Novel pyrroline-5-carboxylate reductase 2 (PYCR2) mutation in an Iranian patient with hypomyelinating leukodystrophy: findings of molecular and in silico investigations

Hypomyelinating leukodystrophy (HLD) is a specific group of leukodystrophies and is characterized by progressive postnatal growth delay that represents a type of clinically overlapping but genetically heterogeneous diseases with autosomal recessive inheritance. Loss-of-function mutations in PYCR2 are one of the main causes of HLD type 10 (HLD10), which is identified by cerebral hypomyelination, inadequate growth, brain atrophy, and movement abnormality. This study aimed to investigate the molecular etiology of HLD10 disorder in an Iranian patient from a consanguineous marriage family. The DNA samples were extracted from the patient, a 9-year-old girl, and her parents. Whole-exome sequencing was conducted for these samples and the results were eventually confirmed and segregated via Sanger sequencing. Our findings demonstrated a novel homozygous frameshift mutation in PYCR2 gene, c.135dup (NM_013328.4). The heterozygous state of this variant was confirmed in parents. Additionally, this mutation was predicted to exhibit damaging effects through protein sequence alteration. Such findings are of importance for understanding the underlying pathogenicity mechanisms and for improving genetic counseling knowledge of HLD patients for families.


Background
Hypomyelinating leukodystrophy (HLD) is a specific group of leukodystrophies and is characterized by progressive postnatal growth delay that represents a type of clinically overlapping but genetically heterogeneous diseases with autosomal recessive inheritance [1]. These patients mainly suffer from symptoms such as cerebral hypomyelination and severe psychomotor developmental delayed with absent speech. In addition, the hyperreflexia, appendicular hypertonia with hyperextensibility of the wrists and ankles, severe muscle wasting, axial hypotonia, and failure to thrive are other reported characteristics of HLD10. This disorder affects one out of every 200,000 to 500,000 individuals and is classified as a rare disease [2]. So far, multiple types of HLD have been recognized based on distinct genes that carry pathogenic variants [1,[3][4][5][6].
Hypomyelinating leukodystrophy type 10 (HLD10) is caused by loss-of-function mutations in Pyrroline-5-Carboxylate Reductase 2 (PYCR2) and its phenotype is identified by cerebral hypomyelination, inadequate growth, brain atrophy, and movement abnormality [7]. Although three isozymes of PYCR -family have been discovered in human (PYCR1-3), only PYCR1 and PYCR2 contribute to human abnormalities. Dysfunction of PYCR2 results in proline deficit as it is responsible for encoding the mitochondrial enzyme pyrroline-5-carboxylate reductase 2 engaged in synthesis of proline, which is vital for central nervous system (CNS) [7][8][9][10]. Importantly, defects in proline synthesis have major deleterious effects on health, resulting in intellectual disability, skin and joint hyperelasticity, osteopenia, and cataracts [11].
Recent nucleotide sequencing technologies with nextgeneration sequencing (NGS) have enabled us to identify both known and novel HLD-responsible genes and variants [2]. NGS has revolutionized medical genetics by improving the chances of receiving a molecular diagnosis for rare genetic diseases. Whole-exome sequencing (WES), as one of the widely used NGS approaches in research and clinical setups, has demonstrated an unprecedented success rate in the recognition of diseasecausing genes in projects such as discovery of the molecular basis of a recognizable syndrome in a homogeneous group of patients and also in pan-genomic sequencing of large heterogeneous cohorts [12]. Moreover, it has been successfully used, as an edge-cutting approach, to study the protein-coding and significant flanking regions of the genome contributing to more than 85% of the monogenic disorders. This high-throughput technique helps to reveal the disease-causing variants with low cost and high resolution in comparison with other diagnostic molecular methods such as single locus-based and panel-based sequencings in a wide range of inherited diseases [13].
In conclusion, since to the best of our knowledge no molecular investigation has been performed on patients with HLD10 in Iran, our goal was to conduct WES analysis for an Iranian patient with intellectual disability and failure to thrive in order to identify the disease-causing variant(s). The results were confirmed and assessed by Sanger sequencing and in silico approach, respectively.

Case presentation
In this study, a clinical examination was carried out on a 9-year-old girl from an Iranian family with consanguineous marriage. The pedigree of the family is shown in Fig. 1 and was drawn based on National Society of Genetic Counselors (NSGC) recommendations [14]. The patient was referred to the hospital for further evaluations and the phenotypes were CP, growth delay, chronic constipation, abnormality of movement without support, hydrocephalus, and tooth abnormalities. All general tests of the patient, including Complete Blood Count, Urine Analysis, Calcium, Phosphorous, Alkaline phosphatase, Total Vitamin D, and cytogenetic banding analysis, were conducted. Written informed consents were obtained from parents, and the study was approved by the Ethics Committee of the Pasteur Institute of Iran.

