Fujian Medical University between June 2016 and February 2024 were identified by Whole Exon Sequencing (WES). This study, conducted in accordance with the Declaration of Helsinki, received approval from the Ethics Committee of Fuzhou Children’s Hospital of Fujian Medical University (approval number 202310). Prior to participation, written informed consent was obtained from the parents of the patients.

Inclusion criteria: (1) The measured height of the children (measured body length of children under 3 years old) was lower than the mean third percentile (-2 SDS) of the population of the same race, age and sex; (2) WES testing; (3) Growth hormone stimulation tests were performed. Patients with a peak GH value < 10 ng/mL were classified as GHD (9); (4) The patient had complete medical records.

Exclusion criteria: (1) hypothalamic-pituitary dysfunction caused by organic lesions such as intracranial tumors and cranial trauma; (2) congenital hypothyroidism; (3) chromosome diseases; (4) inherited metabolic bone diseases, such as congenital chondrodysplasia and mucopolysaccharidosis; (5) other chronic diseases, such as congenital heart disease, chronic hepatitis, henoch-schonlein purpura nephritis, and inflammatory bowel disease.

Detailed clinical data, including age, gender, chief complaint, physical chararteristics, as well as medical and family histories were systematically collected. Growth hormone stimulation tests utilized L-dopa (10 mg/kg) and arginine (0.50g/kg). Serum levels of thyroid hormones, GH, IGF-1, and IGFBP3 were measured using chemiluminescent immunoassay. Bone age was determined using Siemens’ direct digital radiography machine and Agfa’s computer X-ray system, with assessment conducted via the Tanner-Whitehouse III (TW3) method (10). The standard deviation scoreof IGF-1 was calculated relative to levels in healthy children of the same age and gender (11). Throughout the rhGH treatment, patients underwent follow-up examinations every three to six months.

A total of 2 mL of venous blood collected in EDTA anticoagulant was obtained from the proband and their family members. Genomic DNA extraction from peripheral blood followed standard procedures, and whole exome sequencing was conducted using established protocols. The GH1 gene was referenced against the National Center for Biotechnology Information (NCBI) entry NG_011676.1 (NM_000515.5). Pathogenicity analysis of genetic variants was performed in accordance with guidelines from the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (12, 13). Validation and segregation analysis of candidate gene variants were carried out through Sanger sequencing.

Conservation of amino acid substitution positions across species was assessed using ClustalX 1.83 software (14). The pathogenicity of the GHW112R variant was predicted using the PREDICTSNP web server (https://loschmidt.chemi.muni.cz/predictsnp1/), which integrates multiple predictors (PredictSNP, MMAP, PhD-SNP, polyphen1, polyphen2, SIFT, and SNAP) with a confidence score (15).

For the modeling of human GH, the crystal structure model of the wild-type GH (amino acid: 27-217) (PDB entry: 1HUW) was used. Mutant proteins were constructed using the Iterative Threading ASSEmbly Refinement (I-TASSER) server (https://seq2fun.dcmb.med.umich.edu/I-TASSER/) (16). The HOPE server (https://www3.cmbi.umcn.nl/hope/) analyzed the effect of variation on the three-dimensional structure of GH (17). SWISS-MODEL online software (https://swissmodel.expasy.org/) compared the three-dimensional structural alterations between wild-type and mutant GH. The DynaMut server (http://biosig.unimelb.edu.au/dynamut/) was employed to predict the interaction between amino acid residues and analyze protein stability and flexibility. Finally, PyMOL 2.5 was used to visualize the modeled proteins.

IGHD II patients were additionally identified through a comprehensive search of the PubMed database for published clinical cases utilizing keywords such as “isolated growth hormone deficiency”, “GH1 gene”, or/and “growth hormone”.

Based on the distribution of variables, we presented them in this study as mean ± standard deviation for normally distributed variables, or median (min-max) for skewed distributions. Categorical data are typically displayed in the form of both numerical counts (n) and corresponding percentages (%). The One-way Analysis of Variance (ANOVA) test was employed to assess the differences in height SDS and diagnosis age among patients with various GH1 heterozygous variants, utilizing the statistical software SPSS.27. The chart was created using GraphPad Prism 10.1.2 software. * indicates p < 0.05, ** indicate p < 0.01, *** indicate p < 0.001, and **** indicate p < 0.0001.

