Microdeletion 20p12.3 involving BMP2 contributes to syndromic forms of cleft palate

Abstract

Orofacial clefts of the lip and/or palate comprise one of the most common craniofacial birth defects in humans. Though a majority of cleft lip and/or cleft palate (CL/P) occurs as isolated congenital anomalies, there exist a large number of Mendelian disorders in which orofacial clefting is part of the clinical phenotype. Here we report on two individuals and one multi-generational family with microdeletions at 20p12.3 that include the bone morphogenetic protein 2 (BMP2) gene. In two propositi the deletion was almost identical at ~600 kb in size, and BMP2 was the only gene deleted; the third case had a ~5.5-Mb deletion (20p13p12.2) that encompassed at least 20 genes including BMP2. Clinical features were significant for cleft palate and facial dysmorphism in all three patients, including Pierre-Robin sequence in two. Microdeletion 20p13p12 involving BMP2 is rare and has been implicated in Wolff-Parkinson-White (WPW) syndrome with neurocognitive deficits and with Alagille syndrome when the deletion includes the neighboring JAG1 gene in addition to BMP2. Despite a significant role for the BMPs in orofacial development, heterozygous loss of BMP2 has not been previously reported in patients with syndromic clefting defects. Because BMP2 was the sole deleted gene in Patients 1 and 2 and one of the genes deleted in Patient 3, all of whom had clinical features in common, we suggest that haploinsufficiency for BMP2 is a crucial event that predisposes to cleft palate and additional anomalies. Lack of significant phenotypic components in family members of Patient 1, suggests variable expressivity for the phenotype.

INTRODUCTION

Cleft lip and/or palate (CL/P) comprise the most common congenital craniofacial anomalies in humans; over two-hundred syndromes have CL/P as a feature(s) [Gorlin et al., 2001]. The etiology is complex with genetic and environmental factors playing significant roles. With a prevalence of >1/1000 live births, CL/P contributes significantly to physical and psychosocial morbidity in those affected. High familial recurrence and observed concordance rates in monozygotic twins suggest a strong genetic component in the etiology of orofacial clefting.

Normal development of the palate is complex. It requires synergy between growth and patterning of different facial primordia. Developmental processes of the vertebrate face are a well-orchestrated interplay of complex genetic cascades that involves a number of crucial developmental genes such as transforming growth factor-b family and their receptors (TGF-β, TGFBR), including members of the bone morphogenetic protein (BMP) family; fibroblast growth factors and their receptors (FGFs and FGFRs); and T-box transcription factors (TBX) [Schutte and Murray, 1999; Spritz, 2001; Stanier and Moore, 2004; Jugessur et al., 2009].

The TGF-β superfamily of growth factors regulates many components of skeletal development, including cartilage and bone formation, mesoderm patterning, and craniofacial and limb development [Zhang et al., 2002; Wan and Cao, 2005; Meng et al., 2009]. BMPs are members of the TGF- β superfamily and important participants in craniofacial development. Binding to and functioning via BMP receptors, BMPRI and BMPRII, they result in phosphorylation-dependent activation of the SMAD proteins and downstream signaling cascades. In addition to SMAD proteins, BMPs activate multiple kinase pathways (MAPK, PI3 kinase, etc.) [Kishigami and Mishina, 2005]. Starting from early stages of neural crest formation and until development of the complete orofacial musculoskeletal system, BMPs, in concert with other genes, are prominently expressed in the facial primordia and palate. Another significant interaction is that between BMPs and homeobox gene MSX1on 4p16.3. MSX1 is required for BMP4 and BMP2 expression. Msx1 deficiency results in cleft secondary palate and tooth agenesis in mice that can be rescued by Bmp4. Similarly, MSX deficiency causes nonsyndromic cleft palate with tooth agenesis in humans [van den Boogaard et al., 2000; Zhang et al., 2002]. Interestingly, MSX1 appears to be an important genetic element in 4p16.3 microdeletions that result in Wolf-Hirschhorn syndrome [Nieminen et al., 2003]. Additional evidence linking orofacial clefting to members of the Fgf family and their receptors has been obtained from studies in transgenic mice [Abu-Issa et al., 2002; Nie et al., 2006].

