Analysis of mutations in GBA gene in Ukrainian patients with Gaucher disease

N. V. Olkhovych, A. M. Nedoboy, N. O. Pichkur © 2017 N. V. Olkhovych et al.; Published by the Institute of Molecular Biology and Genetics, NAS of Ukraine on behalf of Biopolymers and Cell. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited UDC 616-056.7-07


Introduction
Gaucher disease (MIM 230800) is the most common storage disorder, caused by hereditary deficiency of the lysosomal enzyme of glucocerebrosidase (EC 3.2.1.45)[1].In clinical terms, Gaucher disease is divided into three types depending on the presence and rate of neurological manifestation progress [2].Type I Gaucher disease (non-neuronopathic) is the most common form, characterized by the absence of neurological symptoms.The main Biomedicine ISSN 1993-6842 (on-line); ISSN 0233-7657 (print) Biopolymers and Cell. 2017. Vol. 33.N 1. P 34-47 doi: http://dx.doi.org/10.7124/bc.000940clinical features of this type of disorder are hepatosplenomegaly, pancytopenia and bone system abnormalities in rather a wide spectrum of severity -from asymptomatic cases to early and short manifestation with fatal outcome during the first years of life.Type II Gaucher disease (acute neuronopathic) is the rarest form, the main symptom of which, in addition to organomegaly, is early and rapid development of severe neurological damage, fatality in utero or during infancy.Type III Gaucher disease (chronic neuronopathic) is a transient form which covers all patients with organomegaly and any neurological signs, manifes ted after the second year of life.
The human glucocerebrosidase gene (GBA) is mapped in locus lq21, it is 7.5 kb long and consists of 11 exons.The pseudogene GBAP, remarkable for almost 96 % homology to the functional gene, is located 16 kb downstream [3].The region, surrounding the GBA gene, contains seven other genes and pseudogenes (MTX1, MTX1P1, THBS3, MUC1, PROPIN1, COTE1, CLK2), which may be involved in the process of forming different fusion or recombinant variants with the GBA gene during the crossing-over of this chromosome region [4].
According to the data of the largest human gene mutation databases (dbSNP, 1000 Genomes, HGMD), there are over 300 currently described pathogenic variants of the GBA gene, most of them are related to the development of Gaucher disease.As for other pathological variants, there are some large described insertions/deletions, including the deletion of the whole gene [5], as well as a considerable number of complex alleles (recalleles), which were formed due to the crossing-over between the functional gene GBA and surrounding genetic structures, most frequently the pseudogene GBAP.
Similar to most hereditary human diseases, Gaucher disease is remarkable for considerable variability of the frequencies of pathological variants of the gene in different populations.For instance, the population of Ashkenazi Jews is notable for a very high frequency of several variants -p.N409S, p.L483P, c.84dupG, IVS2(+1)A, RecNciI and RecTL, which together comprise about 90 % of all mutant alleles, with a single replacement p.N409S remarkable for over 70 % mutant alleles [6].At the same time, these pathological variants in non-Jewish patients account for the total of no more than 60 % of mutant alleles, and the frequency of replacement p.N409S is in the range of 10-50 % depending on the population [5].
There have been studies of the spectrum of pathological variants of the GBA gene in patients, suffering from Gaucher disease, in many European countries [7][8][9][10][11][12][13][14][15].This information is of great relevance for understanding the origin and distribution of the pathological variants of this disease as well as for the evaluation of genotype-phenotype correlations among patients.
The aim of our work was to identify pathological variants in the GBA gene which caused the development of Gaucher disease in Ukrainian patients, to compare their spectrum with the variants in the patients from Slavonic and other European countries and to evaluate the genotype-phenotype associations for this disease.

Patients
The genotype was analyzed in 63 patients of age from 6 months to 65 years from different regions of Ukraine, who were diagnosed with Gaucher disease during the period from 2001 to 2016, based on the complex of clinical (anemia, thrombocytopenia, hepatosplenomegaly), morphological (Gaucher-like cells in the bone marrow biopsy sample) and biochemical (glucocerebrosidase activity deficiency in blood leukocytes) data.As the Center of Orphan Diseases of NCSH OKHMATDYT of the Ministry of Health of Ukraine is the only institution in Ukraine to conduct biochemical and molecular-genetic testing of Gaucher patients, the group of patients, examined by us, may be deemed as the representative group of Ukrainian population.The control group to analyze the pathogenicity of new mutations was formed from the blood samples of 50 vo lun teer donors aged from 18 to 60 years without any clinical signs of lysosomal pathology.All patients (patients' parents) and volunteer donors gave their informed consent to the use of their biomaterial for the study.The work has been approved by the Ethics Committee of SI IGRM NAMS of Ukraine.

