Prospective investigation of FOXP1 syndrome

Paige M. Siper; Silvia De Rubeis; Maria del Pilar Trelles; Allison Durkin; Daniele Di Marino; François Muratet; Yitzchak Frank; Reymundo Lozano; Evan E. Eichler; Morgan Kelly; Jennifer Beighley; Jennifer Gerdts; Arianne S. Wallace; Heather C. Mefford; Raphael A. Bernier; Alexander Kolevzon; Joseph D. Buxbaum

doi:10.1186/s13229-017-0172-6

Research

Open Access

Prospective investigation of FOXP1 syndrome

Paige M. Siper†^{1, 8},
Silvia De Rubeis†^{1, 8},
Maria del Pilar Trelles^{1, 8},
Allison Durkin^{1, 8},
Daniele Di Marino²,
François Muratet^{1, 8},
Yitzchak Frank^{1, 8},
Reymundo Lozano^{3, 8},
Evan E. Eichler⁴,
Morgan Kelly⁵,
Jennifer Beighley⁵,
Jennifer Gerdts⁵,
Arianne S. Wallace⁵,
Heather C. Mefford⁵,
Raphael A. Bernier⁵,
Alexander Kolevzon†^{6, 8}Email author and
Joseph D. Buxbaum†^{7, 8}Email author

†Contributed equally

Molecular AutismBrain, Cognition and Behavior20178:57

https://doi.org/10.1186/s13229-017-0172-6

Received: 4 July 2017

Accepted: 27 September 2017

Published: 24 October 2017

Abstract

Background

Haploinsufficiency of the forkhead-box protein P1 (FOXP1) gene leads to a neurodevelopmental disorder termed FOXP1 syndrome. Previous studies in individuals carrying FOXP1 mutations and deletions have described the presence of autism spectrum disorder (ASD) traits, intellectual disability, language impairment, and psychiatric features. The goal of the present study was to comprehensively characterize the genetic and clinical spectrum of FOXP1 syndrome. This is the first study to prospectively examine the genotype-phenotype relationship in multiple individuals with FOXP1 syndrome, using a battery of standardized clinical assessments.

Methods

Genetic and clinical data was obtained and analyzed from nine children and adolescents between the ages of 5–17 with mutations in FOXP1. Phenotypic characterization included gold standard ASD testing and norm-referenced measures of cognition, adaptive behavior, language, motor, and visual-motor integration skills. In addition, psychiatric, medical, neurological, and dysmorphology examinations were completed by a multidisciplinary team of clinicians. A comprehensive review of reported cases was also performed. All missense and in-frame mutations were mapped onto the three-dimensional structure of DNA-bound FOXP1.

Results

We have identified nine de novo mutations, including three frameshift, one nonsense, one mutation in an essential splice site resulting in frameshift and insertion of a premature stop codon, three missense, and one in-frame deletion. Reviewing prior literature, we found seven instances of recurrent mutations and another 34 private mutations. The majority of pathogenic missense and in-frame mutations, including all four missense mutations in our cohort, lie in the DNA-binding domain. Through structural analyses, we show that the mutations perturb amino acids necessary for binding to the DNA or interfere with the domain swapping that mediates FOXP1 dimerization. Individuals with FOXP1 syndrome presented with delays in early motor and language milestones, language impairment (expressive language > receptive language), ASD symptoms, visual-motor integration deficits, and complex psychiatric presentations characterized by anxiety, obsessive-compulsive traits, attention deficits, and externalizing symptoms. Medical features included non-specific structural brain abnormalities and dysmorphic features, endocrine and gastrointestinal problems, sleep disturbances, and sinopulmonary infections.

Conclusions

This study identifies novel FOXP1 mutations associated with FOXP1 syndrome, identifies recurrent mutations, and demonstrates significant clustering of missense mutations in the DNA-binding domain. Clinical findings confirm the role FOXP1 plays in development across multiple domains of functioning. The genetic findings can be incorporated into clinical genetics practice to improve accurate genetic diagnosis of FOXP1 syndrome and the clinical findings can inform monitoring and treatment of individuals with FOXP1 syndrome.

