Efficient Central Nervous System Remyelination Requires T Cells Allan J. Bieber, PhD,1,2 Scott Kerr, BS,1 and Moses Rodriguez, MD1,2,3 General nonspecific immunosuppression often is used as a treatment for demyelinating disease and spinal cord injury. A better understanding of the role of the immune system in CNS injury and repair is essential for the ongoing evaluation and development of these therapies. Materials and Methods Animals We demonstrate a role for immune functions in the spontaneous remyelination of central nervous system (CNS) axons after lysolecithin-induced demyelination in the spinal cord. Rag-1–deficient mice lack both B cells and T cells and show significantly reduced spontaneous remyelination compared with control mice of matching genetic background. Mice lacking or depleted of either CD4ⴙ T cells or CD8ⴙ T cells also exhibit reduced remyelination. These data indicate that T cells are necessary for efficient CNS remyelination. Thus, general nonspecific immunosuppression as a therapeutic approach for the treatment of CNS injury and demyelinating disease may have undesirable effects on subsequent tissue repair. Ann Neurol 2003;53:680 – 684 Spontaneous remyelination is a normal physiological response after myelin damage but the factors that control rate and extent of spontaneous myelin repair after central nervous system (CNS) demyelination are largely unknown. Several recent reports have suggested that various immune cells and immune effector molecules may play a role in myelin repair.1–5 The aim of this study was to establish a role for immune functions in the repair process by examining the extent of remyelination in mice with genetic deficiencies in immune function or after antibody depletion of specific sets of immune cells. CNS demyelination was induced by injection of lysolecithin into the spinal cord, and remyelination was assessed 35 days later. Lysolecithin-induced demyelination occurs independently of immune function and therefore is an excellent system in which to assess the contribution that immune functions make to the spontaneous remyelination process. From the 1Department of Neurology, 2Program in Molecular Neuroscience, and 3Department of Immunology, Mayo Medical and Graduate Schools, Rochester, MN. Received Dec 16, 2002, and in revised form Feb 10, 2003. Accepted for publication Feb 11, 2003. Address correspondence to Dr Bieber, Department of Neurology, Guggenheim 418, Mayo Clinic and Foundation, 200 First St. SW, Rochester, MN 55905. E-mail: email@example.com 680 © 2003 Wiley-Liss, Inc. C57BL/6J, B6-Rag1tm1Mom, B6-CD4tm1Mak, and B6CD8tm1Mak mice were purchased from the Jackson Laboratories (Bar Harbor, ME) or bred in-house. Mice were housed in plastic cages with food and water provided ad libitum. Handling of animals conformed to National Institutes of Health and Mayo Clinic guidelines. Antibody Depletion The hybridoma lines Gk1.5 and Lyt2.43, which secrete function neutralizing rat antibodies to CD4 and CD8, respectively, were obtained from the American Type Culture Collection. Hybridomas were adapted to growth in serum and protein-free medium (Sigma, St. Louis, MO) supplemented with 1% fetal calf serum. Antibody was precipitated from culture supernatant with ammonium sulfate and then purified by gel filtration on a Superose-6 column. Lysolecithin Injection and Quantification of Remyelination Lysolecithin injections were performed on 12-week-old mice as previously described.6 In brief, after dorsal laminectomy a micropipette was inserted into the dorsolateral part of the spinal cord and 2.5l of 1% lysolecithin in saline was injected. The needle was withdrawn and the wound was closed. The day of lysolecithin injection was designated day 0. Thirty-five days after lysolecithin injection, mice were anesthetized and perfused by intracardiac administration of Trump’s fixative (phosphate-buffered 4% formaldehyde with 1% glutaraldehyde, pH 7.4). Spinal cords were removed and cut into 1mm sections, postfixed with osmium, and embedded in araldite plastic (Polysciences, Warrington, PA). Onemicrometer-thick cross-sections were cut from each block and stained with 4% p-phenylenediamine to visualize myelin. For each animal, the block showing the largest demyelinated lesion was used for quantitative analysis. Blocks showing obvious signs of tissue damage due to injection were excluded from the analysis. The average density