Modulation of Macrophage Infiltration and Inflammatory Activity by the Phosphatase SHP-1 in Virus-Induced Demyelinating Disease

Abstract

The protein tyrosine phosphatase SHP-1 is a crucial negative regulator of cytokine signaling and inflammatory gene expression, both in the immune system and in the central nervous system (CNS). Mice genetically lacking SHP-1 (me/me) display severe inflammatory demyelinating disease following inoculation with the Theiler's murine encephalomyelitis virus (TMEV) compared to infected wild-type mice. Therefore, it became essential to investigate the mechanisms of TMEV-induced inflammation in the CNS of SHP-1-deficient mice. Herein, we show that the expression of several genes relevant to inflammatory demyelination in the CNS of infected me/me mice is elevated compared to that in wild-type mice. Furthermore, SHP-1 deficiency led to an abundant and exclusive increase in the infiltration of high-level-CD45-expressing (CD45^hi) CD11b⁺ Ly-6C^hi macrophages into the CNS of me/me mice, in concert with the development of paralysis. Histological analyses of spinal cords revealed the localization of these macrophages to extensive inflammatory demyelinating lesions in infected SHP-1-deficient mice. Sorted populations of CNS-infiltrating macrophages from infected me/me mice showed increased amounts of viral RNA and an enhanced inflammatory profile compared to wild-type macrophages. Importantly, the application of clodronate liposomes effectively depleted splenic and CNS-infiltrating macrophages and significantly delayed the onset of TMEV-induced paralysis. Furthermore, macrophage depletion resulted in lower viral loads and lower levels of inflammatory gene expression and demyelination in the spinal cords of me/me mice. Finally, me/me macrophages were more responsive than wild-type macrophages to chemoattractive stimuli secreted by me/me glial cells, indicating a mechanism for the increased numbers of infiltrating macrophages seen in the CNS of me/me mice. Taken together, these findings demonstrate that infiltrating macrophages in SHP-1-deficient mice play a crucial role in promoting viral replication by providing abundant viral targets and contribute to increased proinflammatory gene expression relevant to the effector mechanisms of macrophage-mediated demyelination.

Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system (CNS) that remains a major cause of disability (76). Several studies demonstrated that MS lesions contain multiple leukocyte cell types, including lymphocytes, macrophages, and dendritic cells, all of which are believed to contribute to lesion formation by various distinct and interacting mechanisms (54, 61). Among these leukocyte subsets, infiltrating macrophages have been identified as major effectors of demyelination in both MS and animal models for MS (17, 44, 59, 102). In accord with these studies, it was recently reported that the dominant mechanism of demyelination in MS is macrophage mediated (15). Indeed, in some models for MS, the requirement for lymphocytes is negligible and macrophages are the sole mediators of demyelination (6, 72).

These findings have stimulated intense interest in the function of macrophages in lesion formation, including signaling events that draw these cells into the CNS white matter and trigger the effector mechanisms by which these cells damage myelin. For instance, macrophages have been identified as the major responders to CNS chemokines and producers of a number of proinflammatory cytokines, chemokines, and toxic molecules known to promote demyelination (32, 44, 63, 85, 90, 91, 93, 103). Interestingly, both the brains of MS patients and the brains of experimental animals with MS-like diseases contain activated transcription factors like NF-κB (34, 40), STAT1 (35, 37, 40), and STAT6 (18, 21, 109), which can lead to enhanced expression of these inflammatory molecules. Based on the findings of our previous work, we propose that the modulation of inflammatory signaling via these transcriptional pathways may be deficient in the leukocytes, including the macrophages, of MS patients and that this deficiency is responsible for susceptibility to inflammatory demyelinating processes within the CNS.

SHP-1 is a protein tyrosine phosphatase with two Src homology 2 domains which acts as a negative regulator of both innate and acquired immune cytokine signaling via NF-κB (52, 67), STAT1 (29, 68), and STAT6 (13, 43, 45, 50). Mice genetically lacking SHP-1 (motheaten mice) display myelin deficiency, which may be mediated by increased innate inflammatory mediators in the CNS (66, 69). Furthermore, motheaten mice are highly susceptible to experimentally induced demyelinating disease (30, 65). Taken together, the findings of these studies indicate that SHP-1 is a key regulator of inflammation in the CNS that may be relevant to the pathogenesis of MS. Indeed, we have recently reported that SHP-1 expression and function are deficient in the leukocytes of MS patients compared to those in the leukocytes of normal human subjects (21).

To further elucidate the function of SHP-1 in inflammatory demyelinating disease, we have been utilizing the Theiler's murine encephalomyelitis virus (TMEV)-induced demyelinating disease animal model of MS (14, 20, 27, 33, 56). When susceptible strains of mice are inoculated intracerebrally with TMEV, the mice develop immune system-mediated demyelinating disease resembling human MS. TMEV infection leads to biphasic disease, with an early gray-matter infection followed by later progressive white-matter disease involving a complex array of leukocytes and proinflammatory molecules that eventually cause demyelination. It is now well-accepted that while lymphocytes are the initiators and perpetuators of demyelination in this model, macrophages are the major effectors of demyelination (8, 24, 57, 87, 88, 94).

Previously, we have shown that SHP-1-deficient mice uniquely display unusually rapid CNS demyelination associated with extensive white-matter cellular infiltration and clinical paralysis within the first week of TMEV infection compared to their wild-type littermates (65). These data suggested that SHP-1 modulates early events in TMEV infections of the CNS which cause inflammatory demyelination. Moreover, the rapidity of demyelination suggests the involvement of innate inflammatory effectors of demyelination. Therefore, it became essential to investigate the cellular and molecular mechanisms that contribute to the early onset of severe CNS demyelination observed in me/me mice.

The findings of the present study demonstrate that a deficiency in SHP-1 leads to an augmented inflammatory gene profile and increased infiltration of peripheral macrophages into the CNS following TMEV infection. In addition, macrophage infiltration into spinal cords is concentrated in areas of demyelination in me/me mice. Accordingly, monocyte/macrophage depletion with clodronate liposomes resulted in a significant delay in the onset of TMEV-induced paralysis and decreased viral loads, inflammation, and demyelination in the spinal cords of me/me mice compared to those in wild-type littermates. Furthermore, we have shown that SHP-1 controls chemokine production by CNS glial cells and the responsiveness of macrophages to these chemokines, which play an important role in CNS macrophage-mediated disease. We therefore propose that SHP-1 is an important regulator of CNS inflammatory demyelination, acting to control inflammatory gene expression and the recruitment macrophages into the CNS in response to viral infection.

MATERIALS AND METHODS

Animals.

SHP-1-deficient motheaten (me/me) mice (with a C3HeB/FeJLe-a/a background) and their phenotypically normal wild-type littermates were produced from heterozygous breeding pairs obtained from Jackson Laboratories (Bar Harbor, ME). The strain designation for heterozygous breeders for motheaten mice is C3FeLe.B6-a/a Hcph^me/+ (stock no. 000225).

Paralysis scores.

Mice were examined daily for signs of paralysis and scored on a 5-point scale, as follows: 1, incomplete hind limb paralysis (dragging of hind limb but ability to move it); 2, complete paralysis of one hind limb; 3, complete paralysis of one hind limb and one forelimb or paralysis of both hind limbs; 4, quadriplegia; and 5, death preceded by paralysis.

Virus inoculation.

The attenuated TMEV strain BeAn 8386 (ATCC, Manassas, VA) was propagated, and titers were determined by plaque assays with BHK-21 cells. Whole-cell glial cultures were inoculated with 10⁶ PFU/ml at a multiplicity of infection of 1.0. Twelve-day-old mice were inoculated intracerebrally in the right hemisphere with 5 × 10⁵ PFU of BeAn 8386 in a volume of 0.005 ml. Mice were observed daily for paralysis. Unless otherwise specified, mice were anesthetized 4 days postinfection (p.i.) and perfused, and the right cerebral hemispheres and spinal cords were suspended in RNA STAT-60 (TEL-TEST, Friendswood, TX) for RNA analysis. Also, brains and spinal cords were used to prepare single-cell suspensions for flow cytometry analysis.

In vivo depletion of macrophages.

