Tissue inhibitors of matrix metalloproteinases are elevated in cerebrospinal fluid of neurodegenerative diseases

Journal of the Neurological Sciences 207 (2003) 71 – 76 www.elsevier.com/locate/jns

Tissue inhibitors of matrix metalloproteinases are elevated in cerebrospinal fluid of neurodegenerative diseases S. Lorenzl a, D.S. Albers a, P.A. LeWitt b, J.W. Chirichigno a, S.L. Hilgenberg c, M.E. Cudkowicz c, M.F. Beal a,* a

Department of Neurology and Neuroscience, Weill Medical College of Cornell University, 525 East 68th Street Room A-501, New York, NY 10021, USA b Department of Neurology, Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, Detroit, MI 48201, USA c Neurology Clinical Trial Unit and Department of Neurology, Massachusetts General Hospital, Boston, MA 02114, USA Received 29 July 2002; received in revised form 10 October 2002; accepted 11 October 2002

Abstract Matrix metalloproteinases (MMPs) are implicated in the pathogenesis of diseases such as Alzheimer’s Disease (AD) and amyotrophic lateral sclerosis (ALS). Increased expression of MMP-9 and TIMPs has been reported in postmortem AD and ALS brain tissue, as well as in ALS cerebrospinal fluid (CSF) and plasma. Although individual studies of MMP and TIMP expression in CSF have included AD and ALS samples, there are no studies comparing the expression of these proteins between neurodegenerative diseases. We measured the levels of matrix metalloproteinases (MMPs)-2 and -9 and the tissue inhibitor of MMPs (e.g. TIMP-1 and TIMP-2) in CSF samples from patients with Parkinson’s Disease (PD), Huntington’s Disease (HD), AD and ALS as compared to age-matched control patients. There was constitutive expression of the proform of gelatinase A (proMMP-2) on zymography gels in all CSF samples. Unexpectedly, there was an additional gelatinolytic band at 130 kDa of unknown etiology in the CSF samples of patients with PD (61% of patients studied), AD (61%), HD (25%) and ALS (39%). Levels of TIMP-1 were significantly elevated in CSF samples from all disease groups. TIMP-2 was significantly increased in CSF of AD and HD patients. MMP-2 levels did not differ significantly between groups. These findings show that TIMPs are elevated in the CSF of patients with neurodegenerative diseases suggesting a potential role of these endogenous inhibitors of matrix metalloproteinases in neurodegenerative diseases. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Neurodegeneration; MMP-2; MMP-9; TIMP-1; TIMP-2; Cerebrospinal fluid

1. Introduction A variety of biochemical, environmental and geneticbased mechanisms of neuronal degeneration have been postulated to be involved in the pathogenesis of neurodegenerative diseases [1]. Among these mediators are cytokines and oxygen-free radicals, both of which can activate astrocytes and microglia to release extracellular matrixdegrading proteases. One family of proteases, namely matrix metalloproteinases (MMPs), can remodel the extracellular matrix and additionally digest pathologically accumulated extracellular proteins like amyloid [2].

* Corresponding author. Tel.: +1-212-746-4565; fax: +1-212-7468276. E-mail address: [email protected] (M.F. Beal).

Matrix metalloproteinases are a family of Zn2 +-containing and Ca2 +-requiring endoproteases. In the central nervous system, they can be released from astrocytes, neurons, oligodendrocytes, microglia, endothelial cells and leukocytes [3,4]. Their target compounds include collagen, gelatin, fibronectin, laminin, elastin and proteoglycans [5]. MMP activity is controlled at many levels including the transcriptional level, where its expression is regulated by growth factors, cytokines and free radicals [6], the activation of its latent form, and inhibition by the endogenous tissue inhibitors of metalloproteinases (TIMPs). The TIMP family consists of four related proteins capable of binding MMPs to form tight noncovalent complexes [7]. Levels of MMPs and TIMPs are increased in neurological disorders such as bacterial and viral meningitis [8,9], stroke [10] and multiple sclerosis [11], and this increased expression may contribute to the resulting tissue damage.

