Disturbed brain ether lipid metabolism and histology in Sjögren‐Larsson syndrome

Abstract Sjögren‐Larsson syndrome (SLS) is a rare neurometabolic syndrome caused by deficient fatty aldehyde dehydrogenase. Patients exhibit intellectual disability, spastic paraplegia, and ichthyosis. The accumulation of fatty alcohols and fatty aldehydes has been demonstrated in plasma and skin but never in brain. Brain magnetic resonance imaging and spectroscopy studies, however, have shown an abundant lipid peak in the white matter of patients with SLS, suggesting lipid accumulation in the brain as well. Using histopathology, mass spectrometry imaging, and lipidomics, we studied the morphology and the lipidome of a postmortem brain of a 65‐year‐old female patient with genetically confirmed SLS and compared the results with a matched control brain. Histopathological analyses revealed structural white matter abnormalities with the presence of small lipid droplets, deficient myelin, and astrogliosis. Biochemically, severely disturbed lipid profiles were found in both white and gray matter of the SLS brain, with accumulation of fatty alcohols and ether lipids. Particularly, long‐chain unsaturated ether lipid species accumulated, most prominently in white matter. Also, there was a striking accumulation of odd‐chain fatty alcohols and odd‐chain ether(phospho)lipids. Our results suggest that the central nervous system involvement in SLS is caused by the accumulation of fatty alcohols leading to a disbalance between ether lipid and glycero(phospho)lipid metabolism resulting in a profoundly disrupted brain lipidome. Our data show that SLS is not a pure leukoencephalopathy, but also a gray matter disease. Additionally, the histopathological abnormalities suggest that astrocytes and microglia might play a pivotal role in the underlying disease mechanism, possibly contributing to the impairment of myelin maintenance.

analyses revealed structural white matter abnormalities with the presence of small lipid droplets, deficient myelin, and astrogliosis. Biochemically, severely disturbed lipid profiles were found in both white and gray matter of the SLS brain, with accumulation of fatty alcohols and ether lipids. Particularly, longchain unsaturated ether lipid species accumulated, most prominently in white matter. Also, there was a striking accumulation of odd-chain fatty alcohols and odd-chain ether(phospho)lipids. Our results suggest that the central nervous system involvement in SLS is caused by the accumulation of fatty alcohols leading to a disbalance between ether lipid and glycero(phospho)lipid metabolism resulting in a profoundly disrupted brain lipidome. Our data show that SLS is not a pure leukoencephalopathy, but also a gray matter disease. Additionally, the histopathological abnormalities suggest that astrocytes and microglia might play a pivotal role in the underlying disease mechanism, possibly contributing to the impairment of myelin maintenance.

| INTRODUCTION
Sjögren-Larsson syndrome (SLS; OMIM #270200) is a neurometabolic disorder caused by fatty aldehyde dehydrogenase (FALDH) deficiency 1 due to biallelic mutations in ALDH3A2. 2 Patients suffer from intellectual disability, spastic diplegia, ichthyosis, and retinopathy. 3,4 The FALDH deficiency results in accumulation of fatty aldehydes and fatty alcohols in plasma and skin. 5,6 FALDH is involved in the degradation of leukotrienes, which leads to increased levels of leukotriene B4 in patients. 7 Furthermore, FALDH plays a role in phytol metabolism, converting phytenal into phytenic acid, apparently without accumulation of phytol or its degradation products in SLS patients. 8 Brain magnetic resonance (MR) imaging studies show a white matter disorder, 9,10 and proton MR spectroscopy reveals abnormal signals between 0.8 and 1.6 ppm, 10 which reflect abnormal lipid accumulation. 9 Only a few postmortem reports on SLS are available, but in none of these patients, the diagnosis was biochemically or genetically confirmed. [11][12][13][14] Consequently, the underlying disease mechanisms causing the brain disorder in SLS, including the most important lipids involved, are not yet known.
Lipids comprise half of the human brain dry weight. They provide membrane structure and are required for membrane trafficking, signal transmission, and synaptogenesis. 15 Cholesterol and phospholipids are the main membrane lipids. Phospholipids have a phosphatecontaining head-group at the sn-3 position of a glycerol backbone, and one or two fatty acyl-groups linked via an ester bond. They are subdivided in subclasses based on the nature of the head group (ie, phosphatidylethanolamine contains an ethanolamine head-group, whereas phosphatidylcholine contains a choline head group). 15 Ether phospholipids, including plasmalogens, constitute a special phospholipid class, characterized by the presence of an ether or a vinyl ether at the sn-1 position of the glycerol backbone instead of an ester (Figure 1). 16 The most abundant lipid species in the human brain, apart from cholesterol, are phospholipids with phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylserine (PS) as the three most abundant classes. 17 We studied the postmortem brain of a 65-year-old female patient with genetically confirmed SLS using targeted as well as untargeted lipidomic approaches in combination with morphological analyses. Our results

