If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Service de Génétique Médicale, University of Lausanne, Lausanne, SwitzerlandLREN–Département des Neurosciences Cliniques, University of Lausanne, Lausanne, Switzerland
Service de Génétique Médicale, University of Lausanne, Lausanne, SwitzerlandLREN–Département des Neurosciences Cliniques, University of Lausanne, Lausanne, Switzerland
Department of Medical Genetics, University of Lausanne, Lausanne, SwitzerlandSIB Swiss Institute of Bioinformatics, University of Lausanne, Lausanne, Switzerland
Department of Genetics and Cytogenetics, Unité fonctionnelle de génétique clinique, Groupe Hospitalier Pitié Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, FranceCentre de Référence “Déficiences intellectuelles de causes rares” and Groupe de Recherche Clinique “Déficience intellectuelle et autisme”, UPMC, Paris, France
SUPEA, and Service of Neuropsychology and Neurorehabilitation, Centre Hospitalier Universitaire Vaudois and University of Lausanne, Lausanne, Switzerland
Department of Genetics and Cytogenetics, Unité fonctionnelle de génétique clinique, Groupe Hospitalier Pitié Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, France
Department of Genetics and Cytogenetics, Unité fonctionnelle de génétique clinique, Groupe Hospitalier Pitié Salpêtrière, Assistance Publique-Hôpitaux de Paris, Paris, FranceDépartement de Génétique, Hôpital Robert Debré, Université Paris VII-Paris Diderot, Paris, France
Laboratoire de Sciences Cognitives et Psycholinguistique, Département d’Etudes Cognitives, Ecole Normale Supérieure, EHESS, CNRS, PSL Research University, Paris, France
Service de Génétique Médicale, University of Lausanne, Lausanne, SwitzerlandSIB Swiss Institute of Bioinformatics, University of Lausanne, Lausanne, Switzerland
LREN–Département des Neurosciences Cliniques, University of Lausanne, Lausanne, SwitzerlandDepartment of Neurology (BD), Max-Planck Institute for Human Cognitive and Brain Science, Leipzig, Germany
Address correspondence to: Sébastien Jacquemont, Département de Pédiatrie, Faculté de Médecine, Université de Montréal, Av. Pierre-Decker 2, Montréal, Québec 1011, Canada.
Deletions and duplications of the 16p11.2 BP4-BP5 locus are prevalent copy number variations (CNVs), highly associated with autism spectrum disorder and schizophrenia. Beyond language and global cognition, neuropsychological assessments of these two CNVs have not yet been reported.
Methods
This study investigates the relationship between the number of genomic copies at the 16p11.2 locus and cognitive domains assessed in 62 deletion carriers, 44 duplication carriers, and 71 intrafamilial control subjects.
Results
IQ is decreased in deletion and duplication carriers, but we demonstrate contrasting cognitive profiles in these reciprocal CNVs. Deletion carriers present with severe impairments of phonology and of inhibition skills beyond what is expected for their IQ level. In contrast, for verbal memory and phonology, the data may suggest that duplication carriers outperform intrafamilial control subjects with the same IQ level. This finding is reminiscent of special isolated skills as well as contrasting language performance observed in autism spectrum disorder. Some domains, such as visuospatial and working memory, are unaffected by the 16p11.2 locus beyond the effect of decreased IQ. Neuroimaging analyses reveal that measures of inhibition covary with neuroanatomic structures previously identified as sensitive to 16p11.2 CNVs.
Conclusions
The simultaneous study of reciprocal CNVs suggests that the 16p11.2 genomic locus modulates specific cognitive skills according to the number of genomic copies. Further research is warranted to replicate these findings and elucidate the molecular mechanisms modulating these cognitive performances.
Deletions and duplications of the 16p11.2 ~600 kb breakpoints 4–5 (BP4-BP5) region are among the most frequent causes of neurodevelopmental and neuropsychiatric disorders (
). Although copy number variations (CNVs) are equally and frequently present in autism spectrum disorder (ASD) cohorts, only duplications are associated with schizophrenia (
). Recent studies in 16p11.2 CNV carriers demonstrated altered global cognition in deletion carriers and duplication carriers showing, respectively, up to 24-point and 15-point decreases in full-scale intelligence quotient (FSIQ) compared with intrafamilial control subjects (
Recurrent reciprocal 16p11.2 rearrangements associated with global developmental delay, behavioural problems, dysmorphism, epilepsy, and abnormal head size.
), and specific assessments showed below-average performance in comprehension, expression, and reading skills with a 71% rate of speech and language disorder as defined in DSM-IV-TR (
). To our knowledge, performance of duplication carriers in specific cognitive domains has not yet been studied.
Most studies of CNVs investigated deletions and duplications separately using a case-control design; however, the simultaneous assessment of reciprocal CNVs of the same genomic region provides a unique opportunity to study how the number of genomic copies (one, two, or three) may modulate clinical phenotypes and endophenotypes. We and other authors previously showed that body mass index inversely correlates to the number of genomic copies at the 16p11.2 locus (
Recurrent reciprocal 16p11.2 rearrangements associated with global developmental delay, behavioural problems, dysmorphism, epilepsy, and abnormal head size.