of myelinated axons is much higher in the dorsal columns of the spinal cord than in the ventral or lateral columns, and mixing quantitative remyelination data from dorsal versus ventral/ lateral lesions creates added variability in the data. Therefore, only ventral and lateral lesions were selected for quantification. Productive lesion formation in the ventral and lateral tracts after lysolecithin injection was observed in 42 of 72 total mice. The total number of animals and the number of lesions that were analyzed in each experimental group are listed in the Table. In most animals, only a single ventral/ lateral lesion was present, but in three animals, two separate ventral/lateral lesions were observed. For statistical analysis, when two lesions existed in the same animal, the data for Table. Central Nervous System Remyelination of Lysolecithin:Induced Lesions in Immunodeficient Mice Experiment C57BL/6 B6-Rag1tm1Mom B6-CD4tm1Mak B6-CD8tm1Mak B6-CD4tm1Mak ⫻ B6-CD8tm1Mak C57BL/6 depleted for CD4 T cells with Gk 1.5 antibody C57BL/6 depleted for CD8 T cells with Lyt 2.43 antibody No. of No. of Animals Lesions 7 6 5 5 7 7 10 6 5 5 7 7 5 5 Mean (median) Lesion Areas (mm2) ⫾ SEM Mean (median) Remyelinated Axons/ mm2, ⫾ SEM Remyelination vs C57BL/6 (P value) 0.0513 ⫾ 0.0070 (0.0546) 0.0468 ⫾ 0.0052 (0.0446) 0.1030 ⫾ 0.0317 (0.0732) 0.0538 ⫾ 0.0155 (0.0418) 0.0384 ⫾ 0.0046 (0.0377) 0.0495 ⫾ 0.0087 (0.0427) 75,447 ⫾ 8,184 (66,833) 26,538 ⫾ 3,893 (29,408) 14,694 ⫾ 3,131 (15,168) 32,772 ⫾ 7,249 (29,235) 63,455 ⫾ 4,976 (62,845) 32,344 ⫾ 6,816 (28,531) 0.001 0.003 0.005 0.259 0.004 0.0683 ⫾ 0.0287 (0.0431) 28,986 ⫾ 4,319 (26,130) 0.003 See Materials and Methods section for additional details. Pair-wise comparison of remyelination to the C57BL/6 control strain was by Mann– Whitney rank-sum test. Comparison of mean lesion area between C57BL/6 and B6-Rag1tm1Mom by Mann–Whitney rank-sum test showed no significant difference ( p ⫽ 0.628). Comparison of mean lesion areas between all experimental groups by Kruskal–Wallis one-way analysis of variance on ranks also showed no significant differences ( p ⫽ 0.564). lesion area and quantification of remyelination were combined, and the two lesions were considered as one. All analysis was done on coded samples without knowledge of the experimental group. Lesions areas were determined using a Zeiss interactive digital analysis system (ZIDAS; Zeiss, Thornwood, NY) as described.7,8 Remyelinated axons were identified by their thin myelin sheaths (Fig) and counted manually on a photographic montage of each individual lesion. Remyelination is expressed as the number of remyelinated axons per square millimeter of lesion (see Table). Pair-wise comparison of remyelination to the C57BL/6 control strain was by Mann– Whitney rank-sum test. Comparison of mean lesion area between C57BL/6 and B6-Rag1tm1Mom was also by Mann– Whitney rank-sum test, and comparison of mean lesion areas for all experimental groups was by Kruskal–Wallis one-way analysis of variance on ranks. elination after lysolecithin injection is independent of immune system function. Morphometric analysis of spinal cord remyelination showed significantly greater remyelination in normal controls compared to Rag-1–deficient mice (see Fig; Table). Control animals showed an average remyelination of 75,447 axons/mm2, whereas B6-Rag1tm1Mom mice had 26,538 remyelinated axons/mm2. This difference was statistically significant with p ⫽ 0.001. Allowing 60 days for remyelination to occur after lysolecithin injection did not increase the extent of remyelination in Rag-1 mice, indicating that the observed effect is a true inhibition rather than a delay in the remyelination process. These data demonstrate that some aspect of immune system function is necessary for efficient remyelination of CNS axons. Results Remyelination of Lysolecithin-Induced Demyelination Is Inhibited in B6-Rag1 Mice To determine whether immune functions play a role in remyelination after lysolecithin-induced demyelination, we induced lesions in B6-Rag1tm1Mom mice, which lack recombination activating gene-1 and therefore produce no mature B cells or T cells.9 Our previous experience with this experimental system has shown that remyelination is well established by 21 days after lysolecithin injection and complete by 35 days.6,7 Therefore, 35 days after lysolecithin injection, we compared extent of remyelination