Mutations Within the Transcription Factor PROP1 in a Cohort of Turkish Patients with Combined Pituitary Hormone Deficiency

Objective: Mutations of the genes encoding transcription factors which play important roles in pituitary morphogenesis, differentiation and maturation may lead to combined pituitary hormone deficiency (CPHD). PROP1 gene mutations are reported as the most frequent genetic aetiology of CHPD. The aim of this study was to describe the phenotypes of Turkish CPHD patients and define the frequency of PROP1 mutations. Methods: Fifty-seven CPHD patients from 50 families were screened for PROP1 mutations. The patients were affected by growth hormone (GH) and additional anterior pituitary hormone deficiencies. Results: All patients had GH deficiency. In addition, 98.2% had central hypothyroidism, 45.6% had hypogonadotropic hypogonadism, 43.8% had adrenocorticotropic hormone deficiency and 7.1% had prolactin deficiency. Parental consanguinity rate was 50.9% and 14 cases were familial. Mean height standard deviation score (SDS) and weight SDS were -3.8±1.4 and -3.1±2.0, respectively. Of 53 patients with available pituitary imaging, 32 (60.4%) showed abnormalities. None had extra-pituitary abnormalities. Eight index patients had PROP1 gene mutations. Five sporadic patients were homozygous for c.301_302delAG (p.Leu102CysfsTer8) mutation, two siblings had exon 2 deletion, two siblings had complete gene deletion and two siblings were homozygous for the novel c.353A>G (p.Q118R) mutation. The frequency of the PROP1 mutations was 16% in our cohort. Mutation rate was significantly higher in familial cases compared to sporadic cases (42.8% vs 11.6%; p<0.01). Conclusion: Phenotype of patients regarding hormonal deficiencies, pituitary morphology, presence of extra-pituitary findings, family history of CPHD and parental consanguinity are important for deciding which pituitary transcription factor deficiency should be investigated. PROP1 mutation frequencies vary in different populations and its prevalence is high in Turkish CPHD patients.

Both in human and mice, pituitary organogenesis and maintenance of its proper functioning necessitates the appropriate expression of a cascade of signalling molecules and transcription factors which are crucial for organ commitment, cell proliferation, patterning and terminal differentiation (2,3,4).
The genes that are related to these transcription factors are PROP1, POU1F1 (PIT1), LHX3, LHX4, and HESX1. In 1998, Wu et al (5) identified homozygous or compound heterozygous inactivating mutations of PROP1 gene as being associated with CPHD. To date, PROP1 mutations are the most commonly reported genetic aetiology of CPHD in humans (4,5,6). Prophet of PIT-1 (PROP1) is a pairedlike homeobox 1 gene, located on chromosome 5q35. 3 and consists of three exons encoding for a 226-amino acid protein which is a late-expressed transcription factor (4). Mutations of PROP1 gene cause autosomal recessively inherited CPHD and clinical phenotypes include GH, TSH, FSH/LH, PRL and rarely ACTH deficiencies together with morphological pituitary anomalies (4,7).
Phenotypes associated with PROP1 gene mutations can be highly variable. Deficiencies of all pituitary hormones may be present with varying severity and at different ages. However, in all cases anterior pituitary function deteriorates over time (4,8). TSH and GH deficiencies have a tendency to occur in early childhood, whereas gonadotropin and corticotropin deficiencies manifest later in life (4,8). As PROP1 is a "later-acting transcription factor", extra-pituitary manifestations are not observed (8). Magnetic resonance imaging (MRI) of the anterior pituitary gland shows normal or enlarged gland in early stages and pituitary involution in later stages whilst size and location of the posterior pituitary is normal and pituitary stalk interruption is not observed (1,4). Rarely, pituitary masses associated with PROP1 gene mutations are reported (4,9,10). Point mutations, small and large deletions and insertions in the PROP1 gene have been reported but there are no associations between any specific variants and specific regions or ethnicities (11,12,13).
The aim of this study was to define patient characteristics and to identify PROP1 gene mutations in our CHPD patient cohort.