Evidence from knockouts of conserved homologs in mouse and drosophila [Dudas et al., 2004; Liu et al., 2005; Le Gloan et al., 2008; Lalani et al., 2009] suggest a critical role in skeletal, cardiac and neural development. Expression of Bmp2–5 in palatal epithelia and mesenchyme is crucial during multiple stages of palatogenesis [Lu et al., 2000; Nie, 2005]. Because BMP signaling is also conserved in developing craniofacial skeleton, changes in expression result in alterations of these structures [Dudas et al., 2004; Nie et al., 2006]. In addition, a significant role for Bmps in cardiovascular development is derived from mouse models deficient in Bmp2 [Zhang and Bradley, 1996; Ma et al., 2005] signaling.

Although no deletions of BMP2 have been reported in individuals with isolated or syndromic CL/P, variably sized microdeletions within 20p12 that include BMP2 have been proposed to result in diverse phenotypes including Alagille (AGS), hypoplastic left heart and Wolff-Parkinson-White (WPW) syndromes, a bypass reentrant tachycardia that results from an abnormal connection between the atria and ventricles [Robert et al., 2007; Le Gloan et al., 2008; Lalani et al., 2009]. Inclusion of neighboring gene JAG1 in one case resulted in a complex phenotype including Alagille and WPW syndrome [Le Gloan et al., 2008]. Deletions involving JAG1 within the AGS critical region have been reported to result in features of AGS and developmental delay, autism, scoliosis, and bifid uvula [Kamath et al., 2009].

Here we report the clinical and molecular characterization of two individuals and one multi-generational family with microdeletions within chromosomal region 20p13p12.

METHODS

Patient ascertainment

Study Patients 1–3 in this study were referred for microarray testing to Signature Genomic Laboratories for physical and intellectual disabilities and/or dysmorphic features. An Institutional Review Board-approved informed consent was obtained from the appropriate parents or legal guardians of the patients to participate in this study.

Molecular analysis

A 105K-feature whole-genome oligonucleotide microarray (SignatureChipOS v1.0, custom-designed by Signature Genomics, Spokane, WA, manufactured by Agilent Technologies, Santa Clara, CA) was used for microarray analysis of DNA from study Patient 1 according to previously published methods [Ballif et al., 2008b]. A 135K-feature whole-genome oligonucleotide microarray (SignatureChipOS v2.0, custom-designed by Signature Genomics, manufactured by Roche NimbleGen, Madison, WI; coordinates based on UCSC Genome Browser build NCBI Build 36.1/hg18) was used for microarray analysis of DNA from Patient 2 according to previously published methods [Duker et al., 2010]. DNA from study Patient 3 was initially tested using a >4,600-clone whole-genome bacterial artificial chromosome (BAC)-based microarray (SignatureChipWG, Signature Genomics) using previously described methods [Ballif et al., 2008a]. Additional microarray analysis was performed for study Patient 3 using the 105K-feature oligonucleotide microarray to characterize the deletion breakpoints in more detail.

Deletions in the three probands and extended family members of Patient 1 were confirmed and visualized by fluorescence in situ hybridization (FISH) with BAC clones from the deletion regions using previously published methods [Shaffer et al., 1994].

RESULTS

Clinical summaries

Patient 1, the proposita in the multi-generational family, is a Caucasian female first referred at 13 days of age for evaluation of cleft palate (Fig 1A, B; Study Patient IV-1, Supplementary Fig 1). She was born at 39 ½ weeks to a 22-year-old G1P0 mother. On physical examination at age 13 days her height was 45 cm (<5^th centile), weight was 2.60 kg (<5^th centile), and head circumference was 33.3 cm (<5^th centile). Besides the cleft palate, micrognathia and glossoptosis suggestive of Pierre Robin sequence, she also had large communicating fontanelles, a long well-formed philtrum, pectus excavatum, diastasis recti, gap between her halluces and second toes with vertical creases, deep palmar flexion creases and short 5th fingers. On examination at age 3 months and 3 weeks, the posterior fontanelle was still open and the anterior fontanelle measured 5 cm × 5 cm. She has a patent foramen ovale. An electrocardiogram was normal.

