Malaria is one of major public health problems at a global level, causing between 300 and 500 million clinical cases and about 1 million deaths each year. In humans, one of five Plasmodium spp. is the etiological agent, but Plasmodium falciparum represents the most dangerous one, especially in Sub-Saharan Africa [97,98]. One of the most important causes of fatal malaria is represented by cerebral malaria (CM), which is the result of accumulation, as well as adhesion, to endothelial cells of parasitized red blood cells in the capillaries and post-capillary venules of the brain with the following hypoxia [99]. Pro-inflammatory cytokines, such as tumor necrosis factor (TNF), interleukin (IL)-1 and -6, elicited by the immune response against the parasite, play an important role in upregulating adhesion molecules on the endothelial cell surface and then aggravating the red blood cell sequestration [100]. Cell adhesion molecules expression and TNF-α and IL-1β levels have been observed to be increased in postmortem brains of children with CM, especially in the cerebellum [101].

The parasite escapes the host immune response with different strategies, one of which is represented by the coding and production of variant proteins present at the surface level, such as, for example, the P. falciparum erythrocyte membrane protein-1 (PfEMP-1), which allows the parasite to bind to different molecules present on the endothelial cell surface, facilitating its sequestration in the blood capillaries [100]. These sophisticated escape mechanisms into various organs cause the mechanical blockage of blood vessels and local inflammation, leading to organ-specific disease syndromes, such as placental malaria and CM [100,102].

Different barrier layers limit and regulate molecular exchange at the interfaces between the blood and the neural tissue or its fluid spaces. The blood brain barrier (BBB), formed by the cerebrovascular endothelial cells between blood and brain interstitial fluid, is a selectively permeable structure regulating ion and nutrient transport into the brain. It represents a filter between the CNS and the blood, limiting the free flow of physiological molecules between the bloodstream and parenchyma. Specialized endothelial cells (kept strictly joined by tight junctions), which line cerebral blood vessels, are surrounded by a basal lamina and astrocyte end-foot processes to form and maintain the BBB [103]. Other barriers are constituted by the choroid plexus epithelium between blood and the ventricular CSF and the arachnoid epithelium between blood and the subarachnoid CSF [103].

Alterations of the BBB integrity can facilitate the passage of potentially harmful substances into the brain, with possible pathological consequences. During CM, vascular dysfunction may cause severe damage to the BBB, enhanced by the inflammatory cascade mentioned above [102].

In many neurological diseases, disruption of the BBB is mediated by the MMPs [reviewed in 96]. In fact, MMPs derived from infiltrating leukocytes or from cells of the CNS cleave matrix proteins, which are essential for the maintenance of BBB integrity and for neuronal survival [104]. In fact, intraparenchymal injection of MMPs causes breakdown of the BBB and the subsequent leakage of capillaries. Furthermore, MMP inhibitors, such as tissue inhibitor metalloproteinase (TIMP)-2, reduce extracellular matrix proteolysis induced by type IV collagenase and protect the BBB [105]. This barrier does not restrict leukocyte diapedesis; in fact, a small number of activated lymphocytes can access the CNS with the aim of developing an immunosurveillance to control possible infectious agents [106].

Both MMPs and TIMPs can play a relevant role in malaria pathogenesis with at least two mechanisms: (i) degradation of BBB substrates and (ii) functioning as effectors and regulators of the |immune response.

The results obtained in malaria patients are contradictory. In fact, while in Kenyan children affected by acute malaria, by means of a genome-wide analysis of expression, an upregulated MMP-9 mRNA expression in blood cells was observed, accompanied by neutrophilia [107]. In Gabonese children affected by either uncomplicated or severe malaria, MMP-9 serum levels were unchanged compared to healthy controls, with TIMP-2 levels even lower. On the contrary, TIMP-1 is associated with signs and symptoms of severe malaria, and MMP-8 levels are elevated in patients with severe or uncomplicated P. falciparum malaria [108]. In the light of these results, it may be argued that TIMP-1 and MMP-8 may predict the severity of malaria. TIMP-1 may prevent further MMP-induced damage by counterbalancing the activity of MMP- 9 and, to a lesser extent, MMP-8 activity.