Background

Haploinsufficiency of the forkhead-box protein P1 (FOXP1) gene has recently been shown to cause a neurodevelopmental disorder with a phenotype characterized by global developmental delay (DD), intellectual disability (ID), speech deficits, mild dysmorphic features, and autism spectrum disorder (ASD) traits [1–5]. Here, we refer to this disorder as FOXP1 syndrome.

Since the first report of a deletion spanning FOXP1 and three other genes in a child with DD, speech delay, hypertonia, dysmorphic features, contractures and blepharophimosis [6], nearly 20 cases have been reported. The identification of deletions limited to FOXP1 [1–3, 7, 8] and of several individuals with loss-of-function and missense variants in FOXP1 [1, 3–5, 9, 10] has delineated FOXP1 haploinsufficiency as sufficient to produce core features including delayed motor and language milestones, global speech impairment, ID, dysmorphic features, and ASD. Additional FOXP1 mutations have also emerged from large-scale targeted sequencing or exome and genome sequencing analyses of cohorts with ID [11–13] or ASD [14, 15]; however, these studies have provided minimal phenotypic information. FOXP1 can be among the genes deleted in cases with 3p14 deletion syndrome, a contiguous gene syndrome that also presents with hearing loss, congenital heart defects, and urogenital abnormalities [16–18]. Interestingly, common variation at the FOXP1 locus has shown association in a cross-disorder meta-analysis of ASD and genome-wide association studies in schizophrenia [19]. Eight additional pathogenic mutations have emerged after the discovery of a de novo mutation in a whole exome sequencing (WES) study on congenital anomalies of the kidney and urinary tract [20]. Notably, all individuals have neurodevelopmental phenotypes compatible with FOXP1 syndrome and the majority (6/8) display upper or lower urinary tract defects [20].

FOXP1 is a transcription factor of the FOX gene family, named for the forkhead-box DNA-binding domain present in the gene family [21]. The FOXP subfamily is comprised of four genes: FOXP1, FOXP2, FOXP3, and FOXP4. The closest homolog to FOXP1, and the best-known member of the FOXP family, is FOXP2. In the brains of zebra finch songbirds and humans, Foxp1 and Foxp2 are co-expressed in the GABAergic medium spiny neurons of the striatum [22, 23], a brain region critical for human language, mouse ultrasonic vocalization (USV), and zebra finch vocal imitation. Both genes are important for language production and comprehension. In addition to the speech impairments observed in individuals with FOXP1 syndrome, maternal uniparental disomy of chromosome 7 (reducing FOXP2 expression) [24], FOXP2 deletions [25], and FOXP2 mutations [26, 27], all result in childhood apraxia of speech and other speech and language defects. In addition, both FOXP1 and FOXP2 are critical during cortical neurogenesis and specification [28–30]. Constitutive FOXP1 knockout is embryonically lethal due to a cardiac defect, but brain-specific conditional FOXP1 null mice display striatal morphological defects, reduced USV (also reported in FOXP2 mutant mice [31]) and social and cognitive deficits [23, 32].

Shared and pervasive clinical features in individuals with FOXP1 deletions and mutations include mild-to-moderate ID, language impairment, and motor delays [1–5, 10, 13]. All reported individuals with FOXP1 syndrome displayed speech and language impairment. While delayed, the developmental trajectory of cognitive, language, and motor skills remains unknown. The majority of individuals described in the literature developed a minimum of phrase speech, all individuals were reported to have difficulty with articulation, and language was often limited to phrases or simple sentences [1, 2, 4, 5]. Several investigators reported that expressive language is more affected than receptive language [1–3]; however, these findings were not based on norm-referenced standardized testing.

Medical features of individuals with FOXP1 mutations reported in the literature vary widely and include brain and cardiac malformations, hypotonia, strabismus, and obesity [2, 4, 5, 7, 16, 33, 34]. Dysmorphic features associated with FOXP1 syndrome appear to be mild and inconsistent. Among the most commonly reported dysmorphic features are a prominent forehead, downslanting or short palpebral fissures, and a short nose with a broad tip [3]. Other features may include widely spaced eyes, frontal hair upsweep, ptosis, and hypertelorism [7]. Behavioral anomalies, including ASD or autistic traits, aggression, anxiety, and obsessive-compulsive symptoms, were present in a majority of reported cases [1, 4, 5, 8, 10, 13].

