Post-mortem analyses in a patient with succinic semialdehyde dehydrogenase deficiency (SSADHD). II. Histological, lipid, and gene expression outcomes in regional brain tissue

Abstract

We have extended previous metabolic measures in post-mortem tissues (frontal and parietal lobes, pons, cerebellum, hippocampus, and cerebral cortex) obtained from a 37 y.o. male patient with succinic semialdehyde dehydrogenase deficiency (SSADHD) who expired from SUDEP (sudden unexplained death in epilepsy). Histopathological characterization of fixed cortex and hippocampus revealed mild to moderate astrogliosis, especially in white matter. Analysis of total phospholipid mass in all sections of the patient revealed a 61% increase in cortex and 51% decrease in hippocampus as compared to (n=2–4) approximately age-matched controls. Examination of mass and molar composition of major phospholipid classes showed decreases in phospholipids enriched in myelin, such as phosphatidylserine, sphingomyelin, and ethanolamine plasmalogen. Evaluation of gene expression (RT² Profiler PCR Arrays, GABA, glutamate; Qiagen) revealed dysregulation in 14/15 GABA_A receptor subunits in cerebellum, parietal, and frontal lobes with the most significant down-regulation in ε, θ, ρ₁ and ρ₂ subunits (7.7–9.9-fold). GABA_B receptor subunits were largely unaffected, as were ionotropic glutamate receptors. The metabotropic glutamate receptor 6 was consistently down-regulated (maximum 5.9-fold) as was the neurotransmitter transporter (GABA), member 13 (maximum 7.3-fold). For other genes, consistent dysregulation was seen for interleukin 1β (maximum down-regulation 9.9-fold) and synuclein α (maximal up-regulation 6.5-fold). Our data provide unique insight into SSADHD brain function, confirming astrogliosis and lipid abnormalities previously observed in the null mouse model while highlighting long-term effects on GABAergic/glutamatergic gene expression in this disorder.

Introduction

Succinic semialdehyde dehydrogenase (SSADH) deficiency (SSADHD) is an orphan heritable disorder of GABA metabolism (for pathway interrelationships, see Fig. 1 of the preceding article in this series; Kirby et al., 2020). Patients with SSADHD manifest a nonspecific neurological phenotype of global developmental delay, neuropsychiatric morbidity, absence of developed speech, and variable epilepsy (Malaspina et al, 2016). Diagnosis based upon clinical characteristics is unreliable but diagnosis can be obtained by metabolic analysis combined with molecular analysis of the ALDH5A1 gene (OMIM 610045; 271980; ALDH5A1=aldehyde dehydrogenase 5A1=SSADH). Although both open-label and placebo-controlled blinded trials have been completed (www.clinicaltrials.gov/NCT02019667; Schreiber et al., 2016), current therapeutics remain symptomatic, and predominantly targeting either behavior (OCD, ADHD) or seizure control. The sole therapeutic targeting the GABA pathway is the antiepileptic vigabatrin (VGB; irreversible inhibitor of GABA-transaminase; Fig. 1, Kirby et al., 2020), yet its clinical efficacy has been mixed and its risk of retinal toxicity complicates long-term consideration (Police et al., 2020). The patient was first reported by Haan et al (1985) and again recently by Kirby et al (2020) post-mortem. A brief summary of clinical details is presented in Materials and Methods. Characterization of the pathophysiology of SSADHD has been significantly enhanced with the development of a knockout mouse model in 2001 (Hogema et al., 2001). Electrophysiological studies in this model identified a pattern of seizure evolution, starting with absence seizures and progressing to generalized tonic-clonic convulsions with eventual status epilepticus (Wu et al., 2006; Buzzi et al., 2006; Cortez et al., 2004). Metabolic characterization has identified significant amino acid disturbances, imbalance of neurotransmitters and neurotransmitter precursors (elevated GABA, the GABA-analogue gamma-hydroxybutyrate, aspartate; depleted glutamine; Kirby et al., 2020) that are accompanied by significant down-regulation of GABA receptors (Gupta et al., 2004; Gibson et al., 2002). Many metabolic abnormalities detected in the murine model have been confirmed in patient physiological fluids (Jansen et al., 2016; Vogel et al., 2018). Vogel and coworkers (2016) verified disruption of GABAergic/glutamatergic receptor gene expression in brain of the mouse model that was responsive to mTOR inhibition. Until now, however, CNS tissue from an SSADHD patient has not been available for confirmatory studies.

Fig. 1. GFAP (glial fibrillary acidic protein) immunohistochemistry of fixed parietal cortex from the patient.

A-subpial and layer I; B-layers V/VI; C(LFB/H&E), D-subcortical white matter with displaced neurons (arrow heads). A, B 200x mag. C, D 400x mag.