Study Design and Patient Selection
This retrospective cohort study was conducted in Çukurova University Research and Education Hospital including 57 patients with combined anterior pituitary hormone deficiency who attended the hospital between January 1997 and August 2019. Exclusion criteria for the patients were isolated GH deficiency (GHD), brain tumour, central nervous system surgery, cranial-neck irradiation, systemic chronic illnesses or chromosomal abnormalities.
Patients who were diagnosed with CPHD were analysed for PROP1 mutations. The patients included in the study were all affected by GHD and at least one additional anterior pituitary hormone deficiencies including TSH, gonadotropins, ACTH or PRL. Diagnosis was based on clinical, laboratory and imaging investigations. Serum GH, insulin-like growth factor-1 (IGF-1), IGF-binding protein-3 and plasma ACTH concentrations were analysed by commercial kits using Siemens immulite 2000 immunoassay system and FSH, LH, oestradiol, cortisol, TSH, free thyroxine and PRL were analysed by commercial kits using Beckman Coulter Unicel Dxl 800 immunoassay system based on electrochemiluminescence immunoassay.
Genomic DNA was isolated from peripheral leucocytes. PROP1 gene (transcript ID: ENST00000308304.2 and protein ID: O75360) was screened by polymerase chain reaction (PCR) amplifications of exons and neighbouring intronic regions. The PCR products were purified and directly sequenced using the Big Dye terminator cycle sequencing ready reaction kit (PE Applied Biosystems, Foster City, Calif., USA) in an ABI PRISM 3130 automatic sequencer (PE Applied Biosystems, Foster City, Calif., USA). DNA sequence data analyses were evaluated with DNA Sequencing Analysis Software-Sequencher 5.0 programme (http://genecodes. com/). All of the variants were investigated using 1000 genomes browser database (http://browser.1000genomes. org/index.html) and the National Center for Biotechnology Information database (https://www.ncbi.nlm.nih.gov/ clinvar) as to whether they were novel or previously reported. Subsequently, mutant variants were interpreted by in silico prediction tools such as Mutation Taster, SIFT and PolyPhen-2 (14,15,16).
Segregation analysis was performed only for the family of patients with the novel variant. It was not possible to test the parents of the patients with known pathogenic variants due to financial limitations. patients from two families gave consent for further testing and multiplex ligation dependent probe amplification (MLPA) assays were performed only for these patients. The other five patients were not included in the calculation of mutation frequency.
Patients from the same family are indicated with the same superscript letter.
The Ethics Committee of the Çukurova University Faculty of Medicine approved this study (approval: #TF2013LTP24), and written informed consent was obtained for each patient or from their legal guardians.

Statistical Analysis
Data obtained from this study were analysed using SPSS statistical software, version 23.0 (IBM Inc., Armonk, NY, USA). The distribution of data was evaluated with the Kolmogorov-Smirnov test. For numerical comparisons, the independent sample t-test or Mann-Whitney U tests were used for parametric and non-parametric distribution of the measured parameters, as appropriate. Descriptive statistics which were not normally distributed were presented as median and range. Frequency distributions and percentages were given for categorical variables.

Results
All 57 of the patients included to the study were affected by GHD and diagnosed in childhood. In addition, 56 patients (98.2%) had central hypothyroidism, 26 (45.6%) had hypogonadotropic hypogonadism, 25 (43.8%) had ACTH deficiency and four (7.1%) had PRL deficiency. More than two-thirds of the patients were male (68.4%). Median age age at diagnosis was 7.7 years (range: 3 months-19.8 years). Mean delay in bone age at diagnosis was 3.3±2.4 years. More than half of the patients (n=29; 50.9%) had parental consanguinity and 14 patients were familial cases. There was no history of perinatal asphyxia or difficult birth. None of the patients had any major dysmorphic findings. Height standard deviation score (SDS) at diagnosis was -3.8±1.4. Weight SDS at diagnosis was -3.1±2.0. IGF-1 SDS at diagnosis was -3.0±1.5). Median age at the start of GH replacement treatment was 8.5 years (range: 3 months -20 years). All of the patients received appropriate treatments for their hormonal deficiencies. Twelve patients achieved their final height and mean final height SDS for these cases was -1.0±0. 7 PCR amplification of second and third exons of PROP1 gene had failed for DNA of patients 1 a -2 a , 7 d -8 d , 9 e -10 e and 15 all of whom had parental consanguinity. No pathogenic mutations were detected within exon 1 for these patients. MLPA assay was only performed for patients 1 a and 2 a from the same family and a homozygous deletion of exon 2 of the PROP1 gene were detected in both siblings. These two brothers have GHD at the time of diagnosis and developed TSH deficiency after approximately one or two years. Both had delayed pubertal development and lack of secondary male sex characteristics due to hypogonadotropic hypogonadism. Eventually, both developed ACTH deficiency (Table 1).
PCR amplification of the whole PROP1 gene had failed for DNA from patients 58 f -59 f . MLPA assays detected complete gene deletion in these siblings. The elder sister showed GHD at the age of two-and-a-half years and developed TSH deficiency four years later. When she reached the age of puberty, she developed both ACTH deficiency and hypogonadotropic hypogonadism. Her younger brother showed both TSH and GH deficiencies at diagnosis; he is currently prepubertal and is not affected by ACTH insufficiency (Table 1). Pituitary imaging revealed pituitary adenoma in patient 58 f but was normal in 59 f . Adenoma did not exhibit progression and remained stable.
Patients 3 b and 4 b from the same family with the same phenotype had homozygous c.353A>G (p.Q118R) variant in exon 3 of the PROP1 gene (Figure 1). This novel variant was predicted to be disease-causing by in silico predictive