DNA extraction and whole-exome sequencing
Peripheral blood was collected from the patient for the extraction of genomic DNA. This process was performed using DNA extraction kit (QIAGEN, Hilden, Germany). The DNA quantity and purity were examined by spectrophotometric measurement of absorbance at 260 and 280 through Thermo Scientific ™ NanoDrop ™ One Microvolume UV-Vis spectrophotometer (Thermo Scientific, Waltham, MA, USA). The DNA samples were analyzed using electrophoresis for verification on 1% agarose gel, and the samples were stored at − 80 °C freezer before sequencing analysis [12].
The obtained DNA was sequenced via WES analysis. The coding exons including genes with known clinical implication were captured by SureSelect Human All Exon V7 kit (Agilent Technologies, Santa Clara, CA, USA). Genomic DNA was fragmented to 151 bp and libraries were produced with Illumina NovaSeq 6000 System (San Diego, CA) in order to generate paired-end reads with the average coverage depth of 100X. The quality assessment of preliminary FASTQ was done by providing quality control (QC) with FastQC, as well as NGS QC Toolkit softwares. Burrows-Wheeler Aligner (BWA) algorithm was applied for alignment of the reads to human genome reference (GRCh37/hg19). Outcomes were, then, assembled in a Sequence Alignment Mapping (SAM) file format. Pic-ardTools was employed for Post-Alignment steps, along with SamTools and Genome Analysis Toolkit version 4 (GATKv4) packages in order to remove duplications. Afterward, the variant calling procedure, comprising both SNV and INDEL variants, was conducted via Haplotype-Caller for production of variant call format (VCF) files of the patient. The quality of attained VCF files was evaluated by commands of variant quality score recalibration In line, the variant filtering and prioritization strategies including total read depth of more than 5X and minor allele frequency (MAF) < 5% were utilized. In-house command lines recruiting compound heterozygous and homozygous variants were applied given the pedigree and phenotype description. Then, by filtering out the synonymous, intronic, and intergenic variants, in addition to selecting the variants in phenotype-related genes, the remaining variants were investigated regarding their pathogenicity. As a supplementary evaluation, the sum of in silico predictions was taken into account in order to consider a pathogenic interpretation for variants. Importantly, the comprehensive classification of variant pathogenicity was adjusted based on American College of Medical Genetics (ACMG) guidelines [15].

Variant confirmation and segregation
Sanger sequencing was used to validate and segregate the detected final variants in the patient and family members, respectively. PCR primers were considered to amplify sequence of the PYCR2 gene. Primers were designed by Primer3 software and checked via Primer-BLAST NCBI (Table 1).

Results
The cytogenetic analysis was performed on peripheral blood lymphocytes and a normal female 46,XX karyotype was found. Gas chromatographic-mass spectrometric screening for Organoacidophathies in urine was performed and no metabolites specific for organoacidopathy were detected. The magnetic resonance imaging (MRI) in the patient suggested hypomyelination, central brain atrophy, severe non-communicating hydrocephalus, and mega cisterna magna, which are aligned with HLD10 phenotypes. Also, WES in affected individual was performed and a total of 56,179,476 reads were generated. Specifically, 98% of these FASTQ file formats had Phred quality scores of > 20 for nucleotides, and 95% of reads had Phred quality scores > 30 (detailed parameters are shown in Table 2). As a whole, 98% of reads passed the filtering steps in the described pipeline. Then, after variant calling step, a VCF file with 178,343 variants was achieved (Table 3).
After implementation of filtering strategies and investigation of phenotype-related gene panels, the novel homozygous frameshift variant of c.135dup (NM_013328.4) in PYCR2 gene was found. The c.135dup (p.R46Q*36) in exon 2 is identified as a pathogenic variant in ACMG classification. Based on this guideline, the discovered variant had a very strong evidence known as PVS1, meaning null variant that can result in a loss-offunction mechanism in PYCR2 gene, involved in Leukodystrophy, hypomyelinating 10. Other introduced criteria for this variant were PM2 and PP3, which relatively refer to significantly low frequency in population databases of Exome Sequencing Project, 1000 Genomes, or ExAC, and multitude computational evidence for prediction of damaging effect of variant on the gene or its product, respectively. Accordingly, in terms of frequency, this variant was not observed in gnomAD (Exome, Genome, and Aggregated) and other relevant databases, such as TOPMed, GME Variome, Iranome, ExAC, 1000 Genomes, ESP6500, 4.7KJPN, and GenomeAsia. One in silico database, namely phyloP, predicted this variant as disease-causing, and no benign predictions were found for this variant in any in silico tools.
Moreover, no report related to this mutation was found in ClinVar. Subsequently, the results of Sanger sequencing in the patient confirmed the homozygous c.135dup variant in PYCR2 gene, and both of her parents were heterozygous for this genetic alteration (Fig. 2).