Screening of major mutations
Genomic DNA was isolated from peripheral blood samples, obtained from EDTA, using commercial sets NucleoSpin ® Blood (Macherey-Nagel, Germany) according to the manufacturer's instructions.Missense replacements p.N409S and p.L483P were determined by the method of nested PCR and subsequent RFLP-analysis as described previously [16].The identification of the mentioned mutations was conducted using restriction endonucleases Xho1 and Msp1 (MBI Fermentas).

Sequencing of GBA gene
Due to the availability of highly homologous pseudogene GBAP, the exon amplification of the GBA gene was conducted using the technology of nested PCR, the first stage of which involved the primers, specific for the sequence of the functional gene (GenBank Accession No. AH006907.2).The sequence of primers, which were used for each stage of nested PCR, is presented in Table 1.The design of primers was independently elaborated using Primer3 software, web-version 4.0.0 (http://bioinfo.ut.ee/primer3/), the synthesis of primers was ordered in Apply Biosystems (USA).
The first stage of nested PCR envisaged obtaining PCR products of 1-4 exons using primers 1F and 4R (product 1873 bp), 4-8 exons using primers 4F and 8R (product 3777 bp) and 8-11 exons using primers 8F and 11BR (product 2267 bp).At the second stage of nested PCR the fragments of the coding part of the GBA gene exons were obtained along with 5'-and 3'-non-translating regions and intron/exon boundary, using exon-specific primers (Table 1).The purification of PCR products was conducted using the commercial kits NucleoSpin ® Gel and PCR Clean-up (Macherey-Nagel, Germany) according to the manufacturer's instructions.
The identification of sequence variants of the GBA gene was conducted by the method of Sanger's direct automatic sequencing using ABI Prism 3130 (Applied Biosystems, USA) and BigDye Terminator sequencing kit (Applied Biosystems, USA) according to the manufacturer's protocol.All the identified rearrangements were confirmed by sequencing using both forward-and reverse-primers.The analysis of sequencing results was made using programs Chromas and BLAST (http://www.ncbi.nlm.nih.gov/blast).The electronic databases of known pathogenic variants -dbSNP, 1000 Genomes, HGMD -were used to characterize the identified rearrangements.The analysis of pathogenicity of new variants was made using programs PolyPhen2 (http://genet-ics.bwh.harvard.edu/pph2),SNPs3D (http:// www.snps3d.org)and Provean (http://provean.jcvi.org/index.php).The description of rearrangements in the GBA gene was based on the nomenclature, recommended by the Human Genome Variation Society (HGVS version 15.11, 2016), with the consideration of the first 39 amino acids of preprotein [17,18].To facilitate the comparison of the obtained results against the previous publications, the first mention of the rearrangement was made with the indication of the traditional name in brackets, without the consideration of 39 amino acids of preprotein.

Results
Our examination involved 63 patients from 63 families with confirmed diagnosis of Gaucher disease from all the regions of Ukraine, including 57 patients, clinically classified as type I Gaucher disease (type I GD), 2 patients with type II Gaucher disease (type II GD), and 4 patients with type III Gaucher disease (type III GD).The severity degree of the clinical course was determined according to the severity index (SSI), suggested by Zimran et al. [19].Table 2 presents the summarized clinical and molecular-genetic characteristics of the patients, examined by us.Three families had two sick siblings each, but only one sibling from each family was used in calculations of allele frequencies.
A total of 124 out of 126 (98.4 %) mutant alleles of the GBA gene were identified.If two previously described rearrangements were identified in a patient, it was assumed that they were in trans-position, which was confirmed by the analysis of parents, if available.The biological material of parents was unavailable The second most frequent pathogenic variant of the GBA gene was the missense replacement in exon 10 p.L483P (L444P), it was identified as an individual mutation in 18 alleles out of 124, which amounted to 14.5 %.This replacement was not identified in homozygous state in any case.Three other patients had the p.L483P replacement as part of the most common recombinant allele RecNciI, which is the result of the conversion of a fragment of exon 10 of the functional gene and a pseudogene, and which, in addition to replacement p.L483p, also carries two replacements p.A495P (A456P) and p.V499V (V460V) in cis-position, remarkable for the pseudogene sequence.Additionally, it was established that three patients had replacement p.L483P in cisposition with single nucleotide replacements c.*92G>A and c.*102T>C, localized in 3'-nontranslating region (3'UTR) of the GBA gene.These replacements were also notable for the sequence of pseudogene GBAP, thus, there is high probability of these patients to have the recombination between a gene and a pseudogene in the site of intron 10 -3'UTR, described as RecG allele [5].