We hypothesized that specific cognitive skills correlate with the number of genomic copies of the 16p11.2 region beyond the decrease in IQ observed in both types of CNV carriers. We also explored whether this relationship is mediated by changes in brain structure. Our findings suggest that this locus modulates specific cognitive domains resulting in impaired as well as preserved or possibly enhanced skills relative to IQ in carriers of reciprocal CNVs. Neuroimaging analysis suggests that brain regions previously identified as correlating with the number of genomic copies are associated with alterations in executive functions in 16p11.2 CNV carriers.
Methods and Materials
Participants
European Cohort
Participants were taking part in a larger research project on CNVs at the 16p11.2 locus with the aim of phenotyping a European cohort of 16p11.2 rearrangement carriers. Included in our analyses were 62 deletion carriers (37 probands/25 nonproband carriers), 44 duplication carriers (21 probands/23 nonproband carriers), and 71 intrafamilial control subjects (Table 1). Nonproband carriers are defined as relatives who carry the CNV identified through family cascade testing but who were not referred for a neurodevelopmental disorder. Inclusion criteria included presence of a recurrent 16p11.2 deletion or duplication comprising the BP4–BP5 region (29.6–30.2 Mb according to the human genome build GRCh37/hg19). Control subjects were noncarriers in the same families. Individuals <3 years old were excluded from the analyses. Also, 24 participants who were unable to perform the cognitive tasks because of low cognitive functioning were excluded (Supplemental Table S1).
Additional deleterious CNVs were identified in one deletion carrier and five duplication carriers (Supplemental Table S2). These participants were not removed from the analyses, but their potentially confounding effects were taken into account in a subanalysis. Deleterious CNVs were defined as 1) CNVs associated with a known recurrent genomic disorder, 2) CNVs encompassing a published critical genomic region or disrupting a gene that is a known cause of neurodevelopmental disorders, or 3) rare (<1/1000) and large (>500 kb) CNVs.
The study was reviewed and approved by the local ethics committee, and signed consent forms were obtained from participants or legal representatives before investigation. Participants were assessed at the Lausanne University Hospital, Switzerland. Of the proband carriers, >87% were referred to the study by the clinical geneticist who had initially established the genetic diagnosis in the context of a neurodevelopmental disorder. Seven probands (four duplication carriers and three deletion carriers) were identified in the population biobank of Estonia as previously described (
), and one duplication carrier was referred for psychiatric problems. All but four participants (n = 2 Asian descent and n = 2 African descent) were of European descent.
Trained neuropsychologists performed all cognitive assessments. Participants underwent age-appropriate and developmentally appropriate neuropsychological tests assessing overall cognitive functioning, fine motor skills, language, memory, and executive functions. All Z scores were derived from standardized scores based on population norms provided by the testing manuals. When normative data were unavailable, we used raw scores that were adjusted for age and sex when appropriate. We specified the tasks where the latter procedure was applied. We used the following psychometric tests.
Overall Cognitive Functioning
The Wechsler Intelligence Scales or Wechsler Abbreviated Scale of Intelligence (
) was used to obtain FSIQ, nonverbal intelligence quotient (NVIQ), and verbal intelligence quotient (VIQ) when available. The FSIQ, NVIQ, and VIQ were estimated using the Differential Ability Scales–2nd Edition (
) (≥5 years old) assessed four conditions: dominant hand, nondominant hand, bimanual, and assembly. Z scores were used as dependent variables.
Language
Phonological skills were assessed with nonword repetition (≥5 years old), oromotor sequences (≥3 years old), and phonological processing (≥3 years old) from the Developmental Neuropsychological Assessment battery (
). Age-adjusted raw scores were the dependent variables. Participants ≥16 years old performed a sentence repetition task including low-frequency words. The score was the number of correct repetitions adjusted for age. Phonology in the VIP cohort was measured by the nonword repetition subtest (≥4 years old) from the Comprehensive Test of Phonological Processing (
) (word comprehension, ≥3 years old), semantic fluency (animal, ≥3 years old), and phonemic fluency (letter M, ≥5 years old). For word comprehension and word definition tasks, Z score were used as dependent variables, and age-adjusted raw scores were used for verbal fluencies.
Comprehension and verbal skills were assessed by selecting 24 items from the Test for Reception of Grammar 2 (
), syntax comprehension (≥15 years old), and the Wechsler Similarity subtest (≥7 years old). Z score (similarity) and age-adjusted raw scores (syntax) were used as dependent variables. Written language (≥12 years old) was assessed through a reading task (PC Robbery) (
). Reading fluency and reading comprehension Z scores were used as dependent variables; age-adjusted and sex-adjusted total raw scores were used for the spelling task.
Memory
Verbal short-term memory was assessed using the forward digit span task (≥6 years old) (
). The total age-adjusted raw score was used as the dependent variable. Verbal long-term memory was assessed using the California Verbal Learning Task (≥17 years old) (
). Two Z scores were used as dependent variables: number of words correctly recalled across all five learning trials (encoding) and number of words recalled after a 20-minute delay (delayed recall).