Discussion
In the current study, we conducted WES on an Iranian 9-year-old female patient with hypomyelinating leukodystrophy type 10 (OMIM 616420), and characterized a novel homozygous variant of c.135dup (p.R46Q*36) in PYCR2 gene, which was confirmed in the patient and segregated in her parents who were heterozygous for this mutation. ACMG criteria classified this mutation as a disease-causing variant, and the in silico tools used in the present research predicted the devastating effects of this variant. As exhibited in Fig. 3, when this mutation happens (shown as yellow spot in the schematic description), the red regions will be lost and a truncated protein (containing only green regions) without functional domains will be produced in the patient. No allele frequency was reported in population databases and, likewise, ClinVar   did not provide any evidence associated with this mutation. Furthermore, its position, 226111403 on chromosome 1, is highly conserved in human according to UCSC genome browser (Fig. 4). An extensive review of literature from various databases did not reveal any more cases of PYCR2-associated HLD 10 beyond other studies in Iran, and to the best of our knowledge, this is the first investigation in PYCR2 gene and HLD10 in Iranian population. Lossof-function mutations in PYCR2 gene result in HLD10, which is identified by hypomyelination, global growth delay, microcephaly, facial dysmorphism, axial hypotonia, and movement disability. It has also been described that these patients could not survive for a long time [9,11,16]. HLDs are a class of rare hereditary neuropathies that their related genes can be determined through latest nucleotide sequencing approaches such as NGS technologies; however, no specific therapeutic method has been developed yet [17]. MRI pattern recognition in the recent years has also been beneficial for diagnosis of different hypomyelinating diseases and diminished white-matter volume. Nevertheless, this procedure is combined with crucial limitations in diagnosis. WES has successfully displayed differential diagnostic efficiency, particularly in rare disease-causing variants, and has a significant potential in identification of de novo mutations [18][19][20].
PYCR2 gene, as a member of Pyrroline-5-carboxylate reductases, is involved in the biosynthesis of proline from glutamate. There have been numerous reports associated with pathogenic effects of PYCR2 mutation [21]. PYCR1 and PYCR2 are 84% similar regarding their protein structure, and biochemically they share notable similarities. However, their genetic roles in human disorders appear to be largely different [3][4][5][6][7][8][9]. PYCR1 mutations cause Cutis Laxa. In contrast, PYCR2 mutations are associated with hypomyelinating leukodystrophy type 10 (HLD10) and intellectual deficits. Importantly, these phenotypes are more severe in PYCR2/HLD10 patients. This may be, in part, given the fact that higher levels of PYCR2 protein exist in the brain than PYCR1. The PYCR2 protein consists of two common domains, known as the dimerization domains, and the NAD(P) binding part, encompassing 268 residues that is equal to 84% of the PYCR2 protein [3]. The NAD(P) binding domain is important for the function of the pyrroline-5-carboxylate reductase (PYCR) enzyme. The pathogenicity of the Arg-119Cys variant in the NAD(P)-binding domain has been recently validated and it was represented that the Arg-119Cys strongly impairs the catalytic efficiency of PYCR2 protein up to 366 times lower than that of the nonmutant enzyme, confirming its pathologic role in HLD10 [22]. This evidence proposes that the molecular alterations in the NAD(P)-binding domain could take part with the PYCR2 function, resulting in the HLD10 phenotypes. The p.R46Q*36 variant reported in our study is located in the P5CR dimerization domain and also is considerably close to another homozygous missense variant, p.Val86Gly, that previously was reported in a PYCR2-deficient patient (Fig. 5). In vitro functional studies of such variations in PYCR2 protein has exhibited that they disrupt protein dimerization, putatively leading to a loss-of-function [11]. By applying whole-exome or targeted sequencing, Zaki et al. found homozygous mutations of PYCR2 in 14 patients from 11 consanguineous families and showed that PYCR2-related syndrome can cause protein dysfunction with distinguishable clinical phenotypes in patients. In terms of genotype-phenotype correlation in PYCR2, reduced levels of protein were seen due to truncating mutations. Additionally, missense mutations were mentioned to have no effects on the levels of protein, however they impaired the protein multimerization [11]. Depletion in PYCR2 has been shown to cause neurodegeneration    [16].
On the other hand, in a study by Meng et al., non-consanguineous family history was reported in three out of five patients with homozygous variants in PYCR2. These observations are consistent with our data and underscore the deleterious impacts of PYCR2 mutations (Fig. 6), as well as the potential application of WES in characterization of novel rare and unknown variants. Noteworthy, more in-depth investigations could be utilized to heighten the knowledge about this specific area of science. The limitation of this study was the unavailability of MRI images for the patient. Finally, as a forward step, we suggest the transgenic animal models and other supplementary examinations for our novel mutation in PYCR2 gene to further strongly confirm these results in multiple biological settings.

Conclusion
In this study, we examined a 9-year-old girl with clinical presentation related to HLD10. WES was performed and assisted us to detect a novel and rare homozygous frameshift variant in the PYCR2 gene, known as c.135dup. The homozygous status of the variant in the patient and its heterozygous status in her parents were confirmed through Sanger sequencing. The pathogenicity of the discovered variant was interpreted via ACMG guidelines and led us to classifying this variant as a pathogenic mutation. These data could be of paramount significance for genetic counselling of patients with HLD and for understanding the underlying mechanisms in this disorder.