Data derived from experimental studies are accumulating, which show unequivocally that MMPs and TIMPs, as well as the TNF-α converting enzyme, play a crucial role in malaria pathogenesis [96].

As in other inflammatory conditions involving the brain, such as lipopolysaccharide (LPS)-injured brain or multiple sclerosis [109,110], MMP-9 might also play an important role in CM pathogenesis, increasing BBB permeability and infiltration of leukocytes. In particular, this gelatinase seems to be induced by the hemozoin, a catabolic product of hemoglobin, at the endothelial cell level, as shown in in vitro studies. Furthermore, this parasite-derived product increases the protein expression of MMP-1, MMP-3 and also of TIMP-2 [111].

In patients infected by P. falciparum, analysis of post-mortem brain samples showed that MMP-1 accumulates in astrocytes of the BBB and in macrophages/microglial cells, which are present in Dürck's granulomas [112], represented by microglial-astroglial nodules surrounding the damaged vessels (see Figure 2 from Deininger et al. [113]).

From all these data, we may argue that in malaria caused by P. falciparum, especially if complicated by CM, the balance between MMP and TIMP expression patterns is altered in favor of the former. This imbalance is responsible for the dramatic disruption of the BBB, which occurs during CM. However, the definitive demonstration of the role played by the MMPs in disrupting the BBB during CM is still not available, and further research on this issue is needed [96].

African Trypanosomosis is caused by protozoa of the genus Trypanosoma (T. brucei), which are responsible for one of the most important parasitic infections in Sub-Saharan Africa. These trypanosomes cause human sleeping sickness (human African trypanosomiasis, HAT) in man and a cattle disease, called Nagana. The disease evolves from a first hemolymphatic stage to a second meningo-encephalitic stage, after parasites cross the BBB and invade the CNS [114].

After injection by the vectors (tsetse fly), the parasites (at the stage of metacyclic trypomastigotes) change morphology and behavior in the circulation; therefore, they acquire the ability to cross the BBB [115]. At this phase of infection, the so-called meningoencephalitic stage occurs. After passing through the highly vascularized epithelium of the choroid plexus, the parasites enter into the CSF, and later, they arrive to the brain parenchyma. At that time, an induced expression of host endothelial receptors, e.g., ICAM-1, is observed. Accumulation of parasites in the brain, followed by the recruitment of leukocytes and activation of astrocytes and microglia, cause chronic encephalopathy, which may be fatal, if not treated [116].

The disruption of BBB during African trypanosomosis is controversial; it was observed, for example [117], that the parasite can pass the barrier by means of a cysteine proteinase, in part by generating Ca²⁺ activation signals. However, the traversal of leukocytes and parasites through the basal lamina into the brain at second stages of HAT may have common mechanisms involving certainly the MMPs.

MMP-2 and MMP-9 were significantly increased; results correlated with the presence of parasites and leukocytes in the CSF of T. b. gambiense-infected patients with second-stage HAT.

It was therefore highlighted that MMP-9, along with ICAM-1, are valuable staging markers for T. b. gambiense HAT and that, alone or in combination, MMPs, as well as ICAMs, can identify the meningo-encephalitic stage of HAT [118]. It has been observed that MMP-2 and MMP-9 create a localized temporary opening of the glia limitans (which is a part of the BBB) by selective cleavage of the β-dystroglycan subunit anchoring the astrocyte end-feet to the parenchymal membrane [119]. In this way, the gelatinases MMP-2 and MMP-9 let leukocyte penetrate the outer parenchymal basement membrane into the brain parenchyma.

In the brain of T. b. brucei-infected mice, a massive increase in parasite and leukocyte numbers was accompanied by a significantly induced mRNA expression of MMP-3, MMP-8 and MMP-12, at thirty days post-infection, whereas levels of MMP-1b, -2, -7, -9, -11, -13, -14 and -19 and TIMP-1 and -2 mRNA were unaltered and MMP-10 was undetectable [120].

We may therefore conclude that, also in trypanosomosis, like in cerebral malaria, increased MMP expression levels are responsible for the passage of the parasites, as well as leukocytes, through the BBB into the brain, where they cause cerebral pathology.