Figure 1.

Facial profiles for the three patients. (A, B) Patient 1 at 13 days and 9.5 months. Note cleft palate in frame (A); micrognathia, long philtrum in frame (B). (C, D) Patient 2 at 7 years. Note deep cleft palate (frame C); micrognathia, long philtrum (frame D). (E, F) Patient 3 at 3 months and 19 months. Note flat facial profile, downslanting palpebral fissures, depressed nasal bridge, small, upturned nose with anteverted nares, long philtrum and micrognathia.

The family history was significant for a mother (III-1, Supplementary Fig 1) with trochlear nerve palsy (CNP); a high palate; small congenital lumbosacral hemangioma; and hands with short 5th fingers and some unusual dermatoglyphic patterns (Table I). Her height was 172.3 cm and her OFC 54.8 cm. The maternal grandmother (II-1, Supplementary Fig 1) measured 154.5 cm (5^th –10^th centile), and had trochlear nerve palsy, high palate, white spots in the irises similar to Brushfield spots, and hands with hard-to-see ridges and short 5th finger. The maternal great-grandmother (I-1, Supplementary Fig 1) had short stature (137.1 cm, <2^nd centile), high palate, trochlear nerve palsy, and hands with single palmar flexion crease (Supplementary Fig 1). There were no other significant physical findings or dysmorphism in these extended family members.

Table I.

Clinical characterization of individuals with microdeletions at 20p12.3 and comparison with Patients 1–3 reported by Lalani et al.,

	Patient IV-1	Patient III-1	Patient II-1	Patient I-1	Patient 2	Patient 3	Lalani Patient 1	Lalani Patient 2	Lalani Patient 3
Age	3 mo				7 y	2y	20 mo	5 y	37 y
Sex	F	F	F	F	F	F	M	M	F
Deletion size	592.7 kb (chr20: 6,222,266–6,814,990)	592.7 kb (chr20: 6,222,266–6,814,990)	592.7 kb (chr20: 6,222,266–6,814,990)	592.7 kb (chr20: 6,222,266–6,814,990)	566.4 kb (chr20: 6,265,253–6,831,640)	5.37 Mb (chr20: chr20: 3,672,605–9,042,183)	1.1 Mb	2.3 Mb	2.3 Mb
Inheritance	Maternal	Maternal	Maternal	Unknown	Unknown	de novo	de novo	Maternal	de novo
Height	53.3 cm (<5th centile)	172.7 cm	154.9 cm (5th–10th centile)	137.2 cm	114.6 cm (10th centile)	10th centile	−2 SD	25th centile	N/A
Weight	4.17 kg (<5th centile)	N/A	N/A	N/A	19.4 kg (10th–25th centile)	<3rd centile	−2 SD	75th centile
Head circumference	38.0 cm (5th–10th centile)	54.8 cm	N/A	N/A	48 cm (<5th centile)	40.3 cm (50th–75th centile)	25th centile	+2 SD	+2.25 SD
Facial dysmorphism	Micrognathia; long philtrum	N/A	Brushfield’s spots	N/A	Large eyes; synophrys; long philtrum; microstomia	Flat facial profile; downslanting palpebral fissures; low nasal bridge; small, upturned nose with anteverted nares; pinpoint hemangioma on tip in middle of nose; long philtrum; transversecrease across chin; micrognathia	Downslanting palpebral fissures; bilateral epicanthal folds; broad nasal root and bridge; malar hypoplasia; full cheeks with microstomia	Frontal upsweep; downslanting palpebral fissures; epicanthal folds; hypertelorism; long philtrum; microstomia; small ears with thickened helices	Hypertelorism; malar hypoplasia
Craniofacial	Cleft palate; open posterior fontanelle; large anterior fontanelle	High palate	High palate	High palate	Submucous cleft palate; bifid uvula, high arched hard palate	Long U-shaped midline cleft of soft palate	N/A	Macrocephaly	Macrocephaly
Heart	Patent foramen ovale	−	−	−			ASD; WPW	WPW	−
Psychomotor and behavioral development	−	−	−	−	Problems with reading comprehension repeated second grade	−	Neurocognitive delay	Neurocognitive delay; motor delay	Neurocognitive delay; motor delay
Neurology	NA	4th central nerve palsy	4th central nerve palsy	4th central nerve palsy	NA	Feeding difficulties; significant head lag and wobble; delayed/poor reflexes; significant central hypotonia	−	−	−
Hearing loss	−	−	−	−	+	Failed newborn hearing exam	N/A	N/A	N/A
Hands/feet	Deep palmar flexion creases, short 5th fingers with mild mesobrachydactyly bilaterally, clinodactyly	Hands with short 5th fingers; central pocket loop in the left 3rd finger; tendency to central pocket radial loop in the right 3rd finger; and to central pocket loop in the right 5th finger	Hands with hard-to-define ridges and short 5th fingers with evident mesobrachydactyly	Hands with tendency to single palmar flexion crease transitional 2 and single flexion crease on both 5th fingers	Zygodactylous triradius between the second and third right toes		Persistent fetal pads	Broad thumbs/toes with persistent fetal pads	Persistent fetal pads
Other	GERD	Small congenital lumbosacral hemangioma	−	−	Cholesteatoma	Lacrimal duct stenosis	−	−	−