In the first paper in this series (Kirby et al., 2020), amino acids, acylcarnitines, guanidine-species (guanidinoacetic acid, creatine, creatinine) and GABA-related intermediates were quantified in human post-mortem frontal and parietal lobes, pons, cerebellum, hippocampus, cerebral cortex, liver, and kidney. Amino acid analyses revealed significant elevation of aspartic acid and depletion of glutamine in the patient, disruption of short-chain fatty acid metabolism, and elevation of GABA-related metabolites (GABA, γ-hydroxybutyrate, succinic semialdehyde, 4-guanidinobutyrate, homocarnosine; Fig. 1 in Kirby et al., 2020), accompanied by disturbances of the creatine biosynthetic pathway. In the current report, we extend these metabolic studies to examine histopathology, lipid content, and gene expression in the same tissues. The following hypotheses were evaluated: 1) patient cortex/hippocampus would reveal astrogliosis (Hogema et al., 2001; Brown et al., 2020); 2) ethanolamine glycerophospholipid content in patient brain sections would be decreased (Donarum et al., 2006; Barcelo-Coblijn et al., 2007); and 3) down-regulation of GABAergic/glutamatergic receptor subunits would be observed (Kirby et al., 2020; Vogel et al., 2016; Buzzi et al., 2006; Wu et al., 2004; Cortez et al., 2004).

Materials and Methods

Informed Consent, Patient and Control Tissue Procurement

The studies herein reported were approved by the Human Research Ethics Committee of the Royal Melbourne Hospital, Parkville, Victoria 3050, Australia, as well as the Institutional Review Board of the National Disease Research Interchange, Philadelphia, PA, USA. The family of the patient consented both to autopsy as well as procurement of tissues for experimental investigations and consented to publication of the data.

The patient was first reported in 1985 (Haan et al., 1985). In childhood, the patient’s clinical course was typical of SSADHD, featuring global developmental delays, hypotonia, ataxia, and an absence of formalized speech. Throughout adolescence and into adulthood (2^nd–3^rd decades of life), his clinical history included epilepsy, autism, and an underactive thyroid and a diagnosis of Hashimoto thyroiditis. He was treated with risperidone. During respite care at the age of 37, hee expired overnight with a post-mortem diagnosis of SUDEP. Because his diagnosis was known, follow-up studies of EEG, MRI nor CT were not undertaken. Following a period of almost 72 hours (cadaver stored at 4°C), autopsy and tissue harvest was undertaken. Description of patient tissues and procurement of control tissues has been reported (Kirby et al., 2020). Samples of the peripheral nervous system, striatal muscle, spinal cord, and peripheral nerves were not obtained at autopsy. More in-depth evaluation of the brainstem, including the nucleus/fasciculus solitarius and the nucleus ambiguus, were also not biopsied at autopsy. It is not possible to make completely confident comparisons when contrasting analytical studies in tissues that were approximately 72 hours of age at harvest (patient) as compared to tissues harvested within 24 hours (controls; Kirby et al., 2020). Since only a single measure was afforded for the patient in each brain region, statistical analysis was not possible apart from gene expression studies. These caveats are considered in data interpretation. Although the patient was an adult, SSADHD is considered a pediatric neurotransmitter disorder and most patients are diagnosed in early childhood.

Neuropathology

Histologic examination of human post-mortem brain tissue was performed as described previously (Ahrendsen et al., 2019). Briefly, formalin fixed paraffin embedded tissue was sectioned and stained with hematoxylin & eosin and luxol fast blue prior to examination. GFAP polyclonal antibody was obtained from Dako (Agilent Pathology Solutions, Santa Clara, CA). Immunohistochemistry staining was performed on a Dako Autostainer Link48 (Agilent Pathology Solutions, Santa Clara, CA), according to standard operating procedure.

Tissue Lipid Extraction

Lipids in the brain samples were extracted using a Tenbroeck homogenizer and a single-phase system of n-hexane/2-propanol (HIP, 3:2 v/v, 4 mL) (Hara and Radin, 1978). The extract was removed and the homogenizer rinsed with HIP (4 mL), which was added to the original extract. The extract was subjected to centrifugation at 2,500 x g for 15 minutes at −10°C to pellet the protein. The supernatant containing the lipid extract was removed and dried under a stream of nitrogen, then dissolved and subjected to ultrafiltration using a 0.2 μm Nylon filter to remove residual proteins. The sample was dried under a stream of nitrogen and dissolved in n-hexane/2-propanol/water (56.7:37.5:5.5 v/v/v) prior to separation by high-performance liquid chromatography.

