SPTSSA variants alter sphingolipid synthesis and cause a complex hereditary spastic paraplegia

Abstract

Sphingolipids are a diverse family of lipids with critical structural and signalling functions in the mammalian nervous system, where they are abundant in myelin membranes. Serine palmitoyltransferase, the enzyme that catalyses the rate-limiting reaction of sphingolipid synthesis, is composed of multiple subunits including an activating subunit, SPTSSA. Sphingolipids are both essential and cytotoxic and their synthesis must therefore be tightly regulated. Key to the homeostatic regulation are the ORMDL proteins that are bound to serine palmitoyltransferase and mediate feedback inhibition of enzymatic activity when sphingolipid levels become excessive. Exome sequencing identified potential disease-causing variants in SPTSSA in three children presenting with a complex form of hereditary spastic paraplegia. The effect of these variants on the catalytic activity and homeostatic regulation of serine palmitoyltransferase was investigated in human embryonic kidney cells, patient fibroblasts and Drosophila. Our results showed that two different pathogenic variants in SPTSSA caused a hereditary spastic paraplegia resulting in progressive motor disturbance with variable sensorineural hearing loss and language/cognitive dysfunction in three individuals. The variants in SPTSSA impaired the negative regulation of serine palmitoyltransferase by ORMDLs leading to excessive sphingolipid synthesis based on biochemical studies and in vivo studies in Drosophila. These findings support the pathogenicity of the SPTSSA variants and point to excessive sphingolipid synthesis due to impaired homeostatic regulation of serine palmitoyltransferase as responsible for defects in early brain development and function.

Keywords: SPTSSA; ORMDLs; hereditary spastic paraplegia; serine palmitoyltransferase; sphingolipids.

Figure 1

SL biosynthesis and HSP-causing variants in the SPTSSA subunit of SPT. (A) SPT catalyses the first and rate-limiting step of sphingolipid synthesis, the condensation of serine with an acyl-CoA (typically, palmitoyl-CoA) to generate the sphingoid base, 3-KDS, which is further modified to form the complex family of SLs. SPT is feedback inhibited by the ORMDL proteins. The dihydroceramide-based SLs [dihydroceramide (dhCer), dihydrohexosylceramide (dhHexCer) and dihydrosphingomyelin (dhSM)] do not usually accumulate to appreciable levels, but excessive SPT activity results in significant elevation of these SLs. (B) The SPT/ORMDL3 complex is a dimer of SPTLC1/SPTLC2/SPTSSA/ORMDL3 tetramers. The enlarged image (right) shows that Thr51 of SPTSSA is in close contact with the luminal end of the first (of four) transmembrane domain (TM1) of ORMDL3. (C) Pedigrees showing de novo occurrence of the heterozygous SPTSSA p.Thr51Ile variant in two families and the inherited p.Gln58AlafsTer10 variant in the third family. Filled symbols denote probands; dots in the centre of the circle or square denote carriers. (D) Thr51 (indicated by triangle) of the evolutionarily conserved SPTSSA subunit is located at the luminal end of the single transmembrane domain (TMD).

Figure 2

Clinical characteristics of patients with SPTSSA-related disorder. (A) Variable findings on brain MRI among affected individuals with SPTSSA variants, including cerebral/cerebellar volume loss and white matter changes. Axial and sagittal images shown are from Patient 1, who has ventriculomegaly and slightly depressed white matter volume. In Patient 2, at age 2 years, magnetic resonance spectroscopy of both grey (caudate) and white (parietal-occipital) matter voxels showed decreased NAA and increased lactate signals. At age 4 years, there was interval development of cerebellar atrophy and mild progressive cerebral volume loss with T₂ hyperintensity in the deep and periventricular white matter bilaterally. In Patient 3, scans at age 12 years showed mild thinning of the corpus callosum but structures were within the normal range. (B) Facial dysmorphisms of affected individuals with the SPTSSA p.Thr51Ile variant. (C) Threshold BAER (brainstem auditory evoked response) evaluation showed elevated thresholds for hearing, as determined by presence/absence of wave V when measured using a 2 kHz tone burst at various stimulation levels, in Patient 1. Shown is response from the left ear with a threshold of 35 dBnHL (decibels normal hearing level).

Figure 3

Effect of the SPTSSA variants on expression and activation of the SPTLC1/SPTLC2 heterodimer. (A) Plasmids expressing HA-tagged WT, SPTSSA^T51I (T51I), SPTSSA^58fs (58FS) or SPTSSA^CΔ14 (C^Δ14) were transfected into HEK SPTSSA KO cells, and microsomal SPT activity was determined as described in the ‘Materials and methods’ section. (B) Microsomal proteins from the HEK SPTSSA KO cells expressing the indicated SPTSSA variants (as in A) were separated by SDS PAGE, and the indicated proteins were analysed by immunoblotting. (C) Microsomes prepared from SPTSSA KO cells co-expressing SPTLC1-FLAG, SPTLC2 and either HA-SPTSSA and SPTSSA^T51I (left) or SPTSSA and HA-SPTSSA^T51I (right) were solubilized and bound to anti-FLAG beads. The immunoprecipitated proteins were analysed by immunoblotting. (D) Microsomes prepared from Patients 1 and 2 (P1 and P2) and age-matched control (Con) fibroblasts were assayed for SPT activity as in A. (E) SPT activity in microsomes from Patient 3 (P3), his carrier mother and his homozygous WT brother was determined as in A.