Patient 2 is a 7-year-old female who was first referred for re-evaluation of cleft palate, deafness and cholesteatoma. She was born at 37 weeks to a 24-year-old mother. Birth weight was 2.84 kg (25^th–50^th centile). Birth length was 45.1 cm (< 10^th centile). Early childhood medical history was significant for bilateral cholesteatoma and chronic middle ear fluid. She underwent surgical removal of the cholesteatomas followed by prosthetic reconstruction of inner ear bones. The surgery improved her hearing, but she requires hearing aids bilaterally. The patient has had no major language deficits, although she is repeating the second grade owing to problems with reading comprehension. On exam at age 7 years her height was 114.6 cm (10^th centile), weight was 19.4 kg (10^th–25^th centile), and head circumference was 48 cm (< 5^th centile). She was noted to have submucous cleft palate, very high-arched hard palate, bifid uvula, large eyes, synophrys, long philtrum, upturned nose and microstomia. There were no significant limb anomalies. She was found to have a zygodactylous triradius between the 2nd and 3rd right toes (Fig 1C, D). She had no cardiovascular problems at primary examination or on continuing follow-up examinations.

Patient 3 is a 2-year-old female born at 41 weeks to a 30-year-old G1 P0–1 mother. Her birth weight was 3.15 kg (25^th centile) and length was 47 cm (10^th–25^th centile). She had a U-shaped cleft palate and micrognathia noted at birth. She failed her newborn hearing exam, although a repeat ABR was normal (Fig 1E, F). On physical exam at 3 months of age, her length was at the 10^th centile, weight was <3^rd centile, and head circumference was 40.3 cm (50^th–75^th centile). Dysmorphic facial features included flat facial profile, downslanting palpebral fissures, depressed nasal bridge, small, upturned nose with anteverted nares, pinpoint hemangioma on the tip of the nose, long philtrum, transverse crease across the chin, U-shaped midline cleft palate and micrognathia (Fig 1E,F). She had a prominent forehead with subtle transient vertical skin gyrata. During the first year of life she was noted to have significant motor deficits with a head lag and wobble, delayed/poor reflexes, central hypotonia and feeding difficulties. She had decreased muscle mass and subtle hyperextensibility of her hands and feet with recurrent hip dislocations and developmental dysplasia of the hips. Cardiovascular exam was normal. High-resolution chromosome analysis revealed a de novo and apparently balanced translocation t(3;5)(q10;q10)dn.