Discussion
In this study, PROP1 gene mutations were detected in eight index patients from a cohort of 57 CPHD patients from 50 families. Segregation analysis of the variants in the pedigrees revealed three patients with the same pathogenic PROP1 mutations. More than half of the patients with mutation were familial cases and positive mutation frequency was significantly higher in familial cases compared to sporadic cases (3/7 familial cases versus 5/43 sporadic cases, p<0.01).
There are several reports of cohorts defining genetic aetiology of CPHD from different parts of the world. PROP1 gene mutations are reported to be the most frequent amongst both sporadic and familial CPHD patients (4,6,8,17). However, the frequency was reported to vary widely between 0% and 70.1% from different populations (10,18,19,20,21). PROP1 mutation frequencies among CPHD patients are highest in Eastern European populations especially Lithuanian, Polish and Hungarian, and also high in Portuguese, Russian and Brazilian cohorts (3,10,12,22,23,24,25,26,27,28). In contrast, PROP1 mutation rates are usually low in Western and Southern European countries, Australia and in cases with Asian origin, especially in sporadic CPHD patients (3,6,18,19,20,21,29). PROP1 gene mutations are not rare among Turkish CPHD patients (13,30 PROP1 mutation prevalence is higher in familial patients compared to sporadic cases in all cohorts (3,6,13,22,24,26,29,30,31,32). Parental consanguinity is known to increase the risk for autosomal recessive conditions. Thus, parental consanguinity would appear to be a risk factor for PROP1 mutations (12,22,30). This hypothesis is supported by evidence from our study, with an overall parental consanguinity rate of 50.9% which increased to 81.8% amongst PROP1 mutated patients. If the cases are sporadic, that is that there is a single affected individual in a family, and there is no parental consanguinity, the aetiology is more likely to be acquired rather than genetic (1,3,4).
The c.301_302delAG mutation was reported to be one of the most prevalent mutations of PROP1 gene (2,8,10,11,26). This mutation is a two base pair deletion which results in a frameshift and early termination of the protein at codon 109. Dusatkova et al (2) investigated this variant and suggested that the reason for the high occurrence rate may be a founder effect rather than a variant hot spot (2). This assumption was made by haplotype analyses and the geographic distribution of the c.301_302delAG variant which was interpreted as suggesting an ancestral origin.  For this reason, MLPA analysis should be a routine part of genetic investigation in MPHD patients.
The previously unreported p.Q118R substitution is interpreted as likely pathogenic considering the concordance of phenotype, parental consanguinity and segregation analyses of the variant. This variant is anticipated to be important as it is highly conserved in different orthologues. In addition, it is located in the homeobox domain (5). In 1998, Wu et al (5), identified p.F117I and p.R120C substitutions and they postulated that these variants allowed protein binding but with reduced affinity. As, p.Q118R variant is present in the sequence between these variants, it is assumed that this variant is also associated with pathogenicity due to altered protein function. In silico analyses with Mutation Taster, SIFT and PolyPhen-2 also indicate likely alteration of protein features and splice site changes (14,15,16).

Study Limitations
Five of the patients with failed PCR amplification were not available for further testing with MLPA analysis. There is a high probability that a large deletion may exist in the PROP1 gene of these familial CPHD cases with parental consanguinity which would have increased the proportion of MPHD patients with PROP1 mutations in our cohort. In this study, we were not able to test the parents of all patients with pathogenic mutations due to financial limitations. For future studies, patients without any mutations identified in the PROP1 gene may be screened for the other genes of pituitary transcription factors and gene panels may be more cost-effective for this purpose.

Conclusion
It is crucial to screen GHD patients regularly for other anterior pituitary hormone deficiencies. With the exact genetic aetiology, the family is able to receive genetic counselling, unnecessary laboratory testing can be avoided and at the same time the opportunity of predicting the typical phenotype and developing hormonal deficiencies can be detected earlier. If the patients are familial and have parental consanguinity, genetic testing would be even more cost-effective.