Cerebral adrenoleukodystrophy: a demyelinating disease that leaves the door wide open: Figure 1

This scientific commentary refers to ‘Brain endothelial dysfunction in cerebral adrenoleukodystrophy’ by Musolino et al. (doi: 10.1093/awv250) Cerebral adrenoleukodystrophy (ALD) can be distinguished from other inherited diseases of white matter (known collectively as ‘leukodystrophies’) by the presence of severe neuroinflammation and blood–brain barrier disruption, evidenced in brain MRI by the leakage of gadolinium-diethylenetriamine pentaacetic acid (Gd-DPTA) at the periphery of demyelinating lesions on T1-weighted sequences. In contrast to multiple sclerosis, blood–brain barrier breakdown occurs downstream of the demyelinating process. Moreover, it is always followed by very rapid progression of demyelinating lesions, likely resulting from myelin destruction by inflammatory cells. Mechanisms by which loss of ABCD1 function in brain endothelial cells, and also in astrocytes, microglia and pericytes, could disrupt the blood–brain barrier in cerebral ALD and enhance neuroinflammation. Loss of ABCD1 function (a fatty acid transporter localized to peroxisomes) in brain endothelial cells increases the expression of ICAM-1 and VCAM, facilitating the adhesion of monocytes at the cell surface. It also displaces CLDN5 (Claudin-5) and ZO-1 from the cell surface to the cytosol, resulting in increased blood-brain barrier (BBB) permeability. Synthesis of MMP9 contributes to complete disruption of the blood–brain barrier, enhancing the entry of monocytes and T and B lymphocytes into the white matter. Loss of ABCD1 function in pericytes could give rise to abnormal crosstalk between pericytes and endothelial cells, mediated by bidirectional TGF-β–TGF-β receptor signalling. ABCD1-deficiency in astrocytes might impair sonic hedgehog (SHH), Tie2 and AT1 signalling, and thus the expression and localization of CLDN5 and occludin. TGF-β1 produced by astrocytes could downregulate the expression of CLDN5 through transcriptional inhibition of MYC. Lastly, ABCD1-deficient microglia release IL-1β, which downregulates ZO-1 expression, which could in turn abolish the protective effect of astrocytes mediated by SHH signalling. In this issue of Brain, Musolino et al. (2015) unravel for the first time some of the molecular mechanisms that underlie blood–brain barrier breakdown in ALD. Using ALD brain tissue, and human brain microvascular cells (HBMECs) in which the ABCD1 (ALD) gene was silenced via siRNA, they demonstrate an upregulation of intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion protein 1 (VCAM1) expression that reflected endothelial activation and that facilitated monocyte adhesion. They also show an increase in the expression of transforming growth factor β1 (TGF-β1), and in the number of vessels, probably due to production of vascular endothelial growth factor (VEGF) by astrocytes and/or microglia. Furthermore, they reveal an increase in the expression of matrix metallopeptidase 9 (MMP9), which contributed to blood–brain barrier disruption, and in turn resulted in leakage of fibrinogen (an exclusively intravascular protein) into the perivascular space, recruitment of monocytes/macrophages and the activation of microglia around vessels in active demyelinating areas. This correlated with the displacement of two members of the endothelial tight junction complex, claudin 5 (CLDN5) and ZO-1, from the membrane to the cytoplasm of endothelial cells: a clear indication that the blood–brain barrier had been breached, boosting the ingress of proinflammatory lymphocytes and monocytes into the white matter. Importantly, ABCD1 silencing in HBMECs also displaced CLDN5 to the endothelial cell cytoplasm, as in vivo. Studies in HBMECs provided two key additional pieces of information: (i) the silencing of ABCD1 resulted in transcriptional inhibition of CLDN5 and MYC (also known as c-MYC); and (ii) deleterious effects in ABCD1-silenced HBMECs were seen before any increase in very long chain fatty acids (VLCFAs) could be observed. ABCD1 encodes one half of