Matrix metalloproteinase-7 facilitates immune access to the CNS in experimental autoimmune encephalomyelitis
© Buhler et al. 2009
Received: 04 September 2008
Accepted: 06 March 2009
Published: 06 March 2009
Skip to main content
© Buhler et al. 2009
Received: 04 September 2008
Accepted: 06 March 2009
Published: 06 March 2009
Metalloproteinase inhibitors can protect mice against experimental autoimmune encephalomyelitis (EAE), an animal model for multiple sclerosis (MS). Matrix metalloproteinase-9 (MMP-9) has been implicated, but it is not clear if other MMPs are also involved, including matrilysin/MMP-7 – an enzyme capable of cleaving proteins that are essential for blood brain barrier integrity and immune suppression.
Here we report that MMP-7-deficient (mmp7 -/-) mice on the C57Bl/6 background are resistant to EAE induced by myelin oligodendrocyte glycoprotein (MOG). Brain sections from MOG-primed mmp7 -/-mice did not show signs of immune cell infiltration of the CNS, but MOG-primed wild-type mice showed extensive vascular cuffing and mononuclear cell infiltration 15 days after vaccination. At the peak of EAE wild-type mice had MMP-7 immuno-reactive cells in vascular cuffs that also expressed the macrophage markers Iba-1 and Gr-1, as well as tomato lectin. MOG-specific proliferation of splenocytes, lymphocytes, CD4+ and CD8+ cells were reduced in cells isolated from MOG-primed mmp7 -/- mice, compared with MOG-primed wild-type mice. However, the adoptive transfer of splenocytes and lymphocytes from MOG-primed mmp7 -/- mice induced EAE in naïve wild-type recipients, but not naïve mmp7 -/- recipients. Finally, we found that recombinant MMP-7 increased permeability between endothelial cells in an in vitro blood-brain barrier model.
Our findings suggest that MMP-7 may facilitate immune cell access or re-stimulation in perivascular areas, which are critical events in EAE and multiple sclerosis, and provide a new therapeutic target to treat this disorder.
Multiple sclerosis (MS) is an autoimmune disorder marked by the infiltration of pathogenic T cells into the central nervous system (CNS) that cause inflammation and oligodendrocyte cell death. In an animal model of MS, called experimental autoimmune encephalomyelitis (EAE), vaccination with CNS-myelin-derived peptides triggers the expansion of oligodendrocyte-specific T cells and a pathological profile that includes CNS inflammation, demyelination, and paralysis. Transmigration of pathogenic T cells across the blood-brain barrier (BBB) is facilitated by the expression of cell adhesion molecules and proteinases that degrade the ECM . The discovery that EAE can be prevented by broad spectrum metalloproteinase inhibitors implicated this large family of enzymes in disease progression [1–5] and has led to recent clinical trials . Matrix metalloproteinases (MMPs) are extracellular enzymes that can cleave ECM and non-matrix proteins, including laminin, collagen, cytokines, other proteinases, and the ectodomains of several membrane proteins. MMPs are usually secreted as pro-enzymes that can be cleavage-activated by plasminogen activators, trypsin, other MMPs, and oxidation. Elevated levels of MMP-2, MMP-7 and MMP-9 have been reported in human MS patients, and in brain and spinal cord extracts from EAE-induced rodents [7–17]. In a delayed-type hypersensitivity model for MS, MMP-7 was found to be the most up-regulated MMP, compared with MMP-2,3,8,9,10,11,12,13,14,15 and 16 .
Within tissues, MMPs usually reside in extracellular spaces as inactive proforms, and factors that activate even a small proportion of those MMPs have significant biological effects. Therefore, determining which factors contribute to MMP activity in MS will be critical to understanding the role(s) these enzymes play in this disorder. Cerebrospinal fluid levels of MMP-9 activity are elevated in MS patients and in rodent models of EAE , and young MMP-9 knockout mice (4 weeks) are resistant to EAE . MMP-2 plays a critical role in angiogenesis and vascular remodeling . Although MMP-2 expression does not increase in MS or EAE, MMP-2 activation may contribute to localized permeabilization of the cerebrovasculature. MMP-2 and MMP-9 are structurally similar gelatinases that can each be activated by MMP-7 . MMP-7 can also cleave many EAE-relevant substrates, including laminin, type IV collagen , β4-integrin , VE-cadherin , E-cadherin [25–27] and the immune suppressor Fas ligand (FasL) . Further, MMP-7 has been reported as necessary for the trans-epithelial efflux of immune cells in bleomycin-treated lungs , which is similar to the extravasation that immune cells must make in EAE and MS.