To deplete wild-type and me/me mice of macrophages, liposome-encapsulated clodronate (clodronate liposomes) were used. Clodronate was purchased from Roche Pharmaceuticals (Germany) and was encapsulated in liposomes by Encapsula Nanosciences (Nashville, TN). The final solution contained 5 mg/ml of the encapsulated clodronate drug. Control liposomes contained phosphatidylcholine and cholesterol without clodronate. Mice weighing approximately 5 g were injected intraperitoneally with 0.2 mg of clodronate 2 days before infection with TMEV. This injection route and the method were previously shown to specifically deplete CD11b⁺ high-level-Ly-6C-expressing (Ly-6C^hi) monocytes/macrophages in multiple tissues and blood (10, 83, 110). As controls, mice were injected intraperitoneally with the same volume of control liposomes 2 days before TMEV infection. Mice further received either control liposomes or 0.1 mg of clodronate liposomes on days 2 and 6 p.i. Macrophage depletion in clodronate liposome-treated mice was evaluated by staining the spleens and CNS tissues of mice for CD45^hi CD11b⁺, CD11b⁺ Ly-6C^hi, and F4/80⁺ cells 4 or 6 days after the first clodronate injection and comparing the samples to those from mice receiving the control liposome injections.

Real-time RT-PCR.

Total RNA was isolated using RNA STAT-60. RNA was quantified spectrophotometrically, and 0.5 μg of total RNA was converted into cDNA. Briefly, total RNA and random primers (Invitrogen, Carlsbad, CA) were incubated at 72°C for 10 min. Reverse transcription was performed using the SuperScript II reverse transcriptase (RT) enzyme according to the specifications of the manufacturer (Invitrogen, Carlsbad, CA). cDNA was diluted with water in a final volume of 200 μl, and 4 μl was used for quantitative real-time PCR with a Sybr green kit (ABgene, Epsom, United Kingdom). The PCR parameters were 15 min for 95°C and 35 cycles of 95°C for 15 s and 60°C for 1 min in an ABI Prism 700 system (Applied Biosystems, Foster City, CA). The primers were used at 10 nM. Serial dilutions of cDNA containing a known number of copies of each gene were used in each quantitative PCR run to generate a standard curve relating the copy number with the threshold amplification cycle (22). Gene expression levels were calculated during the logarithmic amplification phase by determining the initial mRNA copy number according to the standard curve. The amplification of each gene-specific fragment was confirmed both by the examination of melting peaks and by agarose gel electrophoresis. The primer pairs used in this study are shown in Table 1.

TABLE 1.

Sequences of primers used for real-time RT-PCR analysis

Product of target gene	Forward primer (5′-3′)	Reverse primer (5′-3′)
TMEV VP2	TGGTCGACTCTGTGGTTACG	GCCGGTCTTGCAAAGATAGT
TNF-α	TGAACTTCGGGGTGATCGGTC	AGCCTTGTCCCTTGAAGAGAAC
IL-6	CAGAGGATACCACTACCAACAG	TCTCATTTCCACCACGATTTCCC
LT-α	TTCCTCCCAATACCCCTTC	AGGCTCCAAAGAATACACTGC
MMP3	TTGACGATGAACGTAGG	GGCTGAGTGGTAGAGTCCCAG
ADAM8	GCAGGACCATTGCCTCTAC	GACCCAACTCGGAAAAAGC
Arginase I	TTTCTCAAAAGGACAGCCTC	GTGAGCATCCACCCAAATG
RANTES	AGCTGCCCTCACCATCATC	CTCTGGGTTGGCACACACTT
MCP-1	GTATGTCTGGACCCATTCCTTC	GCTGTAGTTTTTGTCACCAAGC
MIP-1α	CAGCCAGGTGTCATTTTCCT	AGGCATTCAGTTCCAGGTCA
Eotaxin	TGCTCACGGTCACTTCCTTC	TTTGCCCAACCTGGTCTTG
GAPDH	ACCACCATGGAGAAGGC	GGCATGGACTGTGGTCATGA

Flow cytometry. (i) Characterization and quantification of immune cell infiltration into the CNS.

The brains and the spinal cords of mice were removed and placed into 2 ml of cold Hanks balanced salt solution (HBSS). The tissues were homogenized with a fine-tip glass pipette and then filtered through a 40-μm-pore-size cell strainer cap as described previously (58) but without collagenase digestion. The single-cell suspensions were washed twice with cold HBSS containing 5% fetal bovine serum (FBS). Aliquots of 10⁶ cells were resuspended in 100 μl of HBSS, and the suspensions were incubated with 10 μl of CD3-fluorescein isothiocyanate (FITC), CD11b-phycoerythrin (PE), CD19-FITC, CD5-PE, CD49b-PE, Gr-1-FITC, or Ly-6C-FITC antibody (all from Becton Dickinson, Mountain View, CA).

(ii) Sorting of infiltrating macrophages.

Aliquots of 2 × 10⁶ cells from single-cell suspensions of the brain, spinal cord, or spleen samples from TMEV-infected mice were stained with CD11b-PE and Ly-6C-FITC antibodies (Becton Dickinson, Mountain View, CA). Two populations, doubly positive cells and doubly negative cells, were sorted by a fluorescence-activated cell sorter (FACSVantage S/E; Becton Dickinson Immunocytometry Systems, Mountain View, CA). Aliquots of 5 × 10⁴ cells were collected, and cells were lysed in the RNA isolation reagent.

(iii) Intracellular staining.

Single-cell suspensions in 1.0 ml of phosphate-buffered saline (PBS) received 100 μl of 16% stock paraformaldehyde for fixation (1.5% paraformaldehyde [final concentration]) for 5 min and were then incubated in 90% methanol at 4°C for half an hour to permeabilize cells for intracellular staining. Cells were washed twice with the staining medium containing 0.5% bovine serum albumin and 0.02% sodium azide in PBS. The levels of several intracellular antigens were concurrently analyzed with CD45-FITC antibody. Fixed and permeabilized cells were incubated overnight at 4°C with 1 μg of goat anti-ADAM8, goat anti-matrix metalloproteinase 3 (anti-MMP3), or goat anti-myelin basic protein (anti-MBP) antibody (all from Santa Cruz Biotech, Santa Cruz, CA) or goat polyclonal immunoglobulin G antibodies for isotype control. Then the cells were incubated for 3 h in 1 μg of swine anti-goat secondary antibody conjugated to PE (Invitrogen, Carlsbad, CA). Similarly, cells were single stained with rat anti-mouse lymphotoxin alpha (LT-α; R&D Systems, Minneapolis, MN), rat anti-mouse F4/80 (Serotec), and mouse anti-arginase I (BD Biosciences) overnight and then stained with swine anti-rat or goat anti-mouse secondary antibodies conjugated to PE (Invitrogen, Carlsbad, CA). Single-cell suspensions were also double stained with CD45 antibody and rabbit anti-TMEV antibody (a gift from Howard L. Lipton, University of Illinois, Chicago), followed by goat anti-rabbit secondary antibody conjugated to PE (Invitrogen, Carlsbad, CA). Cells were analyzed on an LSRII analyzer (Becton Dickinson, Mountain View, CA), and the percentage of positively stained cells and the mean florescence intensity were recorded. Data were analyzed with FlowJo software (TreeStar, Ashland, OR).

Immunohistochemical analysis.

Spinal cords from wild-type and me/me mice before and 4 days after infection with TMEV were used for immunohistochemical analyses. Also, spinal cords of mice that were pretreated with control liposomes or clodronate liposomes and then infected with TMEV were analyzed. Mice were anesthetized and intracardially perfused via the left ventricle first with 10 ml of PBS and then with 20 ml of 4% paraformaldehyde in PBS. Spinal cords were further incubated in 4% paraformaldehyde in PBS for 1 h and then dissected and placed in a 30% sucrose solution in PBS overnight. Spinal cord samples were embedded in Tissue-Tek OCT (optimal cutting temperature) compound Ted Pella, Inc., Redding, CA), frozen on dry ice, and divided into 8-μm-thick sections with a cryostat at −16°C. To double stain for MBP and CD11b, sections were first stained with mouse CD11b monoclonal antibody directly conjugated to biotin (R&D Systems) and then incubated in a streptavidin-alkaline phosphatase conjugate. A blue alkaline phosphatase reaction product was produced using a BCIP (5-bromo-4-chloro-3-indolylphosphate)/Nitro Blue Tetrazolium substrate kit (Zymed/Invitrogen, Carlsbad, CA). The same sections were then stained with a goat anti-MBP antibody (catalog no. sc-13914; Santa Cruz Biotechnology), followed by anti-goat immunoglobulin-horseradish peroxidase (Dako), and a red color was developed using the horseradish peroxidase substrate aminoethyl carbazole chromogen (Zymed). Also, spinal cords were stained with rat biotinylated ani-CD4 and anti-CD8 monoclonal antibodies (Biosource).