Visuospatial short-term memory was assessed using the forward spatial span task (≥6 years old) (
). The total age-adjusted raw score was used as the dependent variable. Visuospatial long-term memory was assessed with the Rey-Osterrieth Complex Figure test (≥5 years old) (
) tasks. The total age-adjusted raw scores were used as dependent variables. Planning skills were assessed with the Tower of London test (≥7 years old) (
) was used to assess motor inhibition. In both tasks, Z scores for response time and number of errors adjusted for age were used as dependent variables.
General Psychiatric and Autism Assessments
Experienced licensed psychologists and psychiatrists performed the Autism Diagnostic Interview–Revised (
Autism Diagnostic Interview–Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders.
) to establish a categorical diagnosis in all participants presenting with ASD symptoms. All adult carriers underwent the Diagnostic Interview for Genetic Studies to screen for major psychiatric disorders (
Variables derived from normative data were converted into Z scores for each participant (mean = 0; SD = 1) (Supplemental Methods and Materials). Raw scores of variables without available normative data were systematically adjusted for age and sex where appropriate. We performed either linear or generalized regression analyses depending on data distribution. Cognitive measures were the dependent variables.
To investigate the effect of CNV on cognition, we used linear models including the following ordinal variables: deletion = 1, control = 2, duplication = 3. For contrasts between groups, we used post hoc t tests. Appropriate linear mixed models or generalized linear mixed models were performed taking the variable “family” as a random factor to account for correlated measures within family. We also included in the statistical design IQ, sex, and their interactions with the number of genomic copies to control for the effects of these variables. These additional covariates were kept in the final models only when the effect was significant. Results of noncarrier participants from deletion or duplication families were pooled, as they did not differ in age, sex, and cognitive performances. Selected models, estimates, and uncorrected p values are reported in Table 2 and Supplemental Tables S4–S6.
Table 2Group Contrasts and Copy Number Variation Effect for Cognitive Measures Adjusting for IQ
Significant p value (corrected threshold, p = .001).
−.59 (.26)
.026
.26
.035
CVLT, California Verbal Learning Test.
Linear mixed models were used to account for the correlations of measures within families. When only one family member was included in the analysis, linear models were performed. Because of space constraints, cognitive variables showing no significant effect of copy number variation or no significant group differences (c, f-l; p values corrected) are presented in Supplemental Table S5.
a Positive estimate: Deletion carrier score < duplication carrier score.
b Negative estimate: Deletion carrier or duplication carrier score < control subject score.
c Phonological processing.
d Raw score.
e Significant p value (corrected threshold, p = .001).
f Word comprehension, semantic and phonemic fluencies.
g Syntax comprehension and verbal reasoning.
h Reading comprehension and spelling.
i Forward spatial span and Rey-Osterrieth Complex Figure (immediate and delayed recall).
j Backward digit span, backward spatial span, and Tower of London (total correct and total move scores).
k Stroop response time, go/no-go response time, and number of successes.
Multiple testing correction is detailed in Supplemental Methods and Materials. It takes into account all 34 cognitive tests, their level of intercorrelation, and the two post hoc contrasts (deletion carriers vs. control subjects and duplication carriers vs. control subjects). Uncorrected p values ≤ .001 are considered significant, and trends are defined as .001 < p values ≤ .05. Statistical analyses were conducted using R 3.0.2 (The R Project for Statistical Computing; http://www.R-project.org/).
Brain Structure and Behavior Correlation Analyses
We tested whether the effects of CNVs on cognitive measures are mediated by changes observed in brain anatomy. Therefore, only cognitive measures (Z scores and raw scores adjusted for age and IQ) with significant differences between CNV carriers and intrafamilial control subjects were subsequently used for regression analysis with brain anatomy. The statistical design also included age, sex, and total intracranial volume as regressors. Voxel-based statistical analysis of the gray matter (GM) regional changes was assessed by creating voxelwise statistical parametric maps for the whole extent of the search volume using the general linear model and random field theory (
), we tested whether the behavioral deficits were associated with an increase of local GM volume. Consequently, one-tailed t tests were used to identify the regions whose volume showed negative correlation with the cognitive score.
We subsequently estimated the degree of overlap between brain areas correlating with behavioral scores and the previously identified regions sensitive to CNV (
). Clusters sharing both effects were obtained with a subtraction of the two statistical maps. We further examined how the number of genomic copies interacted with the brain-behavior correlation in these regions, using a multiple linear regression analysis of the summed voxel values for each group. For all whole-brain analyses, we applied a voxel-level threshold of p < .05 after familywise error correction for multiple comparisons. Trends were assessed by using an auxiliary uncorrected voxel threshold of p < .001 (
Deletion carriers and duplication carriers showed cognitive impairment (FSIQ = 72 for deletion carriers, FSIQ = 75 for duplication carriers) compared with intrafamilial control subjects (FSIQ = 98), consistent with a larger study we recently published (Supplemental Table S4) (
). All verbal and nonverbal subtests from the Wechsler Abbreviated Scale of Intelligence equally contributed to this IQ deficit (Supplemental Figure S1). To estimate the effect of the deletion and the duplication on specific cognitive domains beyond their impact on general cognition, we performed all subsequent analyses adjusting for NVIQ in verbal tasks and FSIQ otherwise (Table 2 and Supplemental Table S5; see Supplemental Table S6 for results not adjusted for IQ).