an ATP-binding transporter that homodimerizes within the membranes of peroxisomes and which is thought to import CoA derivatives of VLCFAs (fatty acids with >22 carbon atoms) into these organelles, wherein they are degraded by β-oxidation. All mutations of ABCD1 result in loss of ABCD1 transporter function and all patients with ALD accumulate VLCFAs in their plasma, cells and tissues. The mechanism by which CLDN5 gene expression is downregulated in ABCD1-silenced HBMECs is puzzling. Using a PCR-array screen of only 84 genes involved in neuroinflammation, Musolino and colleagues found that MYC gene expression was markedly decreased. Western blot confirmed that c-MYC protein expression was also reduced. Silencing cMYC in HBMECs changed the expression of CLDN5 and ICAM1 in a similar manner to that seen in ABCD1-silenced HBMECs. In macrophages, inhibition of MYC blocks the expression of MMP9 and TGF-β1 (which are upregulated in ALD brain tissue and ABCD1-silenced HBMECs) but upregulates the IL4 signalling mediator ‘signal transducer and activator of transcription 6’ (STAT6) (Pello et al., 2012). Through the downregulation of STAT6, a decrease in c-MYC could potentially downregulate the expression of CLDN5 (Dalmasso et al., 2014). c-MYC and TGF-β signalling are mutually antagonistic. c-MYC supresses the activation of TGF-β-induced genes, whereas TGF-β represses MYC transcription (Smith et al., 2009). Thus, c-MYC downregulation in cerebral ALD could be a downstream effect of TGF-β upregulation. In rat brain endothelial cells, pharmacological inhibition of the TGF-β receptor ‘activin receptor-like kinase-5’ (ALK5), which mediates the effects of TGF-β1, increases the expression of CLDN5 (Ronaldson et al., 2009). In contrast, TGF-β produced by pericytes induces CLDN5 expression (Dohgu et al., 2005). Given these conflicting data, it would be interesting to determine the effects of MYC silencing and pharmacological inhibition of ALK5 in HBMECs, possibly also in co-culture with normal and ABCD1-deficient astrocytes. While there is no path linking c-MYC and ABCD1, it would also be worthwhile to determine if the loss of ABCD1 gene function in brain endothelial cells could mediate the downregulation of c-MYC expression through changes in the microRNAs regulating its transcription (Jackstadt et al., 2015). Since the discovery (by serendipity) that VLCFAs accumulate in tissues, cells and plasma of patients with ALD, the accepted dogma has been that all pathogenic processes observed in ALD (cerebral demyelination, spinal cord axonopathy, adrenal insufficiency) are due to the ‘toxic’ accumulation of these fatty acids. There is still only indirect evidence that the ABCD1 transporter imports VLCFA-CoA into peroxisomes. In fact, the ABCD1 transporter seems to import a wide range of fatty acid CoAs. The changes in CLDN5 expression and localization observed in ABCD1-silenced HBMECs were seen before any increase in C26:0 lysophosphatidylcholine (C26:0-LPC) could be detected. Incubation of HBMECs with an excess of C26:0-LPC had no effect. Cerotic acid (C26:0) is the VLCFA that accumulates at highest levels in ALD. As with other VLCFAs, it accumulates mostly as an acetylated form of various lipids. One can argue that in ABCD1-silenced HBMECs, C26:0 accumulated in lipid fractions other than lysophosphatidylcholine, possibly even as an acylated form of certain proteins. Nevertheless, in the absence of definitive evidence that the ABCD1 transporter imports VLCFA-CoA, the question remains as to whether the effects observed in ABCD1-silenced HBMECs could be mediated by a metabolite other than VLCFA-CoA. This seminal work by Musolino and colleagues opens the door to many additional questions. In particular, could the loss of ABCD1 gene function in pericytes and astrocytes also contribute to the blood–brain barrier disorganization? The brain endothelium’s abluminal surface is covered by a basement membrane in which pericytes and their processes are embedded, allowing direct intercellular crosstalk via peg-socket junctions. Is ABCD1 expressed in pericytes and, if