Molecular analysis

Microarray analysis identified a microdeletion within 20p13p12 in each of the three study patients. Patient 1 had a 592.7-kb deletion (chr20: 6,222,266–6,814,990, based on UCSC hg18 coordinates); Patient 2 had a 566.4-kb deletion (chr20: 6,265,253–6,831,640); and Patient 3 had a 5.37-Mb deletion at 20p13p12.2 (chr20:3,672,605–9,042,183) (Fig 2A–C). For Patient 3, high-resolution chromosome analysis had previously revealed a de novo balanced translocation [t(3;5)(q10;q10)dn]. We did not detect any copy number loss or gain at the putative breakpoints in chromosomes 3 and 5 within the limits of resolution of the array used, which has substantial coverage of the pericentromeric regions of chromosomes 3 and 5. The deletions in Patients 1 and 2 encompassed one known gene, BMP2. The deletion in Patient 3 encompassed at least 20 known genes (Fig 2D–F).

Figure 2.

Molecular cytogenetic characterization of microdeletions at 20p12.3. (A) In Patient 1, oligonucleotide microarray analysis identified a single-copy loss of 34 oligonucleotide probes from the short arm of chromosome 20 at 20p12.3, approximately 592.7 kb in size (chr20: 6,222,266–6,814,990; UCSC Genome Browser coordinates, build hg18, 2006). (B) In Patient 2, oligonucleotide microarray analysis identified a single-copy loss of 38 oligonucleotide probes from the short arm of chromosome 20 at 20p12.3, approximately 566.4 kb in size (chr20: 6,265,253–6,831,640). (C) In Patient 3 oligonucleotide microarray analysis identified a single-copy loss of 266 probes at 20p13p12.2, approximately 5.37 Mb in size (chr20: chr20:3,672,605–9,042,183). (D) Chromosome 20 ideogram for segment p13->p12.2.

(E) Schematic of microdeletions at 20p13p12.3 identified in study Patients 1–3, and compared to Patients 1–5 reported by Lalani et al. [2009]. Green bars represent the deleted interval in each of the current and previously reported cases. (F) The blue vertical bars represent genes within the interval. BMP2 highlighted by red circle.

The deletions in Patients 1 and 2 were confirmed by FISH using BAC probe RP11-184L8, and the deletion in Patient 3 was confirmed by FISH using BAC probes RP11-481K14 and RP11-177P14 (Supplementary Fig 2). FISH analysis was performed on PHA-stimulated peripheral blood lymphoblasts from the mother, maternal grandmother and maternal great-grandmother for Patient 1, and all three individuals were identified to have an apparently identical deletion. FISH on parental samples for Patient 3 showed the deletion to be apparently de novo in origin. Parental samples were unavailable for Patient 2. In addition, no CNVs in the deleted region were identified in the Database of Genomic Variants (http://projects.tcag.ca/variation/), suggesting that there are no significant copy-number polymorphisms in this genomic region.

DISCUSSION

CL/P is one of the most common orofacial anomalies occurring as isolated defect(s) or part of complex syndromic disorders. However, identifying monogenic causes for CL/P has been of limited success. The propositi reported on here share some significant clinical features, most notably cleft palate and variable facial dysmorphism including long philtrum and micrognathia. Short stature and developmental delay were variable and not seen in all the patients; microcephaly was obvious in only Patients 1 and 2. In addition, Patients 1 and 3 have features of Pierre Robin sequence. The overlapping phenotypes in Patients 1 and 2, both of whom have a deletion that encompasses only one gene, BMP2, support a causative relationship between BMP2 and syndromic cleft palate. The third patient had a ~5.5-Mb deletion that included 20 OMIM annotated genes including BMP2. The multiple genes involved in the deletion in this case may contribute to the additional dysmorphism. The identification in Patient 3 of an apparently balanced t(3;5)(q10;q10), although unlikely, cannot definitively rule out the possibility of the disruption of one or more genes at the translocation breakpoints. Nonetheless, the three patients reported here, Patient 1 and 2 in particular, provide novel evidence in support of the hypothesis that hemizygosity for BMP2 might be one important event in the pathogenesis of syndromic cleft palate. In addition, a clefting deformity, cleft lip, has been reported for one patient with a 10.7-Mb deletion at 20p12p13 that included the BMP2 gene [Lalani et al., 2009]. The coincidental finding of congenital trochlear nerve palsy in Patients III-1, II-1 and I-1 was intriguing; there was no history of head trauma noted for any of the Patients (Supplementary Fig 1). Any clear correlation of the trochlear nerve palsy to the molecular defect and cleft palate remains remote.