16p.11.2 Locus Modulates Phonology, Written Language, and Vocabulary
The number of genomic copies positively correlated with phonology measures such as nonword repetition (p = 9.9e-05), oromotor sequences (p = 8e-05), and sentence repetition (p = .0001) as well as a closely related task, reading fluency (p = .0009). This effect was mainly driven by the deletion carriers, who performed worse than control subjects and duplication carriers; these two latter groups did not significantly differ (Figure 1A–C). Word definition showed a similar trend (p = .0047) (Table 2). Except for trends, none of the measures assessing spelling, verbal fluencies, verbal comprehension, and verbal reasoning were significantly affected by the 16p11.2 locus (Supplemental Table S5).
Figure 1Copy number variation and group comparisons on language, memory, and executive measures. Box plots represent language measures of phonology (A–C), memory (D–F), and executive functions (G–I) in deletion carriers, duplication carriers, and intrafamilial control subjects. Higher scores translate into better performance except for panel (I), where better performance is represented by a shorter response time. The bold line shows the median, and the bottom and top of the box show the 25th (quartile 1 [Q1]) and the 75th (quartile 3 [Q3]) percentile, respectively. The upper whisker ends at highest observed data value within the span from Q3 to Q3 + 1.5 times the interquartile range (Q3–Q1), and lower whisker ends at lowest observed data value within the span for Q1 to Q1 − (1.5 * interquartile range). Circles are outliers. Copy number variation effect (one, two, or three copies) and group contrasts are estimated using a linear mixed model to account for correlated measures within families (A, B, D, G, H, I) and a linear model when only one family member is included in the analysis (E, F). A nonlinear model was required for panels (C, G, H). We present Z scores when normative data are available in the testing manual. Otherwise, we present raw scores adjusted for age and sex when appropriate. All scores are adjusted for IQ except for panel (I). Significant post hoc group comparisons (p-corrected threshold = .001) are represented by solid lines with exact p values above, and trends are represented by dashed lines with exact p values above. CTRL, intrafamilial control subjects; DEL, deletion carriers; DUP, duplication carriers.
Results from the VIP data set confirmed that the number of genomic copies correlated with performance in the nonword repetition task adjusted for NVIQ (p = 9e-6). The effect was driven by both CNVs, with deletion carriers performing worse than duplication carriers and intrafamilial control subjects (p = 2.7e-6 and p = .002, respectively) and duplication carriers outperforming control subjects (p = .016) (Supplemental Figure S2).
Verbal Short-Term and Long-Term Memory Processes Are Modulated by the Number of Genomic Copies
All verbal memory measures correlated positively with the number of genomic copies. Duplication carriers outperformed deletion carriers and control subjects in measures of verbal long-term memory after adjusting for NVIQ (Figure 1D–F). Scatter plots showing correlations between verbal long-term memory measures and NVIQ confirmed that duplication carriers scored higher than control subjects with similar NVIQ (Supplemental Figure S3). Regarding verbal short-term memory, deletion carriers scored significantly worse than duplication carriers (p = .0002) and control subjects (trend, p = .003). The 16p11.2 locus did not significantly affect either short-term or long-term visuospatial memory.
16p11.2 Locus Modulates Inhibition Skills But Not Working Memory and Planning
Motor and verbal inhibition measures positively correlated to the number of genomic copies (significant for verbal inhibition, trend for motor inhibition) (Figure 1G–I). Deletion carriers who performed or tended to perform worse than control subjects and duplication carriers mainly drove this effect (Table 2). We did not observe any group differences in working memory and planning skills (Supplemental Table S5).
Fine Motor Skills
Deletion carriers performed worse than control subjects in Purdue dominant and nondominant hand conditions, but there were no significant differences in the bimanual and assembly conditions.
Overall Neuropsychological Profile
To illustrate the neuropsychological profiles of deletion and duplication carriers, we summarized a sample of the cognitive tasks adjusted and unadjusted for IQ (Figure 2A, B). Carriers’ data were converted into Z scores relative to the intrafamilial control subjects to highlight preserved skills (performance similar to controls after adjusting for IQ), specific deficits, and enhanced performances (lower and higher performances than expected for IQ level). When adjusted for IQ levels (Figure 2A), the cognitive profile of deletion carriers showed specific deficits in language and inhibition domains, whereas there was no specific impairment in the profile of duplication carriers. The latter group showed enhanced performance in verbal long-term memory. When results were unadjusted for IQ (Figure 2B), both CNV carriers showed similar decreased performance in several tasks consistent with their overall IQ level (e.g., verbal comprehension, working memory, planning, visuospatial skills).
