INTRODUCTION
Sjogren-Larsson syndrome (SLS) is a neurocutaneous disease with an autosomal recessive
inheritance, caused by mutations in the gene that encodes fatty aldehyde dehydrogenase
(ALDH3A2). It affects predominantly the skin and brain, resulting in the classic triad of
ichthyosis, spastic diplegia, and cognitive impairment.[1]
[2] Brain imaging plays an essential role in the diagnosis, albeit with nonspecific
findings, the most prominent of which is a leukoencephalopathy. Previous data from
brain magnetic resonance imaging (MRI) of SLS patients also report brain atrophy in
some individuals.[3]
[4] While white matter involvement is well established, data related to grey matter
involvement is scarce, with some neuropathological reports indicating discordant areas
of affected grey matter. In the present study, we performed a volumetric analysis
of the brain MRI of two siblings with SLS using an automated segmentation software
(Freesurfer) and compared the volumes obtained to the data of a healthy control group,
with the aim of detecting grey and/or white matter volume reduction in these patients.
METHODS
Patients
Patients that had a confirmed diagnosis of SLS and who were followed in our institution
were selected for the case group. Patients with presumptive diagnosis and no confirmatory
test were excluded from the selection.
For the control group, we searched for subjects with no medical background in our
brain MRI database and selected them on the basis of sex and age so as to match the
case group. Subjects with structural alterations in the brain were excluded from the
study.
The Institutional Review Board/Ethics Committee, Research Ethics Committee of the
Faculty of Medical Sciences – Universidade de Campinas (UNICAMP, in the Portuguese
acronym) approved the study, and informed consent was obtained for all participants.
Magnetic resonance imaging
Magnetic resonance imaging was performed on all patients and control subjects on a
high resolution 3T Achieva Intera Scanner (Philips, Amsterdam, Netherlands) using
a standard protocol consisting of routine T1 and T2 weighted sequences. Additionally,
for the FreeSurfer analysis, a volumetric T1-weighted sequence was acquired using
a standard 8 channel head coil, with a sagittal orientation, voxel matrix 240 × 240 × 180,
voxel size 1 × 1 × 1mm3, TR/TE = 7.0/3.2 ms, flip angle 8°.
Volumetry
Cortical and subcortical segmentation and volumetric analysis was performed with the
Freesurfer 6.0 image analysis suite, which is available for download online (http://surfer.nmr.mgh.harvard.edu/). Technical details can be found in prior publications.[5]
[6]
[7]
[8] Briefly, in its image processing, it includes field inhomogeneity correction, skull
removal, and alignment to Talaraich atlas, followed by the labelling of each voxel
as grey matter, white matter, or cerebrospinal fluid and tessellation of the grey
matter/white matter boundaries. The maps created rely on both the absolute signal
intensity and spatial intensity gradients across the tissues and provide sub millimeter
resolution, beyond the voxel resolution. The process was checked for errors and the
need of manual adjustments, although no correction was warranted. We included in our
analysis the volumes of the deep grey matter structures (caudate, putamen, and pallidum
nuclei, as well as the thalamus), the brainstem, and the cerebellar cortex. Due to
the cerebral white matter signal alteration that is inherent to the disease, we did
not include neither supratentorial cortical thickness and volume nor cerebral white
matter volume in the analysis considering the possibility that the software could
misinterpret the cortical grey matter/white matter boundaries.
Statistical analysis
The normality of the distribution of the volumes obtained from the control group was
tested using the Shapiro-Wilk test, confirming a normal distribution with p-values > 0.05. The brain bilateral structures (cerebellar cortex, thalamus, pallidum, caudate
and putamen) had their volumes represented by the average of the volumes of both hemispheres.
The mean volumes of the case group were compared with the control group using a one-sample
Z-test, since we disposed of the population (control group) standard deviation (SD)
and variance values. The test was one-tailed, considering that the previous literature
data indicated only brain volume reductions in SJL patients. The primary significance
level was p < 0.05, adjusted to p < 0.008 after the Bonferroni correction for multiple comparisons, to avoid type I
errors.