792 unrelated short stature patients ascertained in Fujian Province, China, were genetically evaluated by WES–based test. We identified three novel pathogenic variants of the GH1 gene, two recurrent pathogenic variants, and one novel likely pathogenic variant in this cohort. The variants in patients 2, 4, 5, and 6 were inherited from their affected parents, whereas the variants in patients 1 and 3 were de novo.

Four patients (Patients 1-4) underwent rhGH treatment at a dosage of 0.10~0.16 IU/kg/d (Figure 1; Table 2 ). The mean age at the initiation of treatment was 4.5 ± 0.7 years. In the initial year of treatment, the height standard deviation score (ΔHtSDS) increased by 1.79 ± 0.15 SDS. Patient 1, who commenced rhGH treatment at 3.8 years old, experienced a height increase to -0.19 SDS (ΔHtSDS: 2.61 SDS) after 2.3 years of treatment. Subsequently, he suspended rhGH treatment for 3.25 years, returning at 9.4 years of age with a height of 130.4cm (-1.12 SD). To achieve a more favorable height, he resumed rhGH treatment for 4.5 years, reaching a height of 169.5cm (+0.50 SDS) at 13.9 years, after which the treatment was discontinued. At the last follow-up at 16 years, the child measured 175.0cm (+0.54 SDS), surpassing the genetic target height of 169.5cm. Patient 2, treated with rhGH at 5.5 years, initially measured 91.1cm (-5.24 SDS) and achieved a height of -2.02 SDS (ΔHtSDS: 3.22 SDS) after 3 years of continuous treatment. Patient 3, treated at 4.1 years, started at 87.1cm (-4.51 SDS) and reached -2.12 SDS (ΔHtSDS: 2.37 SDS) after 1 year of treatment. Patient 4, treated at 4.9 years, started at 84.5cm (-6.09 SDS) and reached 94.5cm (-4.65 SDS, ΔHtSDS: 1.44 SD) after 6 months of treatment. Patient 5 and 6 had not received GH treatment due to economic reasons. Regular follow-ups during treatment indicated normal levels of blood and urine parameters, calcium and phosphorus metabolism, blood glucose, blood lipids, liver and kidney functions, and IGF-1 within the normal range. Detailed clinical information is provided in Table 2 .

A total of six GH1 gene pathogenic/likely pathogenic variants was detected, thus, the overall prevalence of pathogenic GH1 variants in this cohort was 0.76% (6/792). Among them, four variants were novel. Exome sequencing and quantitative PCR revealed a heterozygous deletion of exons 2-5 in the GH1 gene in Patient 1 ( Figure 2A ). The variant was determined to be de novo, as it was not present in either parent. Sanger sequencing confirmed that Patient 2 carried a heterozygous missense variant of the GH1 gene (c.334T>C, p.Trp112Arg), inherited from the affected father ( Figure 2B ). Patient 3 and Patient 4 carried heterozygous splicing site variants (c.291 + 1G>A, p.Glu58Asnfs*120 and c.291 + 2T>A, p.Glu58Asnfs*120), respectively. The variant in Patient 3 was de novo and the variant in Patient 4 was inherited from affected mother ( Figures 2C, D ). Additionally, the maternal grandmother of Patient 4 also carried the variant. Interestingly, the patient 4’s father carried a homozygous variant of GHRHR gene (c.659T>C;p.Leu220Pro), patient 4 was a heterozygous carrier. According to the ACMG guideline, it is a variants of unknown significance (PP3+PM3_Supporting+PM2). Homozygous variants in the GHRHR gene lead to isolated growth hormone deficiency type IV (MIM#618157), which may be the reason why the father of patient 4 does not carry the GH1 gene mutation but has severe short stature clinical manifestations. Patient 5 carried a large fragment heterozygous deletion of approximately 1.5 kb, containing only the GH1 gene (ex.1_5del), inherited from her mother, and her maternal grandmother (145cm, -3SDS) also carried the same heterozygous deletion ( Figure 2E ). Furthermore, we identified an approximately 1.7 kb heterozygous deletion in patient 6, encompassing only the GH1 gene (ex.1_5del). Her father carried the same deficiency ( Figure 2F ). In summary, we reported six variants of the GH1 gene, four were novel (exon2-5del, c.334T>C, 1.5 kb deletion of GH1, and 1.7 kb deletion of GH1) ( Table 3 ).