High-performance liquid chromatography

Phospholipids were separated by high-performance liquid chromatography (HPLC) and individual phospholipid classes were collected to quantify mass. This method separates all major phospholipid classes including ethanolamine glycerophospholipids (EtnGpl), lysophosphatidylethanolamine (lysoPtdEtn), phosphatidylinositol (PtdIns), phosphatidylserine (PtdSer), choline glycerophospholipids (ChoGpl), sphingomyelin (CerPCho) and lysophosphatidylcholine (lysoPtdCho) with baseline resolution. Phospholipid elution was monitored by UV absorbance at 205 nm. Authentic standards purchased from Avanti Polar Lipids (Alabaster, AL) were used to confirm elution order and retention time of phospholipid classes.

The HPLC system used consisted of a Beckman Coulter System Gold 127 Solvent Module (Fullerton, CA), Beckman Coulter System Gold 166 Detector, Supelco Zorbax silica column (Bellefonte, PA) (25 cm x 4.6 mm, 5 μm), and Kipp & Zonen BD-41 recorder. Solvents were HPLC grade purchased from EMD Millipore-Sigma (St. Louis, MO). Solvent A was n-hexane:2-propanol (3:2 v/v) and solvent B was n-hexane:2-propanol/water (56.7:37.8:5.5 v/v/v). Initial proportions were 70% A/30% B with a flow rate of 1.5 mL/min, increasing proportion of solvent B in a stepwise manner until it reached 100% at 100 minutes, yielding baseline resolution of all phospholipids analyzed.

Plasmalogen mass

Following separation of phospholipid classes by HPLC, the EtnGpl and ChoGpl fractions were quantitatively divided. One half of each fraction was dried under a stream of nitrogen and subjected to mild acid hydrolysis of the plasmalogen vinyl ether linkage (Murphy et al., 1993), yielding an acid-stable fraction containing PakEtn (1-O-alkyl, 2-acyl glyceroethanolamine)/PtdEtn and PakCho (1-O-alkyl, 2-acyl glycerocholine/PtdCho, and an acid-labile fraction containing lysoPtdEtn (1-lyso, 2-acyl glycerolethanolamine) and lysoPtdCho (1-lyso, 2-acyl glycerocholine), which originated from the acid-labile PlsEtn and PlsCho, respectively. The mild acid hydrolysis removes the vinyl ether linked fatty alcohol on the sn-1 position of the plasmalogen, resulting in a lyso position at the sn-1 position. Therefore, the plasmalogens were separated as lysophospholipids using the same solvent system described above. Initial solvent proportions were 45% solvent B and 55% solvent A at a flow rate of 1.8 mL/min, increasing to 100% B over 20 minutes. Plasmalogen mass was calculated using the EtnGpl and ChoGpl masses collected in the major phospholipid class separation and the proportion of lysophospholipid:phospholipid collected in the plasmalogen separation.

Phospholipid mass

Phospholipid and plasmalogen masses were quantified by measuring lipid phosphorus (Rouser et al., 1969). Briefly, lipid fractions were collected and dried overnight in an oven set at 85°C. Phospholipids were digested with 0.5 mL water and 0.65 mL concentrated perchloric acid at 185°C for one hour and allowed to cool. Then, 0.5 mL ascorbic acid in water (10% w/v), 0.5 mL ammonium molybdate in water (2.5% w/v), and 3.3 mL water were added to each tube, vortexed well, and placed in a heating block at 100°C for 5 minutes. Inorganic phosphorus mass was quantified by absorbance at 797 nm and compared to a standard curve of known nmol of phosphorus.

Gene expression analyses

Total RNA was isolated from cerebellum, frontal cortex and partial cortex using the miRNeasy Mini Kit (Qiagen) according to manufacturer’s protocol. RNA was not extracted from the other regions due to lack of tissue. Quality of RNA was assessed by nanodrop and fragment analyzer (Agilent). RNA was quantified by Qubit Assay (Invitrogen). cDNA was synthesized using the RT2 First Strand Kit (Qiagen) and the Human GABA & Glutamate RT² profiler array (Qiagen) was selected for gene expression analysis. RT² profiler arrays were run on a CFX 384 (BioRad) instrument. Data analysis was performed using the GeneGlobe (Qiagen) platform, which calculated relative expression to the geometric mean of the housekeeping genes GAPDH (glyceraldehyde 3-phosphate) and ACTB (β-actin) using the delta delta CT method (essentially “log-fold change”). For data analysis, the fold regulation threshold was 2.0, and the p-value threshold was 0.05.

Results

Histology in fixed cortex and hippocampus of the patient revealed mild to moderate reactive astrogliosis, as assessed by GFAP immunohistochemistry, particularly in white matter areas of the brain (Figs. 1, 2). In the parietal cortex, mild reactive astrogliosis was present in the subpial region as well as cortical layers I, V, and VI (Figure 1A, B). Additionally, increased numbers of white matter neurons and moderate reactive astrogliosis was observed in the underlying cortical white matter (Figure 1C, D). Similar patterns of reactive gliosis were observed in the hippocampus, with mild gliosis in the dentate gyrus and entorhinal cortex and moderate gliosis in the hippocampal polymorphous layer and entorhinal subcortical white matter (Figure 2).

