FURTHER DELINEATION OF THE AUTS2 HX REPEAT DOMAIN-RELATED PHENOTYPE

Abstract

Haploinsufficiency of AUTS2 is associated with a neurodevelopmental disorder characterized by intellectual disability, autistic features, and spasticity. AUTS2 protein interacts with p300, encoded by EP300, through the HX repeat domain of AUTS2, thereby activating transcription. We previously reported two de novo variants in the HX repeat domain of AUTS2. These variants disrupt the AUTS2-P300 interaction, resulting in a phenotype resembling Rubinstein-Taybi Syndrome (RSTS) associated with variants in EP300/CREBBP. Here, we expand beyond the initial clinical description to delineate the HX domain-associated phenotype and compare it to the AUTS2-haploinsufficient phenotype. We reviewed clinical data, photographs, and neuroimaging studies to examine genotype-phenotype relationships. Our review of 80 individuals included 14 individuals we present here and 66 individuals with AUTS2 variants presented in the literature. The clinical features for individuals with variants in the HX repeat domain include severe intellectual disability, severe language disability, distinct craniofacial and skeletal dysmorphic features, and neuroimaging findings. Facial dysmorphisms include wide and prominent nasal bridges with complex nasal shapes and dysmorphic eyebrows. Dysmorphisms include digit anomalies: symphalangism and hypoplasia of distal phalanges, exclusive to the HX domain variant group. Cerebellar anomalies not seen with other AUTS2 variants are seen within this group. Our report delineates a distinct and severe clinical phenotype associated with variants in the AUTS2 HX domain, including an in-depth comparison with the AUTS2 haploinsufficiency phenotype features.

INTRODUCTION

Activator of transcription and developmental regulator AUTS2 (HGNC:14262) spans 1.2 Mb on chromosome 7q11.22 and encodes 19 exons that produces a full-length isoform and at least one smaller C-terminal isoform in humans.^1–4 The AUTS2-related neurodevelopmental disorder (OMIM #615834) is mainly associated with heterozygous single or multiple exonic deletions as well as structural variations and a variable phenotype associated with haploinsufficiency.⁵ Clinical findings include prominent behavioral phenotype, mild to moderate intellectual disability (ID), microcephaly, cerebral palsy, facial dysmorphic features, and variable autistic features.^5–7 AUTS2 is implicated in several neurological and non-neurological disorders beyond the neurodevelopmental phenotype.^3,4,8 It has a wide array of functions including transcriptional activation, RNA metabolism regulation, actin cytoskeleton organization, neuronal migration, and synaptogenesis in the developing human brain.^1–4

Polycomb repressive complexes (PRC) are large protein complexes that regulate gene expression by modifying histones.⁴ Full-length human AUTS2 binds directly to the PRC1.3 and PRC1.5 subgroups of mammalian PRC1 that include Polycomb group proteins PCGF3/5, RING1A/B, RYBP or its homolog YAF2, and the heterotetrameric holoenzyme Casein Kinase 2 (CK2).^4,9 PRC1 acts canonically to repress transcription.⁴ However, PRC1 binding to CK2 represses this inhibitory action while recruitment of p300 (encoded by EP300) by AUTS2, turns the complex into a transcriptional activator.⁴ Recent work showed that p300 is recruited to PRC1 by the HX repeat domain in exon 9 of AUTS2.⁹ The HX repeat domain is a short and highly conserved histidine-rich motif with repeating sequences of histidine (H) and another amino acid (X).⁹ The specific HX repeat domain of AUTS2 encodes a HQHQHQHQHTHQHTHQHT amino acid sequence at positions 525 to 542.⁹

Several PRC1 subunits, including AUTS2, CK2 components CSNK2A1 and CSNK2B, RING1, and YAF2, are reported as causative or strong candidates for neurodevelopmental disorders. FBRSL1, which shares high sequence homology with AUTS2, interacts with PRC1.3 and PRC1.5 and causes a neurodevelopmental disorder.^10,11 The human EP300 gene and its structural and functional twin CREBBP are associated with Rubinstein-Taybi syndrome (RSTS) (OMIM #180849), a neurodevelopmental condition characterized by recognizable facial dysmorphisms, broad thumbs and halluces, short stature, and ID. Variants in the HX repeat in exon 9 in AUTS2, a missense (p.Thr534Pro) and a small in-frame deletion (p.His535_Thr542del), disrupt binding to p300 in PRC1 and produce a phenotype that overlaps with classic RSTS.⁹ This is striking as the phenotype associated with these variants appears to be more severe than the syndrome associated with the AUTS2 haploinsufficiency.⁹ However, the initial report that delineated these findings included scant clinical detail, necessitating the identification of additional patients and description of clinical features.