The third most frequent variant (6 out of 124 alleles, 4.8 %) was found to be a missense replacement in exon 5 p.R159W (R120W), which was identified in six patients: in five -in the compound with p.N409S, and in one patient with type III GD in the compound with a missense replacement in exon 6 p.G241R (G202R).The same number of alleles (6 out of 124, 4.8 %) had a missense replacement in exon 9 p.G416S (G377S), which was identified in four patients: two had it in a homozygous state and two -in the compound with another single nucleotide rearrangement.
It was observed that there was rather a high prevalence of recombinant alleles among our examined patients with the involvement of the functional gene GBA and pseudogene GBAP.For instance, in addition to the abovementioned seven patients with alleles RecNciI and RecG, five patients with a complex of six replace-ments of pseudogene origin in cis-position in exon 6 (p.W223R, p.N227R, p.V230G, p.S235P, p.G241R, p.F252I) were identified, which may correspond to the described recombinant allele RecC with the conversion of a site between intron 5 and exon 7 (g.4179_5042con)[5].It was found that two more patients had cis-position of the abovementioned six replacements of pseudogene origin in exon 6 and replacements p.R159W in exon 5.This combination may correspond to the described recombinant allele RecC5a with the conversion of a site between intron 4 and exon 6 (g.3941_4430con) [5].Furthermore, two patients had the single nucleotide replacements c.*92G>A and c.*102T>C, localized in 3'-nontranslating region (3'UTR) of the GBA gene.These replacements were also notable for the sequence of pseudogene GBAP, thus, there is high probability of these patients to have the recombination between a gene and a pseudogene in the site of intron 11 -3'UTR [5].Therefore, the total frequency of recombinant alleles was 12.9 %.Notably, the final confirmation of the availability and localizations of recombinant alleles requires additional studies.
Six pathogenic rearrangements, which have not been previously described, are identified in the GBA gene -four missense replacements: p.N131I in exon 4, p.F167L in exon 5, p.S390I in exon 8 and p.P430A in exon 9; a nonsense replacement p.Y244* in exon 6 and a single nucleotide deletion c.901delC in exon 7. The pathogenicity of the identified rearrangements was confirmed in silico using programs PolyPhen2, SNPs3D and Provean (Table 3).None of the mentioned rearrangements was found in the database of 1000 Genomes or in 50 samples of donor blood, examined by us.The paternal analysis demonstrated that replacements p.S390I and p.Y244* were inherited by probands from fathers, whereas rearrangements p.N131I and c.901delC were inherited by probands from mothers.The paternal analysis of rearrangements p.F167L and p.P430A was not conducted due to the unavailability of the parental biological material of the probands.
Unfortunately, genotypes of two patients (patients 8 and 17) are yet to be determined completely.It was identified that patient 8 had a previously non-described missense replacement p.P94S in exon 3 of the GBA gene, which was evaluated as a polymorphic variant during the pathogenicity check (Table 4).However, we have no possibility to detect the second pathogenic alleles in these patients due to a lack of material.

Discussion
Our examination of 63 independent Ukrainian patients suffering from Gaucher disease identified a total of 32 pathogenic rearrangements: 22 one nucleotide substitutions, 2 of which were nonsense replacements, 2 duplications, 3 deletions, 1 splicing mutation and 4 recombinant rearrangements.
Replacement p.N409S is known to belong to so called "mild" variants of the gene and is almost not found in patients with neurological disorders [5].In the group of Ukrainian patients replacement p.N409S was also found only in patients with type I GD -41 of them carried this replacement in at least one allele, which amounted to 71.9 % of all the patients with type I GD (Table 4).Moreover, the data, obtained by us, confirm the conclusions of other researchers on the mitigating impact of replacement p.N409S [6].For instance, the availability of this replacement in patients 24 and 60, regardless of the availability of zeromutation p.R202* in another allele, conditioned type I (non-neuronopathic) clinical course of medium severity (Table 2).Three patients, who had the identified recombinant alleles in the compound with replacement p. N409S which impairs the structure and func-  tion of the GBA gene considerably, underwent also a protective impact of the mentioned replacement, which conditioned the development of type I (non-neuronopathic) Gaucher disease.Similar to most populations, replacement p.L483P was ranked the second most frequent among our patients -14.5 % (18/124) alleles contained this replacement as an individual mutation, and 5.6 % more (7/124) -as a constituent of recombinant alleles of distal regions of the GBA gene.Traditionally, the replacement p.L483P is classified as "severe", it is often associated with the availability of neurological disorders in patients, especially in homozygous state [5].No homozygote p.L483P/p.L483P was identified in the Ukrainian patients.17 out of 18 carriers of replacement p.L483P, as an individual mutation, had type I (non-neuronopathic) Gaucher disease i.e. they had no neurological disorders.The severity of this replacement in 13 of them was mitigated by replacement p.N409S in the second allele, in two -by the presence of missense replacements p.I200S and p.R502C, also remarkable just for nonneuronopathic type of the disease [5].One patient had replacement p.L483P, identified in the compound with RecC5a allele which was described predominantly for patients with type I Gaucher disease [5,15].