Fig. 2. GFAP (glial fibrillary acidic protein) immunohistochemistry of fixed hippocampal section from the patient.

A-dentate gyrus; B-polymorphous layer; C-entorhinal cortex; D-entorhinal subcortical white matter (200x mag)

We observed several putative changes, defined as two standard deviations from the mean of the control group, in phospholipid mass between the control group and patient. In the cortex, the patient had a marked 61% increase in total phospholipid mass (nmol phospholipid/gram wet weight) from the mean of the control values (Table 2a). Conversely, total phospholipid mass in the hippocampus was decreased 51% in the patient as compared to the controls (Table 5a). In the cerebellum, frontal lobe, pons, and parietal lobe, total phospholipid masses were not considered different between the patient and control group. Hence, the total phospholipid levels were increased only in the patient’s cortical section and decreased in the patient’s hippocampal section.

Table 2a.

Cortex phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	28099	17735	1395*	40.7%	42%	3.3%
PtdIns	2393	1607	178*	3.5%	3.75%	0.3%
PtdSer	9308	4973	547*	13.5%	11.6%	0.5%*
ChoGpl	21319	14612	1619*	30.9%	34.1%	2.3%
CerPCho	7934	3838	355*	11.5%	9.0%	0.7%*
Total	69053	42765	2905*
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 5a.

Hippocampus phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	11667	24408	2882*	40.1%	41.0%	2.5%
PtdIns	826	1562	389*	2.8%	2.6%	0.3%
PtdSer	3031	8751	2405*	10.4%	14.45%	2.2%
ChoGpl	10824	17730	1810*	37.2%	29.8%	1.6%*
CerPCho	2761	7376	1630*	9.5%	12.2%	1.1%*
Total	29110	59827	8523*
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Because the changes in total phospholipid mass may not reflect changes in every phospholipid class, the masses of individual phospholipid classes were analyzed. As expected from the total phospholipid data, the patient’s cortex had increased levels of each major phospholipid classes as compared to control values. In the patient’s cortex, the EtnGpl mass was increased 58%, the PtdSer mass was increased 49%, the PtdIns mass was increased 87%, the ChoGpl mass was increased 46%, and CerPCho mass was increased 107% as compared to controls (Table 2a). In contrast, the mass of each major phospholipid class was significantly decreased in the patient’s hippocampal sample as compared to the control group, reflective of the large decrease in total phospholipid mass (Table 5a). In the patient’s hippocampus, the EtnGpl mass was decreased 52%, the PtdIns mass was decreased 47%, the PtdSer mass was decreased 65%, the ChoGpl mass was decreased 39%, and the CerPCho mass was decreased 63% as compared to control values (Table 5a). Interestingly, in the patient’s frontal lobe, only the patient’s PtdSer level was decreased 63% compared to the control group (Table 3a). In the patient’s pons, the EtnGpl mass was increased 44% and the ChoGpl mass was increased 42% as compared to the control values (Table 4a). Although we observed large changes in the hippocampus and cortex that were non-specific in terms of phospholipid class, the frontal lobe and pons showed class-specific changes in phospholipid mass.

Table 3a.

Frontal lobe phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	12519	16153	2886	38.2%	39%	1.6%
PtdIns	1549	1840	243	4.7%	4.5%	0.6%
PtdSer	1792	4818	1390*	5.5%	11.5%	2.1%*
ChoGpl	14339	14474	2338	43.7%	35.2%	3.9%*
CerPCho	2614	4036	1032	8.0%	9.7%	1.7%
Total	32812	41321	6911
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 4a.

Pons phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD (control)	Patient	Mean (control)	SD (control)
EtnGpl	26536	18428	2549*	44.8%	41.0%	1.0%*
PtdIns	1031	978	153	1.7%	2.2%	0.7%
PtdSer	6911	6991	1572	11.7%	15.4%	1.0%*
ChoGpl	18820	13290	2543*	31.8%	29.4%	0.8%
CerPCho	5933	5396	852	10.0%	12.0%	0.1%*
Total	59231	45083	7361
	n = 1	n = 2

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Plasmalogens are a subclass of phospholipids that contain a vinyl ether linkage in the sn-1 position. Because plasmalogens are enriched in arachidonic acid, they are considered a putative intracellular signaling molecule with a role in lipid-mediated signal transduction (Farooqui et al., 1995; Rosenberger et al., 2002), we also analyzed plasmalogen mass in each group. In the cortex, the mass of the acid-stable fractions of EtnGpl and ChoGpl were increased 58% and 48%, respectively, while the PlsEtn mass was increased 58% in the patient as compared to the control group (Table 2b). In the pons, the acid-stable fraction of EtnGpl was increased 30%, while the PlsEtn mass was increased 50%, and the PlsCho mass was increased 39% as compared to control values (Table 4b). In the hippocampus, the patient PlsEtn mass was decreased 70% and the acid-stable ChoGpl fraction was decreased 38% compared to the control group (Table 5b). Therefore, plasmalogen mass was reflective of the change in major glycerophospholipid mass except for the hippocampus, where the PlsEtn mass was significantly decreased and the acid-stable fraction of EtnGpl was unchanged.