Myelin-specific T cells can be detected in the blood of MS patients and EAE-induced mice even during periods of remission, when they no longer persist in the CNS. Tight junctions between microvascular endothelial cells within the brain prevent the direct entry of macromolecules and blood-borne cells, forming the BBB. Compromise of BBB integrity facilitates immune cell access to the CNS and is essential for MS and EAE. For example, MRI detection of gadolinium accumulation in the brain lesions of MS patients is an indicator of compromised BBB integrity and a reliable predictor of pending disease activity. Factors that affect the cell-to-cell contacts of cerebrovascular endothelial cells, or their viability, can reduce BBB integrity and increase immune cell access to the CNS. VE-cadherin is an important component of tight junctions between endothelial cells and is also a substrate for cleavage by both MMP-7  and MMP-9 . The two layers or ECM that surround the cerebrovasculature contain laminin and type IV collagen, which are cleaved by MMP-7 , as well as collagens and elastins are cleaved by MMP-9 .
In addition to the BBB, immune cells actions are restricted within the CNS action by the expression of cell death ligands CD95L/Fas ligand/FasL and TRAIL that can trigger apoptosis in activated T cells and myeloid cells. FasL is a potent inducer of apoptosis in activated T cells  and its expression is elevated in response to injuries that compromise BBB integrity [15, 33–35]. Facial nerve damage locally increases FasL expression and T cell apoptosis around the facial nucleus in EAE-induced mice . The availability of FasL is closely regulated at the transcriptional and post-translational levels, which includes proteolytic cleavage of its receptor binding region by MMP-3  and MMP-7 [28, 38]. While membrane-bound FasL has potent pro-apoptotic activity for activated T cells, cleaved or soluble FasL (sFasL) shows differential effects, depending on the disease model and specific cell types [39, 40]. Moreover, sFasL has a chemotaxic effect on mononuclear cells, which can be blocked by anti-FasL antibodies .
Here, we used MMP-7-deficient (mmp7 -/-) mice to investigate the role of this proteinase in EAE. Wild-type (wt) and mmp7 -/- mice were compared for clinical and immunological responses to an encephalitogenic fragment of myelin oligodendrocyte protein peptide (MOG35–55). Localization of MMP-7 expression in the CNS was examined during EAE in MOG-primed mice at the peak of EAE. MOG-specific proliferation of splenocytes, lymphocytes, CD4+, and CD8+ T cells were compared in both strains of mice. We also used adoptive transfer to assess the encephalitogenic potential of T cells from wt and mmp7 -/- mice to induce EAE in recipients of both strains. Finally, we tested the effects of active MMP-7 on endothelial cell connectivity using an established in vitro model of the BBB.
Haematoxylin and eosin staining of brain sections from wt mice isolated 15 days after MOG injection, near the peak of EAE, showed extensive mononuclear cell infiltration (Figure 1b, d). Perivascular cuffing, a classic feature of encephalitis, was not seen in the brains of MOG-primed mmp7 -/- mice (Figure 1c, e). wt mice also showed patches of demyelination adjacent to perivascular cuffs that were infiltrated by mononuclear cells and contained pycnotic nuclei (Figure 1f), all of which were absent in mmp7 -/- mice (Figure 1g). These findings indicate that mmp7 -/- mice are resistant to a stimulus that induces EAE in wt mice.
The choroid plexus consists of highly vascularized structures that produce cerebrospinal fluid (CSF) within brain ventricles. Although choroid plexus has highly fenestrated capillaries, blood components are kept separate from cerebrospinal fluid by tight junctions between specialized ependymal cells in the outer layer, constituting the blood-CSF barrier (BCB). Between fenestrated capillaries and ependymal cell layers is a basal lamina composed largely of collagen in which we observed strong immunoreactivity for MMP-7. Notably, CD4+ T cells were observed between these layers but did not show MMP-7 immunoreactivity (Figure 3e–g). Accumulation of MMP-7 within the choroid plexus may degrade components of the basal lamina and connective tissue, or disrupt cadherin-mediated contacts. Indeed, confocal microscopy revealed patches of MMP-7 that appeared to breach the ependymal layer of the choroid plexus and become continuous with the CSF-containing lateral ventricle (Figure 3c). Significantly, such breaches in the BCB would permit the passage of immunoglobulins and serum proteins into the CSF, an important diagnostic indicator of MS.