Glial cultures.

Mixed glial cultures containing astrocytes, oligodendrocytes, and microglial cells were produced from brains of 8-day-old mice by a modified procedure as described previously (46, 66, 71). Briefly, brains were minced into fine pieces in Kreb's buffer with curved scissors. The minced brain tissue was centrifuged and resuspended in Kreb's buffer containing 0.25% trypsin, and the suspension was incubated at 37°C for 1 min. After the addition of 5% FBS and 40 μg/ml DNase in Kreb's buffer to stop trypsinization, the tissue was pelleted and resuspended in fresh Kreb's buffer-FBS-DNase. The tissue was repeatedly triturated with a fire-polished pipette to dissociate cells. The cells were centrifuged and then resuspended in complete culture medium containing Dulbecco's modified Eagle's medium with 10% FBS, 24.5 mM KCl, and 100 μg/ml of insulin and plated onto polylysine-coated dishes. Every 3 days, cells were fed with fresh medium consisting of Dulbecco's modified Eagle's medium with 10% horse serum heat inactivated at 56°C for 1 h, and cells were used 10 days after plating. By morphological criteria, glial cultures appeared to be composed of 10% microglial cells, 55% astrocytes, and 35% oligodendrocytes.

Macrophage cultures.

Spleens were removed from mice and ground between the frosted surfaces of two glass slides in HBSS. Freed cells were left for 2 min on ice and pelleted by centrifugation, and the supernatant was discarded. The pellet was resuspended in 4 ml of red blood cell lysis buffer (155 mM NH₃Cl, 0.1 mM EDTA, 12 mM NaHCO₂)/spleen and incubated on ice for 1 min. Cells were washed twice with HBSS and resuspended in RPMI medium containing 15% FBS supplemented with 10% (vol/vol) culture supernatant from L929 cells (ATCC) (12). Cells were plated at a density of 5 × 10⁵ cells/ml and were fed every 3 days. On day 9, viability was tested with trypan blue and the adherent cells were infected with TMEV for 48 h. Following culture, more than 97% of the cells were CD11b⁺ Ly-6C⁺ cells as assessed by flow cytometry.

Chemokine-cytokine ELISA.

The levels of the chemokine monocyte chemoattractant protein 1 (MCP-1) in supernatants of glial cultures were measured using DuoSet enzyme-linked immunosorbent assay (ELISA) kits according to the protocol of the manufacturer (R&D Systems). The levels of the cytokines tumor necrosis factor alpha (TNF-α) and interleukin-6 (IL-6) in spinal cord homogenates were quantified. Spinal cords were homogenized in radioimmunoprecipitation assay buffer, the amount of total protein was determined, and the amounts of TNF-α and IL-6 were assessed using the DuoSet ELISA kits according to the protocol of the manufacturer (R&D Systems).

Chemotaxis assay.

The migration of splenic macrophages was measured using the CytoSelect 96-well cell migration assay (Cell Biolabs, San Diego, CA). Briefly, either RPMI medium with 10% FBS containing 40 ng of MCP-1 or the supernatants of 48-h cultures of glial cells infected or mock infected with TMEV served as test chemoattractants. Splenic macrophages were cultured as described above, and aliquots of 7.5 × 10⁵ cells from either wild-type or me/me mice were placed into the 5-μm polycarbonate membrane chambers. Cells were incubated in the different chemoattractant solutions for 6 h, and cells that migrated were lysed and quantified with CyQuant GR dye by evaluating the fluorescence intensities at 480 and 520 nm. In the MCP-1 function-blocking experiments, glial supernatants were incubated with 10 μg/ml of rat anti-MCP-1 neutralizing antibody (catalog no. MAB489; R&D Systems) or 10 μg/ml of rat anti-immunoglobulin G control antibody (catalog no. 6-001-A; R&D Systems) at 37°C for 1 h.

Statistical analysis.

Histograms and tables show the mean values with standard errors. The numbers of samples or individual mice used in each assay are indicated in the figure legends. The P values were generated using the unpaired Student t test, and a P value of less than 0.05 was interpreted to indicate a statistically significance difference between the results for two samples.

RESULTS

Elevated inflammatory gene expression in the brains and spinal cords of me/me mice following TMEV infection.

It was shown previously that cultured CNS glial cells of SHP-1-deficient me/me mice display augmented signal transduction through STATs and NF-κB, resulting in increased inflammatory gene expression. The regulation of these pathways by SHP-1 may therefore account for the increased inflammatory demyelination in the CNS of TMEV-infected me/me mice, in which these signaling molecules become activated (13, 52). Consequently, we investigated whether me/me mice showed increased expression of STAT- and NF-κB-responsive genes in the brain or spinal cord compared to that in wild-type littermates following TMEV infection (Table 2). Wild-type and me/me mice were inoculated with the BeAn strain of TMEV, and 4 days p.i., the spinal cords were analyzed for gene expression by real-time RT-PCR. At 4 days p.i., me/me mice exhibited complete hind limb paralysis, in contrast to wild-type mice, which did not show signs of paralysis. The spinal cords of me/me mice showed a significant 12-fold increase in the amount of TMEV RNA compared to that in wild-type cords.

TABLE 2.

Gene expression in spinal cords of wild-type and me/me mice before and 4 days after infection with TMEV^a

Gene product or target of analysis	RNA transcript level before TMEV infection in:		RNA transcript level after TMEV infection in:
	Wild-type mice	me/me mice	Wild-type mice	me/me mice
TMEV			*1 ± 0.3	12 ± 4†
TNF-α	1 ± 0.1	3.8 ± 0.6†	*129 ± 53	650 ± 190†
IL-6	1 ± 0.3	1.4 ± 0.1	*6.6 ± 2	94 ± 35†
LT-α	1 ± 0.2	3.0 ± 0.2†	*1.6 ± 0.3	3.8 ± 0.8†
MMP3	1 ± 0.3	1.1 ± 0.4	*5.3 ± 3	86 ± 33†
ADAM8	1 ± 0.1	1.2 ± 03	1.1 ± 0.2	12 ± 3†
Arg1	1 ± 0.1	3.5 ± 1.1†	*1.8 ± 0.5	8.1 ± 1.6†
RANTES	1 ± 0.1	3.4 ± 0.8†	*9.2 ± 1.4	6e3 ± 5e2†
MCP-1	1 ± 0.3	8.1 ± 1.0†	*1.1e3 ± 6e2	2e4 ± 9e3†
MIP-1α	1 ± 0.1	5.1 ± 1.1†	*6.7 ± 1.3	524 ± 80†
Eotaxin	1 ± 0.4	2.5 ± 0.6†	*2.2 ± 0.6	7.4 ± 2.4†

^aTotal RNAs from uninfected wild-type (n = 14), infected wild-type (n = 13), uninfected me/me (n = 11), and infected me/me (n = 12) mice were analyzed by real-time RT-PCR. The actual transcript copy number for each gene is reported in relation to the transcript level in uninfected wild-type mice, which is set at 1. Gene expression was normalized to that of the housekeeping GAPDH gene. Data are means ± standard errors. * indicates significance at P of <0.05 for values from before and after virus infection, and † indicates significance at P of <0.05 for values from wild-type and me/me mice.

The expression of the STAT6-responsive arginase I gene (79) in the brains and spinal cords of me/me mice was elevated both constitutively and in response to TMEV infection compared to that in wild-type mice (Table 2). The examination of genes that are regulated by both STAT6 and NF-κB, such as the protease ADAM8 (53, 95), the chemokine eotaxin/CCL11 (70), and the LT-α (81, 106) genes, revealed substantial increases in expression in me/me mice compared to that in wild-type mice following TMEV infection.

