Article Text

Download PDFPDF

Original article
FOXP2 variants in 14 individuals with developmental speech and language disorders broaden the mutational and clinical spectrum
  1. Miriam S Reuter1,
  2. Angelika Riess2,
  3. Ute Moog3,
  4. Tracy A Briggs4,5,
  5. Kate E Chandler4,
  6. Anita Rauch6,
  7. Miriam Stampfer2,
  8. Katharina Steindl6,
  9. Dieter Gläser7,
  10. Pascal Joset6,
  11. DDD Study,
  12. Mandy Krumbiegel1,
  13. Harald Rabe8,
  14. Uta Schulte-Mattler8,
  15. Peter Bauer2,
  16. Stefanie Beck-Wödl2,
  17. Jürgen Kohlhase9,
  18. André Reis1,
  19. Christiane Zweier1
  1. 1Institute of Human Genetics, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, Germany
  2. 2Institute of Medical Genetics and Applied Genomics, University of Tübingen, Tübingen, Germany
  3. 3Institute of Human Genetics, Heidelberg University, Heidelberg, Germany
  4. 4Manchester Centre for Genomic Medicine, St Mary's Hospital, Central Manchester University Hospitals NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK
  5. 5Faculty of Medical and Human Sciences, Institute of Human Development, University of Manchester, Manchester, UK
  6. 6Institute of Medical Genetics, University of Zurich, Schlieren-Zurich, Switzerland
  7. 7Genetikum, Neu-Ulm, Germany
  8. 8Kinderzentrum St. Martin, Regensburg, Germany
  9. 9Centre for Human Genetics, Freiburg, Germany
  1. Correspondence to Dr Christiane Zweier, Institute of Human Genetics, Friedrich-Alexander-University Erlangen-Nürnberg, Schwabachanlage 10, Erlangen 91054, Germany; christiane.zweier{at}


Background Disruptions of the FOXP2 gene, encoding a forkhead transcription factor, are the first known monogenic cause of a speech and language disorder. So far, mainly chromosomal rearrangements such as translocations or larger deletions affecting FOXP2 have been reported. Intragenic deletions or convincingly pathogenic point mutations in FOXP2 have up to date only been reported in three families. We thus aimed at a further characterisation of the mutational and clinical spectrum.

Methods Chromosomal microarray testing, trio exome sequencing, multigene panel sequencing and targeted sequencing of FOXP2 were performed in individuals with variable developmental disorders, and speech and language deficits.

Results We identified four different truncating mutations, two novel missense mutations within the forkhead domain and an intragenic deletion in FOXP2 in 14 individuals from eight unrelated families. Mutations occurred de novo in four families and were inherited from an affected parent in the other four. All index patients presented with various manifestations of language and speech impairment. Apart from two individuals with normal onset of speech, age of first words was between 4 and 7 years. Articulation difficulties such as slurred speech, dyspraxia, stuttering and poor pronunciation were frequently noted. Motor development was normal or only mildly delayed. Mild cognitive impairment was reported for most individuals.

Conclusions By identifying intragenic deletions or mutations in 14 individuals from eight unrelated families with variable developmental delay/cognitive impairment and speech and language deficits, we considerably broaden the mutational and clinical spectrum associated with aberrations in FOXP2.

  • FOXP2
  • language
  • speech
  • developmental delay

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Genetic factors are supposed to play a major role in the aetiology of speech and language impairment,1 ,2 but monogenic forms appear to be very rare. Based on the findings of a missense mutation segregating in a multigenerational family with speech disorder, and of a gene-disrupting translocation in a single individual with a similar phenotype in 2001, FOXP2, encoding a forkhead transcription factor, was the first gene to be implicated in a specific developmental speech and language disorder (SPCH1, OMIM #602081).3

Since then, numerous individuals with large 7q31 deletions or reciprocal translocations, affecting the FOXP2 locus, were identified.4–12 The described phenotypes included language impairment but were generally more complex due to haploinsufficiency for other genes or due to complex rearrangements.4 ,7 Furthermore, several variants of unknown significance, affecting the coding sequence of FOXP2 but not cosegregating with the disease phenotype in the respective families were reported.13–15

