2016, 63 (4), 405-410

Note A novel mutation in HESX1 causes combined pituitary hormone deficiency without septo optic dysplasia phenotypes

Masaki Takagi1), 2) *, Mai Takahashi3) *, Yoshiaki Ohtsu3), Takeshi Sato1), Satoshi Narumi1), Hirokazu Arakawa3) and Tomonobu Hasegawa1)

1) Department of Pediatrics, Keio University School of Medicine, Tokyo, Japan 2) Department of Endocrinology and Metabolism, Tokyo Metropolitan Children’s Medical Center, Tokyo, Japan 3) Department of Pediatrics, Gunma University Graduate School of Medicine, Maebashi, Japan

Abstract. Heterozygous and/or homozygous HESX1 mutations have been reported to cause isolated growth hormone deficiency (IGHD) or combined pituitary hormone deficiency (CPHD), in association with septo optic dysplasia (SOD). We report a novel heterozygous HESX1 mutation in a CPHD patient without SOD phenotypes. The propositus was a one- year-old Japanese girl. Shortly after birth, she was found to be hypoglycemic. She was diagnosed with central adrenal insufficiency based on low cortisol and ACTH at a time of severe hypoglycemia. Further endocrine studies indicated that the patient also had central hypothyroidism and growth hormone deficiency. Using a next-generation sequencing strategy, we identified a novel heterozygous HESX1 mutation, c.326G>A (p.Arg109Gln). Western blotting and subcellular localization revealed no significant difference between wild type and mutant HESX1. Electrophoretic mobility shift assays showed that the mutant HESX1 abrogated DNA-binding ability. Mutant HESX1 was unable to repress PROP1-mediated activation. In conclusion, this study identified Arg109 as a critical residue in the HESX1 and extends our understanding of the phenotypic features, molecular mechanism, and developmental course associated with mutations in HESX1. When multiple need to be analyzed for mutations simultaneously, targeted sequence analysis of interesting genomic regions is an attractive approach.

Key words: HESX1, Combined pituitary hormone deficiency, , Targeted next-generation sequencing

THE PROLIFERATION and terminal differentia- tion to these genes, some causative genes for Kallmann tion of the anterior pituitary gland are strongly influ- syndrome (KS), which is defined by hypogonadotropic enced by the precise spatial and temporal expression hypogonadism with anosmia, have been identified of transcription factors [1-3]. Mutations in these tran- in a small number of CPHD and septo optic dyspla- scription factors result in various types of congenital sia (SOD), a condition characterized by pituitary hor- , ranging from isolated growth hor- mone deficiencies, optic nerve hypoplasia and midline mone deficiency (IGHD) to combined pituitary- hor defects [4-7]. mone deficiency (CPHD). Several Among transcription factor genes responsible for genes have been linked to the pathogenesis of CPHD, CPHD, human HESX1 mutation was first reported including POU1F1, PROP1, LHX3, LHX4, OTX2, in sibling case with SOD in homozygous state [8-9]. , SOX3, GLI2, and HESX1 [2]. Recently, in addi- Subsequently, heterozygous HESX1 mutations were also shown to be associated with CPHD or IGHD, with Submitted Jul. 14, 2015; Accepted Dec. 17, 2015 as EJ15-0409 or without SOD phenotypes [10-12]. To date, more Released online in J-STAGE as advance publication Jan. 15, 2016 than 22 mutations in HESX1 have been described Correspondence to: Tomonobu Hasegawa, M.D., Ph.D., (HGMD; http://www.hgmd.cf.ac.uk). The majority Department of Pediatrics, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. of these are missense mutations (14 missense, 2 non- E-mail: [email protected] sense, 2 splice site, 1 gross insertion, and 3 frameshift *Masaki Takagi and Mai Takahashi contributed equally to this work. mutations). Here, we report a case of CPHD without ©The Japan Endocrine Society 406 Takagi et al.