To determine whether a lack of MMP-7 prevents MOG-specific immune responses, we assayed representative cell populations from wt and mmp7 -/- mice, in vitro. Splenocytes and lymphocytes were isolated from mice 15 days after MOG-injection and then cultured with MOG and 3H-thymidine. After 4 days, splenocytes from both wt and mmp7 -/- mice showed higher 3H-thymidine incorporation in response to MOG (10, 20, or 30 μg/ml) when compared with controls (see Additional file 3). However, MOG-induced proliferation of mmp7 -/- splenocytes was less than wt cells. Primary lymphocytes from MOG-primed mmp7 -/- mice also showed less proliferation in response to MOG in vitro, compared with wt. Higher density lymphocyte cultures showed similar results (not shown). These findings demonstrated that splenocytes and lymphocytes from mmp7 -/- mice have lower antigen-specific proliferation in response to MOG, in vitro.
High MMP-7 activity has been reported in demyelinating MS lesions and in the cerebrospinal fluid of MS patients [7–10, 15], yet the role of this extracellular proteinase in MS is still unclear. Here, we have shown that mmp7 -/- mice are resistant to MOG-induced EAE. Diminished T cell responses to MOG make it less likely that EAE will develop, but mmp7 -/- mice still produce encephalitogenic T cells that can cause disease in wt mice. Immuno-localization of MMP-7 in areas adjacent to vascular structures suggests that it may facilitate immune access at the BBB. A common diagnostic feature of MS is MRI-detectable breaches of the BBB and spokes of demyelination that project outwardly from a cerebrovascular core (Dawson's fingers). Localization of MMP-7 to breaches of the BBB suggests that it may act directly on BBB integrity. We tested this possibility using an in vitro BBB model and found that the direct exposure of brain endothelial cells to recombinant MMP-7 increased the permeability between those cells. In contrast, MMP-7 did not increase permeability on the brain side of the BBB, but it may have other effects.
Inside the BBB, MMP-7 may contribute to EAE by activating MMP-2 and MMP-9 (proform), or by reducing immune privilege by cleaving FasL . As a potent inducer of apoptosis in activated T cells, FasL is crucial to the recovery phase of EAE. Mice with defective FasL undergo more severe and prolonged EAE, compared with WT mice . Indeed, CNS injury models have shown that a localized increase of FasL expression within the CNS is a consistent response to breaches of the BBB [33, 34]. The potent activity of FasL can be attenuated by shedding the Fas-binding domain by MMP-7 [28, 38, 39]. Macrophages that infiltrate the CNS and produce high local concentrations of MMP-7 may cause localized shedding of FasL and create discrete pockets of lowered immune privilege within the CNS. This process could prolong the survival of activated T cells that would otherwise die by activation induced cell death or FasL-induced apoptosis. Interestingly, granulocyte macrophage colony stimulating factor (GM-CSF) knockout mice are also resistant to MOG-induced EAE ; however, their resistance is attributed to attenuated MOG-specific T cell responses, perhaps due to fewer antigen-presenting cells. Variables that affect macrophage availability might be expected to have a direct impact on MMP-7 production near cerebrovascular structures and also affect EAE susceptibility.
Within the brain MMP-7 may also contribute to memory, motor and cognitive problems that often accompany MS episodes. Recent studies have shown that MMP-7 and MMP-9 can disrupt mature dendritic spines, causing them to assume immature morphologies, which greatly reduces the synaptic strength of excitatory synapses and can impact memory and behaviour [53, 54]. Furthermore, EphB receptors are critically important for the formation and maintenance of mature dendritic spines, and MMP-9 has recently been shown to cleave EphB . Infiltrating macrophages in perivascular cuffs or near demyelinating lesions might produce enough MMP-7 to activate extracellular pools of MMP-9 (proform) and disrupt synapse stability in nearby areas – even without demyelination or overt pathology .