Figure 2Neuropsychological profile in 16p11.2 deletion carriers and duplication carriers. IQ-adjusted (A) and IQ-unadjusted (B) neuropsychological profile of deletion carriers and duplication carriers. The y axis represents mean cognitive residual scores for deletion carriers (red circles) and duplication carriers (blue squares) converted into Z scores relative to the intrafamilial control subjects (black dashed line). Error bars represent SEM. When appropriate, scores are adjusted for age and sex. The x axis lists one task per subdomain with the most complete data set: NW repetition (nonword repetition), Word def (word definition), Word compr (word comprehension), V short-term (verbal short-term memory), S short-term (spatial short-term memory), V long-term (verbal long-term memory), S long-term (spatial long-term memory), VW memory (verbal working memory), Planning, V inhibition (verbal inhibition number of success), M inhibition (motor inhibition number of success), and FSIQ (full-scale intelligence quotient). CTRL, intrafamilial control subjects; DEL, deletion carriers; DUP, duplication carriers.
We considered possible confounders including a diagnosis of ASD (three deletion carriers and five duplication carriers) or the presence of additional deleterious CNVs (one deletion carrier and five duplication carriers). The analyses performed after excluding participants with additional CNVs (Supplemental Table S7) and analyses excluding participants with ASD led to the same results. We also considered possible bias secondary to ascertainment for neurodevelopmental disorders, but proband carriers were not outperformed by nonproband carriers in any of the cognitive measures (Supplemental Table S8). Finally, the effect of inheritance was tested. In the deletion group, performances of de novo carriers did not differ from performances of the inherited carriers (Supplemental Table S9).
Correlation With Brain Anatomy
Whole-brain voxel-based morphometry analysis showed a positive correlation between GM volume and the verbal inhibition error rate in bilateral insula and transverse temporal gyri (Figure 3 and Supplemental Table S10). Subsequent regression analyses (left cluster, r2 = .14, p = .007; right cluster, r2 = .24, p = .0003) revealed that the effect was mainly driven by deletion carriers, who showed the greatest variance (Figure 3B, C). We report a trend (uncorrected p value < .001 for all the clusters) for increased GM volume in the left inferior frontal gyrus, bilateral superior temporal gyri, and bilateral caudate associated with deficits in two measures related to phonology: nonword repetition and reading fluency. Measures of memory did not covary with any brain structure.
Figure 3Brain structure-behavior correlation analysis between verbal inhibition score and gray matter (GM) map. (A) Spatial overlap between negative linear correlation with Stroop score (red) and the neuroanatomic structures previously identified as correlating to the number of genomic copies of the 16p11.2 locus (
). Overlapping clusters are represented in orange. Maps are thresholded at the p < .05 familywise error corrected level. Panels (B, C) are scatter plots showing the linear correlation between GM volume and the Stroop performance at the clusters located in the left insula (B) and in the right temporal gyri (C) in panel (A). Both panels include the regression line, correlation coefficient, and p value for each cohort. CTRL, intrafamilial control subjects; DEL, deletion carriers; DUP, duplication carriers.
By assessing carriers of deletion and duplication at the 16p11.2 locus and intrafamilial control subjects, our study characterizes the effect of this genomic region on several cognitive domains. After adjusting for global cognition, which is decreased in both CNVs, specific cognitive functions including verbal memory, executive functions, and phonological skills show a positive correlation with the number of genomic copies. The data suggest that duplication carriers outperform intrafamilial control subjects with the same IQ for measures of verbal memory and phonology.
) comparing recurrent reciprocal CNVs failed to identify any correlations between cognitive functions and copy number state at the 16p11.2 BP4–BP5 and 15q11.2 BP1–BP2 loci. Either the small sample size (n = 7 for 16p11.2 deletion carriers and duplication carriers, respectively) or the small effect size (for 15q11.2 BP1–BP2) could explain the lack of significant correlations between cognitive traits and copy number state.
Although linear models show that the number of genomic copies correlates with performance in several cognitive domains, our study was underpowered to demonstrate formally that deletion carriers perform worse than control subjects and control subjects perform worse than duplication carriers for a specific function. However, the results are suggestive of such a phenomenon; our analysis of phonologic skills in a larger independent data set (VIP) supports the hypothesis that, similar to other complex traits such as body mass index or brain anatomy, cognitive performances may covary with molecular mechanisms. Recent findings on 16p11.2 CNVs mouse models show enhanced memory skills on a recognition task in duplicated mice compared to wild-type animals (
). Memory processes have also been linked to mechanisms regulating long-term synaptic potentiation and depression. These synaptic mechanisms require bursts of local protein synthesis during training and stimulation (
). Both mammalian target of rapamycin signaling and MAPK3 signaling regulate local synaptic protein synthesis, which modulates memory performances in murine models (
). The expression levels of MAPK3, which maps within the BP4–BP5 interval, and of mammalian target of rapamycin pathway members are significantly altered in 16p11.2 CNV carriers (
). These are good candidate genes underlying or mediating the correlation between genomic copy number and memory performances.