To further assess the pathogenicity of the GH^W112R variant, we analyzed interspecific conservation and pathogenicity, comparing GH protein sequences across 12 vertebrate species. Supplementary Figure 1 illustrates the comparison of GH protein sequences in representative vertebrates, such as Homo sapiens, Bos taurus, Ovis aries, Sus scrofa, Felis catus, Equus caballus, Oryctolagus cuniculus, Mus musculus, Monodelphis domestica, Macaca mulatta, Anguilla japonica, and Oncorhynchus mykiss. Notably, the Trp112 residue of the GH protein exhibited high conservation across these represented vertebrates ( Supplementary Figure 1C ). The novel missense variant of GH1, c.334T>C,p.Trp112Arg, was predicted to be deleterious by seven different predictors (PredictSNP, MMAP, PhD-SNP, polyphen1, polyphen2, SIFT, and SNAP) with confidence scores of 0.87, 0.88, 0.88, 0.74, 0.81, 0.53, and 0.89, respectively. These findings strongly suggest the potential pathogenicity of the c.334T>C,p.Trp112Arg variant in GH1.

In the context of homology modeling and structural analysis of human GH, molecular visualization images for GH and growth hormone receptor (GHR) complexes were generated using PyMOL 2.5 ( Figures 3A, B ). GH comprises four inverse parallel α-helices, and their precise spatial arrangement is crucial for GH-GHR binding. The GH^W112R variant was modeled on the I-TASSER online server, revealing the substitution of tryptophan at position 112 with arginine. This substitution significantly reduces the localized flexibility of the α-helix (ΔΔSVib ENCoM: -3.356 kcal·mol^-1·K^-1) ( Figure 3C ). Multiple hydrophobic bond interactions were identified between the surrounding amino acid residues (Phe80, Pro115, Leu119, Leu188, Cys191, Cys192) and Trp112, contributing to the stabilization of relative spatial positions between adjacent α-helices ( Figure 3E ). The highly conserved tryptophan was predicted to be buried in the α-helix, and its replacement by charged arginine was expected to disrupt the local helix structure. The GH^W112R variant interfered with the local hydrophobic bond of amino acid residue 112 and disrupted the interaction between α-helix II and α-helix IV ( Figures 3D, F ).

Including the six children reported in this study, a total of 180 individuals with GH1 gene variants and IGHD II had been reported. The affected individuals most commonly present with isolated short stature, and some have other clinical manifestations, including a prominent or bossing forehead (n=21/180, 11.7%), depressed nasal bridge (n=12/180, 6.7%), obesity (n=7/180, 3.9%), medial face hypoplasia (n=5/180, 2.8%), small penis (n=4/180, 2.2%), undescended testis (n=3/180, 1.7%), high-pitched voice (n=3/180, 1.7%), and neonatal hypoglycemia (n=3/180, 1.7%). The specific phenotypes of all reported patients with other clinical manifestations besides short stature are listed in Supplementary Table 1 .

We conducted a comparative analysis of height SDS in children with various heterozygous variants of the GH1 gene, specifically splice variants (n=62), missense variants (n=73), and large segment deletions (n=3). Following the application of pairwise comparison with Bonferroni-corrected significance levels, it was determined that the height SDS change in patients with GH1 heterozygous splicing variants was significantly lower compared to patients with GH1 heterozygous missense variants and large segment deletion (P=0.0002 and P=0.0104, respectively) ( Figure 4A ). Furthermore, individuals harboring the splice variant exhibited the earliest age of diagnosis, with a median (interquartile range) of 2.2 (1.0-4.3) years, potentially attributable to their severe short stature ( Figure 4B ).