ldentification and functional analyses of a novel FOXL2 pathogenic variant causing blepharophimosis, ptosis,and epicanthus inversus syndrome

Yu-Cheng Yan, Lu Zhou, Jin-Cai Fan

Plastic Surgery Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100041,China

Abstract

INTRODUCTION

Blepharophimosis, ptosis, and epicanthus inversus syndrome (BPES; the Online Mendelian Inheritance in Men #110100), first reported by Ammon in 1841, is a rare congenital disease, whose prevalence is estimated as 1/50 000 according to statistics[1].Depending on the clinical features,BPES is separated into two main phenotypes: the type I patients have premature ovarian insufficiency/failure(POI/POF) and eyelid malformations at the same time while the type II patients only have deformities in eyelids[2].

The forkhead box L2 (FOXL2) gene, encodes a conserved transcription factor, which is a member of the winged helix transcription factor family.FOXL2contains one forkhead domain, serving to affect the promoters of downstream genes, and the poly-Ala tracts which were associated with the intracellular localization[3].FOXL2is the only known gene responsible for BPES[4].

FOXL2 involves in a lot of physiological processes serving as a transcription factor[5-6].During the development of an embryo,FOXL2selectively expresses in the mesenchymal stem cell of the fetal ovary and developing eyelids, which has been proven in animal models.TheFOXL2knockout mice exhibit absent upper eyelid and failure of somatic cell development in the fetal ovary[7].Besides the embryogenetic stage, researchers have proven that FOXL2 is involved in almost all functions of the adult ovary[8].The promoter of the steroidogenic acute regulatory protein (StARorSTARD1) gene has been proven to be the direct target ofFOXL2[9], which plays a significant role in granulosa cells.The regulatory interaction betweenFOXL2and the odd-skipped related 2 transcription factor (OSR2) gene is crucial in the development of facial bones and eyelids, which has been proven by previous studies[10].

Female patients with BPES are usually in childhood when visiting hospitals, without symptom available in favor of forecasting the POI/POF, and the variation of ovarian reserve function make it difficult to make a prompt diagnosis[11].Accordingly, researching the genotype-phenotype correlation is of great value for predicting the classification of BPES.De Baereet al[12]subdivide theFOXL2mutations into nine classes, based on the predicted protein; Moumnéet al[13]once reported that theFOXL2mutations expressing truncated proteins upstream of the poly-Ala tract likely lead to type I BPES, while type II BPES usually caused by theFOXL2mutation resulting in extended proteins.However, these claims cannot explain all the cases reported, such as the patients with heteroplasmy[14-15], and hence the genotype-phenotype relationship of BPES still needs further researches for providing evidences.

In this study, we performed gene sequencing forFOXL2in a sporadic case.The further researches concerning biological functions were carried out to provide follow-up medical advice and clues for determining the genotype-phenotype correlation.

SUBJECTS AND METHODS

Ethical ApprovalThe present study conformed to the Declaration of Helsinki and was approved by the ethics committee of the Plastic Surgery Hospital (Institute), CAMS,PUMC (2023-53).Informed consent was obtained from guardians of the patient.

PatientA 3-year-old girl with eye malformations (Figure 1A) was admitted to the Plastic Surgery Hospital (Institute),CAMS, PUMC seeking surgical treatment.The patient was born at 38wk of gestation after an uncomplicated pregnancy and a normal delivery from a nonconsanguineous healthy Chinese parents both without genetic disorder.Received systematic ophthalmic examinations, the patient was diagnosed with BPES for the classic symptoms of blepharophimosis,ptosis, and epicanthus inversus.Photos were taken before the surgery.The patient was tested with normal 46, XX karyotype,and her ovarian function can not be evaluated because of the age.