0022-510X/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 5 1 0 X ( 0 2 ) 0 0 3 9 8 - 2

72

S. Lorenzl et al. / Journal of the Neurological Sciences 207 (2003) 71–76

Aberrant expression of several MMPs and TIMPs occurs in Alzheimer’s Disease (AD) [12,13] and amyotrophic lateral sclerosis (ALS) [14]. MMP-9 levels are increased in tissue, plasma and cerebrospinal fluid (CSF) of ALS patients [14 – 16]. In other neurodegenerative diseases, such as Parkinson’s Disease (PD) and Huntington’s Disease (HD), expression of MMPs and TIMPs has not previously been investigated in CSF. The present study examined levels of MMP-2, MMP-9, TIMP-1 and TIMP-2 in CSF samples from PD, HD, ALS and AD patients as compared to agematched controls.

2. Materials and methods Samples of CSF from controls (n = 41, age 50 F 13 years), PD (n = 13, age 54 F 14 years), AD (n = 31, age 68 F 11 years), ALS (n = 18, age 52 F 17 years) and HD (n = 20, age 50 F 18 years) patients were provided from the Massachusetts General Hospital. Disease diagnosis was based on established clinical criteria: PD, clinical stages 1 and 2 according to the Hoehn and Yahr criteria [17], not taking L-dopa treatment at the time of lumbar puncture; AD, NINDS/AIREN diagnostic criteria; ALS, El Escorial criteria [18]. HD, patients with a positive family history and clinical diagnosis of HD. Control subjects were free of neurological illness. All patients and controls exhibited normal CSF parameters, i.e. absence of inflammatory signs (pleocytosis, intrathecal production of immunoglobulins) and normal protein concentrations. Following informed consent, 1 – 2 ml of CSF was obtained by lumbar puncture, aliquots were centrifuged and immediately frozen at 80 jC. Samples were stored for less than 6 months before being studied. All experimental procedures were performed by individuals blinded to sample identity and diagnosis. 2.1. Gelatine zymography Gelatinolytic activity was measured in all CSF samples by zymography as described previously [19]. Briefly, CSF samples (14 Al) were diluted with 4 Al of Laemmli buffer (BioRad, Hercules, CA) and loaded on 7.5% polyacrylamide gels containing 2 mg/ml gelatin type A (Sigma G 2500). Proteins were run under nonreducing conditions at 4 jC for 1 h 40 min. After electrophoresis, gels were washed twice for 30 min in 2% Triton X-100 and incubated for 20 h in incubation buffer (50 mM Tris base, 5 mM CaCl2, 1 AM ZnCl2, 0.01% sodium azide, pH 7.5 at 37 jC. After incubation, gels were fixed in 20% trichloroacetic acid (Sigma) for 30 min and stained at room temperature in 0.5% Coomassie brilliant blue (Sigma B-7920) dissolved in 35% methanol and 10% acetic acid for 90 min. After destaining in 35% methanol and 10% acetic acid solution, bands of gelatinolytic activity were visible as clear bands on a blue background. High molecular weight standard protein markers (BioRad), and recombinant human MMP-2 and MMP-9 (Oncogene Scien-