The interference of the missense and small in-frame deletions of the HX-repeat with protein function emphasizes the importance of accurately classifying such variants that may be overlooked in genes that are associated with loss-of-function. Pathogenicity of AUTS2 variants in the context of small amino acid changes has not yet been clarified, and patient findings have not been adequately detailed to aid clinical decision-making and correct grouping of patients. Additionally, nonsense variants causing haploinsufficiency are reported less frequently, while most of the AUTS2 variants reported so far are associated with exon deletions or structural variations.

Here, we present a more comprehensive delineation of the severe AUTS2-related phenotype associated with missense and in-frame variants in the HX repeat domain in exon 9, initially defined by Liu et al.⁹ Our analysis highlights the differences between the haploinsufficiency-related phenotype and the HX repeat-related phenotype -the latter referred to as AUTS2/HX-NDD for distinction. Additionally, we expand the phenotypes associated with variants causing haploinsufficiency by reporting deletion, nonsense, and splice variants across the gene. The HX-repeat domain variants produce structural brain defects, especially cerebellar malformations, not seen with other variants in AUTS2. Additional skeletal and facial dysmorphic features are first described here in detail. The analysis concludes with a review of the pathogenicity of AUTS2 variants and regional constraint analysis.

MATERIALS AND METHODS

A total of 80 individuals with variants in AUTS2 (RefSeq: NM_015570.4) were ascertained through national and international collaborations and the literature. Ten individuals were included in the AUTS2/HX-NDD group: five individuals (LR05–007, LR15–003, LR18– 404, LR19–314, LR19– 506) were briefly discussed by Liu et al.⁹ Another individual (LR23–017) was studied in a brain organoid system by Fair et al.¹², with limited clinical details presented. Additionally, two new patients are reported in this article (LR- CG-1, LR23–023). These eight Individuals with HX repeat variants were analyzed with two other previously published case reports describing patients with variants in the AUTS2 HX repeat domain.^13,14

Seventy individuals were included in the AUTS2-haploinsufficient variant group: two individuals (LR15–004, LR15–097) discussed by Liu et al.⁹; four new patients reported in this article (LR19–321, LR19–315, LR15–081, LR-ER-2) and 64 additional cases identified after literature review with heterozygous variants and sufficient phenotypic details to describe the haploinsufficiency related phenotype as compared to the HX-repeat related phenotype.^{2,5–7,15–30} Clinical details for LR15–004 were excluded from the main analysis and discussed separately in the supplement, due to the variant’s presence in All of Us and gnomAD databases.^31,32 Her clinical data was included in Table 1 for a complete update to the AUTS2-related phenotypes as she did not have any other diagnosis to explain her clinical presentation.

Table 1.

Tabulated clinical information of main cohort reported under 5 main categories: growth parameters, neurodevelopment, brain imaging features, dysmorphic features, skeletal abnormalities.