Table 2b.

Cortex plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	11536	7311	598*	41.1%	41.6%	6.4%
PlsEtn	16563	10424	1861*	58.9%	58.4%	6.4%
ChoGpl	19621	13233	1479*	92.0%	90.6%	0.9%
PlsCho	1698	1379	195	8.0%	9.4%	0.9%
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 4b.

Pons plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD (control)	Patient	Mean (control)	SD (control)
EtnGpl	7044	5407	590*	26.5%	29.4%	0.9%*
PlsEtn	19492	13021	1959*	73.5%	70.6%	0.9%*
ChoGpl	16526	11640	2705	87.8%	87.2%	3.7%
PlsCho	2293	1650	163*	12.2%	12.8%	3.7%
	n = 1	n = 2

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 5b.

Hippocampus plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	6933	8815	1331	59.4%	36.9%	10.2%
PlsEtn	4734	15593	3986*	40.6%	63.1%	10.2%*
ChoGpl	9728	15751	2256*	89.9%	88.5%	4.3%
PlsCho	1096	1979	530	10.1%	11.5%	4.3%
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Phospholipid molar composition was also determined, and values are expressed as a molar percent of the total phospholipid mass. Molar composition is useful to assess changes in metabolism of certain lipid classes, and indeed, several interesting alterations were noted. In the cerebellum, the proportion of CerPCho was slightly increased 6% from control values (Table 1a). Although there was a marked increase in phospholipid mass noted in the cortex (Table 2a), proportional changes were limited to PtdSer (16% increase) and CerPCho (28% increase) compared with control values. Interestingly, in the frontal lobe, the proportion of PtdSer was decreased 53%, as expected from the decrease in its mass, but the proportion of ChoGpl was increased 24% as compared to controls (Table 3a). Further, the acid-stable fraction of EtnGpl was increased 38%, while the proportion of PlsEtn was decreased 26% compared to the control group (Table 3b).

Table 1a.

Cerebellum phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	11965	12228	4572	38.1%	37.8%	1.4%
PtdIns	1173	1445	137	3.7%	4.8%	1.4%
PtdSer	3103	3228	1625	9.9%	9.7%	1.7%
ChoGpl	11947	12100	3660	38.1%	38.1%	1.8%
CerPCho	3208	3123	1130	10.2%	9.7%	0.2%*
Total	31397	32125	11103
	n = 1	n = 3

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 3b.

Frontal lobe plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	7027	6512	1295	56.1%	40.6%	5.8%*
PlsEtn	5492	9641	2191	43.9%	59.4%	5.8%*
ChoGpl	12924	12912	2236	90.1%	89.1%	1.1%
PlsCho	1415	1562	147	9.9%	10.9%	1.1%
	n = 1	n = 4

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

The pons showed compositional changes in several phospholipid classes (Table 4a). The proportion of EtnGpl and ChoGpl were increased 10% and 11%, respectively compared to control patient values. Slight, but potentially important, increases in PlsEtn (4%) and the acid-stable fraction (10%) accompanied the increase in EtnGpl (Table 4b). In contrast, the proportions of PtdSer and CerPCho were decreased 24% and 16%, respectively. In the hippocampus, the proportion of ChoGpl was increased 25%, while the proportion of CerPCho was decreased 22% as compared to controls (Table 5a). Additionally, the proportion of PlsEtn was markedly decreased 36% (Table 5b). Lastly, in the parietal lobe, the proportion of EtnGpl was slightly increased 6% from controls (Table 6a). As opposed to phospholipid mass, in which the observed changes were largely non-specific, the molar composition analysis revealed several proportional changes in PtdSer, CerPCho, ChoGpl, and PlsEtn.

Table 6a.

