Elevated cerebrospinal fluid glial fibrillary acidic protein levels in Smith-Lemli-Opitz syndrome

Abstract

Smith-Lemli-Opitz syndrome (SLOS) is a rare, multiple malformation/intellectual disability disorder caused by pathogenic variants of DHCR7. DHCR7 catalyzes the reduction of 7-dehydrocholesterol (7DHC) to cholesterol in the final step of cholesterol biosynthesis. This results in accumulation of 7DHC and a cholesterol deficiency. Although the biochemical defect is well delineated and multiple mechanisms underlying developmental defects have been explored, the post developmental neuropathological consequences of altered central nervous system sterol composition have not been studied. Preclinical studies suggest that astroglial activation may occur in SLOS. To determine if astroglial activation is present in individuals with SLOS, we quantified cerebrospinal fluid (CSF) glial fibrillary acidic protein using a Quanterix Simoa® GFAP Discovery Kit for SR-X™. Relative to an age-appropriate comparison group, we found that CSF GFAP levels were elevated 3.9-fold in SLOS (3980 ± 3732 versus 1010 ± 577 pg/ml, p = 0.0184). Simvastatin, a 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor, has previously been shown to increase expression of hypomorphic DHCR7 alleles and in a placebo-controlled trial improved serum sterol levels and decreased irritability. Using archived CSF samples from that prior study, we observed a significant decrease (p = 0.0119) in CSF GFAP levels in response to treatment with simvastatin. Although further work needs to be done to understand the potential contribution of neuroinflammation to SLOS neuropathology and cognitive dysfunction, these data confirm astroglial activation in SLOS and suggest that CSF GFAP may be a useful biomarker to monitor therapeutic responses.

Keywords: 7-dehydrocholesterol reductase; Biomarker; DHCR7; GFAP; Glial fibrillary acidic protein; Neuroinflammation; SLOS; Simvastatin; Smith-Lemli-Opitz syndrome.

Figure 1.. Cerebrospinal Fluid Glial Fibrillary Acidic Protein levels in Smith-Lemli-Opitz syndrome.

A) CSF GFAP levels in pediatric individuals with either ASD (n=10) or SLOS (n=37). B) Log₁₀ transformed CSF GFAP levels in pediatric individuals with either ASD (n=10) or SLOS (n=37). Mean log₁₀ CSF GFAP level was significantly elevated in SLOS (p=0.0184) relative to the comparison group. C) Correlation between CSF GFAP levels and age in pediatric individuals with SLOS (n=37, p=0.36, r²=0.02). D) Correlation between CSF GFAP levels and SLOS Severity Score in pediatric individuals with SLOS (n=37, p=0.0069, r²=0.19). Error bars represent one standard deviation. Dotted lines indicate the 95% confidence interval of the regression curve.

Figure 2.. Correlation of Cerebrospinal Fluid Glial Fibrillary Acidic Protein levels and Sterol Levels in Smith-Lemli-Opitz syndrome.

A) Negative correlation between CSF GFAP and concurrent serum cholesterol levels (n=37, p=0.0337, r²=0.12). B) No correlation was observed between CSF GFAP and concurrent dehydrocholesterol (7DHC + 8DHC) levels (n=37, p=0.18, r²=0.05). Isomerization of 7DHC gives rise to 8-dehydrocholesterol (8DHC). C) Positive correlation between CSF GFAP levels and concurrent dehydrocholesterol/total sterol ratio (n=37, p=0.0436, r²=0.11). Dotted lines indicate the 95% confidence interval of the regression curve.

Figure 3.. Cerebrospinal Fluid Glial Fibrillary Acidic Protein levels decrease in response to simvastatin therapy.

A) GFAP levels were measured in CSF collected at baseline, end of the placebo phase and end of the simvastatin treatment phase in a double-blinded cross-over trial. Mean baseline (n=10, 3956 ± 2791 pg/ml) and end of placebo-phase (n=16, 4001 ± 4486 pg/ml) values were similar. Mean end of simvastatin-phase CSF GFAP level was lower (861 ± 1155 pg/ml) B) CSF GFAP data was log₁₀ transformed to better approximate a normal distribution. No difference was observed between baseline and end of placebo-phase GFAP levels (p=0.98, t-test). In contrast, CSF GFAP levels were significantly decreased (p=0.0119. t-test) in samples obtained after the end of the simvastatin treatment phase relative to placebo. Error bars represent one standard deviation.