Deletion and duplication are strongly and equally associated with ASD, but our study shows that their cognitive profiles seem to be distinct, highlighting the heterogeneity of ASD. Our results are consistent with the results of D’Angelo et al. (
), who observed that the duplication is associated with a form of low-functioning ASD, whereas cognition in deletion carriers with ASD is mostly within the normal range. These results are in line with studies highlighting the potential role of the CNVs in the specific patterns of autistic symptoms (
). This study demonstrates that the same genomic region predisposing to ASD may or may not have a deleterious effect on structural language depending on the nature of the mutation. This dissociation between phenotypes observed in reciprocal CNVs is also corroborated by a recent study demonstrating that deletions, but not duplications, encompassing ASD genes are primarily associated with impairments in language domains (
) and may represent the first example of an ASD-related genetic predisposition leading to specific cognitive strengths compared with control subjects with the same IQ. The absence of any specific impairment beyond the IQ shortfall in duplication carriers at risk for schizophrenia echoes the nonspecific cognitive deficit pattern observed in first-episode idiopathic schizophrenia (
Our results also suggest that neuroanatomic structures previously defined on the basis of their correlation to number of genomic copies at the 16p11.2 locus (
) are associated with alterations in measures of language and verbal inhibition. These findings dovetail with previously reported structural findings positing the insula as a key player in verbal inhibitory processes (
). However, larger samples are required to replicate these results and elucidate any specific association within groups. The causal relationship between genomic copies, brain structure, cognition, and behavior also remains unknown.
Increased volume of the caudate has been observed in individuals who carry FOXP2 mutations, which is one of the few genetic forms of specific language impairment studied to date (
) reporting phonological deficits in the context of general language impairment. Other studies reported the presence of childhood apraxia of speech, although the exact characteristics of this deficit are unclear (
The present study aimed at characterizing the effect of the 16p11.2 locus on specific cognitive domains by adjusting for the effect of IQ. We are confident that our findings apply to a broad range of IQ with the exception of very low-functioning participants who were unable to perform these tasks (11% of deletion carriers and 27% of duplication carriers). It is unlikely that the exclusion of individuals with low IQ has biased our results because a much larger study showed that most deletion carriers and duplication carriers have an IQ >55. The duplication and the deletion have been associated with a decrease of approximately 16 points and 22 points of IQ, respectively (
). This finding suggests that the subgroup of very low-functioning carriers may be related to the presence of additional deleterious genetic variants, which are two to three times more frequent in 16p11.2 duplication carriers compared with deletion carriers (
). It is unknown whether low-functioning carriers represent a distinct subgroup with different profiles. Most of the probands were ascertained for neurodevelopmental disorders, which may have also biased our results. However, stratifying groups in proband and nonproband carriers does not affect any of the results. The low frequency of schizophrenia is discordant with the association reported in prior studies (
). This low frequency is likely to be due in part to the young age of our probands and the fact that most of the adults were parents, a fact that selects against a diagnosis of schizophrenia (
). Follow-up of our probands is required to estimate accurately the risk of schizophrenia in this population.
In conclusion, this study suggests that cognitive skills may be modulated by the number of genomic copies at the 16p11.2 locus in humans. Such effects are clouded by the global decrease in cognitive functioning that affects both CNVs. The strength of this study lies in the administration of an extensive cognitive test battery to both CNV carrier groups and the intrafamilial control subjects; this allowed us to assess cognitive function precisely relative to each participant’s global cognitive level. Further research is warranted to elucidate the contribution of specific genes within the 16p11.2 locus by studying the relationship between expression patterns of these genes and cognitive tasks, brain anatomy, and brain function. These approaches may ultimately elucidate the molecular mechanisms that affect specifically phonological, verbal memory, and inhibition skills based on the number of genomic copies. A deeper knowledge of the cognitive strengths and weaknesses of these patients is critical for developing cognitive support strategies. Careful language assessment is recommended in deletion carriers, who might benefit from more emphasis on the use of visuospatial processes when learning. In contrast, verbal methods may improve learning strategies in duplication carriers.
Acknowledgments and Disclosures
This work was supported by the Leenaards Foundation Prize (SJ, ARey, NH), Swiss National Science Foundation (SNSF) Sinergia Grant No. CRIS FN CRSII33-133044 (ARey, SJ), National Center of Competence in Research Synapsy Project Grant Nos. 320030_135679 and Special Program University Medicin 33CM30_140332/1 (BD), Simons Foundation Autism Research Initiative (SFARI) Grant No. SFARI274424 (ARey), Foundation Parkinson Switzerland (BD), Foundation Synapsis (BD), Human Brain Project, a European Union initiative (BD), SNSF Grant No. PP00P3_144902/2 (SJ), a Canada Research Chair (SJ), Jeanne and Jean Louis Lévesque Foundation (SJ), Center of Excellence in Genomics (Estonian Genome Center, the University of Tartu [EGCUT] [AMe, ARei, AK, KM]), University of Tartu Grant No. SP1GVARENG (EGCUT [AMe, ARei, AK, KM]), Estonian Research Council Grant No. IUT20-60 (EGCUT [AMe, ARei, AK, KM]), Swiss Scientific Exchange New Member States Program (KM), and a fellowship from the doctoral school of the Faculty of Biology and Medicine, University of Lausanne (MZ). Laboratory for Research in Neuroimaging is supported by the Roger de Spoelberg and Partridge Foundations. FR received funding from Agence Nationale (contracts ANR-10-LABX-0087 IEC and ANR-10-IDEX-0001-02 PSL). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The Simons Variation in Individuals Project (VIP) work is supported by SFARI. We thank all the families at the participating Simons VIP sites as well as the Simons VIP Consortium. We obtained access to phenotypic data on SFARI Base. Approved researchers can obtain the Simons VIP and SSC population data sets described in this study by applying at https://base.sfari.org.