So far, only three clearly pathogenic intragenic FOXP2 mutations have been published: the initial missense mutation p.(Arg578His) (named p.(Arg553His) in the original publication based on isoform NM_014491/NP_055306), segregating in 15 affected members of a large three-generation family;3 a nonsense mutation p.(Arg353*) (named p.(Arg328*) in the original publication, isoform NM_014491/NP_055306) in two siblings and their mother;14 and a frameshift mutation p.(Gln415Valfs*7) in a sporadic patient.15 The phenotype associated with FOXP2 mutations comprises a broad spectrum of speech and language-related disorders. Among those, a particularly prominent and frequent key phenotypic aspect is childhood apraxia of speech (CAS),16 a neurological speech sound disorder in which the precision and consistency of movements underlying speech are impaired in the absence of neuromuscular deficits.17 While CAS is only a specific aspect of the broad FOXP2-related speech and language phenotypes, vice versa mutations in FOXP2 only underlie a small proportion of cases with CAS.

We now report on 14 further affected individuals from eight unrelated families with deleterious mutations or intragenic deletions in FOXP2, thus further delineating the mutational and clinical spectrum of FOXP2-related speech and language impairment.

Material and methods

All nine index individuals from eight unrelated families were seen and diagnosed at different genetic centres in different countries. Aberrations in FOXP2 were detected by various approaches.

Chromosomal microarray testing and monozygosity testing

Chromosomal microarray testing with an Affymetrix CytoScan HD-Array was performed within routine diagnostic testing for developmental delay in one of the twin sisters of family 1 and subsequently in her parents to test segregation of the FOXP2 deletion. Monozygosity of both sisters was confirmed by multiplex PCR with 15 polymorphic markers located on 13 chromosomes (PowerPlexTM16, Promega), analysed on an automatic capillary sequencer (ABI3100).

Targeted FOXP2 sequencing

Index patients from families 2, 3 and 8 were specifically tested for FOXP2 mutations, prompted by clinical suspicion. All coding exons of FOXP2 (NM_148898.3 or ENST00000350908, respectively, regarding different isoforms and mutation annotations see online supplementary table and figure S1) were sequenced by standard methods. Primer sequences and conditions are available on request. Subsequent testing in the parents and siblings, if applicable, comprised sequencing of the affected exon, respectively. The index patients from families 3 and 8 were tested in routine diagnostic settings. In family 3, additional testing for intragenic deletions by quantitative real-time PCR with six intragenic amplicons was performed, according to the method described previously.18 The index patient from family 2 was included in a group of 15 individuals with mild developmental delay or intellectual disability with more pronounced speech and language impairment selected for research testing of FOXP2 from a larger study group of patients with developmental delay or intellectual disability. This study was approved by the ethics committee of the medical faculty of the Friedrich-Alexander-University Erlangen-Nürnberg, and all participating families had given their informed consent.

Panel sequencing of known intellectual disability (ID) genes

The index individuals of families 4 and 7 were tested by next-generation sequencing (NGS) with the TruSight One Panel (Illumina) containing ∼4800 known disease genes on a MiSeq platform (Illumina) within routine diagnostics for developmental delay. In individual 4, only 525 genes associated with intellectual disability and contained in the panel were analysed and evaluated. Confirmation of the variants and segregation testing in the respective families were performed by Sanger sequencing.

Trio exome sequencing

The DECIPHER database19 (status 04/16) contained three individuals with de novo point mutations in FOXP2, two of them included as families 5 (DECIPHER number 271246) and 6 (DECIPHER number 267249) in more detail in this study. In both index individuals, trio exome sequencing was performed and mutational findings validated as described by the Deciphering Developmental Disorders (DDD) study,20 ,21 having tested the first 1000 of planned 10 000 children with developmental disorders. Both mutations were confirmed by Sanger sequencing in the Manchester Genetics Laboratory.


Clinical and mutational details

Family 1

These monozygotic twin sisters are the second and third children of healthy, non-consanguineous parents. They were born after uneventful pregnancy at 36 weeks gestation with a weight of 2560 g and a length of 49 cm, each. The girls showed mild delay of motor development with walking at age 18 months. Speech development was severely delayed with first words at age 4 years and short sentences at age 8 years. Both girls attended a school for children with special needs. They could calculate up to 1000 but had impaired reading and writing abilities. Formal IQ testing at the age of 10 years resulted in full-scale IQs of 78 and 79, respectively (Wechsler Intelligence Scale for Children). Verbal comprehension indices were 77 and 67, perceptual reasoning indices were 86 and 88, working memory index was 82 in both girls and processing speed indices were 88 and 103, respectively. Tests for reading abilities (ELFE 1–6) and orthography (Salzburger Lese- und Rechtschreibtests (SLRT-II)) showed results significantly below average.

