An autosomal dominant neurological disorder caused by de novo variants in FAR1 resulting in uncontrolled synthesis of ether lipids

Abstract

Purpose: In this study we investigate the disease etiology in 12 patients with de novo variants in FAR1 all resulting in an amino acid change at position 480 (p.Arg480Cys/His/Leu).

Methods: Following next-generation sequencing and clinical phenotyping, functional characterization was performed in patients' fibroblasts using FAR1 enzyme analysis, FAR1 immunoblotting/immunofluorescence, and lipidomics.

Results: All patients had spastic paraparesis and bilateral congenital/juvenile cataracts, in most combined with speech and gross motor developmental delay and truncal hypotonia. FAR1 deficiency caused by biallelic variants results in defective ether lipid synthesis and plasmalogen deficiency. In contrast, patients' fibroblasts with the de novo FAR1 variants showed elevated plasmalogen levels. Further functional studies in fibroblasts showed that these variants cause a disruption of the plasmalogen-dependent feedback regulation of FAR1 protein levels leading to uncontrolled ether lipid production.

Conclusion: Heterozygous de novo variants affecting the Arg480 residue of FAR1 lead to an autosomal dominant disorder with a different disease mechanism than that of recessive FAR1 deficiency and a diametrically opposed biochemical phenotype. Our findings show that for patients with spastic paraparesis and bilateral cataracts, FAR1 should be considered as a candidate gene and added to gene panels for hereditary spastic paraplegia, cerebral palsy, and juvenile cataracts.

Fig. 1. Ether lipids and ether lipid metabolism.

a Molecular structures of glycerolipids and glycerophospholipids. These types of lipids contain a glycerol backbone (red) attached to one or two fatty acid/fatty alcohol (blue) substituents at the sn-1 and sn-2 position and a headgroup R (green) attached at the sn-3 position. When R is a hydrogen or a fatty acid the molecule is a glycerolipid; when R is a phosphate linked to a species characteristic headgroup (choline, ethanolamine, etc.) the molecule is a glycerophospholipid, also known as a phospholipid. In ether phospholipids, the fatty linkage at the sn-1 position is either an ether (plasmanyl) or a vinyl ether (plasmenyl); the latter substitution defines the subclass that is known as plasmalogens. In analogy to ether phospholipids, ether variants of glycerolipids also exist with an ether fatty linkage on the sn-1 position. b Ether lipid biosynthesis. Dihydroxyacetone phosphate (DHAP)-acyltransferase (DHAPAT), esterifies DHAP at the sn-1 position with a long-chain fatty acyl-CoA yielding 1-acyl-DHAP. Next, alkylglycerone phosphate synthase (AGPS) replaces the acyl unit by a fatty alcohol, produced by reduction of an acyl-CoA by fatty acyl-CoA reductase 1 (FAR1), resulting in the formation of 1-alkyl-DHAP. 1-Alkyl-DHAP is then reduced to 1-alkylglycerolphosphate by acyl/alkyl-DHAP reductase (PexRAP), which is converted to 1-alkyl-2-acylglycerol (DG[O]) at the endoplasmic reticulum (ER). Exogenous 1-alkylglycerol (for example 1-hexadecylglycerol [HDG]) can be phosphorylated to 1-alkylglycerolphosphate, which then bypasses the peroxisomal steps in ether lipid biosynthesis. DG[O] can be converted to plasmanyl-PE and plasmanyl-PC by condensation with CDP-ethanolamine and CDP-choline, respectively. Creation of a vinyl ether by desaturation in the ER yields plasmenyl-PE, which can be converted to plasmenyl-PC, both of which are called plasmalogens. PC and PE ether phospholipids, thus including plasmanyl and plasmenyl-species, are collectively called PC[O] and PE[O], respectively. DG[O] can also be acylated which yields 1-alkyl-2,3-diacylglycerol (TG[O]).