To date, no study has prospectively evaluated more than three individuals with FOXP1 syndrome using a battery of standardized measures. The goal of the present study was to comprehensively characterize FOXP1 syndrome by utilizing a multidisciplinary team of biologists and clinicians and objective assessments to prospectively evaluate a cohort of children and adolescents with mutations in the FOXP1 gene. We also characterize one individual with a duplication of 8.4 Mb spanning FOXP1 and 47 additional genes, which has not previously been described and remains of unknown clinical significance (Additional file 1). Evaluating genetic results in conjunction with a robust battery of clinical assessments will better elucidate the genotype-phenotype relationship in this recently described syndrome, while functional dissection of mutations will provide insights into the pathobiological mechanisms underlying the FOXP1 syndrome.

Methods

Participants

Phenotypic characterization was completed in nine children and adolescents (seven female) with mutations in FOXP1 and one female with a large duplication spanning FOXP1 (Additional file 1). Individuals were between the ages of 5–17 (mean ± SD = 11.1 ± 3.6). Evaluations were completed at the Seaver Autism Center for Research and Treatment at the Icahn School of Medicine at Mount Sinai (n = 6, subjects S1-S6) and at the Center on Human Development and Disability at the University of Washington (n = 4, subjects W1-W4). Comprehensive medical, neurological, dysmorphology, and neuropsychological evaluations were completed by teams of child and adolescent psychiatrists, clinical psychologists, neurologists, and clinical geneticists. A battery of standardized assessments was used to examine ASD symptomatology, intellectual functioning, adaptive behavior, receptive and expressive language, fine and gross motor skills, visual-motor integration, and psychiatric features. Individual S2 was previously reported by Lozano et al. (2015) [4]. Individual S4 was previously reported by Sollis et al. (2016) (subject 1) [5]. Individual W1 was previously reported by O’Roak et al. (2011) (subject 12817.p1) [10], but without a clinical description. This study was approved by the Institutional Review Boards of both participating sites. All caregivers provided informed written consent and assent was obtained when appropriate.

Genetic testing

All mutations were validated by Clinical Laboratory Improvement Amendments (CLIA)-certified clinical genetics testing laboratories. The mutation in S1 was identified through clinical WES by the Molecular Genetics Laboratory at the Children’s and Women’s Health Centre in Vancouver. The mutation in S2 was identified as described before [4]. The mutation in S3 was detected through clinical WES performed by Baylor Miraca Genetics Laboratories. The mutation in S4 was identified as described before [5]. The mutation in S5 was identified through clinical WES by GeneDx. Mutations in individuals S1-S5 were further validated by Sanger sequencing in the laboratory. The mutation in W1 was identified by WES as described before [10]. The mutation in W2 was identified by clinical WES performed by the Shodair Children’s Hospital. The mutation in W3 was identified through clinical WES by GeneDx. The mutation in W4 was detected by clinical WES at Victorian Clinical Genetics Services.

The Human Genome Variation Society (HGVS) guidelines for mutation nomenclature were used [35]. In all tables and figures, the cDNA and amino acid positions were annotated according to the most updated FOXP1 RefSeq mRNA and protein sequence (NM_032682.5 and NP_116071.2). Nucleotide numbering referring to cDNA uses +1 as the A of the ATG translation initiation codon in the reference sequence, with the initiation codon as codon 1.

The following control databases were used: Exome Variant Server (EVS, http://evs.gs.washington.edu/EVS/), Exome Aggregation Consortium (ExAC, http://exac.broadinstitute.org), and genome Aggregation Database (gnomAD, http://gnomad.broadinstitute.org).

Review of individuals with previously published FOXP1 mutations

We searched the published literature for mutations in FOXP1 using PubMed, the Human Gene Mutation Database (HGMD) Professional (Biobase), and denovo-db [36]. We retrieved and examined the genetic information from all studies indicated in Additional file 2: Table S1. We also included pathogenic mutations reported in ClinVar (NCBI, http://www.ncbi.nlm.nih.gov/clinvar/). All mutations were annotated on NM_032682.5 and NP_116071.2.

