Three unrelated patients (two males and one female) born to physically healthy and non-consanguineous parents were enrolled in this study. Written informed consent has been obtained from each family.

WES was performed in all three patients as previously described (Wang et al., 2014; Wang et al., 2016). Briefly, genomic DNA was extracted from patients’ peripheral blood and was then sheared to create fragments of 150 to 200 bp. Sequencing library was prepared using the SureSelect XT Human All Exon V6 kit (Agilent Technologies, Santa Clara, CA, United States), and sequencing was performed by the Illumina NovaSeq 6000 System (Illumina, San Diego, CA, United States). After base calling, quality assessment, and alignment of the sequence reads to the reference human genome (GRCh37, dbSNP135), all single nucleotide variants (SNVs) and indels were saved as VCF format file which was then uploaded to the QIAGEN Clinical Insight (QCI) Interpret Translational tool (https://apps.qiagenbioinformatics.cn/) for filtering and annotation. The SOX11 variants identified by WES were validated by Sanger sequencing using the ABI 3700 sequencer (Applied Biosystems, Foster City, CA, United States), in indicated patient and their parents (sequencing primers were available upon request).

HEK293T cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% (v/v) fetal bovine serum (Sigma-Aldrich, St. Louis, MO, United States) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Waltham, MA, United States) in a 5% CO2 incubator at 37°C.

The wild-type (WT) and mutant SOX11 cDNA were synthesized by BGI (Shenzhen, China), and were then cloned into pcDNA3.1(+)-N-MYC vector, respectively. The GDF5 promoter 5′-flanking sequence (−448/+319) was synthesized and cloned into the pGL3-basic vector. Plasmid DNA was transfected into cells using Lipofectamine™ LTX reagent (Thermo Fisher Scientific).

HEK293T cells were cultured on cover slips in 12-well plate at 80% confluence prior for transfection. 24 h after transfection with the WT or mutant SOX11 constructs, cells were washed by 1 x PBS and fixed using 4% paraformaldehyde for 15 min at room temperature. Samples were then washed with 1 × PBS three times and blocked in the blocking buffer (1x PBS/5% goat serum/0.3% Triton X-100) for 1 h. Coverslips were incubated with mouse anti-Myc monoclonal antibody (Cell Signaling Technology, Danvers, MA, United States) at 4°C overnight. The cover slips were then mounted on microscope slides using ProLong^® Gold Antifade reagent with DAPI (Cell Signaling Technology) and analyzed using a Leica DM6000 fluorescence microscope (Leica Microsystems, Wetzlar, Germany).

HEK293T cells were seeded into 6-well plate and transfected with the indicated SOX11 constructs. 24 h after transfection, cells were washed with ice-cold 1x PBS and lysed with SDS lysis buffer (100 mM pH6.8 Tris-HCl, 10% Glycerol, 1% SDS). Whole cell lysates were separated on 10% SDS-PAGE gels, transferred to polyvinylidene difluoride (PVDF) membranes. After blocking in 5% non-fat milk in TBS-T for 1 h, membranes were then incubated overnight at 4°C with mouse anti-Myc monoclonal antibody (Cell Signaling Technology) and mouse anti-beta-Actin monoclonal antibody (Sigma-Aldrich). Proteins were detected using a chemiluminescence system with a horseradish peroxidase conjugated secondary antibody.

GDF5-luc plasmid was used to monitor the regulatory role of SOX11 in modulating the transcription. Briefly, HEK293T cells were seeded into 96-well plate at a density of 3×10⁴ per well. 24 h after culture, the cells were co-transfected with 28 ng GDF5-luc, 2 ng pRL-SV40, and 70 ng control or SOX11 expression vectors. 24 h after transfection, luciferase activity was determined by the Dua-Glo luciferase assay system (Promega Corporation, Madison, WI, United States), and normalized to renilla luciferase.

Patient 1 was a male who was born at full-term gestation with a birth weight of 2100 g (−3.1 SD) by spontaneous vaginal delivery. At the stage of pregnancy, intrauterine growth restriction (IUGR) was noticed. At 3-year old, he was referred to our hospital due to growth retardation. Physical examination showed a height of 79 cm (−4.7 SD) and weight of 9.6 kg (−3.3 SD). His hair was sparse and the anterior hairline was high. He presented with facial deformities consisting of prominent forehead, arched eye brows, flat nasal bridge and broad nose, low-set ears, short philtrum, and micrognathia (Figure 1A). He had left cryptorchidism, short fingers, and clinodactyly of the 5th finger. In addition, finger joint laxity was also noticed in the patient (Figure 1B). X-ray showed delay bone age which was 1.5-year old. No DD or ID was found in the patient. During the last follow up, his parents denied the child have mental development problems and refused further assessment, though learning difficulties were mentioned. No special was recorded for the development of the gross motor skills of the patient. The levels of growth hormone, IGF-1, IGF-BP3, LH and FSH were normal. Cranial magnetic resonance imaging (MRI) showed his pituitary gland was thin and oblate. Abdominal ultrasound was normal.