Toxoplasma gondii is an obligate intracellular parasite capable of infecting virtually any warm-blooded animal [121]. In humans, Toxoplasma infections are widespread and can lead to severe disease, i.e., toxoplasmic encephalitis in individuals with an immature or suppressed immune system [122].

In murine Toxoplasma encephalitis, CD4+ and CD8+T cells are mainly recruited to the brain, where they are crucial in preventing the reactivation of latent infections [123]. This process is mediated by IFN-γ; this cytokine, in fact, stimulates anti-parasitic effector mechanisms, for example, of macrophages and regulates chemokine expression and leukocyte recruitment [124]. Different molecules involved in the immune and inflammatory response during toxoplasmosis, such as IL-1, IL-23, TNF-α and COX-2, increase the MMP production in the brain, reviewed by Clark and colleagues [125].

In the brain tissue of T. gondii-infected mice, an increase in CD4+ and CD8+ T cells producing both MMP-8 and MMP-10 has been recently shown [125]. Furthermore, TIMP-1 is expressed in invading T-cells and CNS-resident astrocytes during infection. In wild-type mice, changes in tissue morphology and signs of astrocyte activation occur, contrary to what happens in infected TIMP-1 KO mice, where an increase in CD4+ T cells along with a significantly reduced parasite burden in the brain was observed, differently from the peripheral amount of parasites. This is not accompanied by any substantial pathology in the brain, as shown by histology, which suggests less focal immune clusters and decreased astrocyte activation.

It may be argued that upregulation of TIMP-1 during infection may inhibit the pathogen clearance by limiting lymphocyte penetration into the CNS, driven by MMPs.

The resulting inhibition of MMPs by an increased expression of TIMP-1 may represent an evasion mechanism by the parasite, with the aim to limit the arrival of immune cells or a host response to downregulate immune-mediated pathology.

In infected TIMP-1 KO mice, in fact, brain inflammation occurred with a reduced or even absent perivascular accumulation, but with a higher number of CD4+ T-cells infiltrating the brain parenchyma, probably reflecting a lack of control of the degradation processes of the basal lamina, played by MMP. All these data suggest that MMPs are beneficial in facilitating the parasite clearance from the brain.

Neurocysticercosis (NCC) is the CNS infection caused by the larva of Taenia solium tapeworm, acquired after parasite egg ingestion (NCC without taeniasis) or raw or poorly cooked pork consumption (NCC with taeniasis) [126]. It is considered the most common cause of acquired epilepsy at the global level.

The clinical picture is pleomorphic, but active seizure is the most frequent manifestation. The disease is slowly progressive, and multiple factors play a role in the severity of the symptoms; one of them is represented by the degree of inflammatory reaction in the host brain. It is also possible, however, that individuals with NCC in many cases could remain asymptomatic, and the exact reasons largely remain unexplained [127].

Studies in experimental animals have shown that the MMP expression plays a crucial role in the differential breakdown of the BBB, from which infiltration of blood leukocytes and, consequently, the production of inflammatory cytokines depend. In this model of NCC, inflammatory infiltrates contain different populations of immune cells, and different MMPs are involved in the mechanisms of cleavage of cytokine, chemokine and adhesion molecules, by which the immune response can be activated [128]. In patients with NCC, serum levels (evaluated by ELISA) and enzymatic activities (evaluated by zymography) of MMP-2 and MMP-9 resulted in association with symptomatic manifestations of the disease. Figure 3 shows an example of gel zymography, where NCC serum samples were loaded.

Mean serum MMP-2 levels were higher both in asymptomatic and symptomatic NCC cases compared to healthy controls; however, there was no difference in the levels of MMP-2 in symptomatic and asymptomatic NCC patients.

On the contrary, MMP-9 serum levels were significantly higher in symptomatic NCC patients than in asymptomatic NCC cases or healthy controls (Figure 4). Levels of both MMPs positively correlated with symptomatic (presence of seizures) NCC [129].

Recent studies have shown that higher levels of MMP-9 correlated with epilepsy [130,131]. In consideration of that, symptomatic NCC patients might underlie the seizures, because of increased levels of MMP-9.