	LR-ER-2	LR19–321	LR19–315	LR15–097	LR15–004	LR-CG-1	LR23–023	LR05–007	LR23–017	LR15–003	LR18–404	LR19–314	LR19–506	LR15–081
Type of AUTS2 aberration	Deletion	Splice	Nonsense	Nonsense	Missense	Missense	Missense	Missense	Missense	Indel	Indel	Indel	Indel	Frameshift
Exon(s) affected (N=19)	5	6	7	9	9	9	9	9	9	9	9	9	9	19
NM_015570.4	c.(660+1_661–1)_(690+1_691–1)del	c.742+1G>A	c.1018C>T	c.1483C>T	c.1550C>T	c.1586A>C	c.1597C>T	c.1600A>C	c.1600A>C	c.1603_1626del	c.1603_1626del	c.1603_1626del	c.1603_1626del	c.2992_3002del
Protein change	p.(Cys221_Lys230)del	NA	p.Gln340Ter	p.Arg495Ter	p.Pro517Leu	p.His529Pro	p.His533Tyr	p.Thr534Pro	p.Thr534Pro	p.His535_Thr542del	p.His535_Thr542del	p.His535_Thr542del	p.His535_Thr542del	Gln998Glyfs*82
Age at last exam	3y 9mo	11y 3mo	2y 7mo	12y 7mo	14y	11y	28y	15y 6mo	10y	18y	5y 6mo	8y	16y	14y
	paternally inherited	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo	de novo
Variant Classification
ACMG Rule Codes	PVS1_Mod	PVS1_Strong, PS2_Strong, PM2_Mod	PVS1_Strong, PS2_Strong, PM2_Mod	PVS1_Strong, PS2_Strong, PM2_Mod	PP3_Mod, BP1_Supp	PS2_Strong, PM2_Mod, PM5_Mod	PS2_Strong, PM2_Mod	PS2_Strong, PM2_Mod	PS2_Strong, PM2_Mod	PS2_Strong, PS4_Mod, PM2_Mod	PS2_Strong, PS4_Mod, PM2_Mod	PS2_Strong, PS4_Mod, PM2_Mod	PS2_Strong, PS4_Mod, PM2_Mod	PS2_Strong, PVS1_Mod, PM2_Mod
ACMG Classification	VUS	Pathogenic	Pathogenic	Pathogenic	VUS	Likely Pathogenic	Likely Pathogenic	Likely Pathogenic	Likely Pathogenic	Pathogenic	Pathogenic	Pathogenic	Pathogenic	Pathogenic
CADD Score	unknown	35	36	42	24.2	29.6	25.8	25	25	20.3	20.3	20.3	20.3	4.6
AlphaMissense Score	na	na	na	na	0.97: pathogenic	0.99: pathogenic	0.99: pathogenic	0.74: pathogenic	0.74: pathogenic	na	na	na	na	na
Growth parameters
Low birth weight	0	0	1	0	0	0	0	0	0	0	0	0	0	nd
Short Stature <3%	0	nd	0	0	0	0	1	1	1	0	0	0	0	1
Microcephaly <3%	0	0	1	1	0	0	0	0	1	0	0	0	0	0
Feeding difficulties (infancy)	1	1	1	1	1	unk	1	1	1	1	1	1	1	1
Notes	None	Birth Weight: Low-normal at 2550 grams. Head size was at 3% at birth and at the low-normal range at the last exam.	Birth Weight: 2485 grams. Height was at 9th percentile at last exam.	None	None	Large anterior fontanelle at birth.	Hypotonic at birth and developed slowly thereafter.	Height was 151 cm at the last exam.	None	None	The feeding difficulties were mild.	None	Temperature dysregulation in infancy.	None
Neurodevelopment
Intellectual disability	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Developmental delay	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Speech-language disability	1	1	1	1	1	1	1	1	1	1	1	1	1	0
Motor delay	0	1	1	1	nd	1	1	1	1	1	1	1	1	1
Autistic features	1	1	1	1	1	0	1	1	1	1	nd	1	1	1
Hypotonia	unk	1	1	1	0	1	1	1	1	1	1	1	0	0
Age of onset Epilepsy	na	na	na	5y	10y	na	na	na	4y	12 to 18 years	na	na	10mo	na
Epilepsy	0	0	0	1	1	0	0	0	1	1	0	0	1	0
Spasticity	0	1	1	1	0	0	1	1	0	0	0	0	0	0
ADHD	0	nd	0	0	1	1	0	0	nd	0	0	1	0	1
Behavioral issues	1	1	1	1	0	0	1	1	1	1	1	1	nd	1
Notes	Impulsive behavior when frustated.	No other remarkable anomaly noted.	Noise sensitivity and long awake phases at night.	No other remarkable anomaly noted.	Severe epileptic encephalopathy.	Poor balance. Severe speech delay, a single word at 3 years.	Non-verbal. Behavioral issues are not prominent beyond the ones related to autism.	Non-verbal. Two febrile seizures.	Severe speech-language disabilities.	Non-verbal.	Severe speech-language disabilities.	Severe speech-language disabilities. A febrile seizure at an unknown age.	Severe speech-language disabilities.	Language is mildly behind. Occasional self isolating behaviors.
Brain imaging features
Structural brain anomaly	0	0	0	0	0	0	1	1	1	1	1	1	1	0
Corpus callosum malformation	0	0	0	0	0	0	1	1	1	0	1	1	0	0
Cerebellar malformation	0	0	0	0	0	0	nd	1	1	1	1	1	1	0