RESULTS
Participants
The search in our institution database returned three patients that fulfilled the
criteria for the case group. They were siblings from non-consanguineous parents from
a small, isolated community where consanguinity is more prevalent and a founder effect
for the SLS could be identified. One of them had to be excluded from the analysis
since the software used could not conclude brain segmentation due to profuse white
matter signal alteration. The patients included in our study were females aged 22
and 37 years old (mean age: 29.5 years old) at the time of investigation. Both of
them presented with the classical triad of congenital ichthyosis, spasticity, and
mental retardation in the first years of life, and had their diagnosis confirmed by
molecular characterization of the ALDH3A2 gene mutation (both had a homozygous mutation c.1108-1G > C in intron 7). They had
moderate cognitive impairment and were dependent on wheelchair.
We included in the control group 41 healthy patients, all females, with ages similar
to the case group (mean age 27.9 years old; range 20 to 37 years old). None had structural
alterations in the brain.
MRI analysis
Both SLS patients presented symmetrical, confluent T2/FLAIR hyperintensity in the
periventricular white matter, predominantly in the frontal lobes, sparing subcortical
fibers. The older patient also presented T1 hypointensity in these regions.
Data analysis
All volumes obtained from the control group were found to be normally distributed
using the Shapiro-Wilk test.
The structures analyzed are represented on [Figure 1]. There was a significant volume reduction of the cerebellum cortex (p < 0.001), the brainstem (p = 0.002), the thalami (p = 0.005), and the pallidum nuclei (p = 0.005) ([Figure 2]). We did not find significant reduction in the volume of the caudate nuclei (p = 0.3) and putamen (p = 0.01). The [Table 1] demonstrates neuroimaging data with FreeSurfer. The severity of white matter disease
is demonstrated in [Figures 3] and [4].
Figure 1 Schematic view of the analyzed structures segmentation.
Figure 2 P-values obtained from each structure represented in a color scale. Structures without
statistically significant volume reduction are represented in orange (putamen: p = 0.01) and yellow (caudate nuclei: p = 0.3).
Table 1
Volumes (mean ± SD) obtained using FreeSurfer, in cubic centimeters.
|
Volumes
|
Control group
|
SLS patients
|
Z-score
|
p-value
|
|
Total brain
|
1071.52 ± 65.82
|
968.79 ± 10.14
|
- 2.20
|
0.013
|
|
Cerebral WM
|
425.70 ± 39.10
|
329.18 ± 41.07
|
- 3.49
|
< 0.001
|
|
Cortex
|
444.43 ± 26.37
|
428.13 ± 11.61
|
- 0.87
|
0.192
|
|
GM
|
601.13 ± 32.50
|
561.95 ± 5.16
|
- 1.70
|
0.044
|
|
Brainstem
|
19.55 ± 1.92
|
15.74 ± 0,14
|
- 2.79
|
<0.002
|
|
Left caudate
|
3.52 ± 0.37
|
3.56 ± 0.03
|
0.15
|
0.440
|
|
Right caudate
|
3.63 ± 0.36
|
3.47 ± 0.01
|
- 0.62
|
0.264
|
|
Left cerebellum cortex
|
49.48 ± 3.45
|
40.51 ± 2.43
|
- 3.67
|
<0.001
|
|
Right cerebellum cortex
|
49.61 ± 3.39
|
41.50 ± 2.60
|
- 3.37
|
<0.001
|
|
Left cerebellum WM
|
14.41 ± 1.42
|
11.53 ± 0.41
|
- 2.85
|
<0.003
|
|
Right cerebellum WM
|
13.59 ± 1.40
|
11.09 ± 1.01
|
- 2.52
|
<0.006
|
|
Left hippocampus
|
4.03 ± 0.30
|
3.95 ± 0.15
|
- 0.37
|
0.351
|
|
Right hippocampus
|
4.14 ± 0.32
|
3.79 ± 0.00
|
- 1.56
|
0.058
|
|
Left pallidum
|
1.86 ± 0.18
|
1.51 ± 0.19
|
- 2.85
|
<0.002
|
|
Right pallidum
|
1.84 ± 0.16
|
1.56 ± 0.17
|
- 2.80
|
<0.003
|
|
Left putamen
|
4.98 ± 0.39
|
4.45 ± 0.66
|
- 1.93
|
0.026
|
|
Right putamen
|
4.96 ± 0.38
|
4.37 ± 0.55
|
- 2.22
|
0.012
|
|
Left thalamus
|
7.14 ± 0.66
|
5.92 ± 0.44
|
- 2.60
|
<0.005
|
|
Right thalamus
|
6.77 ± 0.55
|
5.74 ± 0.07
|
- 2.61
|
<0.005
|
Abbreviations: SD, standard deviation; SLS, Sjogren-Larsson syndrome; WM, white matter;
GM, grey matter. Note: *significant p < 0.008 (Bonferroni corrected one-tailed Z-test).
