Phenotypic variation between siblings with Metachromatic Leukodystrophy

Metachromatic Leukodystrophy (MLD) is an autosomal recessive, monogenic disease caused by mutations in the arylsulfatase A (ARSA) gene, leading to deficiency of the enzyme ARSA and therefore inadequate degradation of sulfatides [1, 2]. Sulfatides accumulate especially in the central and peripheral nervous system, and lead to progressive demyelination and neurological symptoms [1, 2]. The clinical course can be divided into a pre-symptomatic stage with normal development, followed by onset of first symptoms and a period of developmental stagnation. This plateau phase is shorter in early onset forms, and longer and more variable in late onset forms. Finally, rapid disease progression evolves with a relatively invariable rapid loss of gross motor function, and a final stabilization at a low functional level [3].

Genotype-phenotype correlation revealed that null alleles, which cause hardly any residual ARSA activity [4, 5], result in an early onset and rapid deterioration of motor and cognitive function characterizing the late-infantile form of MLD with first symptoms occurring before 2.5 years of age [6]. In later onset forms (juvenile MLD with disease onset between 2.5 and 16 years, and adult MLD with disease onset after 16 years of age), the prevalent genotypes were associated with some remaining residual activity of the enzyme [4, 5]. Although this allows some genotype-phenotype correlation, the exact relationship between genotype, residual enzyme activity, and clinical phenotype remains to be elucidated. Today more than 250 ARSA mutations are known, making it challenging to define more precise genotype-phenotype relationships especially in the later onset forms, which often show compound heterozygosity for different mutations [7‐11].

The phenotypic variability becomes especially relevant when treatment evaluation is based on comparison with an untreated sibling carrying the same mutations [12, 13]. Younger siblings of an affected older sibling are usually diagnosed in an early (or even pre-symptomatic) disease stage and have the opportunity to undergo a treatment like conventional or gene-corrected stem cell transplantation [14‐17]. However, it is unclear how their clinical course would have evolved without treatment. Case reports suggest that there might be some variability of disease onset even between siblings with MLD [18, 19], but a systematic analysis is currently missing.

The aim of this study is to systematically investigate phenotypic variability in a significant number of siblings with MLD, regarding disease onset and progression, and compare them to the phenotypic variability in a large cohort of children with MLD. In addition, case reports in the literature will be discussed.

The results are not only relevant for counseling of families with affected children, but will also allow better interpretation of treatment outcomes, and might help to better understand the complex genotype-phenotype relationships in MLD.

MLD is an inherited disease with an autosomal recessive trait, sometimes affecting several siblings in one family. With regard of the natural disease course in siblings, there exist only case reports. Some of them indicate a very similar clinical course and some show more variability between the siblings. This study represents the first systematic analysis of the phenotypic variability in siblings with MLD.

We were able to show that age of onset may vary considerably between siblings with MLD. Although some sibling pairs share a very similar age of onset, others show a high discrepancy. Within the same family, however, the onset and course is either late infantile or juvenile. Even within late infantile MLD onset in siblings may be up to 12 months apart. In juvenile MLD, around 80% have a disease onset of more than 1 year apart and around 40% more than 2 years. This variability in age of onset is similar between the siblings and between randomly chosen unrelated pairs of the MLD cohort.

Case reports of siblings in the literature, summarized in Table 2, also show this variability. The first description of siblings with MLD was reported by Scholz in 1925 [22]. He described two siblings with the juvenile form and a quite similar age of disease onset at around 8 years of age, both with cognitive and motor symptoms at onset (brother 1: spasticity in legs resulting in rigid and slow walking, optical atrophy, learning and perception problems as well as emotional lability; brother 2: spasticity in extremities, tremor in legs, whiny, unstable mood and cognitive problems (learning, perception)) and a rapidly progressive disease course, both died around 3 years after onset in a very advanced disease stage. Alves et al. [23] reported on four cases of late onset (15, 17, 18, 21 years) MLD in a family with 15 siblings. In accordance with our results the described siblings all had the same symptom cluster at onset - mental deterioration and/or behavioral changes. A quite similar case by Satoh et al. [24] reported on a sibling pair with motor and cognitive symptoms at disease onset at the ages of 19 and 15 years. Koul et al. [25] described two affected siblings with late-infantile MLD both with an onset with deterioration of motor function at 2 years of age. Mahmood et al. reported triplets with the late-infantile form, all of them with motor function starting around 16 months of age after normal psychomotor development [26]. Another report of siblings with late-infantile MLD by Nyberg-Hansen et al. [27] delineated 2 brothers, who both developed motor dysfunction and dysarthric speech within the second year of life with rapid progression in the following years. Aslan et al. [28] reported on two siblings with the juvenile form of MLD, who had their first symptoms at 6 and 7 years of age respectively, also showing a similar symptom constellation at onset with both motor and cognitive signs (the first sibling: ataxia, attention deficit and perceptual difficulties; the second: learning difficulties, mild changes in fine and gross motor function). The later onset case reports by Hoes et al. [29], Cengiz et al. [30] and Manowitz et al. [31] also have in common a similar symptom constellation at onset (Table 2). Hoes et al. reported on two brothers: the first developed first symptoms at the age of 27 years with problems in social interaction soon followed by hypersexual, aggressive and apathetic behavior, he died at the age of 30 years after a kidney biopsy [29]. His younger brother showed first symptoms at the age of 26 (social handicap, psychiatric symptoms) and died within 8 years after onset [29]. Cengiz et al. described three sibling with variability in their age at onset of clinical symptoms (21, 18 and 12 years) but very similar symptom constellation (1: behavioral disturbance, working memory difficulties, epileptic seizures; 2: epileptic seizures and progressive mental changes; 3: poor school performance, disinhibition, hyperactivity and self-awareness) [30]. Furthermore Manowitz et al. described an adult sibling pair with cognitive symptoms at onset and late initiating motor progression [31]. These case reports corroborate our findings by showing some variability in age of onset (maximum 9 years difference in Cengiz et al.), but clear similarity in the type of first symptoms and disease progression (e.g. Scholz 1925: rapid, motor and cognitive, died within a few years; Alves: pronounced severe mental deterioration in disease course). This is especially remarkable as children with juvenile MLD are known to show a heterogeneous presentation and clinical course. One report from 1981 by Yatziv and Russell et al. [32], was difficult to interpret. They described a family with three affected siblings; the eldest showed first symptoms (developmental delay) aged around 1 year, while the two younger siblings were reported with a disease onset at 6 years, with all three developing dystonia as their dominant sign with normal cognitive function into adulthood. As none of the other symptoms usually reported in MLD were described [3, 21] and the diagnosis was based on enzyme measurements, where pseudodeficiency and carrier status for pathogenic mutations may explain relatively low values, diagnosis of MLD seemed uncertain, and we did not include this report in Table 2.