ces, MA) were used to identify the gelatinolytic bands and their approximate weights on the gels (Fig. 2). 2.2. Western blot analysis Western blot analysis was used to further confirm the identity of bands from zymography. In brief, samples of CSF (14 Al) were diluted with 4 Al of Laemmli buffer before being subjected to SDS-PAGE on 7.5% polyacrylamide gels. After electrophoresis, proteins were transferred to polyvinylidene difluoride membranes. After blocking membranes for 1 h in 5% dry milk in PBS (pH 7.4), membranes were incubated overnight at 4 jC with primary antibodies (polyclonal antibody directed against MMP-2 (Ab-7), monoclonal antibody directed against MMP2 (Ab-3), monoclonal antibody directed against MMP-9 (Ab-2), and monoclonal antibody directed against MMP-9 (Ab-7) (Oncogene Sciences). All primary antisera were diluted 1:200 in 5% dry milk in PBS (pH 7.4). After washing, the membranes were further incubated with peroxidase-conjugated secondary antibody (Jackson Laboratories) for 1 h at room temperature. Immunoreactive bands were visualized using an enhanced chemiluminescent system (SuperSignal, Pierce, Rockford, IL). 2.3. ProMMP-2, MMP-9, TIMP-1 and TIMP-2 ELISA The ELISAs (Amersham) were performed according to the manufacturer’s instructions. In short, CSF samples were diluted with assay buffer to a total volume of 100 Al (1:10 for MMP-2, TIMP-1 and TIMP-2) or used undiluted (MMP9) and incubated for 120 min at room temperature. Sulphuric acid (1 mM) was added to stop the reaction and the resulting yellow colour was measured spectrophotometerically at 450 nm (Perkin-Elmer). The ELISA for proMMP-2 recognises proMMP-2 in both its latent form and complexed with TIMP-2. The ELISA for MMP-9 recognises proMMP9 in both its latent from and complexed TIMP-1. The ELISA for TIMP-1 recognises free TIMP-1 and that complexed with MMPs. It does not cross-react with TIMP-2. The ELISA for TIMP-2 recognises free TIMP-2 as well as TIMP-2 complexed with active forms of MMPs, but not TIMP-2 complexed with proMMP-2. 2.4. Statistical analysis Statistical analysis of differences between groups was done using Mann – Whitney U-test with a adjustment ( p = 0.01). Data are expressed as mean F standard deviation.

3. Results 3.1. Zymography and Western blot analysis A single band of gelatinolytic activity at 72 kDa was observed in all CSF samples (Fig. 1A). The inclusion of a

S. Lorenzl et al. / Journal of the Neurological Sciences 207 (2003) 71–76

73

Fig. 1. Gelatin zymography of representative CSF samples from control, PD, AD, HD, ALS and PSP patients (A). The 130-kDA band is present in all selected CSF samples, except in the control sample. Western blot analysis using MMP-2 antisera confirms the 72-kDa band is proMMP-2 (B).

human recombinant MMP-2 standard as well as separate immunoblot analyses using MMP-2 antisera confirmed this band to be the proform of MMP-2 (Fig. 1B). Bands of gelatinolytic activity corresponding to MMP-9 were not observed in CSF from either control or disease patients. However, an additional band of gelatinolytic activity at approximately 130 kDa was observed in CSF samples from patients with PD (61% of patients tested), AD (61%), HD (25%) and ALS (39%) (Figs. 1A and 2). We observed this 130-kDa band of gelatinolytic activity in only 5% of control samples. It was seen more frequently in PD ( p = 0.002), AD ( p = 0.0007), ALS ( p = 0.001), as well as in HD patient samples ( p = 0.05). This band was noted in previous reports of CSF samples from patients with brain tumors, viral meningitis, neuroborreliosis and AD [8,20,21]. It was suggested to be the covalent heterodimer of neutrophil gelatinase B-associated lipocalin (NGAL) with gelatinase B [16,22] or the noncovalent complex of MMP-9 and TIMP1 [23,24]. However, immunoblot analysis with antisera against MMP-2, MMP-9, TIMP-1 and TIMP-2 did not establish the composition of this band. This may be due to unavailability of epitopes within complex, or an abundance of complex below the level of detectability.

p = 0.001; AD, p = 0.003; ALS, p = 0.002; HD, p = 0.003). Levels of TIMP-2 in AD ( p = 0.0001) and HD ( p = 0.0001) CSF were significantly increased from controls. TIMP-2 levels in CSF samples from PD ( p = 0.03) or ALS ( p = 0.04) patients were unchanged. Trace amounts of protein were detected with the MMP-9 ELISA in the CSF from patients