Parietal lobe phospholipid mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	18209	16292	4543	41.7%	39.5%	0.4%*
PtdIns	1093	1209	245	2.5%	3.0%	0.3%
PtdSer	5404	5430	2205	12.4%	12.8%	2.2%
ChoGpl	14766	14015	2866	33.8%	34.6%	3.4%
CerPCho	4214	4297	1624	9.6%	10.2%	1.4%
Total	43685	41243	11383
	n = 1	n = 3

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Gene expression are shown in Tables 7 and 8. Negative values indicate down-regulation, positive values up-regulation. A total of 84 genes were assessed in total RNA from cerebellum, frontal and parietal lobes of the patient. Of 15 GABA_A receptor subunits analyzed, the patient revealed dysregulation of 14 in at least 1 tissue (predominantly down-regulation). Down-regulation of α4 and α5 subunits in cerebellum were accompanied by up-regulation in other tissues (Table 7). The most significant levels of down-regulation levels were observed for ε, θ, ρ1, and ρ2 subunits (7.7–9.9-fold). Unexpectedly, we found little effect on the GABA_B receptor, and minimal dysregulation in the expression of either ionotropic or metabolomic glutamate receptors, apart from consistent down-regulation of the metabotropic glutamate receptor 6 (Table 7). For genes encoding solute carriers, down-regulation of the Na⁺-dependent inorganic phosphate cotransporter 6 and the neurotransmitter transporter for GABA, member 13 (3.4–7.3-fold) were consistent findings. For 38 additional genes associated with GABAergic/glutamatergic activity, 15 showed dysregulation in at least one tissue of the patient (Table 8), including consistent down-regulation of interleukin 1β (up to 9.9-fold) and up-regulation of synuclein α (up to 6.5-fold).

Table 7:

Dysregulation of GABAergic, glutamatergic, and solute carrier genes in cerebellum and frontal/parietal lobes of the patient

	Subunit	Cere	Frontal	Parietal	Comment
GABAAR	α2	−2.56
	α4	−2.77		2.32
	α5	−4.19	2.12	2.63
	α6			−2.34
	β1	−4.84	−2.26	−4.46
	β3			2.49
	δ			−2.56
	ε	−3.95	−4.71	−7.71
	γ1	−2.35
	γ2			2.0
	γ3	−4.37
	θ	−3.08	−3.35	−8.81
	ρ1	−2.73	−3.18	−9.86
	ρ2		−3.69	−9.86
GABABR	R2			−2.3
Glu R	A1	−2.92			Ionotropic AMPA 1
	A4			−2.17	Ionotropic AMPA 4
	K5			−3.31	Ionotropic kainate 5
	N1			−2.22	Ionotropic NMDA 1
	1			−2.11	Metabotropic 1
	5			−2.82	Metabotropic 5
	6	−2.1	−4.0	−5.87	Metabotropic 6
	8				Metabotropic 8
SLC	17A6	−3.37	−2.18	−2.21	Na+-dependent Pi cotransporter member 6
	17A8	−2.62			Na+-dependent Pi cotransporter member 8
	1A3	−3.67	2.26		Glial high affinity glu transporter member 3
	1A6	−2.18			High affinity asp/glu transporter member 6
	6A13	−2.89	−2.02	−7.3	Neurotransmitter transporter, GABA, member 13
	7A11		2.57		Anionic amino acid transporter, light chain, xc− system member 11

All values shown significant at p<0.05. Negative, down-regulation; positive, up-regulation.

Abbreviations: R, receptor; cere, cerebellum; glu, glutamate; SLC, solute carrier; AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; NMDA, N-methyl-D-aspartate; asp, aspartate; P_i, inorganic phosphate; xc⁻, cystine-glutamate antiporter. Of the 51 genes evaluated for expression levels, 29 revealed dysregulations in at least one tissue from the patient.

Table 8:

Dysregulation of miscellaneous genes associated with GABAergic/glutamatergic signaling in cerebellum and frontal/parietal lobes of the patient

Gene	Cere	Frontal	Parietal	Comment
IL1β	−3.78	−2.93	−9.86	Interleukin 1, β
ITPR1	−2.32		−2.4	Inositol 1,4,5-triphosphate receptor, type 1
P2RX7	−2.81	−2.03		Purinergic receptor P2X, ligand-gated ion channel, 7
ADCY7			−2.4	Adenylate cyclase 7
SHANK2		−2.08	−2.97	SH3 and multiple ankyrin repeat domains 2
ADORA2A			−2.98	Adenosine A2a receptor
CACNA1A			−2.21	Calcium channel, voltage dependent, P/Q type, α1A subunit
CACNA1B			−3.02	Calcium channel, voltage dependent, N type, α1A subunit
CLN3			−2.17	Ceroid lipofuscinosis, neuronal 3
PLA2G6			−2.35	Phospholipase A2, group VI (cytosolic, Ca+2-independent
PLCB1			−2.11	Phospholipase C, β1 (phosphoinositide specific)
MAPKI	6.54			Mitogen-activated protein kinase 1
SNCA	5.08	4.88	6.52	Synuclein, α (non A4 component of amyloid precursor)
AVP	−2.34		−2.95	Arginine vasopressin

All values shown significant at p<0.05. Negative, down-regulation; positive, up-regulation.