in B6-Rag1tm1Mom mice with that in control mice with matching genetic background (C57BL/6). There were no significant differences in mean lesion area between any experimental groups in this study ( p ⫽ 0.564, see Table). The mean lesion area in B6Rag1tm1Mom mice was slightly larger than that in C57BL/6 mice, but this difference was not significant ( p ⫽ 0.628), suggesting that the mechanism of demy- Remyelination Is Inhibited in Mice Deficient for Either CD4⫹ or CD8⫹ T Cells To further explore the role of immune functions in remyelination, we assessed remyelination in mice deficient for CD4, which is important in MHC class II– restricted immune responses, or deficient for CD8, important in MHC class I–restricted immune responses. Remyelination was assessed 35 days after lysolecithin injection and compared with that in control mice with matching genetic background (C57BL/6). Morphometric analysis of remyelination in CD4 and CD8-deficient mice showed that absence of either branch of immune function resulted in significantly reduced remyelination (see Fig, Table). B6-CD4tm1Mak mice had an average of 14,694 remyelinated axons/ mm2, whereas B6-CD8tm1Mak mice had an average of 32,772 remyelinated axons/mm2. These numbers were significantly different from those of C57BL/6 control animals ( p ⫽ 0.003 for CD4, p ⫽ 0.005 for CD8). Reduced remyelination in both CD4 and CD8- Bieber et al: Remyelination Requires T Cells 681 682 Annals of Neurology Vol 53 No 5 May 2003 deficient strains was somewhat unexpected. Both strains were created in the same laboratory; therefore, we were concerned that an unknown second site mutation that affects remyelination might have existed in the parent strain. To test this possibility, we crossed the B6-CD4tm1Mak and B6-CD8tm1Mak strains and assessed remyelination in F1 progeny. These animals demonstrated an average of 63,455 remyelinated axons/mm2, which was statistically similar to that observed in the C57BL/6 control group ( p ⫽ 0.259). Antibody-Mediated Depletion of CD4⫹ and CD8⫹ Cells Inhibits Remyelination Mutations often have pleiotropic effects, and we were concerned that CD4 and CD8 mutations might have unknown effects on CNS development that result in an inability to efficiently remyelinate in adult animals. To address this concern, we used antibodies to deplete CD4⫹ or CD8⫹ T cells in adult C57BL/6 mice and then assessed remyelination. Mice were injected on three successive days with 0.5mg of either Gk1.5 (anti-CD4) or Lyt2.43 (antiCD8). One week after injection, the peripheral blood of all animals was assessed for the presence of CD4⫹ and CD8⫹ cells by fluorescence-activated cell sorting. Average depletion for CD4⫹ cells was 98.53 ⫾ 0.41% compared with untreated C57BL/6 controls, and 95.78 ⫾ 0.50% for CD8⫹ cells. By 35 days after lysolecithin injection when the animals were prepared for histological examination, the CD4⫹ or CD8⫹ T-cell populations in peripheral blood generally had recovered to only 20% of their normal levels. Mice depleted of either CD4⫹ or CD8⫹ cells displayed levels of remyelination similar to those seem in animals with a genetic deletion (see Fig, Table). Animals depleted for CD4⫹ cells averaged 32,344 remyelinated axons/mm2, and those depleted of CD8⫹ cells averaged 28,986 remyelinated axons/mm2. These data demonstrate that both the CD4 and CD8 branches of the immune system make independent contributions to CNS remyelination. Discussion The immune response in the adult mouse spinal cord after lysolecithin injection has been characterized,10,11 and our own immunostaining experiments are consistent with the published observations. Lysolecithin in- Š jection induces a rapid but transient influx of T cells and neutrophils into the CNS. Infiltration and activation of macrophages and microglia begins within hours after injury but persists for many days. The role that these cell types play in establishing an environment in which remyelination can occur is unknown. A neuroprotective role for T cells after CNS injury has been suggested.2,12 Systemic injection of myelin-reactive T cells after spinal cord injury results in enhanced accumulation of T cells, B cells, and macrophages at the site of injury.13 This T-cell response promotes the expression of various neurotrophins by macrophages and astrocytes that may play a role in promoting neuronal survival. Several growth factors have been identified that