A causative association between haploinsufficiency for BMP2 and WPW syndrome has recently been suggested from studies in five patients harboring deletions at 20p12 and dysmorphic features and variable neurocognitive deficits [Le Gloan et al., 2008; Lalani et al., 2009]. The four propositi and one parent reported by Lalani et al. had deletions ranging in size from 1.1 Mb to 10.7 Mb. Patient 1 in that study had a 1.1-Mb deletion that encompassed BMP2 but no other genes; this patient had, in addition to WPW, motor and neurocognitive delays and dysmorphic features including downslanting palpebral fissures, bilateral epicanthal folds, broad nasal root and bridge, malar hypoplasia, full cheeks with microstomia, pectus deformity with a carinatum appearance superiorly and excavatum appearance inferiorly, and persistent fetal pads. Some of these features were shared with the patients described in the present study. Interestingly, none of the patients were reported to have had cleft palate [Le Gloan et al., 2008; Lalani et al., 2009], although maxillary hypoplasia was reported in three patients and cleft lip was noted in one patient. Cardiovascular problems were not evident in any of the patients reported here, including normal ECG in Patients 1 and 3. Of note, detailed cardiac electrophysiological studies in a Bmp2+/− mouse model did not recapitulate any conduction defects or dysrhythmias [Lalani et al., 2009]. [Lalani et al. 2009] suggest the neurocognitive and musculoskeletal phenotypes are fully penetrant. However, our results suggest variable expressivity of the musculoskeletal phenotype because short stature was present only in Patient 1 and her maternal great-grandmother. In addition, our results suggest BMP2 may not be associated with neurocognitive delays, because only one of our patients had developmental delay. Mutations of BMP2 may also show incomplete penetrance for clefting anomalies, as three members of the multi-generational family had a milder form, high palate, and none of the patients reported by Lalani et al. [2009] were reported to have clefting anomalies. The clefting abnormality in the three propositi may result from a combination of BMP2 haploinsufficiency and additional genetic or non-genetic factors.

Studies in mice suggest endogenous Bmp2 and Bmp6 cooperatively play significant roles in bone formation under both physiological and pathological conditions. Additionally, Bmp2 and Bmp6 are the main subtypes expressed in hypertrophic chondrocytes that induce endochondral bone formation. Fetal and adult compound-deficient mice (Bmp2 +/−; Bmp6 −/− showed a reduction in the trabecular bone volume with suppressed bone formation with normal bone resorption. However, the single deficient mice (Bmp2+/− or Bmp6−/−) did not show the same effects [Kugimiya et al., 2005]. Therefore, BMP2 haploinsufficiency in these patients might be contributory to growth deficiency, and short stature might be an indirect outcome.

Although haploinsufficiency of BMP2 may be solely responsible for the phenotypes evident for the patients described above, mutational events might be contributory. Nonetheless, based on the phenotypes of our patients and previous expression studies of the mouse homolog, hemizygous loss of BMP2 likely contributes to clefting abnormalities. There are few studies using microarray analysis to identify genomic copy number losses and gains in syndromic and nonsyndromic CL/P. Osoegawa et al. [2008] reported whole-genome microarray analysis in 83 syndromic (including 20 with Van der Woude syndrome) and 103 nonsyndromic CL/P cases with an abnormality detection rate of ~1%. One syndromic and two nonsyndromic cases had pathogenic microdeletions suggestive of a contiguous gene deletion syndrome. Our report demonstrates the utility of comprehensive genomic analysis in the genetic diagnosis of CL/P.

The highly conserved BMP signaling pathway and its functional and regulatory interactions with multiple growth factors and pathways deems it reasonable to speculate that perturbation of the above pathways would results in orofacial and skeletal anomalies. The three cases with cleft palate and additional anomalies presented here support a causative relationship between BMP2 and syndromic forms of cleft palate. Additional cases with similar microdeletion and mutational analysis of BMPs in clefting anomalies with or without short stature will further delineate this genotype-phenotype correlation.