Apart from this case report, siblings of children with late-infantile MLD all had the late-infantile form with the typical rapid disease progression. It may not surprise that within the later onset forms, with a higher residual enzyme activity than in the late-infantile form, the age of onset can vary considerably. For example, Arbour et al. 2000 [19] reported on a Vietnamese family with two affected siblings, both with the same genotype, of whom the first was earlier and more severely affected starting with cognitive symptoms and epileptic seizures at the age of 7 years with a more rapid disease progression, whereas the second sibling had very mild cognitive symptoms at the age of 22 years. Another sibling pair was described by Clarke et al. 1989 [18], with clinical manifestation of MLD at the age of 9 years with cognitive symptoms (learning and behavioral difficulties) for one affected sibling. The second affected sibling presented none of the common symptoms until the age of 23 years (age at description) except sulfatide accumulation in the gallbladder. Also the siblings with juvenile MLD reported by Kappler et al. in 1992 show this kind of difference in the age of onset, with the older one presenting with behavioral changes at the age of 14 years [33].

It can currently only be speculated that there must be unrecognized factors that influence the phenotypic variability beyond the genotype level. It is interesting to note that even when children with the same genotype are investigated, the siblings amongst them (who share a substantial part of their genome) do not have a more similar age of onset compared with non-related children. It is likely that there are other epigenetic, metabolic or unidentified factors which influence the onset of first symptoms of MLD in addition to the genotype. Our results underline, that although probably valid in many cases, comparing treated patients to their non-treated siblings, has to be done with caution. Comparison to a larger non-treated cohort of children with MLD might be a more valid approach [34]. This underlines the importance of natural history data as retrospective controls for treatment trials, in the absence of the feasibility of randomized placebo-controlled studies in rare progressive disorders, like MLD.

There are some limitations in respect to our findings. Most importantly, although being the most comprehensive sibling study, larger numbers of patients would be desirable. While in our study we were able to sustain a high data quality due to the same investigators and clinical standards, a multi-centric international collaboration will increase patient numbers and potentially confirm our results with more statistical power and certainty.

In addition, data on disease onset are often retrospective (as long as neonatal screening is not available), and analysis of clinical parameters can be challenging. For example, age of disease onset might be influenced by a certain parental interpretation and perception. We have, however, used clearly defined clinical parameters and have validated parent reported disease onset by medical records and phone interviews [3, 21]. Furthermore, we have not observed a tendency in recognizing earlier symptom onset in the second sibling when the older sibling was already diagnosed. This is underlined by the fact that in 5 of 12 sibling pairs the first diagnosed (older) sibling had a disease onset at an earlier age than the next (younger) sibling.

Symptom constellations at disease onset may be difficult to identify. Some symptoms might be harder to pinpoint at a specific age (e.g. concentration problems, which might slowly start or be unspecific), other symptoms might start shortly after the first. Therefore, we considered them to belong to the first symptoms even if they appeared a short time after. From a pathophysiological or biochemical perspective, symptom onset might be related to a supra-threshold level of progressive pathological changes, like sulfatide accumulation, demyelination, or axonal damage. Further information is needed from biochemical or MRI biomarkers in order to define such supra-physiological thresholds.

In the most comprehensive study presented so far, we investigated the phenotypic variability of siblings with MLD compared to unrelated pairs of children with MLD. We found that siblings with MLD show a similar type of first symptoms, MRI pattern, and dynamic of disease progression. However, regarding the age of disease onset some sibling pairs show considerable variability. As genotype-phenotype-correlation was not higher in siblings than in unrelated children with the same genotype, this suggests that additional biochemical and epigenetic factors might influence the clinical phenotype.

These data are important for family counseling, but also essential for the evaluation of treatment trials, where untreated siblings are often used as controls.