SYNOPSIS
The brain disorder in Sjögren-Larsson syndrome (SLS) is explained by a severely disrupted lipid profile in both white and gray matter with accumulation of fatty alcohols and ether lipids and histopathological abnormalities in astrocytes and microglia.
show a severely disturbed lipid profile in the SLS brain with accumulation of various species of fatty alcohols, ether(phospho)lipids, and triacylglycerols in both white and gray matter, and confirm the presence of structural white matter abnormalities.

| MATERIALS AND METHODS
Materials and methods are described in the Supporting information.

| ResultsHistopathology of the SLS brain
The SLS brain (weight 1347 g) was macroscopically normal (Figure 2A-C). Microscopically, lack of myelin with tissue rarefaction was seen in hemispheric deep white matter and capsulae ( Figure 2D,E), whereas the U-fibers and corpus callosum were better preserved without signs of demyelination. Oligodendrocyte numbers were not diminished and scattered axonal spheroids were found ( Figure 2D-F). A diffuse isomorphic astrocytic gliosis was present, and some microglia had a phagocytic morphology ( Figure 2G,H). Around blood vessels clustering macrophages contained pigmented PAS-positive ironnegative material. These pathological changes were more F I G U R E 2 Legend on next page. evident in occipital white matter. In the white matter, smaller blood vessels had thickened walls with relative loss of smooth muscle cells in the tunica media and thickened tunica adventitia ( Figure 2I). The cerebral cortex was only mildly gliotic, without neuronal dropout or microglia activation. Basal nuclei, thalamus, and hippocampus were unaffected ( Figure 2J). The long white matter tracts in the brainstem and spinal cord also F I G U R E 2 Neuropathology of SLS. A, Whole coronal slice of the right hemisphere at the level of the pulvinar shows no normal trophism of the cortex and of the white matter. B,C, Whole coronal mounts at the same level show that the periventricular and deep hemispheric white matter is pale (B, hematoxylin and eosin [H&E]; C, Kluver stain for myelin combined with periodic acid Shiff), whereas the U-fibers are relatively better preserved. D, H&E stain of the occipital white matter confirms white matter pallor and reveals preservation of oligodendrocyte numbers and presence in the neuropil of yellowish lipid droplets (inset). E, Stain for the major myelin protein proteolipid protein (PLP) shows that white matter pallor is due to decreased amounts of myelin. F, stain for the pan-oligodendrocytic marker Olig2 further confirms that oligodendrocytes are not depleted. G, Stain against the astrocyte marker glial fibrillary acidic protein (GFAP) reveals a diffuse isomorphic astrogliosis. H, stain against the microglia/macrophage marker CD68 shows that cells with a phagocytic morphology are clustered around blood vessels. I, the histochemical stain Trichrome shows blood vessel wall changes consistent with hyalinosis, including thickening of the walls and loss of cells in the tunica media. J, H&E stain of the occipital cortex shows normal organization with no loss of neurons. K, PLP stain of the descending tracts in the brainstem shows reduced myelin amounts. L, In the same areas, the CD68 stain reveals presence of normal numbers of microglia that do not show an ameboid activated aspect. M,N, H&E stain (M) and immunohistochemistry against high-molecular weight neurofilaments (NF, N) reveals loss of Purkinje cells in the cerebellar cortex with presence of empty baskets. O, Stain against the Bergmann glia marker S100β shows that Purkinje cell dropout corresponds with proliferation of Bergmann glia, some of which are abnormally translocated to the molecular layer showed deficient myelin, however, without microglia or astrocytic activation ( Figure 2K,L). The cerebellar white matter was better preserved compared to the cerebral areas; however, activated microglia and scattered macrophages were found in perivascular spaces. In cerebellar cortex, a mild dropout of Purkinje cells and granular neurons were seen, with some degree of mislocalization of the Bergmann glia nuclei to the molecular layer (Figure 2M-O). No white matter pathology was found in the optic nerve, optic chiasm, and optic tract.