Participants were scanned at the Centre d’Imagerie BioMédicale, which is a research initiative of the following partners: University of Lausanne, Swiss Federal Institute of Technology Lausanne, University of Geneva, Centre Hospitalier Universitaire Vaudois, Hôpitaux Universitaires de Genève, and the Leenaards and Jeantet Foundations.
We thank all families for their contribution to this work. We thank Anne Ruef for her helpful advice on the magnetic resonance imaging data analyses. We thank Zoltan Kutalik for his help with statistical analyses.
The 16p11.2 European consortium collaborators who contributed families to this study include M.-C. Addor, Service of Medical Genetics, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; J. Andrieux, Institut de Génétique Médicale, Hôpital Jeanne de Flandre, CHRU de Lille, Lille, France; B. Arveiler, Université Bordeaux, Maladies Rares: Génétique et Métabolisme, Service de Génétique Médicale, CHU-Bordeaux, Bordeaux, France; G. Baujat, Centre de Référence Département de Génétique, Hôpital Necker-Enfants malades, Paris, France; M. Belfiore, Laboratoire de Cytogénétique, Service de Génétique Médicale, CHUV, Lausanne, Switzerland; F. Béna, Service of Genetic Medicine, University Hospital of Geneva, Geneva, Switzerland; F. Bilan, Service de Génétique, CHU-Poitiers, Poitiers, France; S. Bouquillon, Institut de Génétique Médicale, Hôpital Jeanne de Flandre, Lille, France; O. Boute, Hôpital Jeanne de Flandre, CHRU de Lille, Lille, France; J.-L. Bresson, Centre de Génétique Humaine, CHU-Besançon, Besançon, France; A. Brusco, University of Torino, Department of Medical Sciences, Turin, Italy; J. Bussat, Service of Medical Genetics, CHUV, Lausanne, Switzerland; D. Campion, Service de Psychiatrie, Centre Hospitalier de Rouvray, Sotteville lès Rouen, France; M. Delrue, Université Bordeaux, Maladies Rares: Génétique et Métabolisme, Service de Génétique Médicale, CHU-Bordeaux, Bordeaux, France; G. Dervaux, Unité de Médecine et Chirurgie de l’Obésité, Centre Hospitalier de Béthune, Béthune, France; R. Etienne, Service de Génétique Médicale, CHUV, Lausanne, Switzerland; C. Fagerberg, Department of Clinical Genetics, Odense Universitetshospital, Denmark; F. Fellmann, Service of Medical Genetics, CHUV, Lausanne, Switzerland; A. Ferrarini, Division of Pediatrics, San Giovanni Hospital, Bellinzona, Switzerland; S. Fert-Ferrer, UF de Génétique Chromosomique, Laboratoire de Biologie Médicale, Chambéry, France; F. Forzano, Genetica Medica, Ospedali Galliera, Genova, Italy; N. Fournier, Institute of Social and Preventive Medicine, CHUV, Lausanne, Switzerland; D. Giachino, Genetica Medica, Azienda Ospedaliera Universitaria San Luigi Gonzaga Orbassano, Torino, Italy; B. Gilbert-Dussardier, Service de Génétique Clinique, La Milétrie, CHU Poitiers, Poitiers, France; O. Guillin, Service Universitaire de Psychiatrie, CHU de Rouen-Centre Hospitalier du Rouvray, UFR de Médecine et de Pharmacie de Rouen, Rouen, France; D. Héron, Département de Génétique, Hôpital Pitié-Salpêtrière, Université Pierre et Marie Curie, Paris, France; M. Holder, Hôpital Jeanne de Flandre, CHRU de Lille, Lille, France; K. Iglesias, Institute of Social and Preventive Medicine, CHUV, Lausanne, Switzerland; A. Jacquette, Département de Génétique, Groupe Hospitalier Pitié-Salpétrière, Paris, France; H. Journel, Génétique Médicale, Centre Hospitalier Bretagne Atlantique, Vannes, France; L. Lazaro, CH Côte-Basque, Service de Pédiatrie, Bayonne, France; M.