Because SHP-1 has been shown previously to control NF-κB signaling in CNS glial and hematopoietic cells, it was of interest to examine the expression of NF-κB-inducible genes like those for the cytokines TNF-α and IL-6, the chemokines MCP-1 and macrophage inflammatory protein 1α (MIP-1α), and the matrix metalloproteinase MMP3 (Table 2). The levels of TNF-α, MCP-1, and MIP-1α were constitutively higher in me/me spinal cords than in wild-type spinal cords before infection with TMEV. Following viral infection, TNF-α, IL-6, MCP-1, and MIP-1α were upregulated in both wild-type and me/me spinal cords but were substantially more abundant in the spinal cords of me/me mice than in those of wild-type mice. Similarly, the levels of expression of the corresponding genes in me/me mouse brains were observed to be higher than those in wild-type mouse brains (data not shown). These findings point to the importance of SHP-1 in controlling CNS inflammatory gene expression that may play a role in the progression of disease.

Immune cell infiltration into the CNS.

The observation that TMEV infection induced an exaggerated inflammatory profile with substantially increased chemokine expression in CNS tissues from me/me mice compared to those from wild-type mice led us to examine potential infiltrating inflammatory cells in these animals (Fig. 1). Brains and spinal cords of wild-type and me/me mice before and 4 days after infection with TMEV were homogenized into single-cell suspensions, and the presence of blood-derived leukocytes was quantified using flow cytometry as described previously (58). First, double staining for CD45 and CD11b revealed that following TMEV infection, 10% of the cells in wild-type brains were CD45^hi CD11b⁺ cells but that me/me brains contained 27% CD45^hi CD11b⁺ cells (Fig. 1A and C). The spinal cords of wild-type mice contained less than 2% CD45^hi CD11b⁺ cells following TMEV infection, compared to 30% CD45^hi CD11b⁺ cells in the spinal cords of me/me mice. Because me/me mice had substantially higher levels of infiltration by CD45^hi CD11b⁺ cells, these cells were further characterized based on recently developed criteria for blood-derived monocytes/macrophages. First, staining demonstrated similar percentages of CD45^hi Gr-1⁺ and F4/80⁺ cells (55, 88), indicating that a large majority of the cells were of monocyte/macrophage origin (Fig. 1D and F). Furthermore, the majority of the CD11b⁺ infiltrating cells had the Ly-6C^hi phenotype (Fig. 1E), which was recently established as a reliable marker of blood-derived macrophages (1, 99, 100). Taken together, these data suggest that the CNS of me/me mice displays a dramatic increase in the number of infiltrating macrophages (7, 41, 58, 88).

FIG. 1.

Characterization and quantification of leukocytes in the brains and spinal cords of wild-type and me/me mice before and 4 days after infection with TMEV. Single-cell suspensions of the brain and spinal cord samples from uninfected wild-type (wt; n = 8), infected wild-type (wt-V; n = 8), uninfected me/me (me; n = 6), and infected me/me (me-V; n = 7) mice were prepared. Cells were double stained for CD45 and CD11b (Mac-1), CD45 and Gr-1, and Ly-6C and CD11b. Cells were single stained for F4/80, EPX, and CD3. Histograms represent the mean percentages of positive cells among the total population of cells analyzed.

Since the lack of SHP-1 has been associated previously with increased susceptibility to inflammation in the lung, in which eosinophils play a role (50), we investigated the presence of eosinophils in the CNS of me/me mice by staining for eosinophil peroxidase (EPX) (19, 42, 89). Both in the brains and in the spinal cords of infected me/me mice, we found a small (3.5%) but significant increase in EPX⁺ cells (Fig. 1G). We did not observe a significant level of CD45^hi CD11b⁻ singly positive cells in the CNS of either wild-type or me/me mice infected with TMEV, indicating that most leukocytes infiltrating the CNS were of macrophage origin. However, to directly address whether other leukocyte subsets may infiltrate the CNS following infection, we stained tissue homogenates with antibodies to CD3 (Fig. 1H), CD49b, CD19, and CD5 (data not shown) to examine the levels of infiltrating T cells, NK cells, B cells, and total lymphocytes, respectively. Staining revealed that the levels of T, B, and NK cells in the CNS did not significantly increase following TMEV infection in either wild-type or me/me mice. Taken together, these data indicate that macrophages and smaller numbers of eosinophils are the main leukocytes infiltrating the CNS of SHP-1-deficient motheaten mice. At 4 days p.i. motheaten mice demonstrated dramatic 3- and 15-fold increases in infiltrating macrophages in the brain and spinal cord, respectively, compared to the levels in their wild-type littermates, and thus, these cells constitute the main inflammatory cellular infiltrate, especially in the spinal cords of TMEV-infected me/me mice.

Infected SHP-1-deficent me/me mice display massive macrophage infiltration into inflammatory spinal cord demyelinating lesions.

As discussed above, macrophages are the main inflammatory cellular infiltrate seen in the CNS of paralyzed me/me mice infected with TMEV. Paralysis in TMEV-infected mice was previously reported to relate to the appearance of extensive spinal demyelinating lesions (65). However, the association between macrophage infiltration and spinal demyelinating lesions has not been established. To do so, we performed double-color immunohistochemistry analysis of CD11b and MBP (Fig. 2). Spinal cords from wild-type and me/me mice that either were uninfected or infected with TMEV for 4 days were fixed, frozen, and sectioned. The spinal cords of each group of animals were surveyed in several sections from different levels of each spinal cord. Such staining allowed an examination of the spatial distributions of macrophage infiltration (CD11b accumulation) and demyelination (MBP loss from specific white-matter tracts). Wild-type and me/me mice that were not infected (Fig. 2A and B) did not show areas of focal demyelination and were free of CD11b⁺ cells. Similarly, TMEV-infected wild-type mice did not show any focal demyelination or the presence CD11b⁺ cells (Fig. 2C). In contrast, infected me/me spinal cords displayed extensive areas of demyelination, primarily in ventrolateral tracts that contained large numbers of CD11b⁺ cells concentrated within the lesions. Gray-matter regions and normal-appearing white matter contained fewer or no CD11b⁺ cells, indicating the preferential localization of macrophages within inflammatory demyelinating lesions (Fig. 2D). Focal demyelination associated with macrophage infiltration was seen throughout the white matter of spinal cords of infected me/me mice, with more severe lesions observed in the lumbar section than elsewhere. Greater magnification revealed that the CD11b⁺ cells displayed a simple rounded, nonramified morphology consistent with blood-borne derivation rather than microglial origin. In addition, the spinal cords of infected wild-type and me/me mice were stained with anti-CD4 and anti-CD8 antibodies and showed no immunoreactivity (data not shown).

FIG. 2.

Immunohistochemical analysis of MBP and CD11b in spinal cord sections from TMEV-infected wild-type and SHP-1-deficient mice. Mice were anesthetized and perfused with paraformaldehyde 4 days after TMEV infection or sham infection. Frozen spinal cord samples were divided into 8-μm-thick sections and double stained with anti-MBP antibody that appears red (aminoethyl carbazole product) and anti-CD11b antibody that appears blue (BCIP/Nitro Blue Tetrazolium product). The sections were photographed at a magnification of ×12 (the upper right corner of each panel shows a scale bar corresponding to 70 μm). Representative spinal cord sections from uninfected wild-type (A), uninfected me/me (B), infected wild-type (C), infected me/me (D), clodronate-treated infected me/me (E), and control liposome-treated infected me/me (F) mice are shown.

Quantifying viral loads and gene expression in sorted CNS-infiltrating macrophages.