Validation of the splice site mutation in individual S1

Peripheral blood samples from individual S1 and her parents were collected in PAXgene® RNA tubes (Qiagen) and total RNA was extracted and purified using the PAXgene® Blood RNA kit, v2 (PreAnalytix). Globin mRNA was depleted from the samples using the GLOBINclear-Human Kit (Life Technologies). The quantity and quality of the purified RNA samples were measured on a Nanodrop. Total RNA (500 ng) was used for cDNA synthesis using SuperScript® II reverse transcriptase (Invitrogen). cDNA was amplified using exon-specific primers in exon 6 (SDR4345: 5′-GGACAGCTCTCAGTCCACAC-3′) and exon 9 (SDR4346: 5′-AGGTGGGTCATCATGGCTTG-3′). PCR products were extracted and purified from a 1% agarose gel using the QIAquick gel extraction kit (Qiagen) and subjected to Sanger sequencing with both forward and reverse primers (Genewiz).

Structural analyses

The average three-dimensional structure of the FOXP1 DNA binding domain was extracted from PDB #2KIU, which contains 20 NMR structures [37], and then superimposed to the FOXP2 domain co-crystallized with one double-stranded DNA molecule (PDB #2AO7) [38]. We mapped all missense and in-frame mutations from this and prior studies on to this structure. Visual inspection of the FOXP1 domain structure and the three-dimensional superposition was performed using UCSF Chimera 1.10.1 [39]. Figures were prepared using UCSF ChimeraX.

Clinical evaluation

A detailed clinical evaluation was completed, including medical history, psychiatric and neurological evaluation, and dysmorphology examination by clinical geneticists. Medical records were also reviewed, including magnetic resonance imaging (MRI) and electroencephalogram (EEG).

ASD symptomatology

ASD diagnosis was determined using the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2) [40], the Autism Diagnostic Interview-Revised (ADI-R) [41, 42], and a clinical evaluation with a child and adolescent psychiatrist or clinical psychologist. The ADOS-2 is a semi-structured direct assessment of social communication and restricted and repetitive behavior. The ADOS-2 produces a total score and domain scores in the areas of Social Affect and Restricted and Repetitive Behavior. Two clinically meaningful cut-offs can be obtained: autism spectrum and autism, the latter reflecting a greater number of symptoms. Symptom severity was assessed for both total score and individual domain scores [43] using comparison scores ranging from 1 to 10. The ADI-R is a structured caregiver interview that assesses ASD symptoms in the areas of socialization, communication, and restricted and repetitive behavior. Caregiver questionnaires were completed to further assess everyday functioning and included the Social Responsiveness Scale, Second Edition [44], the Repetitive Behavior Scale-Revised [45, 46], and the Short Sensory Profile [47]. A consensus diagnosis was determined for each individual based on ADOS-2, ADI-R, and clinical evaluation (DSM-5).

Intellectual and adaptive functioning

Global cognitive ability was measured using the Stanford Binet Intelligence Scales, Fifth Edition (n = 5) [48] and the Differential Ability Scales – Second Edition (n = 3) [49]. Records from a Wechsler Preschool and Primary Scale of Intelligence – Third Edition [50] was reviewed for one participant (W4). Full scale IQ (FSIQ), nonverbal IQ (NVIQ), and verbal IQ (VIQ) were obtained for 7/9 participants and ratio IQs were calculated for 2/9 participants (W1, W3). Adaptive behavior was measured using the Vineland Adaptive Behavior Scales, Second Edition, Survey Interview Form [51]. The presence of ID was based on results from cognitive testing and the Vineland-II.

Expressive and receptive language

Language milestones were assessed during the ADI-R and the clinical evaluation. Current expressive and receptive language abilities were measured using the Vineland-II (n = 9), the Peabody Picture Vocabulary Test, 4th Edition [52] (n = 8), and the Expressive Vocabulary Test, 2nd Edition [53] (n = 7).