Both girls were last reviewed aged 11 years. The first girl had a height of 136 cm (10th–50th centile), a weight of 34 kg (25th–50th centile) and a head circumference of 52 cm (10th–25th centile). Her sister had a height of 132 cm (3rd–10th centile), a weight of 29 kg (10th–25th centile) and a head circumference of 51.5 cm (10th–25th centile). Both girls behaved shy, but friendly and did not speak spontaneously. When replying to questions they spoke in short sentences with slurred articulation. Minor morphological aspects included epicanthic folds, mildly upslanting palpebral fissures, prominent incisors and sandal gaps. EEG testing was normal.

Chromosomal microarray testing revealed a heterozygous, de novo deletion of 14 kb within FOXP2 (arr[hg19] 7q31.1(114 296 590–114 310 602)x1), which was excluded in the parents (figure 1A). This deletion affects exons 12–17 and is predicted to result in frameshifting and thus truncation of the protein.

Figure 1

Genetic aberrations in FOXP2. (A–H) Pedigrees of families with identified deletions or mutations (electropherograms of the index individuals) in FOXP2 showing clinical manifestation and mutational status. Black and dark grey boxes and circles represent individuals with speech and language impairment, boxes with lines or rectangles represent individuals with non-FOXP2-related speech and/or developmental delay or Marfan syndrome, respectively.

Family 2

This proband is the first of two children of non-consanguineous parents. During pregnancy, high blood pressure in the mother and placental insufficiency were reported. The boy was born at 40 weeks gestation with a weight of 3000 g, a length of 49 cm and a head circumference of 34 cm. He could sit at age 8–9 months and walk freely at age 14 months. His speech development was delayed with using only single words at age 4 years. Developmental testing at age 8 years (Snijders-Oomen non-verbal intelligence test, SON-R 2½–7) revealed a non-verbal intelligence in the low normal range and severe receptive and expressive language deficits. He attended a school for children with special needs. On physical examination at age 13 years and 9 months, his head circumference was 53.3 cm (25th–50th centile). Minor morphological aspects were a high nasal root, a pointed nasal tip, prominent incisors, an overbite, tapering fingers and broad feet. He behaved friendly and spoke in simple sentences with dyspraxic and slurred articulation. IQ testing (Wechsler Intelligence Scale for Children) at age 14 years and 3 months revealed a full-scale IQ of 72 with obvious discrepancies between subtests. Verbal comprehension index was 57, while indices for perceptual reasoning, working memory and processing speed ranged between 77 and 96. EEG testing as well as conventional karyotyping and testing for fragile X syndrome were normal. Molecular microarray testing revealed a copy number variant of unclear significance that was inherited from the mother (Affymetrix Genechip 6.0 Mapping SNP Array: arr[hg18] 16p13.11(14 897 352–15 035 487)x1; 138 kb, genes: NPIP, PDXDC1).

His younger brother was also referred because of developmental delay. He was born after uneventful pregnancy at gestational week 39 with a weight of 3080 g and a length of 50 cm. He could walk early but showed initially delayed speech development. At age 7 years, his head circumference was 50.5 cm (10th–25th centile). He had mildly asymmetric palpebral fissures and a long philtrum. He was friendly and liked to talk which he did fluently in complex sentences. Developmental testing (Wechsler Intelligence Scale for Children) revealed a full-scale IQ of 83 (verbal comprehension index of 79, perceptual reasoning index of 79, working memory index of 84, processing speed index of 111). Chromosomal microarray testing with an Affymetrix CytoScan HD-Array did not reveal pathogenic copy number variants.

The mother of both boys reported about a history of ‘stuttering’ and learning difficulties during her own childhood. Speech dyspraxia was noted in her. Formal testing of her cognitive and language abilities was not performed.

Targeted sequencing of FOXP2 revealed the nonsense mutation c.1057C>T, p.(Arg353*) in the index patient and his similarly affected mother, but not in the younger brother with previous developmental delay but fluent speech (figure 1B).

