The major findings of this study are as follows: (1) thrombin promoted M1 polarization of meningeal macrophage; (2) thrombin induced vascular inflammation in subarachnoid space; (3) inhibition of PAR1 attenuated the inflammatory response in subarachnoid space after thrombin injection; (4) meningeal inflammation was associated with severity of thrombin-induced hydrocephalus and neuron loss.
Thrombin is an important component of the coagulation cascade and plays an essential role in hydrocephalus [2, 3, 7]. In the in vitro study, doses of thrombin greater than 5 U/ml resulted in dose-dependent lactate dehydrogenase release [21]. It is well known that 50μl of whole blood produces 13 to 18 units of thrombin [22] and intraventricular infusion of 5 units of thrombin caused high mortality in our preliminary experiments. Thus, the amount of thrombin in this current study was 3 unites in 50μl saline, the same as our previous study [2, 3].
Because lateral ventricular volumes in rats were significantly larger in 24 hours rather than in the following day 1 to 28 after intraventricular hemorrhage [2], our aim was to investigate the role of thrombin in the early phase following brain injury. Our previous study found that thrombin disrupted the blood-CSF barrier via VE-cadherin downregulation and led to communicating hydrocephalus [2, 3, 9]. And the inflammation response was previously proved to be closely associated with the formation of hydrocephalus [11, 12, 23, 24]. Systemic inflammation is a therapeutic target in acute ischemic stroke [25, 26]. Also, it was indicated that anti-inflammation treatment could alleviate spontaneous hydrocephalus [27]. In the current study, the role of inflammation in thrombin-induced hydrocephalus was investigated.
Non-parenchymal brain macrophages are mononuclear phagocytes that are increasingly recognized to be critical players in the diseases of the central nervous system [27-30]. There are three types of central nervous system macrophages between parenchyma and circulation- namely, meningeal, perivascular and choroid plexus macrophages [17, 18, 30]. Meningeal inflammation has been identified as a key feature of hydrocephalus [12, 15, 23]and may contribute to the extensive cortical pathology that accompanies progressive disease [31-34]. Meningeal cells are involved in cortical development, fibrotic scar formation, brain inflammation, and neurodegenerative disorders such as Multiple Sclerosis (MS) and Alzheimer's disease (AD) and other brain disorders [33-35]. Here, our study aimed to explore whether meningeal macrophages were involved in the inflammation of subarachnoid space and the formation of hydrocephalus [12, 15]. Two macrophage populations were found in the central nervous system, corresponding to the proinflammatory M1 and “alternatively activated” anti-inflammatory M2 classes [18, 29]. The M1 macrophages are characterized by the high expressions of oxygen intermediates and proinflammatory cytokines which are closely associated with brain injury [27, 36]. While the M2 phenotype is considered to be involved in the promotion of tissue remodeling [18, 36]. In our study, a significant increase in M1 polarization was observed in meningeal macrophages and many of them are CD68 positive after thrombin injection. Whereas, as marked by CD206, we found the change in M2 macrophages were not significant compared with the saline group.
One of the most important features of proinflammatory macrophage is the secretion of pro-inflammatory cytokines [36]. In this study, activated endothelial cells and inflammatory infiltrations were observed after thrombin injection, pro-inflammatory cytokines were also significantly increased. Cytokines (like IL-1β, IFNγ) were shown to decrease VE-cadherin protein levels and induce endothelial barrier dysfunction [37, 38]. Intercellular adhesion molecule 1 (ICAM1), induced by IL-1β, IFNγ [39], is best known for its role in leukocyte adhesion and typically expressed on endothelial cells and cells of the immune system [20, 40]. ICAM1, a biomarker of endothelial dysfunction [41], could mediate Src activation [42, 43] which is involved in the regulation of VE-cadherin [3, 8]. In this study, the increased expression of ICAM1, IL-1β, and IFNγ after thrombin injection were observed, which highlighted the role of the inflammatory response in thrombin-induced hydrocephalus.
As thrombin plays an important role in inflammatory neurotoxin, increased evidence demonstrated that thrombin causes acute but not chronic cell death contributing to neuronal atrophy [44]. In our study, neuron loss was observed after intraventricular injection of thrombin. Furthermore, a growing body of evidence demonstrated that leukocyte-endothelial adhesion interaction could lead to post-hemorrhagic vasospasm which also could exacerbate the neuron loss [45-47]. A growing body of evidence supported that meningeal inflammation in the subarachnoid compartment plays a key role in the pathogenesis of cortical grey matter lesions in multiple sclerosis patients [32]. Unfortunately, in this study, we didn’t detect caspase 3 positive cells in the thrombin group, other assessments will be done in further study.
To investigate the relationship between inflammation and thrombin-induced brain injury, the ventricular volume and the neuron loss in the cortex were analyzed. Interestingly, the level of infiltrating CD68 or CD206 positive macrophages in meninges, correlated modestly with the severity of hydrocephalus and neuron loss. Rats with increased meningeal inflammation had a more severe brain injury. These results suggested that diffuse meningeal inflammation may contribute to the pathological mechanisms driving the progression of hydrocephalus and cortical neuronal pathology. These findings are consistent with other observations in meningeal inflammation [31-34].
The cellular signaling effects of thrombin are mediated by protease-activated receptors (PARs) [2, 3, 48]. PAR1, PAR3, and PAR4 can be activated by thrombin, PAR2 is activated by coagulation factors VIIa and Xa, but not by thrombin [48]. While lesser extent PAR3 and PAR4 are regulated in reactive brain injury [49], PAR1 plays a more extensive role in thrombin signaling [8, 19, 50]. The activation of PAR1 disrupts vascular integrity[3, 8], results in neurotoxicity[51, 52] and increases secondary brain injury after stroke[2, 7, 50]. In our study, inhibition of PAR1 reduced the meningeal inflammation, neuron loss, and hydrocephalus. It is possible that anti-inflammation will provide novel therapeutic targets for hydrocephalus management in the future.
This study indicated that meningeal inflammation played a key role in the thrombin-induced hydrocephalus. Additionally, our previous study found that the choroid plexus damage also contributed to the thrombin-induced hydrocephalus [3], the activation of choroid plexus macrophages was also closely associated with hydrocephalus [53]. These results demonstrated the great importance of immune function of blood- cerebrospinal fluid barrier (BCSFB) in hydrocephalus. Meningeal macrophages and choroid plexus macrophages are localized at the interface between the parenchymal and CSF leading to novel therapies for brain diseases [18].
This study has several limitations. (1) As sufficient research base may be available [3], only four rats were used in each group in this current study. This limitation may affect the CD206 results, in vitro studies will be conducted to confirm this result in future; (2) As a lot of thrombin is released after hemorrhage stroke which is closely associated with brain edema in the early phase, only a single time point (24h) was evaluated in this current study to explore the role of thrombin in early pathophysiology. Thrombin injection induced brain hemorrhage at day-3 and brain atrophy at day-7 through PAR-1 activation [54]. More research will be done in subsequent time points to explore the role of thrombin and PAR1 in brain injury.