CS-associated mutations

In total, 37 distinct families were enrolled with 44 affected males. 31 unique mutations were identified in the cohort (table 1; figure 1). Of the 31 unique mutations, there were: 15 frameshift and/or nonsense mutations; 4 deleterious missense mutations; 9 splice mutations; 2 copy number variant deletions; and 1 mutation that removes the first methionine residue. Of those with known inheritance (29 of 37 families), 16 cases were inherited (55%) and 13 cases were de novo (45%). Notably, 5 mutations were recurrent (ie, occurred in more than 1 family). Each of the following mutations occurred in 2 distinct families: c.899+1 del (mutation 14), c.1024G>A (mutation 16), c.1148G>A (mutation 17) and c.1710G>A (mutation 30). Mutation 25, c.1498C>T, was found in 3 families. These recurrent mutations were inherited in 3 families not known to be related and were de novo in 5 families. Inheritance was unknown in the 3 remaining families with recurrent mutations. No strong genotype–phenotype associations were evident (online supplemental table 1).

Figure 1

Christianson syndrome-associated NHE6 mutations. A total of 31 unique NHE6 mutations were identified in 44 CS male probands across 37 families. Pathogenic NHE6 mutations include: Frameshift and/or nonsense (n=15, cyan or orange), deleterious missense (n=4, black), splice (n=9), CNVs (n=2, green) and one that removes the first methionine residue. Out of CS pedigrees with known inheritance (78%, 29 out of 37 families), 55% were inherited (16 out of 29 pedigrees) and 45% were de novo (13 out of 29 pedigrees). A total of 5 mutations were recurrent across multiple pedigrees: c.899+1delGTAA (mutation 14, 2 pedigrees), c.1024G>A (mutation 16, 2 pedigrees), c.1148G>A (mutation 17, 2 pedigrees), c.1498C>T (mutation 25, 3 pedigrees) and c.1710G>A (mutation 30, 2 pedigrees). The following mRNA isoforms are used: NM_001042537.1, NM_006359, NM_001379110.1 and NM_001177651. CS, Christianson syndrome; CNVs, copy number variants; TM, transmembrane.

CS core and associated symptoms across the lifespan

In total, 44 males were evaluated with SLC9A6 mutations, ranging from age 2 to 32 years. In order to consider CS symptoms across developmental stages, cross-sectional analysis was conducted across four age groups (table 2): 0–5 years at enrolment (n=11); 6–11 years (n=9); 12–16 years (n=13); and 17–32 years (n=11). In terms of race and ethnicity: 11% of the sample identified as Hispanic/Latino (n=5); 89% identified as White (n=39); 5% identified as multiracial (n=2); 2% identified as Black (n=1); and 2% identified as other (n=1). Participants were born in 11 different countries and seven different primary languages were represented in these families.

Table 2

Clinical characteristics of CS males across developmental cohorts

Based on our prior study of 14 patients with CS, we proposed CS core symptoms (present>85%) and secondary symptoms (present in>35%).1 Core symptoms included ID (severe to profound), non-verbal status, epilepsy, ataxia, postnatal microcephaly and hyperkinesis (ie, behavioural hyperactivity).1 The previously proposed core symptoms were upheld here in this larger study on average across ages by combining all age groups (table 2). Specifically, all of the core symptoms were present in greater than 85%, including ID (100%), epilepsy (100%), non-verbal/non-phrase speech (100%), microcephaly (95%), ataxia (95%) and hyperkinesia (93%). Of the 44 participants with CS, 32 never had spoken words. While there were 12 participants with words, generally this reflected five or fewer words and none had phrase speech. For caregivers of the 12 participants with words, nine caregivers also noted the loss of words with time. Notably, we also discerned that high pain tolerance represents a new core symptom when considering all age groups together, present in over 85% (91%).

Secondary symptoms occurred in 35–84% of participants and included current sleep problems (51%), unprovoked laughter (62%), visual acuity problems (52%), current GERD (56%), constipation (55%), eye movement abnormalities (68%), regression (54%), swallowing problems (44%) and contractures (36%) (table 2). Previously identified secondary symptoms such as prior Angelman syndrome diagnosis and Autism Spectrum Disorder (ASD) diagnosis occurred in 29% and 18% of the current sample, respectively. ASD diagnosis and prior Angelman diagnosis were more common in the older groups, perhaps reflecting more access to accurate genetic testing in the younger group. Scoliosis occurred in 39% of participants; two participants had corrective surgeries for scoliosis between 15 and 20 years of age.