Gross motor, fine motor, and visual-motor integration

Motor milestones were assessed during the ADI-R and the clinical evaluation. Current fine and gross motor skills were measured using the Vineland-II (n = 5). Visual-motor integration was measured using the Developmental Test of Visual-Motor Integration, Sixth Edition [54] (n = 7), which required individuals to copy shapes and patterns using pencil and paper. The Developmental Coordination Disorder Questionnaire [55] was completed by all caregivers to measure motor control during movement, fine motor/handwriting skills, and general coordination. Scores below 57 are indicative of a developmental coordination disorder.

Psychiatric features

The presence of internalizing and externalizing psychiatric symptoms was assessed during the psychiatric evaluation and through caregiver report forms including the Achenbach Child Behavior Checklist [56] and the Aberrant Behavior Checklist [57–59].

Results

FOXP1 mutational spectrum

Our cohort consists of nine individuals with FOXP1 mutations (Table 1, Fig. 1) and one individual with a large duplication encompassing the FOXP1 gene (Additional file 1). Three individuals had frameshift mutations introducing a premature stop codon, one individual had a nonsense mutation, one individual had a mutation in an essential splice site, three had missense mutations and one had an in-frame deletion (Table 1, Fig. 1). The p.Tyr470Cys mutation is detected in a control in gnomAD and all other mutations are absent from EVS and gnomAD. Inheritance status is unknown for p.Leu414*, while all other mutations are confirmed de novo. All missense mutations are predicted to be damaging by Polyphen-2 and SIFT. Analysis of the FOXP1 splicing in individual S1 reveals that the NM_032682.5:c.975-2A > C in the acceptor splice site between exon 7 and 8 causes exon 8 skipping and results in a Lys325Asnfs*12 mutation (Fig. 2).

Table 1

FOXP1 mutations in this cohort

ID	Coding DNA change	Protein change	Genomic change	Effect	Inheritance
S1	c.975-2A>C	p.Lys325Asnfs*12	chr3:g.71050212T>G	Splice-site	De novo
S2^a	c.1267_1268delGT	p.Val423Hisfs*37	chr3:g.71027059_71027060delAC	Frameshift	De novo
S3	c.1333_1335delinsAA	p.Val445Asnfs*29	chr3:g.71026992_71026994delinsTT	Frameshift	De novo
S4^b	c.1393A>G	p.Arg465Gly	chr3:g.71026829T>C	Missense	De novo
S5	c.1506C>G	p.Phe502Leu	chr3:g.71026116G>C	Missense	De novo
W1^c	c.1014dupA	p.Ala339Serfs*4	chr3:g.71050171dupT	Frameshift	De novo
W2	c.1240delC	p.Leu414*	chr3:g.71027087delG	Nonsense	Unknown
W3	c.1409A>G	p.Tyr470Cys	chr3:g.71026813T>C	Missense	De novo
W4	c.1590_1601delAGGGGCAGTATG	p.Gly531_Trp534del	chr3:g.71021757_71021768delCATACTGCCCCT	In-frame deletion	De novo

cDNA (NM_032682.5), protein (NP_116071.2/Q9H334–1), and genomic (GRCh37/hg19) changes are shown. Asterisks indicate stop codons

^aIndividual S2 was previously reported by Lozano et al. (2015) [4]

^bIndividual S4 was previously reported by Sollis et al. (2016) (subject 1) [5]

^cIndividual W1 was previously reported by O’Roak et al. (2011) (subject 12817.p1) [10], without clinical description

Fig. 1
*FOXP1* mutations. The mutations described in this study and those described in the literature are shown in the upper and lower panels, respectively. FOXP1 domains are reported as described for Q9H334–1 in Uniprot. The two nuclear localization signals (NLS) are indicated as previously reported [68]. Missense and in-frame mutations are indicated in blue, while loss-of-function (LoF) mutations are indicated in black. Recurrent mutations are indicated in bold. The position of c.975-2A > C reflects the Lys325Asnfs*12 mutation. The positions of the other splice-site mutations reflect the first residue of the exon downstream of the intron