Abbreviations: Of the 38 genes evaluated for expression levels, 15 revealed dysregulations in at least one tissue from the patient. This group of genes contained 38 genes in total, for which the fifteen shown in the table were dysregulated in at least one of the patient’s tissues.

Discussion

The availability of these post-mortem tissues represented a unique opportunity for metabolic and molecular investigations with the potential to further delineate pathomechanisms in SSADHD. There are obvious limitations to the interpretation of our findings, including the delay from time of death to tissue harvest, the inability to accurately match control tissues for age and medications, the limited number of control specimens, and the fact that the patient was receiving risperidone at time of death (Kirby et al., 2020). Fortunately, risperidone does not appear to specifically target GABAergic/glutamatergic receptors, or solute carrier systems (Gao et al., 2018).

For the first time we have shown the presence of astrogliosis in selected brain regions of a patient with SSADHD, confirming earlier studies in the brain of null mice (Hogema et al., 2001; Brown et al., in press). Reactive astrogliosis was present in the patient’s cortex, subcortical white matter, and hippocampus. While gliosis in and of itself is not specific for any specific pathologic entity, it is a useful marker for subtle or early pathologic changes and can also reflect chronic injury to a particular brain region. It is particularly useful to identify pathologic insult in cases such as epilepsy that do not display overt loss of neurons/myelin or have a robust inflammatory response.

Previous studies on lipid alterations in SSADH deficiency have demonstrated alterations consistent with myelin dysfunction in a gene-ablated mouse model (Donarum et al., 2006; Barcelo-Coblijn et al., 2007). In the current study, we focused on phospholipids, one of the most abundant class of lipids in the myelin sheath (Sun and Horrocks, 1968; Morell and Raine, 1984), and identified several putative differences with control values. The patient’s cortical and hippocampal samples had marked changes in total phospholipid mass, which were unexpectedly increased in the cortical sample (61%) and decreased in the hippocampal sample (51%). This was reflected in the mass of each of the major phospholipid classes in these regions (Tables 2a and 5a). In the cortex, there were also significant increases in the acid-stable and plasmalogen fractions of ethanolamine glycerophospholipids (EtnGpl), as expected (Table 2b). However, the molar composition of the cortex revealed proportional increases in phosphatidylserine (PtdSer) and sphingomyelin (CerPCho) (Table 2a). Taken together, the increase in the proportion of PtdSer and CerPCho along with the increase in total phospholipids suggest that the patient’s cortical sample was contaminated with myelin. Myelin contains a greater proportion of phospholipids per unit fresh weight by 40% (Morell and Raine, 1984), and include greater proportions of PtdSer and CerPCho (Horrocks, 1967; Sun and Sun, 1972). Further, these cortical data contradict prior studies in which myelin proteins are downregulated and PlsEtn is decreased in the cortices of SSADH-deficient mice (Donarum et al., 2006; Barcelo-Coblijn et al., 2007). Therefore, since the cortex shows a significant increase in the total phospholipid mass as well as proportions of CerPCho and PtdSer, this is suggestive of myelin contamination in the patient sample as compared to the control group, accounting for the unexpected increase in mass.

Consistent with previous studies on SSADH-deficient mice (Donarum et al., 2006; Barcelo-Coblijn et al., 2007), the remaining brain regions reveal several phospholipid changes that may be related to the observed decreases in myelin components observed in gene-ablated mice. Phospholipid proportions differ in white matter and gray matter, with decreased proportions of PtdIns and ChoGpl, and increased proportions of EtnGpl, PlsEtn, PtdSer and CerPCho seen in white matter (Horrocks, 1967; Sun and Sun, 1972). Interestingly, these same changes were observed in samples from the pons, hippocampus, and frontal lobe. The proportion of CerPCho was decreased in the pons and hippocampus (Tables 4a and 5a), while PtdSer was decreased in the pons and frontal lobe (Tables 4a and 3a). In the frontal lobe and hippocampus, the proportion of ChoGpl was significantly increased (Tables 3a and 5a) while the proportion of PlsEtn was decreased (Tables 3b and 5b). Although the patient group was n = 1, the pattern present in these brain regions suggests brain-wide alterations in lipid metabolism that favor myelin derangement and support the conclusions from previous studies (Donarum et al., 2006; Barcelo-Coblijn et al., 2007).

We formulated our hypothesis of dysregulated GABAergic/glutamatergic receptor subunits based upon earlier studies in the null mouse that revealed decreased expression both in hippocampal regions and in whole brain (Buzzi et al., 2006; Wu et al., 2006;Vogel et al., 2016). Direct comparison of those data with current data in human are not possible, since here we have looked at specific human regions in an RNA-dependent approach. However, the prediction that GABAergic receptor systems would be down-regulated generally held forth (Table 7), especially GABA_A receptors. Conversely, glutamatergic receptors were considerably less impacted, except for the metabotropic glutamate receptor 6 (Table 7). As well, specific solute carrier genes were also impacted, especially the SLC17a6 (inorganic phosphate cotransporter, or vesicular glutamate transporter; VGLUT2) and SLC6a13 (GABA transporter 2; GAT-2). Whether both genes are down-regulated due to GABA, or glutamate, or a combination of both is unknown, but there was no frank elevation of glutamate in brain regions of the patient (Kirby et al., 2020).

