Cerebrospinal fluid matrix metalloproteinases are elevated in cerebral adrenoleukodystrophy and correlate with MRI severity and neurologic dysfunction

Abstract

Background: X-linked adrenoleukodystrophy results from mutations in the ABCD1 gene disrupting the metabolism of very-long-chain fatty acids. The most serious form of ALD, cerebral adrenoleukodystrophy (cALD), causes neuroinflammation and demyelination. Neuroimaging in cALD shows inflammatory changes and indicates blood-brain-barrier (BBB) disruption. We hypothesize that disruption may occur through the degradation of the extracellular matrix defining the BBB by matrix metalloproteinases (MMPs). MMPs have not been evaluated in the setting of cALD.

Methodology/principal findings: We used a multiplex assay to correlate the concentration of MMPs in cerebrospinal fluid and plasma to the severity of brain inflammation as determined by the ALD MRI (Loes) score and the neurologic function score. There were significant elevations of MMP2, MMP9, MMP10, TIMP1, and total protein in the CSF of boys with cALD compared to controls. Levels of MMP10, TIMP1, and total protein in CSF showed significant correlation [p<0.05 for each with pre-transplant MRI Loes Loes scores (R(2) = 0.34, 0.20, 0.55 respectively). Levels of TIMP1 and total protein in CSF significantly correlated with pre-transplant neurologic functional scores (R(2) = 0.22 and 0.48 respectively), and levels of MMP10 and total protein in CSF significantly correlated with one-year post-transplant functional scores (R(2) = 0.38 and 0.69). There was a significant elevation of MMP9 levels in plasma compared to control, but did not correlate with the MRI or neurologic function scores.

Conclusions/significance: MMPs were found to be elevated in the CSF of boys with cALD and may mechanistically contribute to the breakdown of the blood-brain-barrier. MMP concentrations directly correlate to radiographic and clinical neurologic severity. Interestingly, increased total protein levels showed superior correlation to MRI score and neurologic function score before and at one year after transplant.

Figure 1. Boys with cALD have elevated CSF protein levels.

Total protein levels were determined in the clinical chemistry laboratory using a Kodak Ektachem method. Shown are the means ± S.D. A Student’s t-test was performed to generate the p-value.

Figure 2. Boys with cALD have higher levels of CSF MMPs and TIMP1 than controls.

CSF MMPs and TIMPs were evaluated using a Luminex system. Statistical analysis was performed using Prism software. The data displayed are those MMPs and TIMP1 that were significantly different from the control group. The horizontal bar represents the mean.

Figure 3. Boys with cALD have higher levels of plasma MMP9 than controls.

Plasma MMP9 was evaluated using a Luminex system as before. Statistical analysis was performed using Prism software (version 5.0d). The horizontal bar represents the mean.

Figure 4. Levels of MMP10, TIMP1, and CSF total protein correlate with pre-HCT Loes MRI Score.

Regression analysis was performed using the previously determined MMP and TIMP1 levels and each patient’s pre-transplant Loes score. P-values and R² were determined using Prism (version 5.0d).

Figure 5. Levels of CSF TIMP, MMP10, and total protein correlate with neurologic score pre- and post-HSCT.

Regression analysis was performed using the previously determined MMP and TIMP levels and each patient’s pre- and post-HCT neurologic function score. The post-HSCT scores were determined at the one year post-HCT follow-up visit. P-values and R² were determined using Prism software (version 5.0d).