Cell-to-cell signaling between neuronal, glial, and vascular cells of the neurovascular unit, sometimes called the ‘neurovascular module’, is required for normal functioning of the CNS, and is important for the disease process (Zlokovic 2008; Guo and Lo 2009; Winkler et al. 2011). In contrast to the highly permeable capillaries in the systemic circulation (Mann et al. 1985), blood–CNS vascular barriers normally isolate effectively CNS from systemic influences and prevent entry into the brain of circulating blood cells, toxic plasma components, and large molecules and polar solutes without specific transport systems (Zlokovic et al. 1987; Zlokovic and Apuzzo 1997). The concept of neurovascular unit has been initially developed related to an acute CNS injury such as acute ischemic stroke and brain trauma (Moskowitz et al. 2010), but more recently is gaining well-deserved recognition in the fields of chronic neurodegenerative disorders such as, for example, Alzheimer's disease and amyotrophic lateral sclerosis (Zhong et al. 2008; Bell et al. 2010, 2012; Zlokovic 2011; Winkler et al. 2012). Whether cell-to-cell interactions within the neurovascular unit that have been described to take place after an acute CNS injury are comparable with those observed during a chronic neurodegenerative process where injury and repair coexist over longer periods of time remains, however, unclear. In addition, it is still relatively poorly understood whether cellular interactions that occur in the gray matter of the brain can also occur in the white matter, which is importantly involved in neurodegeneration and dementia.
Hayakawa et al. (2012) propose that during white matter disease, activated astrocytes use high-mobility group box 1 (HMGB1) (Malarkey and Churchill 2012), a member of the damage-associated-molecular-pattern family of proteins, to attract circulating endothelial progenitor cells (EPCs) that then contribute to repair (Fig. 1). To investigate the role of astrocytic HMGB1 in this process, Hayakawa et al. injected lysophosphatidylcholine into the corpus callosum of mice to induce focal demyelination. Examination of these brains showed that HMGB1 expression was increased in the damaged white matter region at 5 days, with the majority of signals colocalizing with GFAP-positive reactive astrocytes. At the same time, flow cytometry analysis demonstrated an accumulation of Fkl-1/CD34-positive EPCs in these damaged areas. EPCs can release many neurotrophic factors, thus their accumulation after CNS injury is considered an important step for neurovascular repair and remodeling (Taguchi et al. 2004; O'Neill et al. 2005; Fan et al. 2010). But how do EPCs recognize the need to repair damaged or diseased white matter? Using a combination of molecular and pharmacologic approaches, the authors showed that these EPC responses require binding of HMGB1 onto the receptor for advanced glycation end products (RAGE) on EPCs. Interestingly, RAGE expression is increased in brain endothelium in situ under pathophysiological conditions associated with neurodegeneration (Deane et al. 2012). In cell cultures, HMGB1-RAGE signaling increased EPC migration and angiogenic function. In vivo, siRNA suppression of RAGE expression decreased EPC accumulation and neurovascular repair in the damaged white matter of mice.
Taken together, these experiments support a model whereby cross-talk between brain astrocytes and systemic EPC responses may contribute to the balance between injury and repair in damaged or diseased white matter. There may be two interesting implications. First, this study suggests that the neurovascular module is not an isolated CNS structure, but may also interact with the rest of the body. Second, these findings remind us that reactive astrocytes are not always deleterious and can display beneficial repair properties as well.
Nevertheless, like any interesting study, findings by Hayakawa et al. raise some important follow-up questions. The authors identified HMGB1 as the signal that allows cross-talk between the reactive astocytes and circulating EPCs. But surely, reactive astrocytes release many other factors as well. A full analysis of the astrocyte secretome is necessary before we can properly understand how these mechanisms are regulated. It is unlikely that HMGB1 will work by itself. Second, the in vivo model used here has many limitations. Lysophosphatidylcholine is a convenient way to induce focal white matter injury, but this is not a model that represents any CNS disease at all. Further studies using cerebral ischemia, brain trauma, or white matter inflammation as seen in the experimental allergic encephalitis models will be necessary to perform before conclusions can be made as to the clinical relevance of the present findings. Finally, a critical piece of evidence that is missing from this study pertains to the integrity and function of the white matter myelin and axonal compartments. The authors propose that HMGB1 signaling enables astrocytes to attract EPCs and augment vascular repair. But, they do not truly show whether this neurovascular response actually helps white matter function. Is myelination affected, and what happens to axonal structure and the ability to conduct neuronal signals? In the end, the main purpose of white matter is to connect wide-ranging neuronal networks. Without more information about axonal function, this study cannot truly reach any conclusions about the relevance of this hypothesis to actual white matter function or dysfunction.
The concept of the neurovascular module has served a useful role for investigating acute mechanisms of stroke and brain injury, and more recently Alzheimer's disease and amyotrophic lateral sclerosis. However, most studies to date have focused on the gray matter pathology. Here, Hayakawa et al. propose a gliovascular mechanism that may be important for the dynamic balance between injury and repair in the white matter disease. If proven true, this hypothesis may allow us to explore new directions for developing therapies directed at the white matter. Further studies are warranted.