Two homozygous carriers of replacement p.G416S were identified among the examined patients, both had type I (non-neuronopathic) Gaucher disease, which is in agreement with the data of other authors about "mild" nature of this replacement [20].In two more patients, this variant was identified in the compound with another single nucleotide rearrangementone had cytosine duplication in exon 3 of the gene (c.203dupC) which conditions the shift in the reading frame, and another had the replacement of guanine for adenine in the last triplet of exon 7, which does not result in the amino acid replacement, but impairs the splicing site.Therefore, both rearrangements are "severe" by their phenotypic manifestation.It is noteworthy that there was no mitigating effect from "mild" replacement p.G416S on the impact of "severe" rearrangement in both patients who had type III (chronic neuronopa thic) Gaucher disease (patients 50 and 57).
Replacement p.R159W is rather common for European populations and is usually characterized as "mild" by its phenotypic manifestation [5].Among the Ukrainian patients this variant was identified in 5 patients with type I (non-neuronopathic) Gaucher disease in the compound with p.N409S and in one patient with type III chronic neuronopathic disease (patient 49) in the composition of a complex allele [p.R159W+p.G241R] and replacement p.D448H in the second allele.
While analyzing the phenotype of patients with identified recombinant alleles in the GBA gene, it should be noted that genotype-phenotype correlation in them is ambiguous.On the one hand, it was predictable that all patients with genotype p.N409S/RecNciI would have type I (non-neuronopathic) Gaucher disease.A patient with genotype p.L483P/RecNciI (patient 13) was notable for early manifestation and severe clinical course of the disease, which conditioned the fatal outcome after a hemorrhage, resistant to treatment, at the age of four, regardless of conducted enzyme-replacement therapy.The absence of neurological symptoms in this patient at diagnostics was the reason why this case was classified as type I (non-neuronopathic) Gaucher disease.However, the "severity" of both mutations and early death allow for an assumption that there was no sufficient time for manifestation of the neurological symptoms in this patient], thus, this case may be considered to be type III (chronic neuronopathic) Gaucher disease.
Most described cases of the recombinant alleles, involving exons 5-7 of the GBA gene (RecC and RecC5a), were observed in the patients with no neurological disorders that was also demonstrated by us.It is notable the identification of genotype p.V414L/RecC in the patient with early severe clinical course of Gaucher disease without any neurological disorders but with hepatic gaucheroma and fatal outcome at the age of 6 (patient 51).Taking into consideration the published data about the clinical course in the patients, homozygous by the missense replacement p.V414L, this variant should be related to "mild" ones by its phenotypic manifestation [21].Therefore, in combination with allele RecC, which is also mainly observed in the patients with type I Gaucher disease, it could be expected that the patient 51 has either mild or medium degree of clinical course severity.Such contradiction requires detailed study to identify the genetic, epigenetic, and environmental factors, impacting the phenotypic manifestation of a specific genotype.
The results of genotyping other patients with neuronopathic types of Gaucher disease are in some agreement with the published data.Both mutations, identified in the patient 61 with type II acute neuronopathic disease (genotype p.G241R/p.A423D), described before, are remarkable for neuronopathic forms of Gaucher disease [22].The duplication of guanine c.84dupG in exon 2 of the GBA gene, identified in a patient with type II (acute neuronopathic) Gaucher disease, was also described as a "severe" form according to its phenotypic manifestation and is notable for neuronopathic forms of the disease.As seen from the published data, the phenotypic manifestation of missense replacement p.N227S depends on the rearrangement in the second allele.The combination of this variant with "mild" mutations, such as p.N409S, leads to the development of type I Gaucher disease, whereas the combination with "severe" mutations, such as Rec-alleles or p.L483P, leads to the occurrence of neuronopathic types of the disease [5].Therefore, the nonsense replacement p.Y244*, identified in the compound with p.N227S in patient 54, conditioned the develop ment of type III (chronic neuronopathic) Gaucher disease.