Family 3

The index patient was born at 35 weeks gestation. Her mother was treated with propranolol during pregnancy because of arrhythmias due to a mitral valve prolapse. Apart from that, pregnancy was uneventful. There were no feeding difficulties, but she was reported to be very quiet during infancy. While her motor development was only mildly delayed (sitting unsupported at 9 months, walking at 20 months), her speech development was significantly delayed. On IQ testing at age 3 years and 10 months (SON-R 2½–7), her developmental age corresponded to 24–34 months. On examination at age 5 years and 5 months, she was tall for her age (height 128 cm, >97th centile; weight 25 kg, 90th centile; head circumference 50 cm, 25th centile), mildly hypotonic and did not show dysmorphic features. She had nearly no active language and could speak a few single words only. Psychological testing at age 5 years and 10 months with the non-verbal version of Kaufman Assessment Battery for Children revealed a total score of 53 corresponding to mild ID which was supported by further tests. When last seen at age 10 years and 10 months, her height and weight corresponded to the 75th and 90th centile, respectively, head circumference was 52.5 cm (mean). She was friendly, interested in her surroundings, had a balanced temper but showed autistic features. A formal diagnosis of autism could not be established. IQ testing at that time was reported to have resulted in an IQ of about 70. She still showed significant language impairment and was trained in using communication aids. She was susceptible to airway infections and suffered from pneumonia at least once a year.

Cytogenetic analysis, testing for 7q11.23 duplication by FISH analysis and molecular genetic testing for fragile X syndrome showed normal results. Her father was a late talker but had no formal testing or therapy in childhood, and no further information was available. Her paternal half-brother was reported to show a similar language development disorder and to attend a school for children with special needs. He began to speak at age 6 years, and at age 17 years he was able to speak slowly in full sentences with unclear pronunciation.

Targeted sequencing of FOXP2 in the girl revealed the missense mutation c.1682G>C, p.(Arg561Pro). The variant was also identified in the father but markedly weaker, thus pointing to mosaicism (figure 1C). However, further investigations by testing a buccal swab or segregation in the family, particularly in the half-brother, was not possible. The p.(Arg561Pro) missense mutation affects a highly conserved amino acid within the forkhead domain (figure 2) and is predicted to be deleterious by MutationTaster22 (score 0.999) and SIFT23 (score 0, median information content 3.12) and to be probably damaging by PolyPhen224 (score 0.998 (sensitivity: 0.27; specificity: 0.99)).

Figure 2

Genomic and protein structure of FOXP2. (A) Schematic drawing of coding exons of FOXP2 (isoform 2, NM_148898.3) with all published and novel intragenic deletions and mutations. Exons encoding the forkhead domain are coloured in blue. (B) Alignment using Clustal Omega32 ,33 of the forkhead domain amino acid sequences as predicted by PROSITE34 of human FOXP2 and eight other species, three further human FOXP and representative FOX proteins. Invariant residues are shown in the bottom line. Frames indicate the positions of the three missense variants in FOXP2. Grey shadings indicate residues that are reported to be affected by pathogenic missense variants in ClinVar and/or high confidence missense variants in the Human Gene Mutation Database (HGMD35) in other disease associated FOX proteins. Below, the proposed structure of the forkhead domain with three α-helices, three β-strands (S1–S3) and two ‘wings’ as established by X-ray crystallography36 and as described previously3 is displayed.

Family 4

This girl is the only child of healthy, non-consanguineous parents. The pregnancy resulted from in vitro fertilisation. Amniocentesis due to maternal age revealed normal results. The girl was born by caesarean section at 41 weeks gestation with a weight of 3690 g, a length of 52 cm and a head circumference of 34 cm. She was breast fed for more than a year and parents reported frequent choking. At age 10 months, she had a possible seizure. Age of free walking was 16 months, and age of the first word was 13 months. Subsequently, delay of speech development was noted. At age 2 years, she spoke approximately 30 words. Developmental testing at age 3 years revealed average cognitive performance with speech delay (SON-R 2½–7). Further testing at age 6 years revealed persistent speech impairment and limited auditive memory capacity (differential performance test, DL-KG according to Kleber and Hans; concentration test ‘Schau genau’; preschool testing to assess auditive-kinesthetic perception and speech processing (Heidelberger Vorschultest HVS)). She attends a mainstream school with good grades in maths, German language (written tests) and other subjects. However, as she has difficulties to find classrooms by herself, she is accompanied by a learning support assistant. She gets speech and occupational therapies.

