and blocking CD95 activity prevents motor neurons from undergoing apoptosis in models of facial nerve axotomy and spinal ischemia.38
Since in our model of EAN, T cells were consistently absent from the gray matter of the spinal cord as well as from the (overlying) leptomeninges and Fas was not up-regulated on motor neurons, a T cell mediated neuronal destruction is very unlikely to account for the prominent loss of neurons, neither by CD8 T cells nor by CD4 T cells, which may also damage neurons via expression of TNF-related apoptosis inducing ligand.29
In addition, IFN--activated macrophages/microglial cells may damage neurons by inducing oxidative stress via their production of TNF and iNOS/ nitric oxide.17,39
However, in murine P0106–125-induced EAN, there was no evidence for a direct macrophage-induced damage of spinal cord neurons, since microglial activation was only moderate and iNOS-expressing macrophages/microglial cells were not associated with neurons in the anterior horn. Only rarely, activated microglial cells/macrophages were observed in the vicinity of neurons, which unequivocally showed evidence for apoptosis, and this colocalization was most likely part of the process of neuronophagia. Collectively, these observations indicate that neuronal loss occurs secondarily to inflammation-induced axonal damage of the sciatic nerve. In fact, leukocytic infiltrates were largely confined to the sciatic nerve, the target organ of the autoimmune response. A similar mechanism underlies retinal ganglion cell degeneration in optic neuritis.40
Thus, this study identifies experimental murine P0106–125-induced EAN as a disease involving both the peripheral nervous system and central nervous system. These observations have implications for the pathogenesis of autoimmune T cell-mediated diseases of peripheral nerves, and clinically relevance for the development of specific therapeutic strategies aiming to inhibit the autoimmune attack against peripheral nerves and to prevent a dying back axonopathy, in particular, by conferring axonal protection, which is of upmost importance for the final clinical outcome.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by Köln-Fortune Program (grant no. 60/2006).
Accepted for publication April 3, 2008.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