SOD phenotypes carrying a novel heterozygous mis- PROP1, LHX3, LHX4, HESX1, OTX2, SOX3, SOX2, sense mutation in HESX1. Through molecular analy- GLI2, and 12 genes implicated in KS, including CHD7, ses, we showed that substitution of a conserved, critical FGFR1, FGF8, GNRH1, GNRHR, KISS1, KISS1R, amino acid within the homeobox domain of HESX1 PROK2, PROKR2, TAC3, TACR3 and KAL1 using abrogated DNA-binding, and was unable to repress the MiSeq instrument (Illumina Inc, San Diego, CA, PROP1-mediated activation. This study extends our USA) according to the SureSelect protocol (Agilent understanding of the phenotypic features, molecular Technologies, Santa Clara, CA, USA) as described mechanism, and developmental course associated with before [13]. mutations in the HESX1 . Crystal structure modeling Patient The crystal structure of the paired homeodomain of The propositus was a one-year-old Japanese girl born Drosophila melanogaster ( ID 1FJL; at 38 weeks of gestation after an uncomplicated preg- http://www.rcsb.org/pdb/, whose sequence identity nancy and delivery. The parents were nonconsanguin- with human HESX1 homeobox domain is 67%) was eous and phenotypically normal. The father was 160.5 used as a reference wild-type structure for modeling cm (-0.6 SD) tall and mother was 162 cm (0.8 SD) tall. the structure of p.Arg109Gln HESX1 using the PyMOL Apgar scores were 8 and 9 at 1 and 5 min, respectively. Molecular Graphics System (http://www.pymol.org). At birth her weight was 2,525 g (-1.3 SD), and length was 46.0 cm (-1.1 SD). Frequent apneic spells ensued Functional studies shortly after birth, and she was found to be hypogly- Construction of expression vectors cemic (blood glucose 20 mg/dL). She was diagnosed To generate HESX1 expression vectors, 2001 bp with central adrenal insufficiency based on low cortisol comprising the entire genomic HESX1 was (<1.0 μg/dL) and adrenocorticotropin (ACTH) (<2.0 cloned into pCMV-myc (Clontech, Palo Alto, CA) pg/mL) at a time of severe hypoglycemia. Further and pEGFPN1 (Clontech, Palo Alto, CA). To gen- endocrine studies indicated that the patient also had erate PROP1 expression vectors, PROP1 cDNA central hypothyroidism on the basis of a low free T4 was cloned into pCMV- vector. We introduced (0.44 ng/dL: Ref. 0.99–1.91) with an inappropriately the mutation by site-directed mutagenesis using the normal thyroid-stimulating hormone (TSH) concen- Prime STAR Mutagenesis Basal kit (TaKaRa, Otsu, tration of 3.09 mU/L (Ref. 0.77–7.3), and growth hor- Japan). The primer pairs, forward (F) and reverse mone (GH) deficiency based on no GH response on (R), used for mutagenesis were as follows: for- arginine hydrochloride testing (GH peak <0.1 ng/mL, ward 5′-AGAGGCCAAAGACCAAGAACTGCTTT Ref. 6<). The brain MRI exhibited anterior pituitary TAC-3′, reverse 5′-TTGGTCTTTGGCCTCTATAC hypoplasia, absent pituitary stalk, and ectopic posterior CAACTCAACT-3′. A luciferase reporter vector pituitary. Optic nerve hypoplasia was not