Fig. 2
Exon skipping caused by the c.975-2A > C mutation. a RT-PCR results for exons 6–9 of *FOXP1* mRNA for blood-derived RNA for individual S1 and her parents. The upper band is the PCR amplicon resulting from the mRNA with exons 7 (blue), 8 (black), and 9 (purple); the lower band results from skipping of exon 8. b Sanger sequencing results of the PCR amplicons obtained in a. The nucleotide of the splice site mutated is indicated on the pre-mRNA in red. Exon numbering was annotated against NM_001244812. The indicated exon 8 corresponds to coding exon 8 in NM_032682.5 (exon 13 if including non-coding exons)

We searched through the published literature and ClinVar for pathogenic mutations in FOXP1. Besides the missense mutation with unknown inheritance reported in Worthey et al. (2013) [9], all published mutations included in Additional file 2: Table S1 are confirmed as de novo. Notably, pathogenic missense or in-frame mutations in our cohort and in previous reports cluster to the DNA binding domain of FOXP1 (Figs. 1 and 3), critical for the transcriptional activity of the protein. As shown for FOXP2, six copies of the FOX domains bind to two double-stranded segments of DNA, with two copies as monomers tightly associated with the DNA and four exhibiting domain swapping [38]. In the monomers, the ~ 100 amino acids of the winged-helix DNA binding domain are arranged in five α-helices (H1-H5) and three β strands (β1-β3) (Fig. 3a) [37, 38]. While H3, β2 and β3 are engaged in DNA recognition and binding, H2 and H4 are directly involved in the three-dimensional domain swapping (Fig. 3b).

Fig. 3
Pathogenic missense mutations in the *FOXP1* DNA-binding domain. a Primary sequence and topological representation of the DNA-binding domain (as reported in PDB 2KIU). The five helices (H1-H5), the three β sheet (β1-β3) and the two wings regions (W1, W2) are shown. Residues mutated in the cohort described in this study are in red, while those affected by mutations described in literature are in blue. b Ribbon representation of the DNA binding domain of the FOXP1 monomer interacting with one double-stranded DNA molecule. The surface for the interaction with the DNA and the region involved in domain swapping are indicated by dashed lines. c Ribbon representation of the unbound FOXP1 DNA binding domain showing the missense mutations reported in the literature in blue (Additional file 2: Table S1). d Ribbon representation of the unbound FOXP1 DNA binding domain showing the missense and in-frame mutations reported in our cohort in red (Table 1)

To understand the structural impact of the FOXP1 mutations, we superimposed the structure of FOXP1 DNA binding domain [37] on the structure of FOXP2 bound to the DNA [38] and we mapped missense and in-frame mutations from this study and previous publications. Missense mutations previously reported (Fig. 3c) and identified in this cohort (Fig. 3d) disrupt amino acids located on both the DNA binding surface and on the helices involved in dimer formation through domain swapping. The Arg465 mutated in individual S4 is directly engaged in the DNA binding and the p.Arg465Gly mutation might affect the electrostatic interaction between the N-terminal region of the domain and the negatively charged backbone of the DNA (Fig. 3b, d). Similarly, the p.Gly531_Tpr534del in individual W4 removes four residues necessary for the three-stranded antiparallel β strands, likely compromising the folding of the domain and thus affecting DNA recognition. Within this region are also two previously reported mutations: p.Ala532Val [15] and p.Trp534Arg [13] (Fig. 3c). The Tyr470 residue mutated in W3 is part of an interaction network of aromatic residues located on H1, H3, H4 and H5 (Fig. 3d) and its mutation to Cys might result in the structural destabilization of the hydrophobic core of the protein. The Phe502 residue mutated in S5 is located on H4, which is involved in the domain swapping (Fig. 3d): reduction of steric hindrance in the core of the swapped dimer resulting from the p.Phe502Leu mutation might compromise FOXP1 dimerization.