Patient 2 was a 5.5-year old female who was born at full-term gestation with birth weight of 3500 g (0.7 SD) and birth length of 50 cm (0.2 SD) by cesarean section. She was able to sit at 6-month old and walk alone at 20-month old. At the age of 5-year old, she was found with severe speech impairment and moderate ID. Physical examination showed a height of 113 cm (−0.1 SD) and weight of 18.7 kg (−0.3 SD). She had prominent forehead, high hairline, sparse scalp hair, arched eye brows, short philtrum, flat nasal bridge and broad nose, hypoplastic right nasal alar, auricle malformation, everted upper lip, and cleft palate and lip (Figure 1A). X-ray of the bone age, cranial MRI, and the abdominal ultrasound were normal.

Patient 3 was an 8-year old male patient born at full-term gestation by cesarean section. He was admitted to the Endocrinology department due to a slow increase in height, which was less than 4 cm/year for 4 years. His birth weight was 3200 g (−0.3 SD) and birth length was 49 cm (−0.8 SD). Physical examination showed a height of 118 cm (−2.2 SD) and weight of 17 kg (−2.5 SD). He had prominent forehead, high hairline, flat nasal bridge, and bilateral cryptorchidism. Specialist assessment suggested he had DD and mild ID. X-ray showed a two-year delay in his bone age. The levels of growth hormone, IGF-1, IGF-BP3, LH and FSH were normal. Testosterone levels had a good response after HCG stimulation, increasing from less than 0.01–0.28 ng/ml. The cranial MRI and abdominal ultrasound were normal.

We compared the clinical phenotypes of our patients with those previously reported (Tsurusaki et al., 2014; Hempel et al., 2016; Khan et al., 2018; Okamoto et al., 2018; Sekiguchi et al., 2019; Diel et al., 2021; Wakim et al., 2021; Al- Jawahiri et al., 2022; Cho et al., 2022), which have been summarized by Al-Jawahiri et al. (Al- Jawahiri et al., 2022). Although the clinical features of the three Chinese patients were variable, they still fell within the spectrum of SOX11 syndrome (Table 1). In summary, patients 1 and 2 had highly consistent facial deformities, comprising of prominent forehead, arched eye brow, flat nasal bridge, broad nose, short philtrum, and abnormal ears. In comparison, the facial deformity of patient 3 was much milder. Sparse scalp hair and high hairline were also noticed in both patients 1 and 2. Patients 2 and 3 had DD and ID, and both male patients had short stature and cryptorchidism. In addition, patient 1 harbored the 5^th finger clinodactyly and microcephaly was only found in patient 3.

WES identified three heterozygous SOX11 variants, including the missense variant of c.425C>G, p.A142G in patient 1, the nonsense variant of c.820A>T, p.K274* in patient 2, and the missense variant of c.337T>C, p.Y113H in patient 3. Sanger sequencing confirmed these variants in the patients (Figure 2A), and also revealed the de novo status of each variant from the results that their parents were WT in SOX11 gene. All of the three variants were not reported in previous cases and not included in the known public database (i.e., gnomAD, HGMD and ClinVar), suggesting they were novel.

The K274* is not located in the last exon or within the 3′-most 50 nucleotides of SOX11, so it is very likely to lead to nonsense-mediated mRNA decay (Abou Tayoun et al., 2018). The Y113 residue locates at the HMG domain of SOX11 protein (Figure 2B), which is highly conserved in multiple species (Figure 2C). Crystal structure analysis using the mouse Sox4 HMG domain showed the Y113H variant may not alter the secondary structure of SOX11, but is likely to affect its ability to bind DNA (Figure 2D). By contrast, the other two variants of A142G and K274* locate downstream of the HMG domain (Figure 2B), and the A142 residue is also highly conserved (Figure 2C). In addition, the two missense variants of Y113H and A142G are predicted to be deleterious according to various in silico tools including PolyPhen-2, MutationTaster2, CADD, and ClinPred (Table 2).

To further evaluate the impact of the Y113H and A142G variants on SOX11 function, we first performed luciferase assay using a reporter construct containing a fragment of GDF5 promoter which has been described previously (Tsurusaki et al., 2014). Compared to the with WT SOX11, both Y113H and A142G mutants result in decreased transcriptional activities (Figure 3A). In vitro experiments in HEK293T cells revealed that both missense variants increased SOX11 expression (Figure 3B), which were also observed in previous studies (Tsurusaki et al., 2014; Hempel et al., 2016). Immunofluorescence analysis showed that the WT and two missense mutant SOX11 protein localized in the nucleus (Figure 3C).

It is estimated that near 5% of patients can have more than one molecular findings, and may cause more serious phenotypic features (Posey et al., 2017). We routinely analyze copy number variation (CNV) of each patient by comparing the read-depth with the WES data from the other samples of the same batch (Yao et al., 2017; Yao et al., 2019). As shown in Figure 4A, patient 1 harbored a 4,300 kb deletion involving the region of 1q24.2-q25.1 (hg19,chr1:169,433,149-173,827,682), which includes multiple disease-causing genes, such as EEF1AKNMT (OMIM#617987), FASLG (OMIM#134638), FM O 3 (OMIM#136132), GORAB (OMIM#607983), MYOC (OMIM#601652), PIGC (OMIM#601730), PRRX1 (OMIM#167420), SLC19A2 (OMIM#603941), and TNFSF4 (OMIM# 603,594) (Figure 4B). Deletion of this genomic region has been described in series of patients (Burkardt et al., 2011; Chatron et al., 2015). According to the clinical interpretation of CNV by the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen) (Riggs et al., 2020), the microdeletion of 1q24.2-q25.1 in this patient was classified as pathogenic. No questionable CNVs were identified in the other two patients.