3.2. ELISA for proMMP-2, MMP-9, TIMP-1 and TIMP-2 The levels of MMP-2, MMP-9, TIMP-1 and TIMP-2 determined by ELISA on age- and gender-selected CSF samples are shown in Fig. 3. Mean levels of MMP-2 were unchanged. Levels of TIMP-1 were significantly increased in each patient group as compared to the control group (PD,

Fig. 2. Zymography gel showing a commercially available pre-stained molecular weight (MW) marker, the human recombinant form of proMMP2 and proMMP-9 and in the last lane, a CSF sample from an AD patient (same patient as in lane 3 in Fig. 1). This gel shows that the faint band (white arrow) in the AD CSF is below the 130-kDa marker, but above the molecular weight of human recombinant proMMP-9. The second band has the same molecular weight as human recombinant proMMP-2.

74

S. Lorenzl et al. / Journal of the Neurological Sciences 207 (2003) 71–76

Fig. 3. The graphs show the levels of MMP-2, MMP-9, TIMP-1 and TIMP-2 in the CSF of patients with neurodegenerative diseases as measured by ELISA technique. Control subjects (w) were compared to PD (5), AD (D), ALS (n) and HD (+) patients. Although some samples of AD and HD patients had high levels of MMP-2, the mean value of MMP-2 was not statistically significant different from controls. The mean values of TIMP-1 were elevated in all neurodegenerative diseases and TIMP-2 was elevated in AD and HD patients as compared to controls. Statistical analysis of differences between groups were done using Mann – Whitney U-test with a adjustment *( p <0.01).

with AD, HD and ALS (Fig. 3). CSF from control patients did not have measurable MMP-9 concentrations.

4. Discussion There are three important results of this study: first, TIMP-1 levels in the CSF from PD, AD, HD and ALS patients are increased as compared to levels in agematched controls. Also, TIMP-2 levels are increased in AD and HD CSF. Together, these results implicate a potential role for TIMPs in CSF of patients with neurodegenerative diseases. A second finding to emerge from these studies is that MMP-9 expression in its proform or active form was not seen on zymography gels of the CSF samples, but ELISA for MMP-9 revealed trace amounts of a MMP-9 reactive protein, suggesting cross-reactivity of the ELISA. The cross-reactivity may be associated with the approximately 130-kDA band on zymography gels, which may represent MMP-9/TIMP-1 heterodimers. Lastly, levels of MMP-2 were unchanged in patients with neurodegenerative diseases. Generally, TIMPs inhibit active MMPs by forming noncovalent stoichiometric complexes within the catalytic site. Several recent reports suggest a role of TIMPs in brain and peripheral nerve injury and repair. TIMPs are induced after

kainate-induced excitotoxic seizures in mice [25]. Interestingly TIMP-1 and TIMP-2, but not TIMP-3 and TIMP-4, are induced after cortical stab injury in rats [26]. TIMP-1 is upregulated in astrocytes and TIMP-2 in microglia cells and macrophages. Furthermore, TIMP-1 is expressed by Schwann cells and macrophages after sciatic nerve injury in humans [27]. These studies indicate that TIMPs by virtue of their ability to limit the extent of injury-induced matrix proteolysis may be associated with the remodeling of neuronal circuits after injury. Brain microvascular endothelial cells upregulate TIMP-1 in response to a variety of cytokines, with the strongest effect exerted by the combination of IL-1h and TNF-a [28]. Additionally, TIMP-1 blocks degradation of IL-1h by several MMPs [29]. These cytokines are implicated in neurodegeneration in many of the diseases investigated in this study and are frequently found in CSF. Interestingly, the regulation of TIMP-2, which is constitutively expressed, is not significantly influenced by cytokines and growth factors [30]. In an animal model of stroke, TIMP-2 inhibited MMP2 activity and reduced proteolytic opening of the blood – brain barrier by MMP-2 [31]. We found significant increases of TIMP-2 levels in the CSF of AD and HD patients. TIMPs may also mediate MMP-independent processes in response to injury. They are involved in cell growth and apoptosis [7] and possess mitogenic activity in a number of