DNA Extraction and SequenceThe patient’s genomic DNA was extracted from the leukocytes of peripheral venous blood using QIAamp Blood DNA Mini kit (Qiagen,Germany) following the manufacturer’s instruction.Three sets of premiers were designed to amplify the region ofFOXL2through polymerase chain reaction(PCR):FOXL2-1F: GGTGAGCACAGGAGGACATAA;1 R: G A G C G A G TA C C G G C A G AT T;F O X L 2-2 F: G T G A G C A C A G G A G G A C AT A A; 2 R:A G T T T G A G A C T T G G C C G T A A;F O X L 2-3 F : T G C T T C A T C A A G G T G C C G ; 3 R :GCACAAGCGAACTGCAGG.After purifying, the PCR products were sequenced using the Sanger sequencing(Applied Biosystems, CA, USA) to confirm the pathogenic variant.Then, the outcome was analyzed using GeneMapperⅤ3.5, and the nomenclature of the novel pathogenic variant was determined according to Human Genome Ⅴariation Society Guidelines.The MutationTaster online software was used for predicting the possible pathogenicity.

Plasmid ConstructionThe coding sequence of theFOXL2opening reading frame (ccds3105.1) was cloned into pcDNA3.1 and pEGFP-C1 vectors, producing pcDNA3.1-FOXL2 and pEGFP-FOXL2.Using wild-typeFOXL2as the template, the mutant plasmids with c.274G>T were constructed by site-directed mutagenesis (QuikChange Lightning Multi Site-Directed Mutagenesis Kit, Agilent), producing pcDNA3.1-FOXL2-G274T and pEGFP-FOXL2-G274T.The promoter and 5’-UTR region ofStARandOSR2were inserted into the pGL3- Basic to build the luciferase reporter vectors.All the constructs were verified without additional mutations through Sanger sequencing.

Western BlottingThe murine leydig tumor cell line 1(MLTC-1) cells were grown in 6-well plates 24h before the transfection, maintained in the Roswell Park Memorial Institute-1640 (Procell, China) with 10% fetal bovine serum(Procell, China) and 1% P/S at 37°C in a 5% CO2atmosphere.Then, the medium was changed into Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, USA), and the MLTC-1 cells of each dish were transfected with 2.5 μg pEGFP‐C1, pEGFP‐FOXL2 or pEGFP‐FOXL2‐G274T using 7.5 μL Lipofectamine 3000 reagent and 5.0 μL P3000 reagent (Ⅰnvitrogen, USA).Finishing 48h of culture, the cells were washed with phosphatebuffered saline (PBS; Beyotime, Shanghai, China) twice after removing the media.Then, the cells of one well were lysed with 150 μL mixed liquor of radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China) and phenylmethanesulfonyl fluoride (Beyotime, Shanghai, China).The protein concentrations were determined by using the BCA Protein Assay Kit (Beyotime, Shanghai, China).A total of 35 mg protein from each sample added loading buffer were loaded on 10% sodium dodecyl sulfate-polyacrylamide gel,and next the protein was transferred to nitrocellulose filter membranes (Osmonics, USA).After blotted by TBST with 5%skimmed milk powder, membranes were incubated in primary antibodies at 4℃ overnight and the corresponding secondary antibodies for 1h at room temperature.Using The ChemiDoc MP system (Bio-Rad Laboratories, California, USA) to detect the immunotagged proteins.The primary antibodies we used are anti-GFP antibody (1:1000, Cat No.3h9, Proteintech),anti‐FOXL2 antibody (1.5 μg/mL, ab5096, Abcam), and anti‐GAPDH antibody (1:5000, ab8245, Abcam).Western stripping buffer (Beyotime, Shanghai, China) was used between the chemiluminescent immunoassay of GFP and incubating the anti-GAPDH antibody.

Subcellular Localization ExperimentMLTC-1 cells were seeded into the 35 mm confocal dishes and were transfected with pEGFP-C1, pEGFP-FOXL2 or pEGFP-FOXL2-G274T respectively using the same type and dose of reagents with the Western blotting experiment.After 72h of culturing, the nuclei were stained with 4’,6-diamidine-2’-phenylindole dihydrochloride (DAPI; Roche, Germany) after 48h of the transfection, and subcellular localization was observed by Zeiss LSM 800 confocal laser scanning microscopy.