While the core symptoms were relatively stable across age groups, there were several changes in frequency across ages that were notable. Absence of phrase speech, ID, ataxia and epilepsy were always observed in all patients. We found that 80% of the youngest age group (strictly below our cut-off for core symptoms) had microcephaly which elevated to 100% in all older age groups, reflecting the idea that the microcephaly in CS has a strong postnatal component (table 3). This variation in head circumference data reflecting decreases across ages approached significance (p=0.089). Similarly, 73% of participants had a high pain threshold in the youngest age group (0–5 years old) which elevated to above 85% (our cut-off for core symptoms) in the older age groups therefore, although prevalent in the youngest age group, high pain threshold may not be considered a strict core symptom in the youngest age group (table 3). In the current group, ataxia was present in 82% of participants in the youngest age group and in 100% of subsequent age groups, again suggesting that ataxia worsens with age. In line with these observations, with regard to the ability to walk, 100% of participants report walking unaided by or before age 5; however, the ability to walk unaided is lost in approximately 20% of participants in the 6–16 groups and 30% are reported as not able to walk unaided in the oldest group (17+ years old). Finally, with regard to hyperkinesis, this core symptom was present in 100% of the participants in the two younger age groups and 92% and 82% of the oldest age group suggesting that hyperkinesis does not meet the cut-off for a core symptom in the oldest age group. Concurrent with these data, we observe that the hyperactivity subscale scores on the ABC are most severe in the younger groups and less prominent in the oldest group (table 2), although this does not reach statistical significance.

Table 3

Core symptoms of Christianson syndrome by age group

There were statistically significant differences across developmental groups for some reported symptoms or phenotypic features that are particularly notable. There were significant group differences in the rate of current constipation, prior history of regression, adaptive function (including social, communication and daily living subscores) and number of hospitalisations (p<0.05). Each of these experiences were most common in the adult group (table 2). Importantly, 100% of participants in the oldest age group reported notable problems with constipation (p=0.002). With regard to regression, reports of any regression at some point in their lifetime increased significantly with age from toddlers (30%), children (44%), adolescents (50%) and adults (90%) (p=0.047). With regard to adaptive functioning, age-normed Vineland scores appeared stable in toddlers and children, and then dropped in the adolescent group and again in the adult group (p<0.001 for communication, p<0.001 social and daily living, and p<0.0005 for adaptive functioning). While not representing a statistically significant difference across groups (p=0.305), it is notable that the highest level of inability to walk is experienced in the adult group at approximately 30% (table 2).

Longitudinal assessment of growth and physical stature across the lifespan in CS

Adult patients with CS have been described as having a thin body habitus3; however, systematic analysis of growth in CS across the lifespan has not been conducted. To study growth in CS across the lifespan, raw measurements of height and weight were converted to age-normed percentiles (WHO growth standards for 0–2 years18; CDC growth curves for 2–20 years19) and to z-scores. To model physical stature across development, we used a linear mixed model (with random slopes and intercepts) for each variable of interest (ie, height, weight) with the age of the child as the time basis. The height (n=24) and weight (n=30) obtained at clinical appointments were analysed via mixed linear modelling of age-based percentiles (figure 2). While height and weight naturally increased with time (figure 2A,B), the age-normalised weight, i.e., the percentile (n=30; total data points=116; range=1–12 points per person) linearly declined over time (slope p<0.001) (figure 2C). Height percentile (n=24; total data points=81; range=1–11 points per person) also linearly declined over time (slope p<0.001). Of note, the predicted intercept for the model at birth was within normal ranges; however, as shown the model demonstrates a progressive slower growth in CS relative to normal growth with age (figure 2). Of relevance to challenges in growth and nutrition, approximately 30% (13/44) of participants had G-tubes placed. Approximately 38%, 23%, 23% and 15% had their G-tubes placed during ages of 0–5, 6–11, 12–16 or 17+years of age. Indications for G-tube placement have been failure to thrive, most often at the younger ages, and inability to eat across all ages, generally related to challenges with swallowing.