Metalloproteinases have been identified in a variety of nematodes, such as Brugia malayi, Toxocara canis, Strongyloides stercoralis, Nippostrongylus brasiliensis, Dirofilaria immitis, Trichuris suis, Ancylostoma caninum, Caenorhabditis elegans and Gnathostoma spinigerum (reviewed in Tsai et al., 2008) [132]. However, one of the most studied model in vivo as regards MMPs involvement in the pathogenesis of nematode-induced cerebral infections is certainly represented by Angiostrongylus cantonensis, a nematode parasite responsible for the most common eosinophilic meningitis in the Pacific Islands and Southeast Asia [132].

This rat lung worm has an obligatory intracerebral migration in its hosts [133]. When humans and mice (nonpermissive hosts) are infected with this parasite, the worms migrate to the brain, where they develop into young adults, failing to reach maturity within the heart and lungs [133]. A. cantonensis is a neurotropic nematode that requires the CNS of mammalian hosts for its growth [134].

The infective third-stage larvae orally infect the final host and are carried in the blood to the CNS, where they molt twice to become immature adults and enter the subarachnoid space. In a permissive host (rodents), immature adults migrate from the brain to the lungs [133].

However, in non-permissive hosts, the immature adults remain in the CNS of the host, and this infection is the main cause of eosinophilic meningitis and eosinophilic meningoencephalitis [135]. Most of information on this parasitic infection derives from experimental studies in rodents. Mice infected with A. cantonensis undergo eosinophilic meningitis, which peaks at around three weeks, when CSF eosinophilia reaches a peak [136,137]. It was also shown that the MMP-9 was detected in CSF at day 10 post-inoculation (PI) and reached a high level from days 15 to 25 PI. This increased level was not due to parasitic production of MMP-9, since excretory/secretory (E/S) antigen from A. cantonensis adult worms lacked the enzyme activity. However, analysis of the extracts and E/S products of the parasite larval and adult stages on gelatin substrate zymography demonstrated the presence of distinct gelatinolytic enzymes. In worm extracts, it was found that the metalloproteinases at different molecular weights: in L1 23 kDa, in L3 66, 42 and 30 kDa, in young adult worms, 72 and 94 kDa, and in adult worm, again 72 and 94 kDa. On the contrary, in E/S products, the L1 revealed one low (42 kDa) and two high (105 and 94 kDa) molecular weight gelatinolytic bands. The L3 revealed three low (66, 50 and 30 kDa) and one high (105 kDa) molecular weight proteolytic bands. By using specific inhibitors, the nature of metalloproteinases was confirmed for the 105 and 94 proteolytic bands of the L1 and for the 50 and 30 kDa proteolytic bands of the L3 larvae. These metalloproteinases secreted in the infective larvae might be associated with the parasite dissemination or pathogenesis [138].

By using Immunohistochemistry, the MMP-9 localized within eosinophils and macrophages, present in the subarachnoid space of experimentally infected mice.

These data suggest that infiltrating leukocytes are important sources of MMP-9 in this parasitic meningitis [139]. In particular, by immunogold electron microscopy, it was shown that in eosinophils, MMP-9 was mostly localized in the 'small' granules in the cytoplasm and along the cell membrane and not in the crystalloid-containing secretory granules observed, suggesting that the enzyme is synthesized and/or stored in the small granules of the eosinophils, to be released into the subarachnoid space of the host's brain by secretion or cell rupture [140]. MMP-9 was also found within the endothelial cells lining the vascular spaces of the brain [141].

The increased MMP-9 activity was significantly associated with the rapid increase of CSF eosinophils and the inflammatory reaction of the subarachnoid space. Contrary to what happens for MMP-9, MMP-2 activity did not change during infection [139].

During experimental infection in mice, Purkinje cells in the cerebellum become small and irregular in shape, with degenerative atrophy or partial loss; furthermore, enlarged vacuolar structures and swollen mitochondria within the cytoplasm are visible at electron microscopy. The MMP-9 mRNA expression, which is absent in normal Purkinje cells, precedes the degeneration process. Then, MMP-9 protein level and enzyme activity increased when the Purkinje cells were degenerated. Furthermore, MMP-9 was localized within degenerative Purkinje cells.