Fig. 2. Functional consequences of FAR1 p.Arg480His/Cys variants.

a Plasmalogen levels in cultured skin fibroblasts from a reference population (Ref), controls (C) (Ref and C, both black diamonds) and FAR1 p.Arg480His/Cys patients (gray, P1 = square, P2 = circle, P3 = triangle). The control samples were analyzed in the same experiments as the patient samples. C16:0- and C18:0-plasmalogens are expressed as % of total palmitate (C16:0) and stearate (C18:0), respectively, and were significantly elevated when compared with the control samples (t-test, two sided, p < 0.0001, indicated by **). b FAR1 enzyme activity in cultured skin fibroblasts of controls (C1–C3, black) and patients (P1–P3, gray). Measurements were performed in duplicate (except for C2). c Immunoblot analysis of FAR1 in fibroblast homogenates of controls and patients. The upper panel shows the (cropped) immunoblot using antibodies against FAR1 (59 kDa) and the lower panel the same blot reprobed with antibodies against β-actin as loading control. d Immunofluorescence microscopy analysis in fibroblasts of P2 of peroxisomal membrane protein ABCD3 (green, upper left) and FAR1 (red, upper right) and overlay (lower left) and magnification (lower right). FAR1 p.Arg480His/Cys variants are normally localized to peroxisomes.

Fig. 3. Effect on FAR1 protein level by treatment of fibroblasts with 1-O-hexadecyl-sn-glycerol (HDG).

Fibroblasts of control subjects (n = 6) and p.Arg480His/Cys FAR1 patients (P1–P3) were treated for 24 and 48 hours with 10 μM HDG followed by immunoblot analysis for FAR1 and measurement of the plasmalogen levels. Two independent experiments were performed (^a indicating the first experiment, ^b the second), each including three (×2) control cell lines and the three patient cell lines (P1 = square, P2 = circle, P3 = triangle). a Immunoblot analysis of FAR1 in fibroblast homogenates of controls and patients using antibodies against FAR1 (59 kDa) and β-actin as loading control (same blot reprobed). Blots (cropped) of both experiments are shown. b Quantification of the FAR1 protein level as ratio over β-actin in control and patient cells in untreated and HDG-treated cells. Results are shown for each cell line as % of the untreated condition. c C16:0-Plasmalogen levels (expressed as % of total palmitate [C16:0]) in untreated and HDG-treated control and patient cells.

Fig. 4. Lipidomics in patient fibroblasts with the p.Arg480His/Cys FAR1 variants.

Summation of lipidomic species per major class for controls (n = 3, black diamonds), patients 1, 2, and 3, (P1 = square, P2 = circle, P3 = triangle); mean is shown and x-fold difference of the patient mean compared with that of controls if p ≤ 0.005. a Diradylphospholipid species, b monoradylphospholipid species, and c neutral lipid species (“radyl” means either a fatty acid or fatty alcohol substituent). DG diacylglycerol, DG[O] 1-alkyl-2-acylglycerol, (L)PC (lyso)phosphatidylcholine, (L)PC[O] (lyso)plasm(a/e)nylcholine, (L)PE (lyso)phosphatidylethanolamine, (L)PE[O] (lyso)plasm(a/e)nylethanolamine, TG triacylglycerol, TG[O] 1-alkyl-2,3-diacylglycerols. d LPC[O] levels 16:0–18:0 are elevated in patients with the FAR1 variants (¹⁰log-scale). e ²Log(fold-change) of the average levels of patients and controls for PC[O] and PC species with 40–44 carbon atoms in the fatty acid side chains. PC[O] species with more double bonds (containing polyunsaturated fatty acids) accumulated more in patients whereas the corresponding PC species were reduced. f Overview of the experiment to assess ether lipid and nonether lipid synthesis from exogenous C17:0-acid in control fibroblasts and patient fibroblasts with the p.Arg480His/Cys FAR1 variants. C17:0-acid can be directly incorporated in nonether lipids or converted by FAR1 to a C17:0-alcohol which in turn can be incorporated in ether lipids. g Comparison of incorporation of C17:0-acid and C17:0-alcohol in LPC(17:0) and LPC(O-17:0), respectively. C17:0-acid is incorporated in comparable amounts in LPC(17:0) in both controls and patients whereas C17:0-alcohol incorporation in LPC(O-17:0) is fourfold elevated in patients when compared with control.