Because macrophages were the major infiltrating leukocyte seen in the CNS of infected me/me mice and localized within areas of demyelination, we wanted to assess the contribution of these cells to inflammatory gene expression in infected mice. CD11b⁺ Ly-6C^hi cells (macrophages) from TMEV-infected mice at 4 days p.i. were sorted and compared to CD11b⁻ Ly-6C⁻ cells (CNS resident cells) (Fig. 3). Total RNA was isolated from 5 × 10⁴ cells of each sample, and the amount of virus or inflammatory gene expression per nanogram of total RNA was determined. In addition, CD11b⁺ Ly-6C^hi splenic macrophages were analyzed for comparison. First, the amount of TMEV RNA was measured (Fig. 3B). Although the mean TMEV mRNA level in me/me brains was slightly higher than that in wild-type brains, the viral loads in the infected wild-type and me/me brains and in the macrophage-depleted brain cellular compartments were not significantly different. However, an analysis of the sorted infiltrating-macrophage compartments from these same specimens indicated that me/me brain macrophages contained a fivefold-higher level of viral RNA than wild-type macrophages. This observation suggests that SHP-1 may control selective tropism for these cells (25, 60, 64, 74) and/or increased TMEV replication in the macrophage compartment.

FIG. 3.

Quantification of gene expression by real-time RT-PCR analysis of sorted CNS-infiltrating macrophages. Single-cell suspensions of brain (Br) or spleen samples from TMEV-infected wild-type (wt; n = 10) and me/me (n = 8) mice at 4 days p.i. were stained for CD11b and Ly-6C. Doubly positive cells (macrophages [Mφ]) and doubly negative cells (CNS resident cells from macrophage-depleted brains) were separated by a fluorescence-activated cell sorter. (A) Analysis of a single-cell suspension of the brain sample from an infected me/me mouse showing the gates used to sort out the Ly-6C^hi CD11b⁺ and Ly-6C⁻ CD11b⁻ cells. The efficiency of the sorting was examined by analyzing the postsorting aliquot of doubly positive or doubly negative cells. (B to H) Aliquots of 5 × 10⁴ cells were used to extract total RNA and perform real-time RT-PCR. The levels of expression of TMEV RNA and the mRNA transcripts of the protease ADAM8, LT-α, arginase I, IL-6, and chemokine MCP-1 genes and the housekeeping GAPDH gene per nanogram of total RNA were quantified.

Because infected me/me mice have increased inflammatory gene expression compared to that in wild-type mice, it was important to examine the expression of the inflammatory genes in wild-type and me/me infiltrating macrophages. The level of the protease ADAM8, which has been implicated previously in neurodegeneration (95) and was shown to directly degrade MBP (2), was 25-fold higher in infected me/me brains than in wild-type brains (Fig. 3C). Interestingly, ADAM8 mRNA transcripts were 1,000-fold more abundant in infiltrating macrophages than in CNS tissue depleted of macrophages, indicating the preferential expression of ADAM8 in infiltrating macrophages. Furthermore, there was a substantial upregulation of ADAM8 in CNS-infiltrating macrophages compared to the level of ADAM8 in splenic macrophages. Similarly, the LT-α mRNA message level in me/me brain tissue was higher than that in wild-type brain tissue following TMEV infection. The amount of LT-α mRNA in CD11b⁻ Ly-6C⁻ cells was significantly 10-fold higher than that in infiltrating macrophages, consistent with reports that LT-α is specifically expressed by astrocytes in a macrophage-mediated demyelinating mouse model (81). Furthermore, arginase I mRNA transcripts in the infected brains, infiltrating macrophages, and CNS resident cells of me/me mice were significantly increased compared to those in wild-type littermates (Fig. 3E). Finally, IL-6 and MCP-1 mRNA transcripts were substantially more abundant in me/me brains than in wild-type brains. Importantly, infiltrating CNS macrophages had over 10⁴-fold more IL-6 and MCP-1 mRNA transcripts than splenic macrophages and had more than 10-fold-higher levels of IL-6 and MCP-1 expression than CNS resident cells. Taken together, these data demonstrate that both the increased number of infiltrating macrophages and the augmented inflammatory profile may contribute to the demyelination seen in TMEV-infected me/me mice.

Depletion of infiltrating macrophages using clodronate liposomes.

The above-described observations indicated a role for infiltrating macrophages in the paralysis, demyelination, and increased CNS inflammatory profile seen in TMEV-infected me/me mice compared to wild-type mice. To directly demonstrate this role, macrophages were experimentally depleted in vivo by injecting mice with liposome-encapsulated clodronate. Clodronate liposomes cause transient and selective elimination of macrophages in the spleen, multiple peripheral tissues, and CD11b⁺ Ly-6C^hi monocytes in the blood (47, 98). Also, it has been shown previously that clodronate liposomes do not deplete microglial cells in the brain parenchyma (5, 82). Importantly, clodronate liposomes have been successfully used to stop blood-derived macrophages from participating as effectors in CNS inflammatory disease (38, 48, 101). Hence, wild-type and me/me mice received either clodronate liposomes or PBS control liposomes 2 days before and 2 and 6 days after TMEV infection. Animals were sacrificed 4 days p.i., and the spleens, brains, and spinal cords were analyzed for the presence of CD45^hi CD11b⁺, F4/80⁺, and CD11b⁺ Ly-6C^hi cells (Fig. 4). Clodronate liposome treatment effectively depleted the spleens of macrophages, while there was no effect from control liposome treatment (Fig. 4A). Furthermore, there was a fivefold decrease, from 27% down to 5%, in the level of CD45^hi CD11b⁺ cells in the spinal cords of me/me mice that were treated with clodronate liposomes compared to the level in me/me mice treated with control liposomes (Fig. 4B). TMEV-infected brains from wild-type animals treated with clodronate liposomes displayed a twofold decrease and those from me/me mice displayed a threefold decrease in the levels of CD45^hi CD11b⁺ macrophages compared to those in the brains of control liposome-treated animals (Fig. 4C). Similarly, CD11b⁺ Ly-6C^hi cells, which were shown previously to be specifically depleted in multiple tissues by clodronate treatment (99), were significantly decreased in wild-type and me/me CNS tissues by clodronate (Fig. 4D). Taken together, these data indicate that clodronate liposomes were effective in substantially reducing the infiltration of blood-derived macrophages into the CNS of TMEV-infected mice.

FIG. 4.

In vivo depletion of CNS-infiltrating monocytes/macrophages by using clodronate liposomes. Wild-type (wt) and SHP-1-deficient mice were treated with either control liposomes or liposomes encapsulating clodronate. Two days after liposome (Lip.) treatment, the mice were infected intracerebrally with TMEV, and they were sacrificed at day 4 p.i. Single-cell suspensions of the brain and spinal cord samples from control liposome-treated wild-type (wt-C; n = 8), clodronate liposome-treated wild-type (wt-CLD; n = 11), control liposome-treated me/me (me-C; n = 8), and clodronate liposome-treated me/me (me-CLD; n = 9) mice were prepared. (A) Splenocytes were stained with PE and FITC isotype antibodies or with CD11b and CD45 antibodies. The doubly positive cells representing the macrophage populations in uninfected or infected mice pretreated with either control or clodronate liposomes were quantified. (B) CD11b and CD45 antibodies were used to stain single-cell suspensions of spinal cord samples from wild-type or me/me mice. (C and D) CD45^hi CD11b⁺ and CD11b⁺ Ly-6C^hi cells in the brains and spinal cords of TMEV-infected mice were quantified.

In addition, to assess the effect of macrophage depletion on demyelination in the spinal cords of TMEV-infected me/me mice following treatment with clodronate liposomes, double-color immunohistochemistry analyses of MBP and CD11b were performed (Fig. 2). Spinal cords of TMEV-infected wild-type mice that were treated with either control or clodronate liposomes were free of CD11b⁺ macrophages and showed no signs of demyelination (data not shown). In contrast, spinal cords of infected SHP-1-deficient me/me mice that were treated with control liposomes showed the typical outcome of TMEV infection, with substantial white-matter infiltration by macrophages in association with extensive demyelination (Fig. 2F). Importantly, macrophage depletion in infected me/me mice by using clodronate liposomes resulted in a disappearance of both CD11b⁺ macrophages and focal white-matter demyelination in stained sections of the spinal cords (Fig. 2E).

Paralysis scores of wild-type and me/me mice following infection with TMEV.