Luciferase Reporter AssayHuman embryonic kidney 293(HEK-293) cells were seeded into 24-well plates 24h before the transfection, maintained in the Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and 1% P/S at 37°C in a 5% CO2atmosphere.The 2.0 μL/well of Lipofectamine 2000 reagent (Invitrogen, USA) was used to transfect pcDNA3.1(vector), pcDNA3.1-FOXL2 (WT), pcDNA3.1-FOXL2-G274T (MT), and the reporter constructs described above.The double luciferase assays of every single reporter plasmid were divided into 6 groups including 1) 500 ng vector; 2) 500 ng WT;3) 500 ng MT; 4) 250 ng WT plus 250 ng vector; 5) 250 ng MT plus 250 ng vector; 6) 250 ng WT plus 250 ng MT, together with 500 ng luciferase reporter plasmids (either pGL3-StAR or pGL3-OSR2) and 25 ng PRL-TK vectors (Promega) per well.The total DNA content was maintained at 1025 ng/well.The cells were incubated with plasmids in DMEM for 8h, and then the culture was continued for another 48h with the complete media.Following the protocol of Dual-Luciferase Reporter Assay System (Promega, USA), the luciferase intensities were measured by EnSpire Muitilable Plate Reader (PerkinElmer,USA).

Quantitative Real-time PCRHEK-293 cells were plated in 12‐well plates and transfected with 1.0 μg of pcDNA3.1‐FOXL2, pcDNA3.1-FOXL2-G274T, and empty pcDNA3.1 vector respectively using 2.0 μL of P3000 reagent and 3.0 μL of Lipofectamine 3000 reagent each well.After 48h of culture,total RNA was extracted with TRIzol LS Reagent (Cat No.10296010, Invitrogen) following the operating instructions,and the final RNA concentrations were measured by NanoDrop 2000 (Thermo, USA).cDNA was synthesized using the Universal RT‐PCR Kit (M‐MLⅤ; Promega, USA).Then, the SYBR Green real-time PCR was performed on LightCycler 96 (Roche).GAPDH gene was chosen for normalization.The primer pairs for qRT-PCR are listed as follows:GAPDH-F: 5’-CAACAGCCTCAAGATCATCAGCA-3’;R: 5’-GTCATGAGTCCTTCCACGATAC-3’; StAR-F:5’-TTTAGTACCAAGAAAACAGGGATG-3’; R:5’-CACACCCATATCAGCCACT-3’; FOXL2-F:5’-G T C C G G C AT C TA C C A G TA C AT C A-3’; R:5’-CCAGCCCTTCTTATTCTTCTCG-3’.

RESULTS

Identification of a Novel Pathogenic Variant in FOXL2The sequences of PCR products derived from the patient’s genomic DNA revealed a novel pathogenic variant (c.274G>T), that resulted in a truncated protein with 92 amino acids (p.E92*)due to the introduction of a premature stop codon (Figure 1B and 1C).The variant was confirmed absent by Sanger sequencing in dbSNP database (http://www.ncbi.nlm.nih.gov/SNP).

As the online software (https://www.mutationtaster.org/)predicted, the truncated protein translated by theFOXL2pathogenic variant is only composed of 92 amino acids,without poly-Ala tract and partial FKH domain (the 3rdα-helical structure).The PhyloP and PhastCons scores were 3.821 and 1, which means the pathogenic variant was highly conserved among species.

Effects of the Novel Pathogenic Variant on Subcellular LocalizationAs is shown in the Figure 2A, an approximate 35 kDa band, 8 kDa higher than the EGFP, was observed with the GFP antibody after being transfected with the pEGFPFOXL2‐G274T, which verified that the mutatedFOXL2could be translated into protein in MLTC-1.An approximate 70kDa protein was detected clearly using either the GFP or FOXL2 antibodies when the MLTC-1 transfected with the pEGFPFOXL2.However, no 70 kDa band was observed in other tracks, which confirmed that theFOXL2expression was absent in MLTC-1 and the protein translated by the mutatedFOXL2was truncated.For the reason that the antibody of FOXL2 we used here was bound with the C-terminal of wild-type FOXL2.To further investigate the changes in biological functions of thisFOXL2variant, we conducted the subcellular localization experiment using the MLTC-1 cells.As the Figure 2B exhibits, the wild-type FOXL2 connected with EGFP tag was distributed exclusively in the nucleus region stained with DAPI diffusely, while the mutant FOXL2 can be detected in both the cytoplasm and nucleus.