show increased levels of expression during remyelination,14 and it is possible that T cells play a similar role in supporting oligodendrocyte remyelination, either directly or by stimulating the activity of CNS glia. Depletion of macrophages impairs oligodendrocyte remyelination, suggesting that these cells also may be important for support of the myelin repair process.4 Immunochemical staining of any of the immunodeficient mice used in this study showed normal accumulations of CD45⫹ and CD11b⫹ macrophages at lesion sites, and the number and morphology of these cells was not different after remyelination. Whether these cells were expressing the full array of normal functions is not clear. Thus, we demonstrate that the immune system does provide functions necessary for remyelination and that CD4⫹ and CD8⫹ T cells are both required for efficient CNS remyelination to occur. This work was supported by grants from the NIH (NS24180, M.R., A.B. and NS40209, M.R.), the National Multiple Sclerosis Society (RG 3172-A-6, M.R., A.B). We thank Eugene and Marcia Applebaum for their generous support to M.R. We thank Dr P. C. O’Brien for assistance with the statistical analysis. References 1. Njenga MK, Murray PD, McGavern D, et al. Absence of spontaneous central nervous system remyelination in class IIdeficient mice infected with Theiler’s virus. J Neuropathol Exp Neurol 1999;58:78 –91. Fig. Remyelination of lysolecithin-induced lesions in control and immunodeficient mice and in mice depleted for CD4⫹ or CD8⫹ T cells. C57BL/6 normal white matter is shown in A. Panels B to H show remyelination in lysolecithin lesions: C57BL/6 (B), B6Rag1tm1Mom (C), B6-CD4tm1Mak (D), B6-CD8tm1Mak (E), B6-CD4tm1Mak ⫻ B6-CD8tm1Mak (F), C57BL/6 depleted of CD4⫹ T cells (G), and C57BL/6 depleted of CD8⫹ T cells (H). Remyelinated axons are identified by their relatively thin myelin sheaths (arrows in B) compared with the thicker and more darkly staining normal myelin sheaths as seen in normal white matter (A). A decrease in remyelination of approximately two- to fourfold can be seen in the lesions of immunodeficient mice (C–E, G, and H) as compared with the control animals (B, F). Bieber et al: Remyelination Requires T Cells 683 2. Schwartz M, Moalem G, Leibowitz-Amit R, Cohen IR. Innate and adaptive immune responses can be beneficial for CNS repair. Trends Neurosci 1999;22:295–299. 3. Arnett HA, Mason J, Marino M, et al. TNF␣ promotes proliferation of oligodendrocyte progenitors and remyelination. Nat Neurosci 2001;4:1116 –1122. 4. Kotter MR, Setzu A, Sim FJ, et al. Macrophage depletion impairs oligodendrocyte remyelination following lysolecithininduced demyelination. Glia 2001;35:204 –212. 5. Bieber AJ, Warrington A, Pease LR, Rodriguez M. Humoral autoimmunity as a mediator of CNS repair. Trends Neurosci 2001;24:S39 –S44. 6. Bieber AJ, Warrington A, Asakura K, et al. Human antibodies accelerate the rate of remyelination following lysolecithininduced demyelination in mice. Glia 2002;37:241–249. 7. Pavelko KD, van Engelen BG, Rodriguez M. Acceleration in the rate of CNS remyelination in lysolecithin-induced demyelination. J Neurosci 1998;18:2498 –2505. 8. McGavern DB, Murray PD, Rodriguez M. Quantitation of spinal cord demyelination, remyelination, atrophy, and axonal loss in a model of progressive neurologic injury. J Neurosci Res 1999;58:492–504. 684 Annals of Neurology Vol 53 No 5 May 2003 9. Mombaerts P, Iacomini J, Johnson RS, et al. RAG-1 deficient mice have no mature B and T lymphocytes. Cell 1992;68: 869 – 877. 10. Ousman SS, David S. Lysophosphatidylcholine induces rapid recruitment and activation of macrophages in the adult mouse spinal cord. Glia 2000;30:92–104. 11. Ousman SS, David S. MIP-1␣, MCP-1, GM-CSF, and TNF-␣ control the immune cell response that mediates rapid phagocytosis of myelin from the adult mouse spinal cord. J Neurosci 2001;21:4649 – 4656. 12. Hohlfeld R, Kerschensteiner M, Stadelmann C, et al. The neuroprotective effect of inflammation: implications for the therapy of multiple sclerosis. J Neuroimmunol 2000;107: 161–166. 13. Barouch R, Schwartz M. Autoreactive T cells induce neurotrophin production by immune and neural cells in injured rat optic nerve: implications for protective autoimmunity. FASEB J 2002;16:1304 –1306. 14. Franklin RJ, Hinks GL, Woodruff RH, et al. What roles do growth factors play in CNS remyelination? Prog Brain Res 2001;132:185–193.