Several cytokines, including TNF-α, have been shown to increase MMP-7 expression and have been repeatedly linked with EAE and MS inflammation [57, 58]. IFN-β is a promising treatment for MS and has been shown to repress MMP-7 and MMP-9 expression for patients with relapsing-remitting MS [18, 51]. Our findings suggest that MMP-7 plays a role in the extravasation of immune cells during EAE. MMP-7 expression adjacent to the cerebrovasculature and ependymal cells of the choroid plexus suggests that metalloproteinase inhibitors could prevent EAE by blocking MMP-7 activity at these sites. Interestingly, hydroxamate-based metalloproteinase inhibitors can prevent EAE, but may not be able to cross an intact BBB, so their ability to prevent EAE would depend on actions after the BBB has been compromised unless they work by inhibiting MMPs outside of the BBB, such as MMP-7 on circulating monocytes .
This study demonstrates that MMP-7 plays a critical role in MOG-induced EAE in C57Bl/6 mice, although it is possible that stronger encephalitogenic stimuli (e.g. myelin basic protein) in other mouse strains may respond differently to MMP-7 deficiency. The reduced response of CD4+ or CD8+ T cells to EAE, did not preclude the production of encephalitogenic T cells, and may reflect an important role for MMP-7 in the antigen-presentation or the specialized microenvironment of lymphoid organs. Immuno-localization of MMP-7 to myeloid cells and cerebrovascular structures, along with its ability to increase the permeability between endothelial cells, implicates this enzyme in compromising BBB and/or BCB integrity, and suggests that MMP-7 may have additional effects in the brain parenchyma that extend beyond lymphocyte infiltration.
The generation and characterization of mmp7 -/- mice on the C57BL/6 background has been reported . Mice that had been back-crossed for at least 10 generations, onto the C57Bl/6 background, were bred in the University of California, Riverside (UCR) vivarium in accordance with Institutional and NIH animal care and use guidelines. Age-matched mmp7 -/- and wt (B6) mice were injected at 7–18 weeks of age with 100 μl of CFA emulsion containing 250 μg of recombinant MOG35–55, divided into three sites on shaved backs. Recombinant MOG was synthesized at the University of California Los Angeles (UCLA) peptide core and was 97–99% pure. One and 3 days after MOG injections, mice were given intra-peritoneal injections of 200 ng pertussis toxin each. Each day, mice were scored for clinical signs of EAE as follows: 0 = no EAE; 1 = total loss of tail tonicity; 2 = hind limb weakness, impaired righting reflex, or forelimb impairment alone; 3 = total paralysis of one or both hind limbs; 4 = hind and forelimb paralysis; 5 = moribund, death . Statistical analysis was done by paired Student-t Tests (2-tailed) of wt and mmp7 -/- clinical scores at each time-point.
Spleen and draining lymph nodes were isolated from experimental and control mice 15–24 days after post MOG injection. Tissues were disrupted in complete RPMI with 5–7 strokes of a glass Dounce (15 mL Pyrex). Single-cell suspensions were prepared by passing homogenates over a 70-μm sieve filter (BD, Franklin Lakes, NJ) and rinsing with 2 mL of complete RPMI. Cells were pelleted and red blood cells lysed by resuspension in cold buffer (Red blood cell lysis buffer, Sigma, St. Louis, MO) for 5 min on ice. Next, cells were pelleted and resuspended in RPMI 1640 media containing 10% FCS (Hyclone, Logan, UT), L-glutamine, sodium bicarbonate, sodium pyruvate, essential and non-essential amino acids, vitamins, penicillin/streptomycin, and β-mercaptoethanol (complete RPMI), and maintained in a humidified CO2 incubator at 37°C. Viable cells were counted by trypan blue exclusion.
Cells were isolated from 3 wt and 3 mmp7 -/- mice 15 days after vaccination with MOG, as described. Quadruplicate wells of each condition consisted of 2 or 4 × 105 cells/well in a 96-well plate, and cultured in the presence of 0, 10, 20 or 30 μg/mL MOG35–55 overnight at 37°C. Cells were then pulsed with 1 μCi/well of 3H-thymidine (ICN Pharmaceuticals, Costa Mesa, CA) for the last 4 h of culture, and processed using a cell harvester (PHD, Cambridge Technologies). Each sample was cut from the filter and radioactivity measured with a scintillation counter. Mean values from quadruplicates were used to generate proliferation responses that were normalized to the control condition (0 μg/mL MOG) for cells from each animal. These normalized proliferation responses were averaged for all mice of each genotype.