The analysis in silico of new rearrangements in the GBA gene, described by us, demonstrated that nonsense replacement p.Y244* and missense replacement p.P430A might impair the structure and function of the gene product the most.The nonsense replacement p.Y244* is responsible for it due to the occurrence of a stop-codon in exon 6 and, as a result, the formation of a shortened gene product, which lost over half of amino acid residues and, first and foremost, the active site of the enzyme (E379) [23].The effect of missense replacement p.P430A is due to the fact that amino acid residue P430 is localized in the loop, contacting with β8-strand of the first domain TIM barrel in GBA molecule.This loop is localized on the upper surface of the active site and contacts the substrate molecule, which makes any replacements in this region critical for normal functioning of the active site of the enzyme [24].Notably, in 1998 Cormand et al. described another replacement of proline 430 -p.P430L in two patients with type I Gaucher disease [25].However, in both patients this replacement was in the compound with p.N409S, which, taking into consideration the mitigating nature of this variant, made it impossible to evaluate the correlation of replacement p.P430L and the patients' phenotype.Replacement p. P430A was identified by us in two patients -in the patient 26 with type I Gaucher disease, conditioned by the presence of replacement p.N409S in the second allele, and in the patient 31 with type II Gaucher disease, with the duplication c.84dupG in the second allele.The development of severe acute neuronopathic form of the disease in patient 31 (genotype p.P430A/c.84dupG)confirms that the replacement of proline 430 is a zero-mutation, conditioning a considerable loss in the enzymatic activity of glucocerebrosidase.
It is probable that missense replacement p.F167L might also impact the functional activity of GBA.Phenylalanine 167 is one of seven aromatic amino acids, forming a chain on one side of the active site pocket of a GBA molecule and is involved into the substrate recognition [27].Leucine is not an aromatic amino acid, thus the replacement of phenylalanine with leucine leads to the impairment in the aromatic chain structure and therefore to the impairment of spatial organization of the active site pocket of the enzyme, which may have negative impact on the catalytic activity of glucocerebrosidase.Unfortunately, we were unable to evaluate phenotype-genotype association of this replacement because it was identified in the compound with "mild" replacement p.Q453R, which conditioned type I Gaucher disease in the patient 21.
Replacements p.N131I and p.S390I, described by us, influence amino acid residues, localized in domain III, the catalytic domain of GBA molecule, but do not participate directly in the recognition and binding of the substrate or activator of saposin C [23,24].This is the most likely factor, conditioning the "mild" impact of conformational changes in GBA molecule, caused by these replacements, on its catalytic activity.This is also confirmed by the clinical data of the patients, described by us, -replacement p.N131I in the compound with p.N409S led to the development of type I Gaucher disease with late manifestation (patient 43, the first symptoms were identified at the age of 15), and replacement p.S390I mitigated the impact of "severe" replacement p. L483P in patient 5 and caused the development of type I Gaucher disease, albeit of medium severity with early manifestation (first symptoms at the age of 1 year), but without neurological disorders.
Therefore, the application of different molecular and genetic approaches, including direct gene sequencing, allowed us to identify 96.8 % of mutant alleles in Ukrainian patients with Gaucher disease.Also six new and previously not described rearrangements of the GBA gene sequence were identified.The comparison of genotypes with clinical form of the disease, identified in patients, demonstrated that at present there are recognized genotype-phenotype correlations for this disease, which allow predicting the type and clinical course of the disease to some degree.In most patients, described by us, the combination of genotype and phenotype corresponded to the data of numerous previous studies.The information about clinical signs of the disease in patients with new, previously not described mutations, allowed us to complement current information about genotype-phenotype correlations for Gaucher disease.However remarkably, there is considerable heterogeneity in the clinical course of the disease even among people with a similar genotype.This depends on different factors, influencing the realization of genetic information in a certain individual -the availability of complex alleles, the influence of adjacent or modifying genes, the impact of environmental factors, etc.These factors impact the possibilities of the clinical prognosis and require additional studies.
Therefore, the determination of molecular and genetic nature of a disease in a specific patient is the mandatory information, which allows confirming the diagnosis, predicting the clinical course of the disease and envisaging the response to specific therapy.

Table 1 . The design of primers for the GBA gene analysis
* Primers specific to the sequence of the functional GBA gene