After removing cases ascertained or reported multiple times, we found seven instances of recurrent mutations (Fig. 1, Additional file 2: Table S1). The NM_032682.5:c.975-2A > G mutation resulting in the Lys325Asnfs*12 mutation is detected in an individual in our cohort (S1, Fig. 2) and in an individual from an ID cohort screened by targeted sequencing [11]. In addition, a mutation in the first nucleotide of the splice donor site (NM_032682.5:c.974 + 1G > C) was found by WES in a DD/ID cohort [12] (Fig. 1, Additional file 2: Table S1), but the consequences of this mutation on splicing have not been assessed. A p.Tyr439* mutation was reported in a Dutch individual included in Sollis et al. (2016) (individual 3) [5] and independently identified in an individual screened by the Emory Genetics Laboratory at Emory University and deposited in ClinVar (#194566) (Fig. 1, Additional file 2: Table S1). Another nonsense mutation (p.Arg503*) was found in at least two independent cases reported in ClinVar (#211038). The mutation was also reported in a large-scale WES analysis [14], but it is unclear whether this case overlaps with one of those in ClinVar. A third nonsense mutation (p.Arg525*) was reported in two independent cases [1, 12] (Fig. 1, Additional file 2: Table S1). The other recurrent mutations are all missense mutations of residues located in the DNA binding domain (Pro466, Arg514, and Ala532) (Fig. 3a). Notably, the p.Pro466Leu is equivalent to the p.Pro505Leu in FOXP2 that has been associated with language and speech impairment [60]. Also, the p.Arg514His is equivalent to the FOXP2 p.Arg553His mutation identified in childhood apraxia of speech [26]. Arg514 and Arg553 in FOXP1 and FOXP2, respectively, are located in the DNA binding surface that intercalates with the major groove of the DNA (Fig. 3b, c) and p.Arg553His has been shown to cause protein mislocalization, with the formation of nuclear and cytoplasmic aggregates, and abolish the binding to DNA and transactivation [61].

Neuropsychological phenotype of FOXP1 syndrome

An extensive battery of clinical assessments was employed to delineate the core phenotype of FOXP1 syndrome. To examine the neuropsychological phenotype, gold-standard assessment of ASD symptomatology, cognitive functioning, adaptive behavior, receptive and expressive language, fine and gross motor skills, and visual-motor integration was completed for participants S1-S6 and W1-W4 (Table 2). Results for individual S6, who carries a large duplication spanning FOXP1, interpreted conservatively as of unknown significance (Additional file 1: Figure S1), are discussed in Additional file 1.

Table 2

Neuropsychological and psychiatric manifestations in individuals with FOXP1 syndrome

Neuropsychological assessments	S1	S2	S3	S4	S5	W1	W2	W3	W4	Total(%)/Average
Age (months)	71	200	133	138	96	191	134	118	117	133.1
ASD symptoms
ASD on ADOS-2^a	–	+	++	–	–	++	–	++	n/a	4/8 (50%)
ADOS-2 Comparison Score	3	5	7	3	3	10	1	6	n/a	4.8
ASD on ADI-R	+	+	+	–	–	+	–	–	–	4/9 (44%)
Consensus diagnosis of ASD	–	–	–	–	–	+	–	+	n/a	2/8 (25%)
Cognitive functioning (SS)
Nonverbal IQ	95	44	56	61	73	19	59	31	70	56.4
Verbal IQ	92	44	51	59	58	15	40	24	74	50.8
Full scale IQ	93	42	51	59	64	17	52	29	68	52.8
Adaptive behavior (SS)
Vineland-II communication	81	59	69	67	74	42	61	57	67	64.1
Vineland-II daily Living	64	47	58	58	78	38	52	62	66	58.1
Vineland-II socialization	74	54	64	69	83	42	68	55	69	64.2
Vineland-II composite	69	53	62	63	77	39	59	59	67	60.9
Language (AE in months)
Expressive vocabulary^b	55	78	94	95	59	n/a	81	35	n/a	71.0
Receptive vocabulary^c	56	69	75	93	54	44	81	29	n/a	62.6
Vineland-II expressive	42	48	59	54	55	25	52	23	46	44.9
Vineland-II receptive	41	35	47	47	30	34	26	34	30	36.0
Motor skills
Visual-motor integration (SS)^d	78	45	<45	45	70	n/a	47	<45	n/a	< 53.6
Vineland-II gross motor (AE)	46	n/a	47	82	59	n/a	n/a	37	n/a	54.2
Vineland-II fine motor (AE)	34	n/a	69	66	68	n/a	n/a	36	n/a	54.6
Psychiatric features
Anxiety	+	+	+	+	+	+	+	–	+	8/9 (89%)
Compulsive behaviors	+	+	+	+	+	+	+	+	+	9/9 (100%)
Attention problems	+	+	+	+	+	+	+	+	+	9/9 (100%)
Externalizing symptoms	+	+	+	+	+	+	+	+	+	9/9 (100%)