The patient was physically last reviewed aged 5 years and 4 months, when her height was 122 cm (>97th centile), her weight was 25.3 kg (97th centile) and her head circumference was 52 cm (75th–90th centile). Minor morphological aspects included narrow palpebral fissures, a high arched palate and mild finger pads. The girl had a broad-based gait at younger age. EEG and audiometry were normal. Ophthalmological examinations revealed strabismus, hyperopia and astigmatism. Testing for fragile X syndrome and chromosomal microaberrations (SNP Array) were normal.

NGS panel of 525 genes implicated in intellectual disability revealed the de novo nonsense mutation c.1507C>T, p.(Arg503*) in FOXP2 (figure 1D).

Family 5

The proband was the third child of healthy, non-consanguineous parents. His eldest brother had Marfan syndrome, secondary to a de novo Fibrillin 1 mutation, but no developmental delay. The proband was born at 41 weeks gestation, following an uneventful pregnancy, and weighed 3570 g. He fed well and there were no early developmental concerns, with independent sitting achieved at 9 months and walking at 15 months. However, concerns regarding speech became apparent and parents commented that he made unusual coughing sounds, like he was choking. By 4.5 years he was only speaking two words, ‘yes’ and ‘no’ and these were spoken quietly. His hearing was normal, as was his vision apart from a left exotropia. Concerns were raised about his fine motor development and he used a palmer grasp until age 5 years. At 6 years and 8 months, he had difficulties dressing himself, could write his name and letters, but no other words and was in mainstream school with an 18-hour statement of educational needs. On examination, his height was 129.6 cm (91–98th centile), weight 28.2 kg (91–98th centile) and head circumference 53.5 cm (50th centile). He had rather full lips and a slightly upturned nose, he was cooperative, but did not speak. He had a normal microarray, fragile X testing and metabolic screen, including homocysteine level testing.

At age 9 years and 11 months, he was diagnosed with autism spectrum disorder, and his school support increased to 25 hours a week. He continued to speak only two words consistently, did not put words together and continued to make unusual vocalisations. He continued to have fine motor delay, struggling with buttons and unable to tie shoes. Assessment at this time considered his receptive language at an age equivalent score of 4 years and 7 months. He used a V-Pen and iPAD to improve communication and had limited use of descriptive and emotionally expressive gestures. He did not and had not ever undertaken imaginative play, tended to watch rather than play with other children, had a number of preoccupations and flapped his wrists when excited. Trio exome sequencing within the DDD study demonstrated the de novo variant c.1765C>T, p.(Arg589*) in FOXP2 (figure 1E).

Family 6

This patient was the third child of healthy, non-consanguineous parents. His elder sister had speech delay and significant anxiety problems with obsessive compulsive traits. He has another healthy sister. Neither parent had any developmental, learning or behavioural difficulties. He was born after a normal pregnancy with a birth weight of 4250 g. His early development was normal and he walked at 15 months. However, concerns were raised about his speech development as he was slow to progress from babble to proper words. As a young child he communicated through sign language and he had no formal speech until 7 years old. When reviewed aged 15 years, he still did not have fluent speech and his pronunciation was poor. However, his understanding of language was good, and he used appropriate responses to questions although these would be stilted and literal in context. When reviewed aged 19 years, parents reported ongoing speech difficulties with stuttering, and stilted intonation and a reluctance to speak. As a child he had poor communication and social interaction skills and tended to be loner, playing alongside rather than with other children. He developed obsessive and ritualistic behaviours. At age 15 years, he was diagnosed with autistic spectrum behavioural disorder. He attended mainstream school throughout his education but required learning support through a Statement of Special Educational Needs. He was good at maths and computers and was able to read and write. His general health was good. He continued to be under review by the optician for myopia. When last examined aged 15 years, his head circumference was 57.1 cm (50th–75th centile), height was 165.25 cm (25th centile) and weight 60.75 kg (50th–75th centile). He had minor facial dysmorphism with small simply folded ears, a prominent nose and chin and a large protruding tongue. He has a 6 cm diameter café au lait macule in his lumbar region and multiple freckles in keeping with his pale complexion and red hair. His neurological examination and gait were normal. Investigations included a normal chromosome microarray, fragile X testing and urinary metabolic screen. Trio exome sequencing within the DDD study demonstrated the de novo variant c.1689delT, p.(Phe563Leufs*28) in FOXP2 (figure 1F).