Figure 3 Axial T2 weighted image demonstrates confluent and symmetrical cerebral white matter
hyperintensities in the periventricular regions and in the posterior limb of the internal
capsules.
Figure 4 Axial T2 weighted image at the level of the centrum semiovale demonstrates confluent
and symmetrical deep white matter hyperintensities.
DISCUSSION
Sjogren-Larsson syndrome is an inherited autosomal recessive neurocutaneous disease
caused by an inborn error of the lipid metabolism resulting from mutations in the
ALDH3A2 gene that encodes fatty aldehyde dehydrogenase (FALDH). Fatty aldehyde dehydrogenase
is an enzyme involved, among other lipid metabolism pathways, in the removal of aldehydes
from both endogenous and exogenous sources through oxidation into fatty acids. Its
activity deficiency leads to abnormal lipid accumulation in the membranes of the skin
and brain. In the skin, it leads to disruption of the epidermal water barrier and
transepidermal water loss, while in the brain it may alter the structure and/or the
function of myelin membranes.[1]
[2]
[4]
[9]
[10]
[11] There are several mutations that result in SLS, some being more common in certain
geographical distributions. The c.1108-1G > C variant was initially described in the
Iberian Peninsula and is probably the most common pathogenic mutation in Brazilian
patients with SLS.[1]
Clinically, it is characterized by a classic triad of ichthyosis, spastic diplegia,
and cognitive impairment. The ichthyosis is typically congenital in onset with erythema
and scaling apparent in the neonatal period.[12]
Neurologic symptoms become evident in the first or second year of life, with a delay
in motor milestones as a consequence of spasticity. This motor disability tends to
be more prominent in legs, leading to difficulty in ambulation, uncommonly affecting
upper limbs. The majority of patients with SLS become wheelchair dependent, although
a smaller number of them is able to walk with the use of walking assistive devices.[1]
[11]
[12]
[13]
Cognitive deficits range from mild to moderate, and, unlike other neurometabolic diseases,
do not progress over the decades.[11] Usually, patients are able to reach a certain level of independency.[1]
Other manifestations include preterm birth, pseudobulbar dysarthria, seizures, and
ophthalmologic alterations. The latter are present after 3 years of age, with features
that are pathognomic when present, characterized by retinal glistening white dots
representing crystalline inclusions, and microcysts in the fovea.[1]
[2]
[11]
Brain imaging plays an important role in the evaluation of patients with SLS, even
though there are no pathognomonic findings. Computed tomography (CT) scans depict
nonspecific hypodensities in the cerebral white matter.[14] Brain MRI, however, is the imaging modality of choice, due to its lack of ionizing
radiation and better characterization of grey and white matter abnormalities, as well
as morphological changes. The prominent finding is diffuse, confluent signal alteration
in the periventricular and deep white matter sparing the subcortical fibers, reflecting
gliosis, delayed myelination, dysmyelination, demyelination, and lipids accumulation.[14]
[15] A frontal or parieto-occipital predominance has been described.[4]
[11] Some case studies also report brain atrophy in SLS patients, although there is not
sufficient data to conclude if it is progressive.[3]
[4]
[16] Magnetic resonance spectroscopy of the white matter abnormality is another useful
resource and depicts peaks assigned to lipids in 0.8 to 0.9 ppm and in 1.3 ppm, which
probably reflects accumulation of fatty alcohols or their metabolites.[17]
[18]
It is well recognized that the main imaging findings are related with the leukoencephalopathy.