S. Lorenzl et al. / Journal of the Neurological Sciences 207 (2003) 71–76

cell lines [32] which protects them from pro-apoptotic stimuli [33]. In AD, TIMPs are localized to neuritic senile plaques, neurofibrillary tangles, and Purkinje cells [13]. We found elevated levels of TIMP-1, but not TIMP-2 in the substantia nigra of postmortem brain tissue of PD, similar to our CSF data (Lorenzl et al., unpublished data). The expression of MMP-9 in CSF has been linked to blood – brain barrier disruption in various neurological disorders, including meningitis [9] multiple sclerosis [34] and stroke [10]. Further, decreased levels of MMP-9 are found in CSF samples from patients with ALS [15,16]. Zymography is generally considered to be a more sensitive assay of MMPs, because it is capable of detecting gelatinases in the picogram levels, as compared to the ELISA assay, which detects MMPs in nanogram levels. In the present study, however, MMP-9 expression was not observed using zymographic analysis of CSF, but MMP-9 ELISA revealed low levels of this protein. MMP-9 and TIMP-1 are secreted in the CSF as complex heterodimers (MW approximately 130 kDa), and the MMP-9 ELISA that we used is capable of detecting MMP-9 either alone or complexed with TIMP1 (cross-reactivity for the MMP-9/TIMP-1 complex = 100%). Therefore, the 130-kDa band which is present in CSF of patients with neurodegenerative diseases may represent the MMP-9/TIMP-1 heterodimer, as has been suggested by others [24]. Constitutively released MMP-9 may be immediately scavenged by TIMP-1. This hypothesis is supported by the increased TIMP-1 levels, which we found in the CSF of patients with neurodegenerative diseases. In summary, our findings demonstrate that TIMPs are elevated in the CSF of patients with neurodegenerative diseases. Especially TIMP-1, which has been shown to play an important role in responses to experimental brain injury, is elevated in the CSF of patients with neurodegenerative diseases. This compensatory response may prevent further tissue damage and preserve the integrity of the extracellular matrix in neurodegenerative diseases. Whether additional functions of TIMPs apart from their MMP inhibitory potential are also involved in an endogenous counterregulation in neuronal injury of these diseases needs to be further investigated. Our findings raise the possibility of a role of TIMPs in neurodegeneration, especially a role in modulating the accumulation of extracellular/intracellular proteins, and ultimately cell death.

Acknowledgements The secretarial assistance of Sharon Melanson and Greta Strong is gratefully acknowledged. This study was supported by the Deutsche Forschungsgemeinschaft (S.L.), the NIH, the Department of Defense, the Alzheimer Disease Association, the ALS Association, the Huntington’s Disease Society of America and the Hereditary Disease Foundation (M.F.B.).