Family 7

The index patient was the first child of non-consanguineous parents originating from the Philippines. After an uneventful pregnancy, he was born without complications at 38+2 weeks of gestation with 2650 g, a length of 47 cm and head circumference of 32.5 cm. Apgar scores were 5/9/9. Motor development was within normal range (unsupported sitting at age 8 months and walking at 14 months). Speech development was severely delayed with first words at age 4 years and only five words at the time of referral at age 6.5 years. At that time his weight was 20.4 kg (31st centile, WHO charts), his length was 109.2 cm (4th centile, WHO charts) and his head circumference was 52.5 cm (50th–75th centile). He showed minor dysmorphism such as abnormal hair whorls, backwards rotated, relatively large ears with uplifted earlobes, horizontal eyebrows, wide intermammillary distance and fetal pads (finger and toes). He interacted well, but had tendencies to hyperactivity and temper tantrums.

His younger sister was 4.5 years old at time of referral and also showed mild speech delay. She was born after uneventful pregnancy and delivery with a birth weight of 3050 g, a length of 46 cm and a head circumference of 34 cm. Motor development was mildly delayed with unaided walking at 18 months which was explained by hip dysplasia. She spoke first words at 8 months and three-word sentences with 4.5 years. She also had mildly backwards rotated, relatively large ears with uplifted earlobes and horizontal eyebrows.

The father was 36 years old and received disability pension. He spoke first words at 7 years with limited further speech development. At the age of 15 years, he migrated to Switzerland and received special education and learned to write and read well. He lived with his parents and was able to help with simple housekeeping tasks, but was slow and could not plan ahead. He had good memory and orientation and could take care of the children, but was too impatient to play with them. He spoke in simple sentences with slurred pronunciation. His head circumference was 58.7 cm (90th centile) and his height was 165 cm (7th centile, WHO charts). He had rather horizontal eyebrows and mildly uplifted earlobes.

High-resolution chromosomal microarray testing, RAI1 sequencing and multiplex ligation-dependent probe amplification (MLPA) in the index patient revealed normal results. Gene panel sequencing of 4813 disease genes revealed the heterozygous FOXP2 stop mutation c.1507C>T, p.(Arg503*) in the index patient and his father. Sanger sequencing confirmed the mutation in both children and the father (figure 1G).

Family 8

The index patient was the first child of non-consanguineous parents of German–Thai descent. He has an unaffected younger sister. He was born after an uneventful pregnancy at term with a weight of 2990 g and a length of 52 cm. His motor development was normal (sitting unsupported and crawling at 7 months, walking at 13 months), and he spoke first words at age 13 months. At age 2.5 years, delay of speech development became apparent. The proband attends a nursery school specialised in speech and language therapy. At last examination at age 6.5 years, he weighed 22.3 kg (50th centile), was 125 cm (75th centile) tall with a head circumference of 51 cm (10th centile) and did not show any dysmorphic features. Even though his vocabulary was age appropriate, his speech was mostly unintelligible and his phrases were incomplete. He reached a t-value of 36 (corresponds to 8%) in the TROG-D-Test (a test to examine grammatical understanding) and thus ranges beneath average for his age. There were also gaps in his phonemic inventory. He has borderline cognitive abilities (IQ 72) and is planned to attend a school for children with language-related learning disabilities. He uses an electronic communication device at home and in nursery school.

His father is the third child of non-consanguineous parents of German descent. His speech development was significantly delayed, while motor development was normal. Additionally, he presented with strabismus since childhood which resulted in an exotropia of the right eye as an adult and was treated with surgery. Due to the speech delay and articulation disorder, he attended a nursery school and school for children with language-related learning disabilities. Nonetheless, he was able to acquire an advanced technical college entrance qualification and is working as an industrial mechanic. Neuropsychological testing at age 29 showed an average intelligence level. Cranial MRI was normal. At age 40 years, he still receives speech therapy.

Chromosomal microarray testing by SNP Array revealed normal results. Targeted diagnostic sequencing of FOXP2 resulted in the heterozygous FOXP2 missense mutation c.1589C>T, (p.Pro530Leu) in the index patient and his father (figure 1H). Testing in the father's brother and father is currently ongoing. The p.(Pro530Leu) missense mutation affects a highly conserved amino acid within the forkhead domain (figure 2) and is predicted to be deleterious by MutationTaster22 (score 1) and SIFT23 (score 0, median information content 3.12) and to be probably damaging by PolyPhen224 (score 0.999 (sensitivity: 0.14; specificity: 0.99)).