evident. She was constructed by inserting six P3 sequences was diagnosed as CPHD and replacement therapy with (5′-AGCTTGAGTCTAATTGAATTACTGTAC-3′) L-thyroxine, hydrocortisone and recombinant human into a pGL4.24 [luc2P/minP] vector (Promega, GH was started. Examination by experienced oph- Madison, WI). thalmologists revealed no eye abnormality. At the last Western blotting examination at age of 17 months, she measured 71.9 COS1 cells transfected with the myc-tagged HESX1 cm (-2.4 SD), and weighed 7.24 kg (-2.5 SD). No other were harvested, and nuclear protein was isolated with family members showed growth disorders. the NE-PER nuclear extraction reagent kit (Pierce, Rockford, IL). Western blotting was performed with Mutation screening a mouse anti-myc monoclonal antibody (Invitrogen). After obtaining informed consent, and with the Subcellular localization analyses approval of the Institutional Review Board of Tokyo We visualized and photographed COS1 cells trans- Metropolitan Children’s Medical Center, genomic fected with GFP-tagged HESX1 using a BZ-X700 flu- DNA was extracted from peripheral blood leuco- orescence microscope (Keyence, Osaka, Japan). cytes of the propositus and her parents. We sequenced EMSA experiment 9 genes implicated in CPHD, including POU1F1, The sequences of the biotin-labeled double stranded A novel HESX1 mutation 407 oligonucleotide used as probe in the EMSA experiment was 5′-AGCTTGAGTCTAATTGAATTACTGTAC-3′ (P3 sequence). Five microgram of nuclear protein extraction was incubated at room temperature in 20-μL binding reaction mixture contained 20 fmol probe, 50 mM KCl, 5 mM MgCl2, 2.5% glycerol, 0.05% NP-40, and 1 μg poly (dI-dC) for 20 min. For competition experiments, a large excess (200x) of unlabeled com- petitor oligonucleotides was included in the binding reactions. The protein-DNA complexes were subject to gel electrophoresis and transferred to a nylon mem- brane. The biotin-labeled probe was detected with the Lightshift chemiluminescent EMSA kit (Pierce). For super-shift assay, we used a mouse polyclonal anti- HESX1 antibody (ab67728, Abcam, Cambridge, MA). Transactivation assay HESX1 has been shown to function as a repres- sor of PROP1-mediated gene stimulation [14]. To assess the ability of the mutatnt HESX1 to repress transcription, wild-type or mutant HESX1 pCMV- myc expression vectors were transfected into COS1 cells together with PROP1, in the presence of the 6xP3Luc plasmid that contains six copies of consen- sus P3 target sites common to PROP1 and HESX1 (5′-AGCTTGAGTCTAATTGAATTACTGTAC-3′, underlined) [8, 15], and pRL-CMV vector used as an internal control for the transfection. Forty eight hours after transfection, cells were analyzed using a dual- luciferase reporter assay system (Promega). Each result is representative of three independent experiments that yielded similar results. The results are expressed as mean ± standard error of the mean (SEM), and statisti- cal significance was determined byt -test.

