Supplementary information
Supplementary Figure 1: Detail of the two phases of inclusions.
(a, b) Data previously published in Melke et al study (17), from 43 individuals with ASD, 34 parents and 48 controls. (c, d) Independent data collected for this study from 235individuals with ASD, 343 parents, 129 unaffected siblings and 368 controls. The control group was compared to other groups using the Wilcoxon two-sample test. Boxes indicate medians and quartiles.
Supplementary Figure 2: Effect of age on biochemical parameters and age-paired comparisons.
The ASD group was compared to the control group for each parameter and each age category (children below 16 years old and individuals above 16 years old) using the Wilcoxon two-sample test. Wilcoxon two-sample tests were also performed to determine if each biochemical parameter was dependent on age. Boxes indicate medians and quartiles. (a) Whole blood serotonin, (b) Platelet N-acetylserotonin,(c) Plasma melatonin. (d) Age distributions in each group.
Supplementary Figure 3: Correlations between the three biochemical parameters in controls (a-c), and correlations between NAS and melatonin in first-degree relatives (d: parents; e: sibs) of patients with ASD.
Supplementary Figure 4: Sleep problems in patients with ASD.
Only items displaying significant differences in frequencies between patients and controls are shown.Groups were compared using the Wilcoxon two-sample test or the Cochran Armitage trend test. Boxes indicate medians and quartiles. (a) Sleep disorders were assessed by self reporting in 147 patients and 96 controls. In addition, one subgroup of children and one subgroup of adults underwent a detailed sleep evaluation. (b-c)28 children with ASD and 35 age- and sex- matched control children completed the Children’s Habits Sleep Questionnaire (CSHQ) (43), including a «bedtime resistance» score (b) and a «sleep onset delay» item (c). (d-f)20adult patients with normal IQ ASD and 8 controls were assessed by actigraphy and with sleep questionnaires including validated French versions of the Pittsburgh Sleep Quality Index (PSQI) (44)(d), the Horne and Ostberg questionnaire (45) and the Epworth sleepiness scale (46). Participants wore the actigraph device (AW-7 CamNtech®) continuously on the non-dominant wrist for 21 consecutive days. They were instructed to press the event-marker when they went to bed to sleep and when they got out of bed to start the day. They concurrently kept a sleep diary. Data were sampled in one-minute intervals and analyzed with the sleep detection algorithm provided by the Actiwatch software (Actiwatch Activity & Sleep Analysis Ltd CamNtech® 7.28). Sleep scoring was performed using the event-marker and the sleep diary, with visual inspection, and17 actigraphy parameters were automatically computed, including sleep latency (duration from bedtime to sleep start)(e)and relative amplitude (difference of activity levels between daytime - 10 most active hours - and nighttime - five less active hours) (f).
Supplementary Figure 5: Association between sleep difficulties and biochemical impairments in patients with ASD.
(a) 145 patients with ASD were assessed for sleep difficultiesby questionnaire. (b-d) Actigraphic recordings were performed for 20 adult patients with high-functioning autism. The association between 17 actigraphic parameters and the 3 biochemical impairments were assessed. (b)diurnal motor activity M10 (activity levels during the 10 most active hours of the 24-hour cycle), (c)amplitude (difference between the minimum and maximum of the motor activity)(d) sleep latency. Boxes indicate medians and quartiles. Groups were compared using the Wilcoxon two-sample test.
Supplementary figure 6: Specificity of serotonin and melatonin impairments in ASD vs. ADHD and ID patients.
(a) Whole blood serotonin and (b) plasma melatonin were assessed in patients with attention-deficit/hyperactivity disorder (ADHD - according to DSM-IV check-list; 23 males, 15 females; mean age 25 ± 8.3 years) or intellectual disability without autism (ID - defined as verbal IQ and performance IQ 70with a negative screen of ASD according to ADI-R and/or ADOS; 39 males; mean age 16 ± 4.2 years), and compared with patients with ASD and controls presented in this study. (c) ROC curves.
Supplementary Table 1: description of the populations investigated.
Controls / Probands with ASD / Parents / Unaffected siblingsn (total) / 416 / 278 / 377 / 129
Gender / M / 231 / 225 / 178 / 62
F / 182 / 53 / 199 / 67
Age / <16y / 111 / 135 / 0 / 72
≥16y / 305 / 143 / 377 / 57
n per parameter / Melke et al study (20)
(serotonin, melatonin) / 48 / 43 / 34 / 0
this study / Blood serotonin / 368 / 235 / 343 / 129
Platelet N-acetylserotonin / 230 / 230 / 226 / 60
Plasma melatonin / 222 / 227 / 225 / 91
Supplementary Table 2: Frequencies of biochemical impairments (taking the 95th or 5th percentile of controls as pathological thresholds), with 95% confidence intervals and comparison (Fisher’s exact test) test between first-degree relatives of patients and controls and between first-degree relatives of patients and patients with ASD.
High serotonin / High NAS / Low melatoninPatients with ASD / Freq. (CI 95%) / 40% [35-46%] / 47% [41-54%] / 51% [45-57%]
p vs controls / <0.0001 / <0.0001 / <0.0001
Mothers / Freq. (CI 95%) / 17% [11-22%] / 24% [16-32%] / 31% [23-39%]
p vs controls / <0.0001 / <0.0001 / <0.0001
p vs ASD / <0.0001 / <0.0001 / 0.0001
Fathers / Freq. (CI 95%) / 10% [5-14%] / 15% [8-22%] / 21% [14-28%]
p vs controls / 0.0003 / 0.005 / <0.0001
p vs ASD / <0.0001 / <0.0001 / <0.0001
Unaffected siblings / Freq. (CI 95%) / 10% [5-15%] / 18% [9-28%] / 25% [16-34%]
p vs controls / 0.02 / 0.002 / <0.0001
p vs ASD / <0.0001 / <0.0001 / <0.0001