Small posterior fossa	0	0	0	0	0	0	nd	1	1	1	1	1	0	0
Other	FLAIR hyperintesity: occipital gliosis / artefact	Normal MRI.	White matter thinning, isolated small gliosis.	Normal MRI.	Normal MRI.	Normal MRI.	MRI was done in 1996 and destroyed 20 years after. The report commented on corpus callosum.	Small brainstem.	Concerning for brachycephaly. Underdeveloped Sylvian Fissure. Mild ventriculomegaly at lateral and third ventricle.	Cerebellum and posterior fossa borderline small.	Small cerebellum is vermis predominant.	Posterior fossa is mildly small. Small cerebellum is vermis predominant.	A borderline small cerebellum and an arachnoid cyst in the posterior fossa.	Normal MRI
Dysmorphic features
Down slanting palpebral fissures	0	0	0	1	0	0	1	1	0	0	1	0	0	0
Short palpebral fissures	0	0	0	1	0	0	1	1	0	0	1	1	0	0
Telecanthus	0	0	1	0	0	1	1	1	0	0	1	1	1	0
Strabismus	0	0	1	0	0	1	0	0	1	1	1	0	nd	0
Eye closure with smile	unk	nd	0	nd	0	0	0	1	nd	0	0	0	nd	0
Synophrys	0	0	0	0	0	1	0	0	0	1	0	1	0	0
Horizontal eyebrows (mild) or low-set	1	1	0	0	0	1	1	1	1	1	1	1	1	0
Thick eyebrows	0	1	0	0	1	0	0	0	0	1	0	0	0	0
Eyebrows sparse (total or partly)	1	0	0	1	0	1	1	1	1	0	1	1	1	0
Prominent (high) nasal bridge	1	nd	0	0	0	1	1	1	1	1	1	0	0	0
Wide (broad) nasal bridge	0	0	1	0	0	1	1	1	1	1	1	0	1	0
Convex nasal ridge	0	nd	0	0	0	0	1	1	0	1	0	0	0	0
Anteverted nares	0	0	0	0	0	1	0	1	0	1	1	1	1	0
Low hanging columella	0	1	0	0	0	1	1	1	0	0	0	0	0	0
Short philtrum	1	1	1	1	0	0	0	0	1	1	1	0	0	1
Narrow (small) mouth	1	0	1	1	0	0	0	0	1	1	1	1	1	0
Thick vermillion of the upper lip or lower lip (Thick lips)	0	1	0	1	0	0	0	0	0	0	0	1	0	0
Small jaw	0	1	0	0	0	0	0	1	1	0	1	1	1	0
Ear malformation or low-set	0	0	0	0	0	1	1	1	1	1	1	1	1	0
Other	No other remarkable anomaly noted.	Overhanging nasal tip. Also, the distinction between low hanging and low instertion columella was not possible due to the low quality of the side image.	No other remarkable anomaly noted.	No other remarkable anomaly noted.	Mild maxillary hypoplasia.	Thin upper lip vermillion.	Mild blepharophimosis. Eyebrows sparse laterally. Medial flaring of eyebrows Nasal tip mildly low-hanging. Nose asymmetric and nasal ridge angulated to the left. Sharp upper nasal ridge. Thin upper lip vermillion. Wide philtrum. Mild maxillary hypoplasia. Mildly prominent jaw. Narrow shoulders.	Mild synophrys. Mildly overhanging nasal tip. Narrow nasal ridge. Thin lips. Blepharophimosis. Deep set eyes.	Narrow nasal ridge. Thin lips. Blepharophimosis. Deep set eyes.	Wide nasal base. Additional dysmorphologic features at earlier ages.	Short palpebral fissures: mild. Short philtrum: mild.	Hypoplastic columella. Long and smooth philtrum.	A right-sided ptosis and squint. Low anterior hairline. A short nose, narrow nasal ridge. Prominent earlobes. A long philtrum. Thin lips.	No other remarkable anomaly noted.
Skeletal abnormalities
Broad terminal phalanges (thumbs/halluces) (mild)	1	0	0	1	0	1	0	1	1	1	1	0	1	0
Kyphosis/scoliosis	unk	0	0	1	1	0	1	1	1	0	0	0	0	0
Tight heel cords	unk	1	1	1	0	unk	0	1	1	1	0	0	0	0
Angulated hallux	0	0	0	0	1	unk	0	0	0	1	0	0	0	0
Other	Bilateral 5th toe clinodactyly	Mild contractures at elbows and ankles.	Bilateral clubfeet.	Bilateral clubfeet.	No other remarkable anomaly noted.	5th finger clinodactyly bilaterally and proximally placed thumbs	Short and small middle and distal phalanges of L2, R2, R4.	Persistent toe walking.Bilateral patellar dislocation. Symphalangism.	Bilateral genu valgum and hip dysplasia. Probable symphalangism.	5th finger clinodactyly, fetal pads, short 4th metatarsals overlapping 4th and 5th toes on both feet.	Chiari I. Cranial skeletal anomalies. Pes planus.	Pes planus along with short fingers and clinodactyly of the 5th finger.	Over-riding 2nd toes.	No other remarkable anomaly noted.