Results Fig. 1 Identification of sequence variation of HESX1 A Partial sequence of PCR product and schematic Mutation screening diagrams of the HESX1 protein are shown. The chromatogram represents a heterozygous substitution of We identified a novel heterozygous HESX1 muta- glutamine (CGA) in place of arginine (CAA) at codon tion, c.326G>A (p.Arg109Gln), being the only gene 109. The red arrow indicates the mutated nucleotide. among 9 CPHD and 12 KS/HH-related genes for which Arginine 109 is located within the homeobox domain. unknown variants were identified. We used Sanger B Homology study showed arginine at codon 109 is sequencing of PCR products from genomic DNA to highly conserved through species in HESX1. C Three-dimensional structure of the paired confirm the HESX1 variant (Fig. 1A). Arginine 109 homeodomain and its target DNA (left panel). Modeled is located in the homeobox domain, which is critical structure of the p.Arg109Gln in comparison with the for binding to target DNA sequences, and is a highly wild-type structure (right upper panels). Modeling of evolutionarily conserved amino acid (Fig. 1B). We mutant was performed using a built-in mutagenesis function of the PyMOL Molecular Graphics System. checked the coverage of exon output from next-gener- Crystal structural modeling showed Arg109Gln HESX1 ation sequencing (NGS) as previously described [16], was predicted to lose a residue-DNA contact (arrow in and found no gross or exon-level deletions/duplica- right lower panel). 408 Takagi et al. tions in 9 CPHD and 12 KS/HH-related genes. The revealed no significant difference between wild type p.Arg109Gln was not detected in 150 healthy Japanese and mutant HESX1, indicating that expression level of controls and was absent from database, including protein and nuclear targeting was not affected by the dbSNP, the 1000 Genomes Project, ExAC database, mutation. Arginine 109 is a highly conserved amino Exome Variant Server, NHLBI Exome Sequencing acid located in the homeobox domain, suggesting that Project and the Human Genetic Variation Database substitution of arginine 109 to glutamine, which is pre- in Japanese. The mutation was submitted to in silico dicted to lose a residue-DNA contact, results in defec- analysis. The results exhibited that this mutation was tive interactions with DNA. Indeed, EMSA studies predicted to cause functional damage by PolyPhen-2 showed that the mutant HESX1 protein had abrogated http://genetics.bwh.harvard.edu/pph2/ (damage score DNA-binding affinity. Dasenet al. have proposed that 0.999, sensitivity 0.09 and specificity 0.99). Genetic HESX1 and PROP1 function as opposing transcrip- analyses showed that the father of the patient carried tion factors and that the carefully regulated temporal the same heterozygous HESX1 mutation. Evaluation sequence of their expression is vital for normal pitu- of the hormonal data for the father was refused. itary development [17]. HESX1 and PROP1 bind to the same DNA response elements, for example, the P3 Crystal structural modeling sequence or the PRDQ9 sequence [18], which includes The p.Arg109Gln HESX1 was predicted to lose a PROP1 and HESX1 binding site, and, in vitro, HESX1 residue-DNA contact (Fig. 1C). can antagonize PROP1 activation. Transactivation assay revealed that cotransfection of mutant HESX1 Functional studies with PROP1 showed no repression activity, clearly Western blotting and subcellular localization analyses demonstrating the deleterious effect of this mutation Western blotting and subcellular localization on HESX1 function. revealed no significant difference between wild type We used a NGS strategy to analyze 122 genes asso- and mutant HESX1 (Fig. 2A, B). ciated to congenital endocrine disorders. Recently, EMSA experiment increasing evidence shows that overlapping geno- Wild type HESX1 specifically bound to the DNA types/phenotypes exist between CPHD, SOD, and KS. and this binding was competed by 200-fold excess Indeed, mutations in FGFR1, FGF8, WDR11, PROKR2, cold competitor and anti-HESX1 antibody. In con- and KAL1, the genes responsible KS, have been iden- trast, mutant HESX1 had abrogated DNA-binding tified in a small number of CPHD and/or SOD [4-7, ability (Fig. 2C). 13, 19-20]. On the other hand, mutations in HESX1, Transactivation assay and SOX3 responsible for CPHD have been identified In COS1 cells, transfection of PROP1 resulted in a in a small number of hypogonadotropic hypogonadism 12 fold stimulation of luciferase activity. As expected, (HH) [19, 21]. These findings strongly suggest that cotransfection of wild type HESX1 with PROP1 the genetic overlap between CPHD, SOD, and KS/HH resulted in an inhibition of the PROP1-dependent acti- is significant, and the genetic basis of CPHD is grow- vation of the P3 sequence. Cotransfection of mutant ing further complex and heterogeneous. When mul- HESX with PROP1 showed no repression activity (Fig. tiple genes need to be analyzed for mutations simul- 2D). Repressing capacities measured by wild type- taneously, targeted sequence analysis of interesting mutant cotransfection (wild type 100 ng; mutant 100 genomic regions is an attractive approach. ng) were comparable with that derived from 100 ng The father of our patient, carrying the same mutation of wild type (Fig. 2D). This indicates that the mutant in a heterozygous manner, was of normal adult height HESX1 did not interfere with the repression of wild and clinically normal. Even though this report is not type in wildtype-mutant cotransfection. the first description of the wide phenotypic spectrum in heterozygous HESX1 mutation carriers [10], it is note- Discussion worthy that HESX1 mutation carriers can clinically present as normal, even though the mutation is non- We characterized a novel mutant (p.Arg109Gln) of functional. The phenotypic variation within the same the HESX1 transcription factor that is associated with pedigree could be partly due to the impact of other CPHD. Western blotting and subcellular localization genes that are important but have not been recognized A novel HESX1 mutation 409