Figure 2

Changes in height and weight across development. (A) Raw score of height (inches) by age (years) from 24 CS probands spanning <1–19 years. Height increases over time (Pearson’s correlation, R2=0.92). (B) Raw score of weight (pounds) by age (years) from 30 CS probands spanning <1–19 years. Weight increases over time (Pearson’s correlation, R2=0.90). (C) Estimated age-normed height and weight based on liner mixed model analysis. Both age-normed height (slope p<0.001) and age-normed weight (slope p<0.001) significantly decline over time. Raw height (n=24, total data points=81) and weight (n=30, total data points=116) measurements were converted to age-normed percentiles (WHO growth standards for 0–2 years18; CDC growth curves for 2–20 years19) and z-scores. Our linear mixed model (random slopes and intercepts) of age-based percentiles for height and weight modelled the intercept and age at time point as fixed and random effects, respectively. Covariance parameters for height were large relative to their standard errors with statistically significant intercept (slope p=0.003), slope (p=0.016) and covariance of intercept and slope (p=0.046). Covariance parameters for weight were large relative to their standard errors with statistically significant intercept (slope p=0.002), slope (p=0.016) and covariance of intercept and slope (p=0.023). CS, Christianson syndrome; CDC, Centers for Disease Control and Prevention.

Adaptive, behavioural and motor functioning across the lifespan

Cross-sectional findings. Parent-reported adaptive and behavioural data were also available for a subsample of the males (n=27–31). When examining differences across the cohort by current developmental stage (cross-sectional analysis), a statistically significant difference across groups was identified (p<0.0005) in age-normed overall adaptive functioning. Specifically, older participants had lower age-normed abilities than younger participants, in that the younger participants had abilities more similar to their same-aged peers while the abilities of the older participants were further below those of their same-aged peers (table 2). In further examining across these cross-sectional age groups, there were no significant differences between age groups in hyperactivity or irritability at baseline (table 2; ABC).

Adaptive/motor findings. Importantly, we also conducted a longitudinal follow-up analysis through investigation of change in adaptive function in individuals (n=22) at baseline and at a 1 year follow-up time point. The change in skills between baseline and follow-up was correlated with age at enrolment. Results showed that participant age was negatively correlated with a change in fine motor skills (r=−0.50; p=0.028) and gross motor skills (r=−0.63; p=0.002; figure 3). Specifically, 5 of 6 adults (83%) experienced a loss in multiple fine motor skills such as picking up small objects with thumb and fingers and moving objects from one hand to the other. Expressive language and interpersonal skills change were not significantly correlated with age (figure 3). These longitudinal studies suggest that older participants are at risk for losing previously acquired skills, particularly in the motor domain, which is observable by parents/guardians at a 1 year follow-up. Therefore, these longitudinal studies augment the above cross-sectional studies.

Figure 3

Changes in motor, language and social functioning as a function of baseline age over a 1 year period. Parent-rated changes in gross motor (A), fine motor (B), expressive language (C) and interpersonal/social functioning (D) from enrolment to 1 year follow-up (n=22). Each point represents an individual proband. Raw score change between enrolment and 1 year follow-up is plotted by age at enrolment. Changes in gross (A, R2=0.40, p=0.002) and fine (B, R2=0.25, p=0.028) motor is significantly correlated with age; older age is associated with greater loss of motor skills over a 1-year period. There is no association between age and changes in expressive language (C, R2=0.002, p=0.83) or social functioning (D, R2=0.00, p=0.96) over this time span. Pearson’s correlation.

Neurological examination. Standardised, in-person motor examination was performed by a board-certified neurologist (JSL). 12 patients were directly examined (online supplemental table 2). The cross-sectional direct assessment of patients additionally supports the interpretation that there is age-related worsening of motor function. Subjects have evidence of cerebellar dysfunction including tremor and truncal ataxia as evidenced by wide-based gait. Importantly, also older patients (ie, age 20 and greater) present with evidence of motor dysfunction with signs of corticospinal tract, i.e., upper motor neuron damage such as increased tone, weakness, increased and abnormal reflexes. Thus, the clinical phenotype includes upper-motor neuron involvement that emerges particularly in adulthood as well as cerebellar dysfunction that is also progressive but presents first in early adulthood.