so, does the loss of ABCD1 gene function result in dysfunction of these cells too? Interaction between endothelial cells and pericytes is mediated by bidirectional TGF-β–TGF-β receptor signalling, which results in the upregulation of endothelial cadherin-2; this in turn leads to firm adhesion between endothelial cells and pericytes, and the deposition by pericytes of extracellular matrix components that contribute to basement membrane formation. To be effective, the crosstalk between endothelial cells and pericytes must be normal. Loss of ABCD1 gene function in astrocytes might also contribute to increased blood––brain barrier permeability. Astrocytes release sonic hedgehog (SHH), which activates Hedgehog signalling in brain endothelial cells through the receptor patched 1 (PTCH1) resulting in the upregulation of occludin and CLDN5 (Alvarez et al., 2011). Netrin 1, which is a downstream effector of SHH signalling, is also an important regulator of blood–brain barrier maintenance (Podjaski et al., 2015). Activation of Src-suppressed C-kinase substrate (SSeCKS, encoded by AKAP12) in astrocytes stimulates angiopoietin 1 (ANG1) production, which signals back to endothelial Tie2 receptors, upregulating and controlling the subcellular distribution of tight junction proteins (Lee et al., 2003). Processing of angiopoietin 1 by angiotensin-converting enzyme in astrocytes additionally results in the production of effector molecule angiopoietin 2 (ANG2), which binds to the endothelial AT1 receptor. Activation of AT1 modulates blood–brain barrier function through threonine-phosphorylation of occludin that modifies its subcellular localization (Wosik et al., 2007). Other key players in the maintenance of blood–brain barrier function include the microglial cells. Abnormal interleukin 1 (IL1) immunoreactivity is observed in microvessels and astrocytes from ALD brain samples (Powers et al., 1992), while ALD lymphoblasts produce increased levels of IL-1β (Uto et al., 2008). Activated microglia accumulate around vessels and release of IL-1β could abolish the protective effect of astrocytes on blood–brain barrier integrity by suppressing the astrocytic activation of hedgehog signalling (Wang et al., 2014) and downregulating the expression of ZO-1, which binds to CLDN5 and occludin. Given the interplay between brain endothelial cells, pericytes, astrocytes and microglia required to maintain blood–brain barrier integrity, experiments using co-culture of normal and ABCD1-deficient endothelial cells, astrocytes, pericytes and even microglia will likely be necessary to decipher the mechanisms that open the blood–brain barrier in cerebral ALD. One of the most interesting findings of Musolino and co-workers relates to processes that could trigger the initial demyelinating event. They observed that, in contrast to multiple sclerosis, microglias are recruited along the vessels in cerebral regions in which there is no obvious demyelination or blood–brain barrier breakdown. Why many, but not all, males with ALD (∼65%) develop cerebral demyelination remains an enigma. Neither the identity of the ABCD1 mutation nor VLCFA levels in blood, fibroblasts or tissue can predict the risk of developing cerebral demyelination in a given patient with ALD. The mechanisms by which allogeneic haematopoeitic stem cell transplantation (or haematopoeitic stem cell gene therapy) can arrest the progression of cerebral demyelination are also unknown. Both therapeutic approaches ultimately replace ABCD1-deficient microglia with normal microglial cells. Loss of ABCD1 function in oligodendrocytes is necessary but not sufficient to develop cerebral demyelination and there is no animal model that mimics the cerebral demyelination of ALD. It is possible that abnormal interaction between ABCD1-deficient microglia and brain endothelial cells could be a key event that triggers the failure of oligodendrocytes to maintain the integrity of the myelin sheath. Hopefully, co-culture of ALD iPS-derived oligodendrocytes, microglia and ABCD1-silenced HBMECs could address this important issue.