Clinical spectrum

Clinical phenotypes are summarised in table 1 and pedigrees are displayed in figure 1. All nine index patients in this study showed various manifestations of language and speech impairment. Seven of them spoke first or single words between age 4 and 7 years. Only two index individuals from families 4 and 8 spoke first words with 13 months, and speech delay was only noted after 2 years of age. Usage of aids to support communication was required for three patients. In eight individuals, articulation difficulties such as slurred speech, dyspraxia, stuttering and poor pronunciation were noted. In comparison to expressive speech, verbal comprehension was better, but tested to be not appropriate for age in six of the individuals.

Table 1

Summary of molecular and clinical data in 14 individuals with mutations in FOXP2

Motor development was normal or only mildly delayed with walking ages between 14 and 20 months. In one individual, mild fine motor difficulties were reported. Mild cognitive impairment was reported for all but two individuals. Formal full-scale IQs were obtained in seven individuals and ranged from 70 to 79 in five of them. Five index individuals attended schools for children with special needs, and three individuals attended mainstream schools but required additional educational support.

Behavioural anomalies were reported in four individuals, two of them formally diagnosed with autism spectrum disorder. Tall stature with body heights above the 90th centile was noted in three individuals, while growth was normal in the remaining. No major malformations, dysmorphism or other general health issues were reported.

Mutational spectrum

Identified mutations are summarised in table 1 and displayed in figure 2. The missense mutations p.(Pro530Leu) and p.(Arg561Pro) are localised within the forkhead domain of FOXP2. The Pro530 residue is located upstream of the first α-helix of the forkhead domain and is invariably conserved across different species, all four FOXP proteins and all displayed other FOX proteins. The Arg561 residue is located in the second α-helix and is highly conserved across different species and all four FOXP proteins. All other variants are deletions, stop or frameshifting mutations, thus predicted to result in truncation of the protein. p.(Arg503*) was observed in two unrelated families, and p.(Arg353*) had previously been reported in another family.14 Only p.(Pro530Leu) has been reported in the ExAC browser once (frequency 8×10−6). None of the other mutations was listed, though at amino acid position 561 three other missense changes to tryptophan, glycine and glutamine were reported in very low frequencies of 1–3 of 121 240 alleles (p.(Arg561Trp), 2.4×10−7; p.(Arg561Gly), 8×10−6; p.(Arg561Gln), 1.6×10−7). Also at amino acid position 578, the location of the first reported mutation,3 another missense variant (p.(Arg578Cys)) was listed with a frequency of 8×10−6.


By various approaches, we identified mutations or deletions in FOXP2 in 14 individuals with speech and language disorders from eight unrelated families. In index patients from families 2, 3 and 8, targeted Sanger sequencing of FOXP2 was initiated prompted by clinical suspicion. In all other index individuals, the respective FOXP2 aberration was detected by unbiased testing with chromosomal microarray testing or NGS.

Six of the herewith reported mutations are novel, while the p.(Arg353*) mutation has previously been reported as p.(Arg328*) (isoform NM_014491/NP_055306) in another family.14 This mutation as well as p.(Arg503*) are so far the only recurrent mutations, observed in two families, each. The initially reported missense mutation p.(Arg578His) (originally named p.(Arg553His) based on isoform NM_014491/NP_055306)3 is located within the forkhead DNA-binding domain of FOXP2, thus expected to result in severe loss of function of the protein. Functional studies confirmed severe impairment of FOXP2 function by disrupting nuclear localisation and DNA-binding properties.25 Also the missense mutations p.(Arg561Pro) and p.(Pro530Leu) in families 3 and 8 affect highly conserved amino acids within the forkhead domain of the protein. Residue Pro530 is invariable across several species and FOX proteins, and residue Arg561 is conserved across several species and all four FOXP proteins. Though not mutated in as many other disease associated FOX proteins as the Arg578 residue, also the highly conserved Pro530 residue is reported to carry a pathogenic missense mutation: p.(Pro339Ala) in FOXP3 in a patient with immunodysregulation, polyendocrinopathy, and enteropathy, X-linked (IPEX) syndrome26 (MIM#304790). Neither in ClinVar nor in the Human Gene Mutation Database pathogenic missense variants were reported for the Arg561 position, but for adjacent or close residues. Therefore, also for the missense variants p.(Pro530Leu) and p.(Arg561Pro), a severe effect on the protein function, similar to the p.(Arg578His) mutation, might be likely. However, experimental proof of potential deleterious consequences is currently lacking. For the published p.(Arg353*) truncating mutation, an unstable, predominantly cytoplasmic product that lacks transactivation capacity was demonstrated.25 In combination with the observation of larger deletions of FOXP2, this pointed to haploinsufficiency as the most likely cause for the related language and speech phenotypes. This is now further supported by the herewith observed intragenic deletion and five nonsense or frameshifting mutations, all predicted to result in truncation of the protein with loss of the forkhead domain. Interestingly, five of the eight point mutations known to date affect arginine residues. This might be due to the susceptibility of CpG sites to sequence alterations.