| Free fatty alcohols and fatty aldehydes
The total free fatty alcohols as a group (C16-24) were increased in SLS brain compared to control brain by 4.1fold in white matter and 3.8-fold in gray matter. A similar fatty alcohol profile was observed for white and gray matter of the SLS brain but fatty alcohols accumulated in much higher concentrations in white matter than in gray matter ( Figure 3A). The chain lengths of the accumulating alcohols ranged from C18 to C24 and comprised both even-and odd-length chains. Notably, the long-chain fatty alcohols C21 to C24, both saturated and monounsaturated were highly elevated in SLS brain, more so in white matter than gray matter. In SLS white matter, this subgroup of fatty alcohols was increased by 17-fold over control brain. C23-fatty alcohols were the most abundant species in both white and gray matter, with fold changes of respectively 166 and 76. There was no parallel accumulation of free very long chain fatty acids. Surprisingly, although small amounts of free fatty aldehydes (C14:0, C16:0, C17:0, C18:0, and C18:1) were present in both brains, there was no accumulation of aldehydes in the SLS brain (data not shown).

| Lipid analysis
The lipidome of white and gray matter samples of SLS and control brain was using untargeted lipidomics. A total of almost 1700 individual lipid species from 33 different lipid classes could be evaluated by our lipidomics technique. The most abundant lipid classes in the control brain in descending order are cholesterol and its esters,  Figures 3B and 4). The profile was similar for white and gray matter. Most individual lipid species were found in remarkably similar concentrations in SLS and control brain, suggesting that the quality of both brain samples and the analysis was comparable. The following major classes, including subspecies, did not change significantly when comparing SLS and control brain: sphingolipids; ceramides, ceramide-1-phosphates, hexosylceramides (HexCer), lactosylceramides, sulfatides, hydroxysulfatides, sphingosines, sphingosine-1-phosphates and sphingomyelins, and phospholipids; (lyso) phosphatidic acids, (lyso)phosphatidylglycerols, phosphatidylinositols, phosphatidylserines, and monoand dilysocardiolipins.
Several lipid classes accumulated in SLS brain, both in white and gray matter (Figure 4) when comparing SLS and control brain for both white and gray matter (Supporting information). Other non-ether phospholipid classes including PS, LPE, and PE also showed a trend toward lower levels in SLS white matter, respectively, 77%, 59%, and 55% of the control tissue abundancy. In gray matter, CE and DG showed a trend toward lower levels in SLS, with, respectively, 51% and 64% abundancy of the control tissue. The concentration of accumulating ether lipids was highest in SLS white matter.

| Individual lipid subspecies
To evaluate which specific lipid species predominantly accumulate in SLS brain, we considered individual lipids with a fold change >3 in SLS compared to control and a difference in relative abundance >10.  . The total CE level in control white matter was only moderately elevated in SLS brain, but this was not seen in gray matter ( Figure 4). Remarkably, specific CE species with polyunsaturated long-and very-longchain acyl-groups were highly elevated in SLS white matter (20:2, 22:3, 24:5, 22:5, 24:4) and to a lesser extent also in gray matter (Supporting information).

| Mass spectrometry imaging
During sectioning of brain tissue for mass spectrometry imaging, two blocks from each brain were prepared; one containing mostly gray matter, the other containing mostly white matter. As expected, large differences in the lipid content were visible between white and gray matter, both in the SLS and control tissues (Supporting information). A total of 1872 m/z features were detected from the overall mass spectrum (including both white and gray matter regions, from SLS and control tissues). We searched for the most discriminating peaks between the SLS and control tissues. The lipid features that distinguish the SLS and control profile could be tentatively