^aADOS-2 classification: + = autism spectrum; ++ = autism

^bExpressive vocabulary measured by the Expressive Vocabulary Test, 2nd Edition

^cReceptive vocabulary measured by the Peabody Picture Vocabulary Test, 4th Edition

^dVisual-motor integration measured by the Test of Visual-Motor Integration, 6th Edition

ADI-R Autism Diagnostic Interview-Revised, ADOS-2 Autism Diagnostic Observation Schedule, 2nd Edition, AE age equivalents, SS standard score, n/a information not available

ASD symptoms. Participants S1-S5 and W1-W3 received ASD diagnostic testing and a clinical evaluation to assess DSM-5 criteria for ASD (Additional file 3: Table S2). All individuals displayed ASD symptoms, however only two of the eight individuals evaluated for ASD (W1 and W3) received a diagnosis of ASD (25%) based on expert clinical consensus.

Four individuals (S1–3, W2) were administered a Module 3 of the ADOS-2 (indicating the presence of fluent speech), two individuals (S4–5) were administered a Module 2 (indicating the presence of phrase speech), and two individuals (W1, W3) received a Module 1 (indicating use of single words or no words). Four of eight individuals met criteria on the ADOS-2; three individuals (S3, W1, W3) were above the cutoff for an autism classification and one (S2) was above the cutoff for an autism spectrum classification. The two individuals who met criteria for a clinical diagnosis of ASD received a Module 1 of the ADOS-2, suggesting limited language ability. ADOS-2 overall comparison scores ranged from 1 to 10 (4.8 ± 2.9). Comparison scores for individual ADOS-2 domains indicated higher scores in the Restricted and Repetitive Behavior domain (6 ± 2.7) as compared to the Social Affect domain (4.8 ± 1.8).

On the ADI-R (n = 9), four individuals were above the cutoff for ASD on all domains, six were above the cutoff on the Social domain and seven were above the cutoff on the Communication and Restricted and Repetitive Behavior domains.

ASD symptoms were further assessed through caregiver report questionnaires. Results from the Social Responsiveness Scale, 2nd Edition (n = 9) indicated deficits across all domains (76.4 ± 7.8). Social motivation fell within the mild range of impairment (65.2 ± 8.3) as compared to moderate impairment in social communication (74.8 ± 9.10), social awareness (74.9 ± 6.6), and repetitive behavior (75.6 ± 8.6), and severe impairment in social cognition (76.8 ± 9.5).

In the area of repetitive behavior and restricted interests, scores on the Repetitive Behavior Scale-Revised (n = 9) were all within one standard deviation of norms published in individuals with ASD [62]. Within this cohort, the greatest number of symptoms was reported for insistence on sameness (5.6 ± 2.7, range: 2–10). Average self-injurious behaviors scores were higher than the reference sample (4.6 ± 4.0), however scores varied across participants (range: 0–13). On the Short Sensory Profile (n = 5), total scores indicated definite sensory differences in three individuals, possible differences in one individual, and typical performance in one individual. An examination of individual domains indicates the greatest number of symptoms in the area of underresponsive/seeks sensation with four of five individuals (S1, S3–4) meeting criteria for a definite sensory difference in this domain.

Results from the psychiatric evaluation provided additional detail on restricted and repetitive behavior in individuals with FOXP1 syndrome. Common repetitive behaviors included intense and highly restrictive interests, stereotypic movements such as hand flapping, insistence on sameness, or adherence to routines. Caregivers also reported compulsive behaviors in all individuals, which included hoarding of small objects. Of note, five out of the nine (56%) participants were reported to engage in compulsive picking of the skin and nails.