Figure 4

The SPTSSA variants increase SPT activity. (A) SLs were extracted from serum of patients (P1, P2 and P3) and 14 unaffected controls (Con) and the indicated SL species were quantified by LCMS as described in the ‘Materials and methods’ section. (B) HEK SPTSSA KO cells were transfected with a plasmid expressing WT SPTSSA or empty vector and after 24 h fresh medium containing d2-serine was added; 24 h later newly synthesized d2-labelled SLs were quantitated by LCMS. (C) Plasmids expressing WT, p.Thr51Ile (T51I), or p.Gln58AlafsTer10 (58fs) SPTSSA were transfected into the HEK SPTSSA KO cells and de novo SL synthesis was analysed as in B. (D) Patient P1 and P2 and control (Con) fibroblasts were incubated for 24 h in medium containing d2-serine, and labelled SLs were quantitated by LCMS as in B. (E) Fibroblasts from Patient 3 (P3), his carrier mother and his homozygous WT brother were incubated with d2-serine and SLs were extracted and quantitated as in D. Unless indicated as not significant (ns), all differences were significant (P < 0.05).

Figure 5

The SPTSSA variants impair ORMDL-mediated inhibition of SPT. (A) Plasmids expressing SPTLC1, SPTLC2 and WT, p.Thr51Ile (T51I), or p.Gln58AlafsTer10 (58fs) SPTSSA were transfected into HEK cells along with increasing amounts (0, 50, 100 or 200 ng) of ORMDL3-expressing plasmid. After 16 h, d2-serine was added, and 24 h later cells were harvested and deuterated SLs were quantitated by LCMS. (B) HEK SPTSSA KO cells were transfected with plasmids expressing WT, p.Thr51Ile (T51I), or p.Gln58AlafsTer10 (58fs) SPTSSA. After 24 h, fresh medium without (−) or with (+) 3 mM serine was added, and 24 h later cells were harvested for SL analyses by LCMS. (C) Microsomes prepared from patient (P1, P2) and control (Con) fibroblasts were assayed for SPT activity without (−) or with (+) added C8-ceramide(Cer)/BSA as described in the ‘Materials and methods’ section. (D) HEK SPTSSA KO cells were transfected with plasmids expressing WT, p.Thr51Ile (T51I) or p.Gln58AlafsTer10 (58fs) SPTSSA and SiRNAs (+Si) directed to ORMDL1, 2 and 3 or a control scrambled siRNA. After 24 h, the medium was changed, d2-serine was added for 24 h, cells were harvested and deuterated SLs were quantitated by LCMS. Unless indicated as not significant (ns), all differences were significant (P < 0.05).

Figure 6

Increased SPT activity causes neurological defects in Drosophila. (A) Schematic of the constructs used in Drosophila that overexpresses human SPT and ORMDL3. (B) Neuronal overexpression of scSPT causes shortened lifespan. ORMDL3 expression rescues the lifespan in flies expressing WT scSPT (scSPTRef) but not in flies expressing the p.Thr51Ile scSPT variant (scSPTp.T51I). (C) Flies co-expressing ORMDL3 and scSPTp.T51I have a shortened lifespan when compared to those co-expressing ORMDL3 and scSPTRef. (D) Neuronal expression of scSPT causes severe climbing defects. ORMDL3 expression recovers the climbing capability in flies expressing scSPTRef but not those expressing scSPTp.T51I. The elav>Empty flies were tested to document the climbing capability of the proper control flies. (E) Flies co-expressing ORMDL3 and scSPTp.T51I exhibit severely compromised climbing capabilities when compared to flies co-expressing ORMDL3 and scSPTRef. The da>Empty flies were tested as controls. (B–E) n values are indicated in each panel. (F) Expression of scSPT causes elevated production of C18, but not C14 and C16 LCBs. Co-expression of ORMDL3 fully suppresses the C18 LCB production induced by scSPTRef, but only partially suppresses the C18 production induced by scSPTp.T51I. (G) Expression of scSPT causes production of C18 ceramide and C18 CPE, two complex SL species commonly observed in Drosophila. Data are presented as Kaplan–Meier curves and analysed using Gehan–Breslow–Wilcoxon test and log-rank test (B and C) or represented as mean ± SEM and analysed using unpaired Student’s t-test (D–G). Unless indicated as not significant (ns), all differences were significant (P < 0.05).