| DISCUSSION
Our study describes a comprehensive biochemical as well as histopathological evaluation of the postmortem brain of a patient with genetically confirmed SLS; the results contribute to our understanding of SLS, and illustrate the power of novel techniques to clarify the biochemical abnormalities of the brain in neurometabolic disorders.
The white matter of the SLS brain showed lack of myelin and perivascular macrophages containing pigmented lipoid material, with a posterior hemispheric predominance. The absence of numerous, scattered macrophages in the parenchyma argues against ongoing demyelination and rather suggests an impairment in the process of myelin maintenance. The white matter abnormalities and accumulating lipoid material correspond to the results of cerebral MR imaging and spectroscopy in living SLS patients. 10 Interestingly, in the more affected occipital lobes, we also found astrocyte and microglia activation, suggesting that the disease mechanisms underlying the leukodystrophy are ongoing. Dysfunction of astrocytes and microglia could contribute to an impairment in myelin maintenance. 18 The direct involvement of these cell types is in line with the findings in previous retinal studies in SLS, which suggested a similar role for Müller cells, the retinal counterpart of cerebral astrocytes. 19 It is known that fatty alcohols accumulate in body fluids of SLS patients, 1 but their accumulation in the central nervous system has never been shown. We found a striking accumulation of fatty alcohols in SLS brain with a predominance of C18-C24 chain length species. Interestingly, although the C18:0 alcohol was elevated, the C16:0 alcohol was not. This long-chain fatty alcohol pattern differs from that seen in SLS plasma, cultured fibroblasts and keratinocytes in which only C16-C18 fatty alcohols accumulate. 5,20 The longer chain fatty alcohols (C21-C24) therefore likely are unique to brain. Longchain fatty alcohols are synthesized from their corresponding fatty acids of similar chain length but can also be formed in the fatty alcohol cycle by reduction of aldehydes that originate from catabolism of metabolites, which will be discussed below. 21,22 The fatty acyl-CoAreductase (FAR) enzyme in mouse and bovine, involved in the de novo synthesis of fatty alcohols, has highest activity with C15-C18 acyl-CoA substrates and very low activity with C20-C22-acyl-CoA substrates. 23,24 Humans possess two distinct FAR enzymes (FAR1 and FAR2); it is unknown which of these two enzymes is responsible for synthesizing the longer C21-C24 alcohols identified in SLS brain, and whether or not they are synthesized via this pathway at all. 25 Nevertheless, the normal fatty acid composition of the SLS brain argues that the accumulation of the fatty alcohols is not due to an increase in their rate of synthesis via either FAR1 or FAR2. Another potential source of fatty alcohols is the catabolism of different metabolites that lead to the production of fatty aldehydes via the fatty alcohol cycle. Fatty aldehydes arise from the degradation of plasmalogens, sphingosine-1-phosphate, branched-chain fatty acids and 2-hydroxyfatty acids ( Figure 6A). 22,26 Aldehydes that are usually oxidized to fatty acids by FALDH would accumulate and be reduced to fatty alcohols, a phenomenon that was previously observed in SLS. 1 In this respect, the predominant accumulation of the odd-chain C23-alcohol was intriguing and unexpected. We speculate that this originates from the breakdown of 2-hydroxy-fatty acids derived from sphingolipids carrying these 2-hydroxy-fatty acids, which are very abundant in brain, especially in white matter. 17,27 When liberated, 2-hydroxy-fatty acids are activated to their corresponding CoA-ester and the enzyme 2-hydroxyacyl-CoA lyase 1 cleaves this molecule F I G U R E 6 Legend on next page. yielding formyl-CoA and an n-1 aldehyde. 28 In brain, the most abundant 2-hydroxy-fatty acid is C24, 27 which results in the formation of C23-aldehyde when brokendown via this route ( Figure 6B). As FALDH has the ability to metabolize aliphatic aldehydes ranging from 6-to 24-carbons long, with substrate specificity toward longchain fatty aldehydes we believe that C23-aldehyde could also be a substrate for FALDH and that this thus accumulates in SLS. 29,30 FALDH functions as part of a complex that, together with fatty alcohol dehydrogenase, sequentially converts fatty alcohols to fatty aldehydes and fatty acids. 1,31 Since fatty alcohols are metabolic precursors for synthesis of ether lipids, it is likely that the accumulation of LPC[O] and TG [O] in SLS brain is a direct consequence of increased amounts of fatty alcohol substrates ( Figure 6B) and that this distribution reflects the accumulating alcohols and those resulting from reduction of catabolically formed aldehydes. There was no increase in plasmalogens in SLS brain based on the dimethyl acetal peaks seen on fatty acid analysis. Although the C18:0 alcohol accumulated in the SLS brain and is a substrate for plasmalogen synthesis, it does not seem to drive the synthesis of plasmalogens by itself. This was supported by the lipidomics experiment which showed comparable levels of plasmenyl-species in SLS and control brain (Supporting information).