nd: not determined.

na: not applicable.

1: present

0: absent

CADD:Combined Annotation Dependent Depletion Score

AlphaMissense: Algorithm-generated Score

VUS: variant of uncertain significance

Mod: Moderate, Supp: Supporting

Previously published patients are openly cross-referenced in Supplementary Table 2, and their updated and expanded clinical information with facial photographs and MRIs were included for a complete demonstration of the AUTS2 phenotypic spectrum along with the patients from the literature. Clinical details were obtained on all individuals through a standardized datasheet, including photographs and original neuroimaging files of the main cohort, and used to construct a database of features. Phenotyping was led by an experienced child neurologist and clinical geneticist (W.B.D.) based on the principles of Elements of Morphology.³³ A complete medical history for all AUTS2/HX-NDD patients is included in the Supplement in the form of a case series. Each individual was re-evaluated based on their clinical characteristics and AUTS2 variant.

Written informed consent was obtained from all participants through protocols approved by the Institutional Review Boards at Seattle Children’s Hospital or the local institutional review board at each site.

All variants were identified by clinical or research exome sequencing. One individual’s (LR-ER-2) variant was detected using exome based CNV analysis. Genomic data is presented in Table 1 for the main cohort and Supplementary Table 1 for patients from the literature. All variants were heterozygous. Our cohort included 13 individuals with de novo variants and one with paternal inheritance (LR-ER-2). Two AUTS2/HX-NDD cases from the literature were de novo, while the rest of the literature included 42 de novo, 2 paternal, 6 maternal cases. An additional 14 cases from the literature were of unknown inheritance, 4 of these being nonmaternal and 1 of these being nonpaternal. Variants reported here were verified using VariantValidator and are reported according to HGVS Nomenclature.^34,35

RESULTS

We detailed the phenotypic information of 13 unrelated individuals with heterozygous, de novo variants in AUTS2 (NM_015570.4), one with paternal inheritance (Table 1), and 66 from the literature.

A summary of the medical history of the index case (LR05–007) and the remaining AUTS2/HX-NDD cases are provided in the Supplementary Data, and their clinical data together with the four new AUTS2-haploinsufficient cases reported in this article and the case LR15–097 described by Liu⁹ are summarized in Table 1.

AUTS2 Haploinsufficiency

An analysis of the clinical features of 69 patients with AUTS2-haploinsufficient variants (LR15–004 is not included) revealed high frequencies of ID (96%) and autistic features (63%). These patients had mild to moderate ID and delayed language and motor development. They also presented with spasticity (32%) and ADHD (32%) which defined the main features of this phenotype. They had varying degrees of microcephaly (56%, usually mild) and facial dysmorphisms (81% of the patients had periorbital dysmorphology, 59% had nose and philtrum dysmorphology, 59% had lips, mouth and oral region dysmorphology, 44% had hand and feet dysmorphology, 35% had ear dysmorphology). (Figure 1, Supplementary Table 1) Additional clinical information for these individuals may be found in the supplementary.

Figure 1.

Tabulated clinical information of individuals with variants in the HX repeat (AUTS2/HX-NDD, including patients reported by Palumbo et al. [1 individual] and Martinez-Delgado et al. [1 individual]) compared to individuals with variants reported throughout the gene (AUTS2/HI, HI: haploinsufficiency, including the patients we report [5 individuals] and patients from the literature [64 individuals]). Colors on the heatmap denote present features, while black denotes absent, and gray denotes unknown features. Each box represents one individual. The frequency of certain features, such as dysmorphisms and structural brain malformations, differed between the two groups. Two groups are compared with a Fisher’s Exact test for each category, and p scores below <0.5 are shown in bold.