Fig. 2 Functional characterization of p.Arg109Gln HESX1 A Western blotting analysis. Western blotting analysis showed that the expression of the mutant HESX1 was comparable to that of the wild type. B Subcellular localization analysis. For subcellular localization analyses, we visualized and photographed COS1 cells transfected with GFP-tagged HESX1 using a BZ-X700 fluorescence microscope, after mounting the cells in Vectashield-DAPI solution. The wild type and mutant HESX1 are localized to the nucleus. C EMSA experiments. Wild type HESX1 showed specific binding to the elements, which was competed by excess amount of (200 times) cold competitors and anti-HESX1 antibody. The mutant HESX1 showed abrogated DNA-binding ability. D Transactivation assay. In COS1 cells, transfection of PROP1 resulted in a 12 fold stimulation of luciferase activity. Cotransfection of wild type HESX1 with PROP1 resulted in an inhibition of the PROP1-dependent activation of the P3 sequence. Cotransfection of mutant HESX with PROP1 showed no repression activity. Repressing capacities measured by wild type-mutant cotransfection (wild type 100 ng; mutant 100 ng) were comparable with that derived from 100 ng of wild type, indicating that the mutant HESX1 did not interfere with the repression of wild type in wild type-mutant cotransfection. The data are mean ± SEM of at least three independent experiments performed in triplicate transfections. **P<0.01. in pituitary development. Furthermore, the possibil- tal disease course associated with mutations in HESX1. ity that heterozygous p.Arg109Gln HESX1 mutation could have no significancein vivo effect in human, and Acknowledgments that she had a de novo mutation in some other yet-to- be- identified gene causative for CPHD, could not be We thank Kazue Kinoshita for technical assis- excluded completely. tance. This work was supported by a Grant-in-Aid In summary, this study expands the range of known for the Health Science Research Grant for Research molecular defects in HESX1, and extends our under- on Applying Health Technology (Jitsuyoka (Nanbyo)- standing of the phenotypic features and developmen- Ippan-014 (23300102)) from the Ministry of Health, 410 Takagi et al.