-P. Lemaître, Service de Neuropédiatrie, Centre Hospitalier Régional Universitaire, Lille, France; J. Lespinasse, Service Génétique Médicale, C.H. Chambéry, Chambéry, France; M. Malcarne, Laboratory of Human Genetics, Ospedali Galliera, Genova, Italy; G. Mandrile, Genetica Medica, Azienda Ospedaliera Universitaria San Luigi Gonzaga Orbassano, Torino, Italy; D. Martinet, Laboratoire de Cytogénétique, Service de Génétique Médicale, CHUV, Lausanne, Switzerland; N. Migone, Genetica Medica, Struttura Complessa a Direzione Universitaria, Ospedaliero-Universitaria san Giovanni Battista, Torino, Italy; L. Olivier-Faivre, Centre de Référence Anomalies du Développement et Syndromes Malformatifs de l’Interrégion Est, Hôpital d’Enfants, CHU de Dijon et Université de Bourgogne, Dijon, France; F. Petit, Hôpital Jeanne de Flandre, CHRU de Lille, Lille, France; G. Plessis, Service de Génétique Clinique, CHU Caen, France; F. Prieur, Service de Génétique Médicale, CHU St. Etienne, St. Etienne, France; G. Ramelli, Department of Pediatrics, San Giovanni Hospital, Bellinzona, Switzerland; C. Rooryck Thambo, Service de Génétique Médicale, CHU de Bordeaux and Université Bordeaux, Maladies Rares: Génétique et Métabolisme, Bordeaux, France; D. Sanlaville, Service de Cytogénétique Constitutionnelle, Hospices Civils de Lyon, CHU de Lyon and Centre de Recherche en Neuroscience de Lyon, Lyon, France; V. Siffredi, Service of Medical Genetics, CHUV, Lausanne, Switzerland; J. Thonney, Department of Psychiatry, CERY Hospital Centre Hospitalier Universitaire Vaudois and University of Lausanne, Lausanne, Switzerland; M.M. Van Haelst, Department of Medical Genetics, University Medical Center, Utrecht, Netherlands; and L. Van Maldergem, Centre de Génétique Humaine, CHU-Besançon, France.
We thank the coordinators and staff at the Simons VIP sites.
Contributors to the Simons VIP Consortium include Benjamin Aaronson, Sean Ackerman, Hanalore Alupay, Katy Ankenman, Ayesha Anwar, Constance Atwell, Arthur L. Beaudet, Marta Benedetti, Jessica Berg, Jeffrey Berman, Leandra N. Berry, Audrey L. Bibb, Lisa Blaskey, Alexandra Bowe, Jonathan Brennan, Christie M. Brewton, Randy Buckner, Polina Bukshpun, Jordan Burko, Phil Cali, Bettina Cerban, Yishin Chang, Qixuan Chen, Maxwell Cheong, Vivian Chow, Zili Chu, Darina Chudnovskaya, Wendy K. Chung, Lauren Cornew, Corby Dale, Deborah D’Angelo, John Dell, Allison G. Dempsey, Trent Deschamps, Rachel Earl, James Edgar, Jenna Elgin, William Faucett, Jennifer Endre Olson, Yolanda L. Evans, Anne Findlay, Gerald D. Fischbach, Charlie Fisk, Brieana Fregeau, Bill Gaetz, Leah Gaetz, Silvia Garza, Jennifer Gerdts, Orit Glenn, Sarah E. Gobuty, Rachel Golembski, Marion Greenup, Kory Heiken, Katherine Hines, Leighton Hinkley, Frank I. Jackson, Julian Jenkins III, Rita J. Jeremy, Kelly Johnson, Stephen M. Kanne, Sudha Kessler, Sarah Y. Khan, Matthew Ku, Emily Kuschner, Anna L. Laakman, Peter Lam, Morgan W. Lasala, David Ledbetter, Hana Lee, Kevin LeGuerre, Susan Levy, Alyss Lian Cavanagh, Ashlie V. Llorens, Katherine Loftus Campe, Tracy L. Luks, Elysa J. Marco, Alastair J. Martin, Christa L. Martin, Stephen Martin, Gabriela Marzano, Christina Masson, Kathleen E. McGovern, Rebecca McNally Keehn, David T. Miller, Fiona K. Miller, Timothy J. Moss, Rebecca Murray, Srikantan S. Nagarajan, Kerri P. Nowell, Julia Owen, Andrea M. Paal, Alan Packer, Patricia Z. Page, Brianna M. Paul, Alana Peters, Danica Peterson, Annapurna Poduri, Nicholas J. Pojman, Ken Porche, Monica B. Proud, Saba Qasmieh, Melissa B. Ramocki, Beau Reilly, Timothy P.L. Roberts, Dennis Shaw, Elliott Sherr, Tuhin Sinha, Bethanny Smith-Packard, Anne Snow Gallagher, John Spiro, Vivek Swarnakar, Tony Thieu, Jennifer Tjernagel, Christina Triantafallou, Roger Vaughan, Nicole Visyak, Mari Wakahiro, Arianne Wallace, Tracey Ward, Julia Wenegrat, and Anne Wolken.
The authors report no biomedical financial interests or potential conflicts of interest.
Recurrent reciprocal 16p11.2 rearrangements associated with global developmental delay, behavioural problems, dysmorphism, epilepsy, and abnormal head size.
Autism Diagnostic Interview–Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders.
00917-8&id=gr1.jpg){kind=link}
00917-8&id=gr2.jpg){kind=link}
00917-8&id=gr3.jpg){kind=link}