All affected index individuals show pronounced impairment of expressive and receptive language, including significant delay in speech development, orofacial dyspraxia, articulation deficits and compromised production of fluent speech. This is similar to the phenotypes reported in previous patients.3 ,14 ,15 However, detailed and systematic comparison of specific speech phenotypes is limited by the circumstance that the herewith reported individuals were seen by different clinicians in different genetic centres and not examined with the same diagnostic instruments/indices. Standardised tests of speech and language in these individuals would be an interesting task for future research. Of note, considerable intrafamilial and interfamilial variability regarding both speech and cognitive abilities could be observed, particularly in families 3, 7 and 8. Interestingly, the index patients from family 4 and the affected girl in family 7, both carrying the truncating mutation p.(Arg503*), had rather mild speech delay and impairment compared with all other affected individuals and particularly compared with the index patient and affected father from family 7 with the same mutation. So far, no genotype–phenotype correlations can be deduced regarding nature or localisation of known mutations. If and to which extent the autism spectrum disorder or autistic features or tall stature observed in three individuals, respectively, might be associated with the FOXP2 mutation or with other, unrelated causes, is currently unclear.

Studies in humans and other species indicate that FOXP2 is a conserved transcription factor with important roles in the developing brain, in neurogenesis, neurite outgrowth and synaptic plasticity.27–31 An implication of FOXP2 specifically in language and speech and in more global or other neuronal functioning might be reflected in the observation that most of the herewith reported individuals as well as previously published patients3 ,15 show dyspraxia and other speech and language phenotypes and often impairment of non-verbal abilities and that all but one of the herewith reported individuals show variable degrees of cognitive impairment. However, in families 2 and 6, additional factors might have contributed to the cognitive impairment in the affected individuals, as also siblings not carrying the FOXP2 mutation were reported to have speech and/or developmental delay or behavioural anomalies, however, milder and without motor speech impairment.

Our findings considerably expand the genotypic spectrum of aberrations in FOXP2 and confirm that the associated phenotypes may include significant speech and language defects and mild neurodevelopmental and cognitive deficits.


The authors are grateful to the participating individuals and their families. They also thank Christine Suchy for excellent technical assistance.



  • Contributors MSR, AR, UM, TAB, KEC, AR, MS, KS, DG, PJ, DDD Study, MK, HR, US-M, PB, SB-W, JK, AR and CZ provided clinical and mutational data. MSR and CZ wrote the manuscript, AR, UM, TAB, KEC, AR, MS and KS wrote the respective case reports. All coauthors read and agreed with the text.

  • Funding CZ was supported by a grant from the German Research Foundation (DFG, ZW184/1-2) and by the IZKF (Interdisziplinäres Zentrum für Klinische Forschung, E26) Erlangen. TAB acknowledges funding from the National Institute of Health Research, UK. The DDD study presents independent research commissioned by the Health Innovation Challenge Fund (grant number HICF-1009-003), a parallel funding partnership between the Wellcome Trust and the Department of Health and the Wellcome Trust Sanger Institute (grant number WT098051). The views expressed in this publication are those of the author(s) and not necessarily those of the Wellcome Trust or the Department of Health. The study has UK Research Ethics Committee approval (10/H0305/83, granted by the Cambridge South REC, and GEN/284/12 granted by the Republic of Ireland REC). The research team acknowledges the support of the National Institute for Health Research, through the Comprehensive Clinical Research Network.

  • Competing interests None declared.

  • Ethics approval Institutional review boards of the Medical Faculty of the Friedrich-Alexander-University Erlangen-Nürnberg.

  • Provenance and peer review Not commissioned; externally peer reviewed.