Concomitantly with the assessment of CNS macrophage depletion, the effects of clodronate liposomes on paralysis and mortality in TMEV-induced disease were determined. TMEV-induced paralysis scores for eight groups of mice that had received either clodronate or control liposomes were recorded (Fig. 5A). No groups of wild-type mice, regardless of liposome treatment, showed early signs of paralysis following TMEV infection. Of the 25 wild-type mice observed daily for paralysis for up to 3 months p.i., only 7 (25%) showed paralysis, with a much later onset than that in the me/me littermates (4 to 5 weeks p.i.), consistent with the typical TMEV-induced disease pattern in wild-type mice. In sharp contrast, TMEV-infected me/me mice that received either no treatment or control liposomes showed initial signs of paralysis at day 3 p.i., and by day 5 p.i., almost all the mice displayed terminal morbidity similar to that described previously (65) and were therefore sacrificed (scored as death preceded by paralysis). Importantly, infected me/me mice that were pretreated with clodronate liposomes had significantly delayed signs of paralysis appearing at day 6 p.i. and remained alive until day 12 p.i. Therefore, depleting me/me mice of macrophages resulted in a delayed onset and reduced severity of disease compared to those in nondepleted mice following TMEV infection.

FIG. 5.

Paralysis scores for wild-type and motheaten mice following infection with TMEV. Wild-type (wt) and SHP-1-deficient (me) mice were intracerebrally infected with 5 × 10⁵ PFU of BeAn TMEV. (A) Wild-type or me/me mice received either control liposomes or clodronate liposomes 2 days before TMEV infection. In total, there were eight groups of mice: uninfected wild-type mice (n = 20), uninfected me/me mice (n = 15), infected wild-type mice (wt virus; n = 35), infected me/me mice (me virus; n = 16), control liposome-treated wild-type mice (wt Lip virus; n = 14), control liposome-treated me/me mice (me Lip virus; n = 13), clodronate liposome-treated wild-type mice (wt CLD virus; n = 19), and clodronate liposome-treated me/me mice (me CLD virus; n = 14). Mice were scored on a 5-point scale (1, incomplete hind limb paralysis; 2, complete hind limb paralysis; 3, complete hind limb and partial forelimb paralysis; 4, quadriplegia; and 5, death preceded by paralysis). (B) The CD45^hi CD11b⁺ cells in the spinal cords of TMEV-infected SHP-1-deficient mice were quantified between days 3 and 5 p.i., and the results were correlated to the paralysis scores exhibited by individual mice.

In addition, the quantity of CD45^hi CD11b⁺ cells present in the CNS at the time of sacrifice was correlated with the severity of clinical disease. The spinal cords of TMEV-infected me/me mice at different stages of clinical disease between days 3 and 5 p.i. were analyzed for the presence of CD45^hi CD11b⁺ cells (Fig. 5B). The paralysis scores showed a direct correlation with the quantities of infiltrating macrophages detected in the spinal cords in TMEV-infected me/me mice. More specifically, when the level of infiltrating macrophages was below 10%, no signs of disease were observed. When the level of spinal cord-infiltrating macrophages rose above 20% and up to 35%, then severe paralysis was observed in infected SHP-1-deficient me/me mice. Taken together, these data indicate that the number of infiltrating macrophages isolated from spinal cords of TMEV-infected me/me mice correlated with the severity of paralytic disease and that pharmacological depletion of these macrophages resulted in the delayed onset and reduced severity of disease following TMEV infection.

CD45^hi infiltrating cells contribute to both increased viral loads and inflammation in SHP-1-deficient mice.

CNS-infiltrating macrophages are known targets for TMEV infection and significantly contribute to both the viral burdens and the process of inflammatory demyelination in the spinal cords of mice (25, 60, 64, 74). Therefore, the contribution of infiltrating macrophages to the viral burden seen in the CNS of TMEV-infected mice was determined. The amounts of viral antigen (Fig. 6A) and viral RNA (Fig. 7A) in the spinal cords of TMEV-infected mice that had been pretreated with either control or clodronate liposomes were measured. The levels of both the TMEV antigen and RNA were more than 10-fold higher in the spinal cords of me/me mice than in those of wild-type mice 4 days p.i. In addition, double immunofluorescence staining for CD45 and TMEV antigens revealed that CD45^hi cells in me/me spinal cords contained abundant TMEV antigens (Fig. 6A). Importantly, clodronate treatment resulted in a significant decrease in the levels of both TMEV⁺ CD45^hi cells and TMEV RNA in me/me spinal cords (Fig. 6A and 7A). In summary, the levels of TMEV antigens and RNA in the spinal cords of TMEV-infected me/me mice were dramatically higher than those in the spinal cords of infected wild-type mice and the depletion of macrophages effectively decreased these levels.

FIG. 6.

Levels of the TMEV antigen (Ag), the protease ADAM8, and TNF-α in CD45⁺ cells in the spinal cords of wild-type (wt) and SHP-1-deficient (me/me) mice. Mice were infected with TMEV, and 2 days before infection, they received either control liposomes or clodronate liposomes. Spinal cords were removed 4 days p.i., and single-cell suspensions were stained and analyzed by flow cytometry. Cells from the spinal cord were double stained for CD45 and either TMEV antigen (A), the protease ADAM8 (B), or TNF-α (C). The left lower quadrants of the graphs for uninfected mice in panels B and C display the data for the isotype control, shown in gray. Lip., liposomes.

FIG. 7.

Quantification of TMEV RNA and inflammatory gene expression in the spinal cords of TMEV-infected mice. Wild-type and SHP-1-deficient mice were treated with either control liposomes or liposomes encapsulating clodronate. Two days after liposome treatment, the mice were infected with TMEV, and they were sacrificed on day 4 p.i. Spinal cords from uninfected wild-type (wt; n = 7), control liposome-treated wild-type (wt-V; n = 9), clodronate liposome-treated wild-type (wt-V CLD; n = 8), uninfected me/me (me; n = 7), control liposome-treated me/me (me-V; n = 9), and clodronate liposome-treated me/me (me-V CLD; n = 10) mice were analyzed. (A and B) The amounts of TMEV RNA (A) and MCP-1 mRNA (B) per nanogram of isolated total RNA were determined using real-time RT-PCR. (C and D) The levels of TNF-α (C) and IL-6 (D) in homogenized spinal cord tissue were quantified using ELISA. (E and F) The levels of expression of the protease ADAM8 (E) and MBP (F) in single-cell suspensions of spinal cord samples were quantified by fluorescence cytometry, and the mean florescence intensity (MFI) was normalized to that for spinal cord samples from uninfected mice.

CNS-infiltrating macrophages mediate CNS inflammation and demyelination via the secretion of multiple inflammatory mediators (9, 39, 44, 75, 92). Therefore, the contribution of infiltrating macrophages in the local production of these mediators in the CNS of TMEV-infected me/me mice was assessed. Recent evidence indicates that the matrix metalloproteinase ADAM8 plays an important role in both transendothelial migration and proteolysis of myelin proteins by macrophages in inflammatory demyelinating diseases (2, 84, 108). According to flow cytometry results, ADAM8 was very highly expressed in CD45⁺ cells in TMEV-infected me/me CNS tissues (Fig. 6B), and these cells were specifically depleted by clodronate liposomes. The spinal cords of wild-type mice did not show an increase in ADAM8 expression following infection, in contrast to the eightfold increase in expression in me/me spinal cords. Macrophage depletion resulted in a significant decrease in ADAM8 expression in infected me/me spinal cords (Fig. 6B and 7E). In parallel with ADAM8 expression, the expression of another macrophage effector molecule, TNF-α, showed significant reduction as a result of macrophage depletion (Fig. 6C). Taken together, these observations indicate that clodronate-mediated depletion of macrophages may ultimately act by reducing macrophage effector molecules that cause spinal cord demyelination.