Data related to grey matter involvement is scarce, with some neuropathological studies
reporting conflicting results concerning neuronal population loss of the putamen,
the substantia nigra, and the thalamus.[19]
[20]
[21]
[22]
[23] A limitation for most of these studies is that there was no definitive confirmation
of the SLS diagnosis. Jagel and Jagel et al.[24]
[25] mention some that presented cases with typical SLS features[19]
[20]
[21] and other with atypical features.[22]
[23] More recently, Staps et al.[26] conducted a postmortem analysis of a genetically confirmed SLS patient. Taking in
consideration only the reports with typical clinical features and the one with confirmed
diagnosis, there are still conflicting results. Sylvester[19] describes fat droplets in grey matter as well as an increased number of astrocytes
in the cortex, caudate, and lentiform nuclei; also noted were a slight loss of Purkinje
cells in the cerebellum and focal areas of cerebellar atrophy. Silva et al.[20] reports normal neuronal population and myelination in the cerebral cortex, the basal
ganglia, and the internal capsule; the only finding was demyelination of the corticospinal
tracts in the medulla. Wester et al.,[21] however, analyzed the brain of two SLS patients and reported a slight loss of neurons
in the telencephalic cortex, the putamen, and the striatum, with the most striking
degeneration in the substantia nigra. Finally, Staps et al.[26] found no grey matter neuronal loss but concluded that the disturbed lipid profile
affects both white and grey matter, thus suggesting that it is not a pure leukoencephalopathy.
In the present study, using brain volumetry, we were able to noninvasively evaluate
and compare deep grey matter, brain stem and cerebellar grey matter volume of SLS
patients with a healthy control group. It should be noted that the reduced grey matter
areas found in the present study are directly or indirectly involved in motor functions.
Clinically, there are motor deteriorations throughout the lives of SLS patient, not
accompanied by cognitive deterioration, thus leading to the postulation that they
are secondary to limitations during the motor development age.[13] However, with proofing of atrophy of structures related to movement control, it
can be questioned whether these deteriorations are not directly caused by the disease
itself.
Interestingly, it is well-known that the thalamus is part of pathways between the
cerebellum, the basal ganglia, and the frontal cortex. Targeted thalamotomies and
deep brain stimulation have been successfully used to treat involuntary movements,
confirming its essential role in motor functions. There are cerebellum-recipient thalamus
nuclei[27] and basal ganglia-recipient (pallidum and substantia nigra) thalamus nuclei, with
afferent pathways to the frontal cortex.[28] Coincidentally, the structures involved in these pathways are the structures with
significant volume reduction in the present study (that is, the cerebellum, the thalamus,
and the pallidum). The substantia nigra was not evaluated in our analysis, but, as
noted before, was shown to be strikingly affected in a previous report.[21] Perhaps, disruption of these pathways is accountable for both the spastic diplegia
present in these patients and the volume reduction observed in the present study.
To our knowledge, no article specifically addresses brainstem alterations in SLS patients,
but there is evidence of demyelination of corticospinal tracts.[20] In addition to that, we can postulate that cerebellothalamic tracts involvement
is also related to the brainstem volume reduction observed in our analysis.
The present study brings light to the possibility that Sjogren-Larsson syndrome has
a more pronounced involvement in the thalamic motor pathways. With this new knowledge,
more studies could better investigate grey matter alterations in these patients in
order to try to establish if these are related to the motor disability clinically
observed, whether as the cause of it or as a long-term consequence.
We consider the hypothesis that motor function deterioration might be related to the
reduced grey matter areas found in the present study, as these structures are related
to movement control; in particular, the present study brings light to the possibility
that SLS has a more pronounced involvement in the thalamic motor pathways, while striatal
structures would be less interested. Further investigation on a possible selective
involvement of grey matter areas or white matter tracts in these patients are needed,
maybe integrating DTI studies, which may provide information about white matter lesions.[30]
A critical inquiry arises as to whether these deteriorations stem directly from the
disease itself or if they represent long-term consequences. The present study carries
a limitation rooted in the application of FreeSurfer, which holds the potential to
introduce errors, especially within regions of white matter hyperintensity, consequently
leading to inaccuracies in the segmentation of gray matter.[29] It is important to acknowledge further limitations within our study, notably the
relatively small sample size and the examination of MRIs of the SLS patients at a
single time point, involving individuals in their adult years --by which juncture
symptoms had already experienced substantial progression. Although reports of cognitive
impairment are present, definitive conclusions regarding the evolution of imaging
findings remain constrained by the paucity of data. It is crucial to recognize that
SLS encompasses a broad spectrum of genotypes and phenotypes, which might account
for the seemingly inconsistent data present within the literature.
Data obtained from this automated segmentation software using only a T1-weighted volumetric
sequence can noninvasively lead to new understandings of the disease.