Included in this group are five individuals we present in detail (Table 1), four of them reported here for the first time, each carrying a heterozygous AUTS2 variant predicted to cause loss-of-function. This group demonstrated features typical of haploinsufficiency-related AUTS2-neurodevelopmental disorder. ID (moderate to severe in our cohort), autistic features, generalized hypotonia, feeding difficulties, spasticity, and variable but mild dysmorphic facial features were seen. Brain MRIs were all normal with no structural anomalies in the main group. Total frequency of structural brain defects including the literature was only 9%. 5% of those were corpus callosum malformations with nonspecific MRI changes without any cerebellar anomalies.

Exon 9: AUTS2/HX-NDD

This group of ten individuals had a distinct phenotype characterized by severe ID, striking dysmorphic facial features, and structural brain defects not seen with variants outside this domain (Figure 1). A wide and prominent nasal bridge with complex nasal shapes, down-slanting palpebral fissures, and dysmorphic eyebrows are observed (Figure 2). Dysmorphisms also include digit anomalies that are seen only in this group (Supplementary Figure 1). The brain malformations observed include hypoplasia of the cerebellum, corpus callosum, and posterior fossa making this phenotype consistently more severe than the phenotype observed in individuals with AUTS2-haploinsufficient variants (Figure 1 and 3). This group of patients consistently had severe language deficiencies. Three of the main group (LR23–023, LR05–007, LR15–003) never developed speech and the rest were severely delayed.

Figure 2:

Facial appearance in 10 individuals with AUTS2 variants. A,B:LR-CG-1 (p.His529Pro); C,D: LR23–023 (p.His533Tyr); E,F: LR05–007 (p.Thr534Pro); G,H: LR23–017 (p.Thr534Pro); I,J: LR15–003 (p.His535_Thr542del); K,L: LR18–404 (p.His535_Thr542del). Individuals with variants in the HX domain (A-L) all have striking dysmorphic facial features with deep-set eyes and abnormal eyebrows always prominent. M,N: LR19–321; O,P:LR19–315; Q,R: LR15–097; S,T: LR15–081. Minimal to no dysmorphic features were seen in these (M-T) individuals with variants outside of the HX repeat.

Figure 3.

Brain magnetic resonance images in 6 individuals with variants in the HX domain including LR05–007 (A-B), LR18–404 (C,D), LR23–017 (E,F), LR15–003 (G,H), LR19–506 (I,J), LR19–314 (K,L). Midline sagittal images (A, C, E, G, I, K) show mildly small and dysplastic corpora callosum (black arrows, A, C, E, K) and mildly small vermis in all individuals. The vermis typical extends from about the mid-tectum to the obex, marked here with short horizontal white lines. The small vermis and hemispheres can be seen on the axial image in LR05–007 (B) but not the others.

AUTS2/HX-NDD Missense (p.His529Pro, p.His533Tyr and p.Thr534Pro)

All (LR05–007, LR-CG-1, LR23–023, LR23–017) had low-set eyebrows that were mildly horizontal and sparse, with a high and broad nasal bridge, and malformed ears. They had unique facial features exclusive to the missense group. These features include mild blepharophimosis and a narrow nasal ridge in two (LR23–023, LR23–017 with p.Thr534Pro variants) and a sharp nasal ridge in one (LR23–023, p.His533Tyr), a very thin vermillion of the upper lip, and deep-set eyes in all four individuals.

Both individuals with the p.Thr534Pro variant and the individual with p.His529Pro variant had mildly broad halluces. Two also had kyphoscoliosis and tight heel cords. Three individuals with missense variants (LR05–007, LR23–023, LR23–017) had abnormal fingers including variable hypoplasia of the distal phalanges and occasionally the middle phalanges and variable symphalangism with absence or fusion of the proximal interphalangeal (PIP) joint (Supplementary Figure 1). This striking digit phenomenon did not occur with other reported AUTS2 variants. Brain MRI features of three individuals were strikingly similar. All three had hypoplasia of the corpus callosum and two (p.(Thr534Pro)) also had hypoplasia of cerebellum and small posterior fossae. Interestingly, neuroimaging of one individual in this group (LR-CG-1) was normal.

AUTS2/HX-NDD in-frame deletion (p.His535_Thr542del)

Individuals with the HX-repeat in-frame deletion (p.His535_Thr542del) had typical features of AUTS2/HX-NDD including feeding difficulties in infancy, severe developmental delays most pronounced in speech and severe intellectual disability with autistic features. Individuals were not spastic and ADHD was only seen in one individual (LR19–314). Neuroimaging features in the 4 individuals with recurrent p.His535_Thr542del in the distal HX motif were similar but less severe. Two had mild hypoplasia of the corpora callosum, while all four had mild vermis-predominant cerebellar hypoplasia. It was noticeable that in corpus callosum defects the middle part was most affected. Other neuroimaging findings are discussed in detail in Table 1 and Figure 3.