To corroborate and expand the above-described macrophage effector analysis, we examined the expression levels of two inflammatory cytokines, TNF-α (80, 96) and IL-6 (23, 78), in the spinal cords of TMEV-infected mice by ELISA. Similar to the fluorescence-activated cell sorter analysis, this assay revealed that TNF-α and IL-6 were not significantly upregulated in wild-type spinal cords. In sharp contrast, we observed a fivefold upregulation of both TNF-α and IL-6 in me/me spinal cords following TMEV infection (Fig. 7C and D). Notably, macrophage depletion with clodronate significantly decreased the expression of both TNF-α and IL-6 in infected me/me spinal cords. Similar changes in ADAM8, TNF-α, IL-6, arginase I, and LT-α after these treatments were observed when levels were quantified by real-time RT-PCR (data not shown). In agreement with previous observations, the chemokine MCP-1 (CCL2) expression in the CNS was increased following TMEV infection in wild-type mice but was substantially greater in SHP-1-deficient me/me mice (Fig. 7B). Importantly, macrophage depletion before TMEV infection resulted in a 50-fold decrease in the expression of MCP-1 in me/me spinal cords, implicating macrophages as a principle source or stimulus for MCP-1 production in the CNS of TMEV-infected me/me mice.

To corroborate the immunohistochemical analysis of demyelination (Fig. 2), we used flow cytometry to quantify the expression of MBP (Fig. 7F), which we previously showed to be substantially reduced in the spinal cords of me/me mice following TMEV inoculation (65, 66, 104). Consistent with previous observations (65), the spinal cords of TMEV-infected me/me mice showed significantly lower MBP contents than those of the infected wild-type littermates. In contrast, when macrophage-depleted me/me mice were infected with TMEV, MBP expression in the spinal cords was not significantly reduced compared to that in the spinal cords of uninfected me/me mice, suggesting that macrophages are important in mediating the reduction of MBP in the spinal cord. Taken together, these data suggest that CNS-infiltrating macrophages in me/me mice play a crucial role in promoting viral replication by serving as a viral pool, contributing to increased gene expression relevant to the mechanisms of inflammatory demyelination, and reducing MBP in the spinal cords of SHP-1-deficient animals.

SHP-1-deficient glial cells express more chemokines and SHP-1-deficient macrophages are more sensitive to chemoattractive stimuli than wild-type counterparts.

Considering the essential role that infiltrating macrophages play in mediating CNS inflammation in SHP-1-deficient mice, it became important to investigate the specific mechanisms of increased macrophage infiltration into the CNS of me/me mice following TMEV infection. It was shown previously that me/me glial cells have exaggerated NF-κB signaling (52, 67), and we have shown that the CNS of infected me/me mice had increased amounts of the NF-κB-inducible chemokine MCP-1 (Table 2; Fig. 3G and 7B). In order to further investigate whether SHP-1 controls MCP-1 secretion in CNS resident cells, primary glial cultures from wild-type and me/me mice were infected with TMEV for 48 h and the secretion of MCP-1 was quantified by ELISA (Fig. 8A). MCP-1 was highly inducible following TMEV infection, in glial cells of both wild-type and me/me mice; however, SHP-1-deficient me/me glial cells secreted three times the amount of MCP-1 (40 ng/10⁶ cells) secreted by wild-type glial cells following TMEV infection.

FIG. 8.

(A) Mixed primary glial cultures from wild-type (wt) and SHP-1-deficient (me) mice were infected with TMEV for 48 h (multiplicity of infection, 1), and the supernatants were used to quantify the levels of the chemokine MCP-1 by ELISA. (B) Results of a chemotaxis assay of wild-type and SHP-1-deficient macrophages. Supernatants from uninfected wild-type (wt-C), TMEV-infected wild-type (wt-V), uninfected me/me (me-C), and TMEV-infected me/me (me-V) primary glial cultures were used as chemoattractants for cultured splenic macrophages isolated from either wild-type or me/me mice. In addition, medium containing 40 ng/ml of MCP-1 served as a positive control to chemoattract wild-type and me/me macrophages (mφ). Cell migration is reported as the number of cells migrating divided by the total number of cells placed per well, multiplied by 100 (percent migration). (C) Results of an MCP-1 neutralization assay. Supernatants from uninfected wild-type, TMEV-infected wild-type, uninfected me/me, and TMEV-infected me/me mice were first incubated at 37°C for 1 h with either control antibody or function-blocking anti-MCP-1 antibody. Antibody (Ab)-treated supernatants were used to chemoattract cultured splenic macrophages from wild-type and SHP-1-deficient mice.

To test whether MCP-1 and perhaps other virus-induced chemokines from CNS glial cells were biologically active, supernatants from wild-type and me/me mouse glial cultures before and after TMEV infection were used to chemoattract cultured splenic macrophages in a dual-chamber assay system (Fig. 8B). Cell migration was reported as the percentage of cells migrating to the lower chamber out of the total number of cells seeded into the upper chamber. Figure 8B shows that both exogenous MCP-1 and the supernatants of TMEV-infected glial cultures exhibited significantly greater chemotactic activity toward splenic macrophages than the supernatants of uninfected cultures. Further, chemotactic activity was induced to higher levels in infected me/me mouse glial cultures than in wild-type mouse glial cultures. Supernatants from infected wild-type glial cultures induced the migration of 23% of the me/me macrophages, in contrast to supernatants from infected me/me glial cultures, which induced the migration of 40% of the me/me macrophages (Fig. 8B).

Further, we wanted to examine whether the splenic macrophages of SHP-1-deficient me/me mice were more or less responsive to a set chemokine concentration than wild-type macrophages. Indeed, medium containing exogenous MCP-1, used as a standard positive control, induced the migration of 27% of the wild-type macrophages but a significantly higher level of migration of 43% of the me/me macrophages. Second, the same supernatant from infected me/me glial cultures stimulated a significantly increased migration of me/me macrophages compared to that of wild-type macrophages. Therefore, not only do me/me glial cells show increased production of chemokines/chemoattractant stimuli, but also me/me macrophages are more sensitive to a particular concentration of a chemoattractive stimulus.

Because MCP-1 has repeatedly been shown to have a central role in chemoattracting macrophages and modulating the severity of disease in both the TMEV (7, 51, 62) and experimental autoimmune encephalomyelitis (4, 49) mouse models of MS, we wanted to assess the contribution of glia-derived MCP-1 in the chemoattraction of macrophages. Supernatants from wild-type and me/me glial cultures before and after TMEV infection were incubated at 37°C for 1 h with either control antibody or function-blocking anti-MCP-1 antibody. Then, supernatants were used to chemoattract cultured splenic macrophages from wild-type and SHP-1-deficient mice (Fig. 8C). Incubation in medium containing the MCP-1 standard with the anti-MCP-1 antibody significantly decreased the migration of wild-type and me/me macrophages, by sixfold. Furthermore, anti-MCP-1 antibody added to medium from infected me/me glial cultures effectively decreased the migration of wild-type and SHP-1-deficient macrophages. Altogether, these data indicated that MCP-1 is a predominant chemokine induced in CNS glial cells by TMEV, in agreement with the findings of previous studies, and that SHP-1 modulates the production of MCP-1 in CNS glial cells and controls the responsiveness of macrophages to MCP-1-induced chemotaxis.

DISCUSSION

This study aimed to characterize the immunological and molecular mechanisms governing the increased susceptibility of SHP-1-deficient mice to virus-induced CNS inflammatory demyelinating disease (65). Herein, we showed that the spinal cords of SHP-1-deficient mice displayed an augmented inflammatory profile and increased viral loads following TMEV infection compared to those of their wild-type littermates. Importantly, the CNS tissues of infected me/me mice contained significantly higher numbers of infiltrating macrophages than those of wild-type mice, and the number of infiltrating macrophages directly correlated with disease severity. Moreover, infiltrating macrophages localized within the white matter and were concentrated within areas of focal demyelination in infected SHP-1-deficient mice. In turn, macrophage depletion resulted in a significant delay in the onset and a reduction in the severity of TMEV-induced paralytic disease seen in SHP-1-deficient mice. Furthermore, the reduction of CNS macrophage infiltration resulted in an attenuated inflammatory profile and reductions in the viral burden and demyelination. Taken together, the findings of this study point to the essential role of SHP-1 in controlling virus infection, macrophage infiltration, and macrophage effector functions in virus-induced CNS inflammatory demyelination.