Labour and Welfare of Japan, a grant from Yamaguchi Disclosure Statement Endocrine Research Foundation, and a grant from Takeda Science Foundation. The authors have nothing to disclose.

References 1. Romero CJ, Nesi-França S, Radovick S (2009) The tion of HESX1 causing pituitary and optic nerve hypo- molecular basis of hypopituitarism. Trends Endocrinol plasia and combined pituitary hormone deficiency in a Metab 20: 506-516. Japanese patient. J Clin Endocrinol Metab 88: 45-50. 2. Kelberman D, Rizzoti K, Lovell-Badge R, Robinson 12. McNay DE, Turton JP, Kelberman D, Woods KS, IC, Dattani MT (2009) Genetic regulation of pituitary Brauner R, et al. (2007) HESX1 mutations are an gland development in human and mouse. Endocr Rev uncommon cause of septooptic dysplasia and hypopitu- 30: 790-829. itarism. J Clin Endocrinol Metab 92: 691-697. 3. Pfäffle R, Klammt J (2011) Pituitary transcription fac- 13. Takagi M, Narumi S, Hamada R, Hasegawa Y, tors in the aetiology of combined pituitary hormone Hasegawa T (2014) A novel KAL1 mutation is asso- deficiency. Best Pract Res Clin Endocrinol Metab 25: ciated with combined pituitary hormone deficiency. 43-60. Variation (in press). 4. Raivio T, Avbelj M, McCabe MJ, Romero CJ, Dwyer 14. Cohen RN, Cohen LE, Botero D, Yu C, Sagar A, et AA, et al. (2012) Genetic overlap in Kallmann syndrome, al. (2003) Enhanced repression by HESX1 as a cause combined pituitary hormone deficiency, and septo-optic of hypopituitarism and septooptic dysplasia. J Clin dysplasia. J Clin Endocrinol Metab 97: E694-E699. Endocrinol Metab 88: 4832-4839. 5. Reynaud R, Jayakody SA, Monnier C, Saveanu A, 15. Brickman JM, Clements M, Tyrell R, McNay D, Woods Bouligand J, et al. (2012) PROKR2 variants in multiple K, et al. (2001) Molecular effects of novel mutations in hypopituitarism with pituitary stalk interruption. J Clin Hesx1/HESX1 associated with human pituitary disor- Endocrinol Metab 97: E1068-E1073. ders. Development 128: 5189-5199. 6. McCabe MJ, Gaston-Massuet C, Gregory LC, 16. Mitsui T, Narumi S, Inokuchi M, Nagasaki K, Alatzoglou KS, Tziaferi V, et al. (2013) Variations in Nakazawa M, et al. (2014) Comprehensive next-gener- PROKR2, but not PROK2, are associated with hypopi- ation sequencing analyses of hypoparathyroidism: iden- tuitarism and septo-optic dysplasia. J Clin Endocrinol tification of novel GCM2 mutations. J Clin Endocrinol Metab 98: E547-E557. Metab 99: E2421-2428. 7. Fukami M, Iso M, Sato N, Igarashi M, Seo M, et al. 17. Dasen JS, Martinez Barbera JP, Herman TS, Connell (2013) Submicroscopic deletion involving the fibro- SO, Olson L, et al. (2001) Temporal regulation of a blast growth factor 1 gene in a patient with paired-like homeodomain repressor/TLE corepressor combined pituitary hormone deficiency. Endocr J 60: complex and a related activator is required for pituitary 1013-1020. organogenesis. Genes Dev 15: 3193-3207. 8. Dattani MT, Martinez-Barbera JP, Thomas PQ, 18. Sornson MW, Wu W, Dasen JS, Flynn SE, Norman Brickman JM, Gupta R, et al. (1998) Mutations in the DJ, et al. (1996) Pituitary lineage determination by the homeobox gene HESX1/Hesx1 associated with septo- Prophet of Pit-1 homeodomain factor defective in Ames optic dysplasia in human and mouse. Nat Genet 19: dwarfism. Nature 384: 327-333. 125-133. 19. Izumi Y, Suzuki E, Kanzaki S, Yatsuga S, Kinjo S, et 9. Carvalho LR, Woods KS, Mendonca BB, Marcal N, al. (2014) Genome-wide copy number analysis and sys- Zamparini AL, et al. (2003) A homozygous mutation in tematic mutation screening in 58 patients with hypogo- HESX1 is associated with evolving hypopituitarism due nadotropic hypogonadism. Fertil Steril 102: 1130-1136. to impaired repressor-corepressor interaction. J Clin 20. McCabe MJ, Hu Y, Gregory LC, Gaston-Massuet C, Invest 112: 1192-1201. Alatzoglou KS, et al. (2015) Novel application of lucif- 10. Thomas PQ, Dattani MT, Brickman JM, McNay D, erase assay for the in vitro functional assessment of Warne G, et al. (2001) Heterozygous HESX1 mutations KAL1 variants in three females with septo-optic dys- associated with isolated congenital pituitary hypoplasia plasia (SOD). Mol Cell Endocrinol 417: 63-72. and septo-optic dysplasia. Hum Mol Genet 10: 39-45. 21. Newbern K, Natrajan N, Kim HG, Chorich LP, Halvorson 11. Tajima T, Hattorri T, Nakajima T, Okuhara K, Sato K, LM, et al. (2013) Identification of HESX1 mutations in et al. (2003) Sporadic heterozygous frameshift muta- Kallmann syndrome. Fertil Steril 99: 1831-1837.