75

References [1] Albers DS, Beal MF. Mitochondrial dysfunction and oxidative stress in aging and neurodegenerative. J Neural Transm, Suppl 2000;59: 133 – 54. [2] Backstrom JR, Lim GP, Cullen MJ, Tokes ZA. Matrix metalloproteinase-9 (MMP-9) is synthesized in neurons of the human hippocampus and is capable of degrading the amyloid-beta peptide (1 – 40). J Neurosci 1996;16:7910 – 9. [3] Harkness KA, Adamson P, Sussman JD, Davies-Jones GA, Greenwood J, Woodroffe MN. Dexamethasone regulation of matrix metalloproteinase expression. Brain 2000;123:698 – 709. [4] Oh LY, Larsen PH, Krekoski CA, Edards DR, Donovan F, Werb Z, et al. Matrix metalloproteinase-9 gelatinase B is required for process outgrowth by oligodendrocytes. J Neurosci 1999;19:8464 – 75. [5] Gottschall PE, Deb S. Regulation of matrix metalloproteinase expressions in astrocytes, microglia and neurons. Neuroimmunomodulation 1996;3:69 – 75. [6] Yong VW, Krekoski CA, Forsyth PA, Bell R, Edwards DR. Matrix metalloproteinases and diseases of the CNS. Trends Neurosci 1998; 21:75 – 80. [7] Brew K, Dinakarpandian D, Nagase H. Tissue inhibitors of metalloproteinases: evolution, structure and function. Biochim Biophys Acta 2000;1477:267 – 83. [8] Kolb SA, Lahrtz F, Paul R, Leppert D, Nadal D, Pfister HW, et al. Matrix metalloproteinases and tissue inhibitors of metalloproteinases in viral meningitis: upregulation of MMP-9 and TIMP-1 in cerebrospinal fluid. J Neuroimmunol 1998;84:143 – 50. [9] Paul R, Lorenzl S, Koedel U, Sporer B, Vogel U, Frosch M, et al. Matrix metalloproteinases contribute to the blood – brain barrier disruption during bacterial meningitis. Ann Neurol 1998;44:592 – 600. [10] Clark AW, Krekoski CA, Bou SS, Chapman KR, Edwards DR. Increased gelatinase A (MMP-2) and gelatinase B (MMP-9) activities in human brain after focal ischemia. Neurosci Lett 1997;238:53 – 6. [11] Leppert D, Ford J, Stabler G, Grygar C, Lienert C, Huber S, et al. Matrix metalloproteinase-9 (gelatinase B) is selectively elevated in CSF during relapses and stable phases of multiple sclerosis. Brain 1998;121:2327 – 34. [12] Asahina M, Yoshiyama Y, Hattori T. Expression of matrix metalloproteinase-9 and urinary-type plasminogen activator in Alzheimer’s disease brain. Clin Neuropathol 2001;20:60 – 3. [13] Peress N, Perillo E, Zucker S. Localization of tissue inhibitor of matrix metalloproteinases in Alzheimer’s disease and normal brain. J Neuropathol Exp Neurol 1995;54:16 – 22. [14] Lim GP, Backstrom JR, Cullen MJ, Miller CA, Atkinson RD, Tokes ZA. Matrix metalloproteinases in the neocortex and spinal cord of amyotrophic lateral sclerosis patients. J Neurochem 1996;67:251 – 9. [15] Beuche W, Yushchenko M, Mader M, Maliszewska M, Felgenhauer K, Weber F. Matrix metalloproteinase-9 is elevated in serum of patients with amyotrophic lateral sclerosis. NeuroReport 2000;11:3419 – 22. [16] Yushchenko M, Weber F, Mader M, Scholl U, Maliszewska M, Tumani H, et al. Matrix metalloproteinase-9 (MMP-9) in human cerebrospinal fluid (CSF): elevated levels are primarily related to CSF cell count. J Neuroimmunol 2000;110:244 – 51. [17] Hoehn MM, Yahr MD. Parkinsonism: onset, progression and mortality. Neurology 1967;17:427 – 42. [18] Brooks BR. El Escorial World Federation of Neurology criteria for the diagnosis of amyotrophic lateral sclerosis. Subcommittee on Motor Neuron Diseases/Amyotrophic Lateral Sclerosis of the World Federation of Neurology Research Group on Neuromuscular Diseases and the El Escorial ‘‘clinical limits of amyotrophic lateral sclerosis’’ workshop contributors. J Neurol Sci 1994;124:96 – 107 [Supplement]. [19] Kleiner DE, Stetler-Stevenson WG. Quantitative zymography: detection of picogram quantities of gelatinases. Anal Biochem 1994;218: 325 – 9. [20] Friedberg MH, Glantz MJ, Klempner MS, Cole BF, Perides G. Spe-