When a specific MMP inhibitor was used, MMP-9 enzyme activity decreased by 41.6%. In addition, the number of degenerated Purkinje cells was reduced, as well [142].

Proteolysis depends on the balance between the proteinases and their inhibitors. MMP-9 and its specific inhibitors, TIMPs, contribute to eosinophilic inflammatory reaction in the subarachnoid space of the A. cantonensis-infected mice. If MMP-9 levels increase during infection, those of TIMP-1 did not change, remaining at basal levels at all time points. Immunohistochemistry demonstrated that also TIMP-1 is localized in eosinophils and macrophages infiltrating the CNS tissue. These results show that MMP-9/TIMP-1 imbalance in angiostrongyloidosis may be associated with eosinophilic meningitis [143].

As already stated, blood–central nervous system barrier breakdown is an important pathophysiological event occurring in meningitis, which is the result of extravasation of leucocytes into subarachnoid space. Two enzymatic systems are essentially responsible for this disruption, the plasminogen activators (PAs) and MMPs [143]. In mice experimentally infected with A. cantonensis, it was shown that during eosinophilic meningitis the activities of tissue-type PA (tPA), urokinase-type activator (uPA) and MMP-9 in cerebrospinal fluid (CSF) were significantly increased, compared to uninfected animals. Furthermore, eosinophilia present in the CSF significantly correlated with tPA, uPA and MMP-9 activities, as well as with albumin, content in the fluid. In addition, when infected mice were treated with a specific MMP blocker, MMP-9 activity and total protein concentrations declined significantly, suggesting that the PAs and MMP-9 proteolytic cascade may be involved in blood–CNS barrier disruption, during eosinophilic meningitis [144].

CSF levels of MMP-2, MMP-9 and TIMP-1 resulted in significantly higher levels in patients infected by A. cantonensis in Taiwan, suffering eosinophilic meningitis, compared with healthy controls. On the contrary, TIMP-4 levels were significantly lower in the same patients. In contrast to MMP-2, proteolytic activity of MMP-9 detected by gelatin zymography was only observed in patients with eosinophilic meningitis. Higher MMP-9 levels were found in the CSF of patients with eosinophilic meningitis, contrary to what happens for MMP-2. CSF MMP-9 increases in patients in parallel with CSF leukocyte counts and CSF/serum albumin ratio (QAlb) values. During recovery from eosinophilic meningitis, after treatment with mebendazole and dexamethasone, a gradual decrease in levels of MMP-9 (decreased more than 50% in six patients two weeks after treatment), QAlb and TIMP-1, as well as an increase in those of TIMP-4, were observed. TIMP-4 levels were lower during the acute phase of infection, persisting as such even when MMP-9 and TIMP-1 have already decreased, until eosinophil meningitis was not definitely recovered, suggesting that TIMP-4 plays a crucial role in the proteolytic balance of BBB damage, in these patients. These results confirmed what was already known from studies in experimental infection models.

In the CSF of patients, activity of MMP-9 was not completely inhibited, because of the simultaneous decrease of TIMP-4, with the resulting BBB dysfunction, as shown by the higher CSF/serum albumin ratio (QAlb) values observed in these patients. Modification of the cytokine milieu has a major impact on modulation of TIMP expression and may be responsible for changes in the levels of these proteins (MMPs/TIMPs) in eosinophilic meningitis [132].

Also, MMP-12 and its substrate, elastin, participate in the inflammatory response. MMP-12/TIMP-1 ratio was significantly increased in the CSF of A. cantonensis-infected mice from day 10 p.i. (post-infection) and reached high levels on days 20 and 25 p.i. The production of MMP-12 resulted correlated with several parameters, such as elastin degradation, eosinophil count, blood–CSF barrier permeability and pathological changes in the subarachnoid space.

MMP-12 might be involved in elastin degradation in the meningeal vessel of the subarachnoid space. After treatment with both albendazole (an antihelmintic) and doxycycline (used in this study as a non-selective MMP inhibitor), a significant reduction of the levels of MMP-12, elastin and Evans blue accumulation in the CSF in mice with meningitis was observed, suggesting that MMP-12 contributes to elastin degradation, which is reduced by the action of doxycycline on the inflammatory reaction mediated by MMP-12 [145].