We showed that the SLS brain lipidome is profoundly altered as fatty alcohols cause accumulation of almost all ether lipid classes in both white and gray matter with a concomitant reduction in non-ether lipids. The accumulating fatty alcohols most likely drive the enhanced synthesis of ether lipids and etherphospholipids with a characteristic pattern that is likely dictated by the tissuedependent aldehydes that are normally processed by FALDH ( Figure 6). Ether lipids, including plasmalogens, are ubiquitously found throughout human tissues and are known to be especially important in brain, heart and spleen. 16 For most accumulating ether(phospho)lipid classes, the concentration in SLS white matter was much higher than in SLS gray matter (Figure 4) 32 The plasmalogen synthesis is regulated by post translational degradation of FAR1, 33 which is the rate-limiting enzyme for fatty alcohol production from acyl-CoAs. As the supply of fatty alcohols is high in SLS, plasmalogen synthesis will proceed normally, which leads to degradation of FAR1 that in turn limits PE [O] production. This possibly explains the relatively normal levels of PE [O] in SLS brain. If, as we suggest above, fatty alcohols are formed from reduction of aldehydes originating from catabolism of 2-hydroxy-fatty acids, sphingosine-1-phosphate and plasmalogens, this formation is not FAR1-regulated and represents an uncontrolled influx of fatty alcohols into ether lipid synthesis which could explain the expansion of the ether lipid pool in SLS ( Figure 6). The accumulation of TG[O] has also been seen in cultured SLS keratinocytes, 20 raising the possibility that this unique lipid profile links the pathogenesis of brain and skin symptoms that are so characteristic of this disease. The critical functions of these two organs depend on the formation of multilamellar membranes in myelin and in the stratum corneum. In SLS skin, ultrastructural studies demonstrate that these stacked membranes are reduced in number and interrupted by lipid deposits, resulting in a leaky epidermal water barrier and the dry scaly appearance of ichthyosis. 34 Myelin membranes may be similarly perturbed by accumulation of these same lipids.
In addition to the abnormal ether phospholipid metabolism, neutral lipid metabolism was also disturbed in the SLS brain. We found increased concentrations of TG and CE which, in addition to TG [O], may also contribute to the "lipid peaks" on brain MR spectra of SLS patients and the lipid droplets seen in the histological study of the SLS brain. Patients with other disorders that lead to TG accumulation, that is, defects in DDHD2 and Chanarin-Dorfman syndrome, have similar lipid resonances in brain MR spectra. 35,36 An abnormal low concentration of ether lipids is well known in some peroxisomal diseases with obvious neurometabolic consequences for the brain. 36 In contrast, increased concentrations are found in SLS. Recently, Vaz et al described PCYT2 mutations that also heavily impact ether (phospho)lipid biosynthesis. 37 Patients developed a progressive para-or tetraparesis with intellectual disability. In line with the metabolic role of the PCYT2 enzyme ). The fibroblast lipid profile in PCYT2 deficiency is reminiscent of that found in SLS brain. The brain lipid profile in patients with PCYT2 deficiency has not been analyzed but specific differences with SLS may be anticipated.
The levels of the regular phospholipids PC, PS, PI, PG, and PE and their lyso-forms were, if at all, mildly affected in SLS brain. This is to be expected as fatty alcohols do not play a role in the biosynthesis of these phospholipids. Some of these phospholipids, like PS, LPE and PE even showed a tendency toward a decreased concentration in SLS which could be caused by the consumption of CDP-ethanolamine/CDP-choline to synthesize PE [O] and PC [O]. Another possible explanation might be fatty aldehyde toxicity, causing inactivity of specific enzymes responsible for the production of these lipids. This mechanism was described in a mouse study in SLS, where the enzymatic activity of fatty acid 2-hydroxylase (FA2H activity) was affected by fatty aldehyde accumulation. 38 We detected no increase in free fatty aldehydes in the SLS brain and very long chain aldehydes were not seen. The storage time of the brain samples before analysis may have played a role in this. It is possible that these highly reactive lipids have formed covalent adducts with amino-containing molecules and thus escape detection 39 or were metabolized to fatty alcohols. 40 Lastly, we did not see any leukotrienes and phytol, lipids that have been implicated in SLS pathophysiology. Phytol is probably processed in the liver and may never reach the brain. 41 Leukotrienes are biologically very active lipids and presumably have concentrations below the detection limit and are not detected in our lipidomics platform.
The main limitation of this study is the fact that only one SLS brain and one replica as control could be included. Obtaining a single SLS-brain that also was suited to perform biochemical studies was a unique opportunity. Equally so it has been very hard to obtain suitable control brain tissue (with same sex, age, no brain disease, and the brain postmortem not processed in formalin or otherwise).
In conclusion, we have found a severely disturbed lipid profile in both white and gray matter of the SLS brain, affecting both neutral ether lipids and ether phospholipids along with changes in neutral lipid metabolism. Taken together our data suggest that the brain disorder in SLS results from the accumulation of many different lipid species, especially fatty aldehydes, fatty alcohols, triglycerides and ether (phospho)lipids. Our data show that, at the biochemical level, SLS is not only a white matter but also a gray matter disease. Additionally, the histopathological study suggests that astrocyte dysfunction might play a pivotal role in the underlying disease mechanism, especially contributing to an impairment of myelin maintenance.
DATA AVAILABILITY STATEMENT Data of the study were uploaded as Supporting Information.