76

[21]

[22]

[23]

[24]

[25]

[26]

[27]

S. Lorenzl et al. / Journal of the Neurological Sciences 207 (2003) 71–76 cific matrix metalloproteinase profiles in the cerebrospinal fluid correlated with the presence of malignant astrocytomas, brain metastases, and carcinomatous meningitis. Cancer 1998;82:923 – 30. Perides G, Charness ME, Tanner LM, Peter O, Satz N, Steere AC, et al. Matrix metalloproteinases in the cerebrospinal fluid of patients with Lyme neuroborreliosis. J Infect Dis 1998;177:401 – 8. Opdenakker G, Van den Steen PE, Van Damme J. Gelatinase B: a tuner and amplifier of immune functions. Trends Immunol 2001; 22:571 – 9. Backstrom JR, Miller CA, Tokes ZA. Characterization of neutral proteinases from Alzheimer-affected and control brain specimens: identification of calcium-dependent metalloproteinases from the hippocampus. J Neurochem 1992;58:983 – 92. Moutsiakis D, Mancuso P, Krutzsch H, Stetler-Stevenson W, Zucker S. Characterization of metalloproteinases and tissue inhibitors of metalloproteinases in human plasma. Connect Tissue Res 1992;28:213 – 30. Rivera S, Tremblay E, Timsit S, Canals O, Ben-Ari Y, Khrestchatisky M. Tissue inhibitor of metalloproteinases-1 (TIMP-1) is differentially induced in neurons and astrocytes after seizures: evidence for developmental, immediate early gene, and lesion response. J Neurosci 1997;17:4223 – 35. Jaworski DM. Differential regulation of tissue inhibitor of metalloproteinase mRNA expression in response to intracranial injury. Glia 2000;30:199 – 208. La Fleur M, Underwood JL, Rappolee DA, Werb Z. Basement membrane and repair of injury to peripheral nerve: defining a potential role for macrophages, matrix metalloproteinases, and tissue inhibitor of metalloproteinases-1. J Exp Med 1996;184:2311 – 26.

[28] Bugno M, Witek B, Bereta J, Bereta M, Edwards DR, Kordula T. Reprogramming of TIMP-1 and TIMP-3 expression profiles in brain microvascular endothelial cells and astrocytes in response to proinflammatory cytokines. FEBS Lett 1999;448:9 – 14. [29] Ito A, Mukaiyama A, Itoh Y, Nagase H, Thogersen IB, Enghild JJ, et al. Degradation of interleukin 1beta by matrix metalloproteinases. J Biol Chem 1996;271:14657 – 60. [30] Hanemaaijer R, Koolwijk P, le Clercq L, de Vree WJ, van Hinsbergh VW. Regulation of matrix metalloproteinase expression in human vein and microvascular endothelial cells. Effects of tumour necrosis factor alpha, interleukin 1 and phorbol ester. Biochem J 1993;296: 803 – 9. [31] Rosenberg GA, Kornfeld M, Estrada E, Kelley RO, Liotta LA, Stetler-Stevenson WG. TIMP-2 reduces proteolytic opening of blood – brain barrier by type IV collagenase. Brain Res 1992;576:203 – 7. [32] Hayakawa T, Yamashita K, Ohuchi E, Shinagawa A. Cell growth-promoting activity of tissue inhibitor of metalloproteinases-2 (TIMP-2). J Cell Sci 1994;107:2373 – 9. [33] Valente P, Fassina G, Melchiori A, Masiello L, Cilli M, Vacca A, et al. TIMP-2 over-expression reduces invasion and angiogenesis and protects B16F10 melanoma cells from apoptosis. Int J Cancer 1998; 75:246 – 53. [34] Maeda A, Sobel RA. Matrix metalloproteinases in the normal human central nervous system, microglial nodules, and multiple sclerosis lesions. J Neuropathol Exp Neurol 1996;55:300 – 9.