Gene expression results were notable in patient tissues that were not dysregulated (glutamic acid decarboxylase, GABA-transaminase, glutaminase (predominantly neuronally located) and glutamine-ammonia ligase (predominantly glial located); Vogel et al., 2016). The enzymes encoded by these genes are responsible for maintenance of the GABA-glutamine-glutamate cycle, which is disrupted in the brain of the null mouse but has not been studied in human SSADHD (Chowdhury et al., 2007). As well, all four genes were down-regulated in the null mouse brain (Vogel et al., 2016). Further, GABA_B receptor subunits (R1, R2) were not down-regulated in the patient, which is at variance with our past animal work (Wu et al., 2006; Cortez et al., 2004). GABA_B receptor subunit findings may also help to explain the lack of efficacy the GABA_B receptor antagonist SGS-742 which was recently administered in a double-blind, placebo-controlled study in SSADHD (Schreiber et al., submitted).

We observed down-regulations of specific purinergic/adenosine receptors, as well as calcium channels and receptor-signaling proteins (phospholipase A2, C), but these changes were not ubiquitous across tissues and predominantly centered on parietal lobe (Table 8). Ren and Mody (2003) demonstrated that exogenous GHB, which is highly elevated in patient brain (Kirby et al., 2020), induces phosphorylation/activation of MAPK1 via GABA_B receptor function, and this is consistent with up-regulation of MAPK1 in cerebellum (Table 8). Other genes of interest dysregulated in the patient included IL1β, synuclein α, and arginine vasopressin. Emmanouilidou and coworkers (2016) demonstrated that GABA transmission regulates α-synuclein secretion in mouse striatum via ATP-dependent K⁺ channels, consistent with the findings of up-regulation for SNCA in the patient across regions. Moreover, earlier studies demonstrated that brain neutral lipid mass was increased, as was turnover of brain phospholipids, in α-synuclein gene-ablated mice (Barcelo-Coblijn et al., 2007; Golovko et al., 2007). This suggests that the upregulation observed for SNCA in patient brain (Table 8) may be associated with the alterations of lipid and plasmalogen classes we observed for the patient.

Down-regulation of IL1β across all patient brain regions was interesting yet challenging to explain. Bianchi and colleagues (1995) documented that administration of IL1β to mice significantly reduced hippocampal GABA. Supraphysiological GABA levels, as seen in SSADHD (Kirby et al., 2020), may conversely lead to down-regulation of IL-1β, underscoring the interrelationships of GABA metabolism in the central modifications induced by IL-1β. Finally, arginine vasopressin was massively down-regulated in previous gene expression studies on the null mouse brain (Vogel et al., 2016), but only slightly down-regulated in our patient’s tissues. The elevation of 4-guanidinobutyrate (a putative derivative of GABA and arginine; Jansen et al., 2006; Kirby et al., 2020) in the brain regions of our patient may provide insight into this down-regulation (Kirby et al., 2020).

In sum, the current results provide a broader understanding of the underlying pathophysiology of human SSADHD. To assist potential future post-mortem studies, should they unfortunately occur, we are developing a protocol that will allow rapid and extensive collection of brain tissue, peripheral tissue biopsies, and physiological fluids to add to our biorepository of specimens that is a component of our ongoing natural history study of SSADHD.

Table 1b.

Cerebellum plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	5052	5999	1693	42.2%	50.3%	6.5%
PlsEtn	6913	6229	3063	57.8%	49.7%	6.5%
ChoGpl	10464	10853	3079	87.6%	90.0%	1.8%
PlsCho	1483	1247	595	12.4%	10.0%	1.8%
	n = 1	n = 3

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.

Table 6b.

Parietal lobe plasmalogen mass in SSADH deficient patient and control group

	nmol/g ww			mol %
	Patient	Mean (control)	SD	Patient	Mean (control)	SD
EtnGpl	5960	7314	732	32.7%	46.8%	10.2%
PlsEtn	12249	8977	3812	67.3%	53.2%	10.2%
ChoGpl	13303	12727	2934	90.1%	90.4%	2.8%
PlsCho	1463	1288	100	9.9%	9.6%	2.8%
	n = 1	n = 3

Mass values are expressed as nmol phosphorus/gram wet weight tissue and represent mean±SD of control and patient values. Molar composition values are expressed as mol % of total phosphorus. n = 2–4.

^*indicates significance between patient and control group, 2 SD from control mean.