As reported previously, SHP-1-deficient mice do not display the typical biphasic TMEV-induced disease characterized as an acute gray-matter disease followed by chronic white-matter demyelination (56, 77). In contrast, the infection of SHP-1-deficient mice results in an acute white-matter disease in which inflammatory demyelination occurs very early (within 4 days p.i.). Moreover, in contrast to TMEV-induced demyelinating disease in wild-type mice, in which agents of both acquired and innate immunity, including lymphocytes and macrophages, play an important role, TMEV-induced demyelination in SHP-1-deficient mice involves primarily innate immune effectors that we identified presently as blood-borne macrophages. The overwhelming presence of macrophages within the demyelinated areas of the white matter in the spinal cords of infected me/me mice underscores the importance of SHP-1 in controlling macrophage trafficking into the CNS and controlling the effector functions of macrophages in mediating demyelination. Thus, the results of the present study are relevant to CNS inflammatory diseases in which either adaptive or innate immune responses play an important role but in which macrophages constitute the main effectors of tissue destruction. Indeed, infiltrating macrophages have been demonstrated previously to be essential effectors of demyelination in both the TMEV (7, 25, 86, 88) and the experimental autoimmune encephalomyelitis (26, 48, 49, 101) mouse models of MS. Also, while the role of T cells in disease pathogenesis may or may not be a predominant feature of demyelinating lesions in both humans with MS and animal models, the importance of macrophages as essential effectors of demyelination and paralysis is now well-established. Previous TMEV studies have shown that the extent of CNS macrophage infiltration correlates with disease severity (58, 88). The present study has characterized a relatively simple and rapid model for analyzing macrophage effector functions in virus-induced inflammatory demyelination in the CNS.

TMEV antigen (28), virions (11), and viral RNA (3) have been detected previously in macrophages from demyelinating lesions in mice, indicating that macrophages further promote disease progression by providing an essential reservoir of antigenic and innate inflammatory stimulation in the CNS (25, 59, 60, 64, 74). The findings of this study demonstrated that the spinal cords of SHP-1-deficient mice had higher viral antigen and RNA levels following TMEV infection than wild-type mice, which correlates with a substantially higher number of infiltrating macrophages. Further, macrophage depletion resulted in significantly decreased viral loads in the spinal cord compared to those in nondepleted mice, indicating that the increased number of infiltrating macrophages was important in contributing to the viral levels in SHP-1-deficient mice. In agreement with this observation, macrophages isolated from the brains of TMEV-infected me/me mice contained significantly higher quantities of TMEV transcripts per cell than wild-type macrophages. We have shown previously that SHP-1 deficiency in CNS glial cells results in higher levels of viral replication both in vivo and in vitro (13, 65), correlating with increased arginase I and decreased nitric oxide (NO) production. Similarly, cultured splenic macrophages of SHP-1-deficient mice display increased viral antigen and RNA levels compared to wild-type macrophages after infection in vitro with TMEV (our unpublished observations). Taken together, these findings show that SHP-1 controls both the number of macrophages entering the CNS and viral replication within macrophages, which in turn both contribute to the increased viral burdens seen in the spinal cords of me/me mice.

As macrophages are essential effectors of inflammatory demyelinating disease, it was important to characterize the inflammatory profiles of infiltrating macrophages in wild-type and SHP-1-deficient mice. We found that several inflammatory genes, including those for TNF-α, IL-6, MCP-1, arginase I, ADAM8, eotaxin, and LT-α, were significantly upregulated in the CNS of infected me/me mice. Importantly, reducing the number of macrophages entering the CNS by depleting peripheral macrophages resulted in a significant decrease in the expression of several inflammatory genes, suggesting that the increased number of macrophages in the CNS of infected me/me mice significantly contributed to the inflammatory profile. Nevertheless, although infiltrating macrophages are major sources of inflammatory molecules in the CNS, infected resident glial cells also contribute to the increased levels of these inflammatory molecules in TMEV-infected CNS tissues (Fig. 3). These data are in accordance with the results of previous studies showing that SHP-1 controls the activation of several transcription factors, including STATs and NF-κB, and corresponding proinflammatory genes in resident CNS glial cells and macrophages in response to either TMEV infection or microbial materials that act via Toll-like receptors (13, 29, 52, 67). Therefore, both the increased number of CNS-infiltrating macrophages and the increased inflammatory profile are likely to contribute to the severe TMEV-induced inflammatory demyelination seen in SHP-1-deficient mice.

Although macrophage depletion was able to delay paralysis in infected me/me mice, it was not sufficient to completely abolish disease. There was a substantial amount of macrophage infiltration into the spinal cords of clodronate-treated me/me mice at 8 days p.i. (data not shown). Therefore, the increased levels of macrophages in infected clodronate-treated animals at later time points (8 days p.i.) could certainly account for the clinical paralysis seen in those mice. That said, we cannot rule out a possible role for parallel mechanisms independent of macrophage activation that may additionally affect increased demyelination and clinical signs at later time points after infection. The eventual progression of disease in macrophage-depleted me/me mice can also be explained by multiple roles for SHP-1 in the CNS that are independent of macrophage functions, including the control of virus replication in CNS glial cells (13, 65) and the regulation of oligodendrocyte pathology (66).

We considered the specific mechanisms of increased macrophage recruitment in the CNS of me/me mice following TMEV infection. It was shown previously that the CNS of me/me mice maintains an intact blood-brain barrier and does not contain increased numbers of inflammatory cells (105). However, studies have shown that in response to inflammatory stimuli, cells of me/me mice secrete increased levels of chemokines compared to cells of wild-type animals, which may promote increased monocyte infiltration and inflammation under certain conditions such as microbial invasion (36, 107). In accord with the results of these studies, we found increased expression of the chemokines RANTES, eotaxin-1, MCP-1, and MIP-1α in the CNS tissues of TMEV-infected me/me mice compared to that in their wild-type littermates. Furthermore, mixed glial cultures lacking SHP-1 showed increased chemokine secretion compared to wild-type cultures following TMEV infection, which may promote enhanced macrophage chemotaxis. In addition, we showed that SHP-1 controlled the expression of MCP-1, an important mediator of CNS inflammatory disease (4, 7, 8, 31, 36, 49), and that blocking MCP-1 in vitro significantly diminished macrophage migration in response to mixed glial supernatants. Importantly, we demonstrated that SHP-1-deficient cultured macrophages responded to the chemokine MCP-1 or to supernatants from TMEV-infected glial cultures with enhanced migration compared to that of wild-type macrophages. The increased sensitivity of me/me macrophages may be due in part to the increased expression of chemokine receptors, including CCR2, on SHP-1-deficient macrophages (51, 80) or the enhanced signaling potential of individual chemokine receptors.

In MS, intense macrophage infiltration into active demyelinating lesions is present and both the numbers and the differentiation/activation of macrophages correlate with disease severity (16, 17, 61, 102, 104). Macrophages can mediate myelin degradation, oligodendrogliopathy, and axonal loss through both cell-mediated processes and the secretion of inflammatory mediators (44, 76, 104). A key area of interest is the stimulus responsible for macrophage trafficking/infiltration into CNS demyelinating lesions. For instance, the chemokine MCP-1 is elevated in MS plaques and in the cerebrospinal fluids of MS patients (62, 73, 97). The idea of an essential role for MCP-1 in the infiltration of macrophages in CNS demyelinating diseases is supported by observations made with various animal models for MS. These findings point out that CNS-derived chemokines and peripherally derived macrophages have a prominent role in the initiation and progression of demyelination in MS.

We have recently shown that peripheral blood mononuclear cells of MS patients have a deficiency in the expression of SHP-1 that leads to an augmented inflammatory profile compared to that of peripheral blood mononuclear cells from normal subjects (21). Since numerous studies have shown previously that several inflammatory mediators/effectors are upregulated in MS lesions and it has been shown that SHP-1 is a master modulator of inflammatory signaling, the SHP-1 deficiency in the leukocytes of MS patients may significantly contribute to MS disease processes. In relation to the present study, preliminary data indicate that macrophages of MS patients exhibit significantly lower levels of SHP-1 expression and function than those of normal subjects (our unpublished observations). Therefore, further studies on the extent and level of control that SHP-1 exerts on macrophage differentiation, activation, and chemotaxis will be important. The results of this study demonstrate that the protein tyrosine phosphatase SHP-1 controls macrophage migration/trafficking into the CNS and the expression of several inflammatory mediators within the CNS, making SHP-1 an attractive target in the treatment of MS.