Predicted Pathogenicity of AUTS2 Variants

AUTS2 is predicted to be intolerant to loss of function based on two primary metrics: oe (observed/expected ratio)/LOEUF (Loss of function Observed/Expected Upper bound Fraction), and the probability of being loss-of-function intolerant (pLI).³⁶ Oe metric in gnomAD based on the canonical AUTS2 transcript (ENST00000342771.10, gnomAD v4)³⁷ demonstrates (oe= 0.15 [0.1–0.23] and LOEUF =0.232) strong selection against loss of function variants possibly associated with haploinsufficiency. Similarly, the pLI for AUTS2 is 1.0, while for any loss-of-function causing variants, pLI >= 0.9 implies an extremely intolerant transcript.³⁶ In contrast, AUTS2 is predicted to possibly tolerate missense variants based on two main metrics: missense oe/Z score ³⁷and AlphaMissense ³⁸ scores. In contrast to the loss of function metrics, the missense oe metric (oe=0.87 [0.84–0.91]) demonstrates a lack of constraint to missense variants. A missense Z score of 1.86 was calculated based on the expected rare variants per transcript (minor allele frequency, %0.1 ENST00000342771.10, based on gnomAD v4), which is less than the expected missense Z score (constrained if Z>3.09, p< 0.001) for a highly constrained gene. AlphaMissense predicts the effects of missense variants by giving a pathogenicity score between 0 and 1 (0: least likely to be pathogenic, 1: most likely to be pathogenic) and a prediction for every possible amino acid substitution in the proteome building on the AlphaFold protein visualization. ³⁸The mean predicted pathogenicity per amino acid substitution for AUTS2 calculated by the AlphaMissense algorithm was 0.437, demonstrating the presence of benign variation that potentially does not affect the protein.

In contrast, genomic constraint analysis of a smaller region within AUTS2 that includes the HX repeat domain (1kb, chr7:70766000–70767000) resulted in a Z score of 4.84 (gnomAD, constrained if Z>3.09, p< 0.001) which shows that area is highly constrained for missense variants (minor allele frequency, %0.1 ENST00000342771.10, based on gnomAD v4) (Figure 4). A larger local regional genomic constraint Z score (40kb, chr7: 70746234–70786234) showed that the neighboring Z scores ranged between −0.84 to 2.76. This is striking as the 1kb region encompassing exon 9 including the HX repeat was the only constrained part within a larger part of AUTS2 (40 kb, from the end of intron 6 to beginning of the intron 16). While the mean predicted pathogenicity per amino acid substitution (AlphaMissense) score for the AUTS2 gene was 0.437, the mean score for the HX repeat domain was 0.97, highest among known AUTS2 domains (Figure 4). AlphaMissense also evaluated the patient variants presented here as pathogenic.

Figure 4.

AUTS2 gene is shown with exon numbers 1 to 19 specified. Local genomic constraint analysis of a region in exon 9 (1kb, chr7:70766000–70767000) show a high Z score of 4.84. The neighboring Z scores ranged between −0.67 to 2.76. Average AlphaMissense scores for each known AUTS2 motif is shown on the gene diagram below. Highest average scores belonged to PY domain in exon 9 and the HX repeat domain. HX-Repeat variants associated with a phenotype are shown below, including c.1584_1613del p.Gln532_His541del variant reported by Martinez-Delgado et al.

Methods and algorithms mentioned here support the pathogenicity of reported AUTS2 variants and they may be used for additional variant scrutiny. These results together with our genotype-phenotype analysis suggest that AUTS2 is likely disrupted by missense variants in the HX repeat domain in exon 9.

DISCUSSION

The key features of AUTS2/HX-NDD included multiple comorbidities: a combination of ID with severe language deficit, motor delay, prominent facial dysmorphology, and skeletal especially digit anomalies. We demonstrate that the AUTS2/HX-NDD phenotype overlaps with RSTS, as it was suggested by Liu et al.⁹, but is not close enough to define a subtype of RSTS. The most distinctive features of RSTS – the low-hanging columella and broad thumbs/halluces – are mild in all individuals. However, clinical diagnosis of RSTS was suggested prior to genetic testing in three individuals (LR05–007, LR15–003 and LR-CG-1). While AUTS2 is generally not highly constrained for missense variants, we show that the exon 9 and the HX domain are highly constrained. Our data also suggest that the three individuals with missense variants in the middle of the HX domain (positions 533–534) had more severe features than the four individuals with the more terminal deletion (positions 535–542) and the single individual with an amino acid change in the beginning of the repeat (position 529).