Effects of the Novel Pathogenic Variant on Transcriptional ActivityThe results of the reported assays showed that the repression ability of WT protein onStARpromoter was obvious, while the MT protein failed to inhibit the luciferase activity of the reporter vectors.The luciferase activity didn’t exhibit a decreasing trend as we increase the transfection amount of the MT plasmids (Figure 3A).To simulate the physiological conditions, equal concentrations of MT/WT vectors were transfected testing the dominant negative effect of the novel pathogenic variant.The results showed that the activity of wild-type FOXL2 in binding the promoter ofStARdid not be diminished when co-transfected with mutant FOXL2 (Figure 3C).We next test the capacity of the mutant FOXL2 in activating theOSR2promoter.As is shown in the (Figure 3B), the MT lost the function of activating theOSR2promoter for the reason that no statistical difference of luciferase activity was detected between the cells transfected with the empty vectors and the MT plasmids.The WT FOXL2 activated theOSR2promoter in a dose-dependent manner.In addition, an obvious dominant negative effect was detected when the cells were co‐transfected with the same dose of WT and MT (Figure 3D).

Figure 1 The novel pathogenic variant and its clinical presentation A: Facial photograph of the sporadic patient; B: Sequencing result revealed the novel FOXL2 pathogenic variant c.274G>T (p.E92*).The arrow indicates the position of the nonsense pathogenic variant.C: Schematic representation of the truncated protein.

Figure 2 Protein expression and distribution A: Western blot analysis showed the expression of the wild-type FOXL2 and the pathogenic variant FOXL2 with the anti-FOXL2 antibody and the anti-GFP tag antibody.The position of the bands was consistent with the predictions.GAPDH served as the reference gene.B: Subcellular localization of EGFP, FOXL2 and the FOXL2-G274T.The first line shows the nuclear stained with DAPI; the second line exhibits the subcellular localization of EGFP or FOXL2 as the fusion protein tagged with EGFP.The third line contains merged images of the above pictures.Scale bars=5 μm.

Effects of the Novel Pathogenic Variant on Expressions of Downstream GenesTo further confirm the effects exerted by this novelFOXL2variant on ovarian functions, we test the endogenous mRNA expressions ofStARusing the qRTPCR.The results (Figure 3E) showed that the expression levels of the target gene were similar when transfected with empty pcDNA3.1 vectors and the pcDNA3.1-FOXL2-G274T plasmids.In contrast, the cells transfected with wild-typeFOXL2exhibited obvious decrease in the expression level ofStAR.The results of qRT‐PCR verified the functional absence of this mutant FOXL2 protein in serving as the repressor of theStAR.

DISCUSSION

Figure 3 Transcriptional activity of p.E92* mutant FOXL2 The pcDNA3.1 (vector), pcDNA3.1-FOXL2 (WT) or pcDNA3.1-FOXL2-G274T (MT)plasmids were co-transfected with the luciferase vectors driven by the StAR promoter (A) and OSR2 promoter (B).Co-transfected the same dose of WT and MT to detect the potential dominant negative effect of the pathogenic variant on StAR promoter (C) and OSR2 promoter (D).Relative expression level of StAR, measured by qRT-PCR, was compared in the cells transfected with empty vectors, WT and MT plasmids.aP<0.05, bP<0.01, cP<0.001, dP<0.0001.NS: P>0.05.

The FOXL2, serving as a conserved transcription factor, plays an important role in the ovary and eyelid from embryonic stages to adult stages[16].As for blepharophimosis, ptosis, and epicanthus inversus syndrome, surgical treatments for eye malformations and predicting the occurrence of premature ovarian insufficiency/failure are major demands for clinical work.Many mature operations have been used in the treatments of BPES, which give patients better appearances and achieve a lower incidence of amblyopia among the diseased cohort.However, no feasible and correct laboratory index guiding type I BPES patients in childhood and their family to make preparations, like tissue banking, for the upcoming POI/POF[8].Support and advice should be offered to relative families.Accordingly, the significance of the genotypephenotype relationship goes without saying.As is mentioned in the preceding part of this paper, because of the lack of a golden standard, BPES type I and II still cannot be distinguished through simple gene sequencing.For the sporadic cases without family medical histories, the prediction of phenotype is even harder.Hence, research concerning the biological functions of new protein pathogenic variants is still needed to be performed.In this research, a novelFOXL2nonsense mutation 274G>T (p.E92*) was reported, located in the forkhead domain.For this 3-year-old female patient,the prediction of the subtype is of great importance for future treatments.