Freshly isolated splenocytes and lymphocytes were pooled, pelleted, and then resuspended in PBS at 6 × 107 cells/ml. An equal volume of PBS containing 10 μM carboxy-fluorescein diacetate succinimidyl ester (CFSE; Molecular Probes, Inc. Eugene, OR) was added, then cells were gently mixed and incubated for 20 min at 37°C. Cells were pelleted and washed twice with PBS. Immediate FACS analysis of an aliquot determined cell labelling efficiency, typically >99%, and those values were used as baseline labelling of the parental generation. CFSE-labelled cells were then resuspended (2.5 × 106 cells/ml) in complete RPMI with 0, 10, or 15 μM MOG, or 25 units/ml IL-2 and incubated at 37°C for 4 days. Aliquots of each preparation were then labelled with fluorescently-conjugated (PE) anti-CD4, anti-CD8, or isotype-specific controls (BD Biosciences, San Diego, CA) at 1 μg/106 cells, 25 min at 4°C. Cells were then rinsed with PBS (3×) and fixed with an equal volume of 2% paraformaldehyde. Data was acquired using a FACScan® (Becton Dickinson, San Jose, CA), and analyzed using ModFit™ software proliferation wizard (Becton Dickinson, San Jose, CA). Typically, 15–50,000 events were collected for each sample.
Primary splenocytes and lymphocytes were isolated from MOG-primed mmp7 -/- mice 10 days after vaccination and cultured with 10 μg/ml MOG and 10 ng/ml IL-2 (2.5 × 106 cells/ml) for 5 days. Recipient mice received 20 × 106 cells in 0.2 mL of PBS by tail-vein injection.
Immediately following the removal of spleen and lymph nodes each mouse was perfused (cardiac) with saline containing heparin, saline alone, and then fresh 4% paraformaldehyde in PBS. The isolated cranium and spinal column were post-fixed overnight in paraformaldehyde, after which the brain and spinal cord were carefully removed. Tissue was cryo-protected with increasing sucrose concentrations and then sectioned on a cryostat (7 μm), and stored at 4°C until use. Slide-mounted sections were warmed to 37°C for 10 min, rinsed with PBS, and non-specific antigens blocked with either 10% BSA or 10% normal goat serum (NGS) for 1 h, depending on the sources and specificities of the antibodies to be used. Antibodies used were: mouse anti-MMP-7 (1:100; Dr. Carole Wilson), rabbit anti-MMP-7 (1:100; Dr. Carole Wilson), mouse anti-CD4-PE (1:100; BD Biosciences), mouse anti-GFAP-PE (1:2000; Sigma), mouse anti-Gr-1 (BD Biosciences), mouse Iba-1-FITC (1:100; Genetex, #1022-5). Secondary antibodies used were FITC-conjugated anti-rabbit IgG (1:100; Sigma) and Cy3-conjugated anti-rat IgG (1:100; Molecular Probes); Alexa-488 conjugated anti-mouse (1:1000; Invitrogen).
The system used has been described in detail previously [61, 62]. Briefly, primary rat-brain micro-vascular endothelial cells were cultured on the lumen surface of porous tubes, and grown with constant perfusion of media through the lumen of the tubes. Outside the tubes were grown primary rat astrocytes, which supplied signalling molecules necessary to create tight junctions between endothelial cells sufficient to be recorded as resistance. Experiments were done at week 3 of EC-astrocyte co-culture in this dynamic in vitro BBB model when all tubes were measured for resistance and only those showing significant resistance were used. Active recombinant human MMP-7 (Millipore, Temecula, CA) was added to the extra capillary space (astrocyte) side of the system at 1 or 5 units per mL (1st group of experiments), or to the lumenal (endothelial) side at 1 unit per mL (2nd group of experiments). In both experimental series MMP-7 was added with continuous perfusion in one set of experiments, or with a 1.5 h pause of flow followed by reperfusion in separate experiments. Permeability of the endothelial sheets was measured by standard electrophysiological recording as the trans-endothelial electrical resistance (TEER) between the ECS and lumen.
carboxy-fluorescein diacetate succinimidyl ester
central nervous system
experimental autoimmune encephalomyelitis
glial fibrillary acidic protein
granulocyte macrophage colony stimulating factor
myelin oligodendrocyte glycoprotein
normal goat serum
soluble Fas ligand
trans-endothelial electrical resistance
University of California Riverside
We thank Dan Welch and Xiaoqiao Xie for technical assistance, Dr. Laurie Owen for discussions and Dr. William Parks for critical comments. This work was supported by grants from the National Multiple Sclerosis Society (PP0842) and the National Institutes of Health (AG21652) to DE.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.