Structural brain malformations were frequently reported in neurodevelopmental disorders associated with other PRC1.3/5 components but not with AUTS2 haploinsufficiency (Supplementary Figure 2). Abnormal brain MRIs were described in 14 of 25 (56%) individuals with CSNK2A1-related NDD.^39–49 However, the findings were variable and only two had hypoplasia of the corpus callosum, one of those also having cerebellar vermis hypoplasia. Similarly, abnormal brain MRI findings were reported in 8 of 26 (31%) individuals with CSNK2B-related NDD, another component of PRC1.3/5, one with microcephaly and typical secondary features such as a thin corpus callosum, two with cerebellar vermis hypoplasia, one with undefined “mega-cisterna magna” one with Chiari malformation and syringomyelia, and three with other findings.^44,50–56 RSTS (EP300/CREBBP) is also associated with agenesis or hypoplasia of the corpus callosum and cerebellar hypoplasia.⁵⁷ These structural brain defects, including the ones we report here, hint at earlier arrest in brain development that is common among these disorders. The analysis of the specific functional interactions between Polycomb proteins and further phenotypic data may uncover other key features shared between Polycomb component-related disorders. A list of all PRC1 related disorders is found in the supplementary files (Supplementary Table 3).

Seven genes containing HX repeats longer than 8 residues were identified.⁵⁸ Among those, AUTS2, ATN1, FBRSL1, and RERE are associated with neurodevelopmental disorders. De novo missense or insertion variants in the HX repeat have been reported in ATN1-related CHEDDA syndrome (congenital hypotonia, epilepsy, neurodevelopmental delay, digit abnormalities).⁵⁹ CHEDDA (OMIM #618494) is non-progressive and differs markedly from the previously described ATN1-related dentatorubral-pallidoluysian atrophy (DRPLA, OMIM #125370) caused by gain of function due to repeat expansion. Therefore, CHEDDA and DRPLA are examples of two distinct phenotypes associated with two different mechanisms where a gene harbors an HX repeat. The role of HX repeats in the context of human disorders need to be investigated further.

Interestingly, there are individuals with in-frame indels encompassing the HX repeat domain reported in gnomAD (c.1588_1599del p. Gln530_His533del and c.1594_1599del p. Gln532_His533del) who have higher allele frequencies (allele count:13, allele frequency: 0.000008055 and allele count: 73, allele frequency: 0.00004523, respectively). The phenotype information for these individuals is not provided. Therefore, it is unclear whether these variants are associated with a phenotype. The HX repeat may include or border other regulatory elements. Hence, the specific amino deletions may have unique properties in their effect on the AUTS2/HX-P300 interaction that do not apply to every in-frame HX repeat deletion.

It was previously suggested that the variability of the phenotype associated with AUTS2 haploinsufficiency differed based on whether a shorter AUTS2 transcript starting in exon 9 was affected (C terminus vs N terminus).² Our literature review and cohort revealed milder phenotypes with exonic deletions encompassing exons 1–5 but not a more severe presentation associated with variants in and beyond exon 9 (Supplementary Table 1). Notably, there is a higher frequency of pathogenic variation seen with variants affecting exon 6, with somewhat more severe features. Functional analysis of variants may reveal unique properties of exon 6 and may predict which variants encompassing the HX repeat may result in the deleterious features we report. Further analysis of AUTS2 interactors and downstream targets is needed to understand other diverse functions of AUTS2 and their contributions to clinical phenotypes.

In conclusion, our analysis further delineates a distinct and severe phenotype associated with HX repeat variants in AUTS2 compared to the patients we report and previously published reports of AUTS2 haploinsufficiency. The brain malformations and severe features seen with missense or small in-frame deletions warrant further variant scrutiny in genes associated with loss of function, such as AUTS2. We believe this work will inform clinicians and researchers about the distinct genetic effects of AUTS2 variants on phenotypic presentation.

DATA AVAILABILITY

Any data not available in the supplementary material are available upon request.