To begin with, the subcellular localizations were tested, for the reason that the wrong location and aggregation could reflect dysfunction of the mutant protein, and the aggregation may predict the dominant negative effec[3,17].The results of the Western blot ensured the validity of the subcellular localization experiments.In addition, the results verified the FOXL2 pathogenic variant was truncated.As a transcription factor, the FOXL2 was transferred into the nucleus for functioning, which corresponded with the result.In contrast, the mislocalization of the MT protein demonstrated the dysfunction of the mutantFOXL2[18-19].

Next, the bioactivities of mutant FOXL2 as a transcription factor were researched by detecting the interactions with the downstream genes[15].StAR is indispensable in the process of delivering large amounts of cholesterol into the mitochondria,where the cholesterol could switch to pregnenolone[20].The expression ofStARimplies the differentiation of granulosa cell.FOXL2 serving as the main passive regulator, could preserve primordial follicles and prevent premature depletion of ovarian follicles by suppressing the promoter ofStAR[9].Thus some FOXL2 mutations failing to bind its expression would induce the BPES type I[5,14].OSR2knockout mice exhibited eyelid malformations caused by the growth defects of the palate shelf[21], and the FOXL2 was proven that could activate its promoter directly.The perturbation of this regulatory interaction could contribute to eyelid malformation of the BPES[22].What’s more,OSR2was also detected expressing in not only the eyelid dermis but also the uterus of neonatal infants[10].Thus, we chose these two downstream genes in vitro cell experiments to predict the pathogenicity in ovarian function and periorbital development.The pathogenesis of BPES could attribute to haploinsufficiency, and the dominant negative effect, FOXL2 pathogenic variant obstructing the action of the wild-type protein, which is more pathogenic[23].

The dual-luciferase reporter assay showed that the novel variant lost its biological function as a transcription factor, and less amount of wild-type protein could achieve a complete reaction with the promoter ofStAR, which could explain the formation of the two subtypes: the development of eyelids is more sensitive to the variation in the quantity of FOXL2.When the cells were co-transfected with WT and MT plasmids,the protein variant obstructed the action of wild-type protein in activating the promoter ofOSR2.However, a similar result was not observed when detecting the luciferase activity of theStARpromoter reporter vectors.This result indicated that the dominant negative effect of this novel pathogenic variant would aggravate the dysplasia of eyelid structures.Therefore,considering that the pathogenic variant lost its capacity in binding the promoter ofStAR, adding its wrong subcellular localization, we speculated that the patient who harbors this pathogenic variant is in a high risk of ovarian insufficiency.However, the occurrence of POI/POF still cannot be predicted accurately for the reason that no dominant negative effect was observed in the dual-luciferase experiments ofStARpromoter.Hence, periodical clinical follow-ups and ovarian function assessments are of great importance[24].

In conclusion, we reported a novelFOXL2pathogenic variant(c.274G>T) which could be translated into truncated protein(p.E92*).The followingin vitrobiofunctional experiments indicated that the mutant protein cannot distribute correctly,and lost its function as a transcription factor ofStARandOSR2.The patient with this novel pathogenic variant has a high risk of ovarian insufficiency.The present study not only expands the spectrum ofFOXL2mutations but also contributes to the understanding of molecular pathogenesis.

ACKNOWLEDGEMENTS

The authors thank the patient and guardians for cooperating our researches.

Foundation:Supported by Funds of Plastic Surgery Hospital of the Chinese Academy of Medical Sciences, Peking Union Medical College (No.YS202010).

Conflicts of Interest:Yan YC,None;Zhou L,None;Fan JC,None.