Investigating the role of non-synonymous variant D67N of ADGRE2 in chronic myeloid leukemia

Ayesha Afzal; Harooma Jamshaid; Yasmin Badshah; Maria Shabbir; Janeen H Trembley; Sameen Zafar; Ghulam Murtaza Kamal; Tayyaba Afsar; Fohad Mabood Husain; Suhail Razak

doi:10.1186/s12885-024-13108-6

. 2024 Nov 5;24:1354. doi: 10.1186/s12885-024-13108-6

Investigating the role of non-synonymous variant D67N of ADGRE2 in chronic myeloid leukemia

Ayesha Afzal ¹, Harooma Jamshaid ¹, Yasmin Badshah ^1,^✉, Maria Shabbir ¹, Janeen H Trembley ^2,^3,⁴, Sameen Zafar ¹, Ghulam Murtaza Kamal ¹, Tayyaba Afsar ⁵, Fohad Mabood Husain ⁶, Suhail Razak ^5,^✉

PMCID: PMC11536965 PMID: 39501172

Abstract

Background

Chronic myeloid leukaemia (CML) is a type of blood cancer that begins in the hematopoietic stem cells. It is primarily characterized by a specific chromosomal aberration, the Philadelphia chromosome. While the fusion gene is a major contributor to CML, several other genes including ADGRE2, that are reported as highly expressed in hematopoietic stem cells and could be utilized as a therapeutic marker in leukemic patients are implicated in the disease’s progression. Until recently, little research had been conducted to identify single nucleotide polymorphisms (SNPs) associated with CML. Therefore, this study aims to investigate the influence of non-synonymous variants on the structure and function of the gene encoding adhesion G protein-coupled receptor E2, ADGRE2, and to evaluate their association with CML and its clinical and pathological characteristics.

Methods

Non-synonymous SNPs of ADGRE2 were retrieved from the ENSEMBL, COSMIC, and gnomAD genome browsers, and the pathogenicity of deleterious variants was assessed using several established computational tools, including SIFT, CADD, REVEL, PolyPhen, and MetaLR.

Results

Various in silico analyses explored the impact of damaging SNP on the function, stability, and structure of EGF-like modules containing mucin-like hormone receptor-like2 (EMR2) protein encoded by the ADGRE2 gene. Genotype analysis was performed on collected blood samples, revealing that altered genotype TT of variant rs765071211 (C/T) was associated significantly with CML patients compared to the control. Further in vitro and in vivo analyses suggest that this SNP holds potential for clinical translation.

Keywords: Non-synonymous, CML, Bioinformatics tools, ARMS PCR, In silico analysis

Introduction

CML is a type of myeloproliferative malignancy, described by the unregulated growth and proliferation of myeloid cells at various stages of their development. The characterization and diagnostic criteria of CML defined by European Leukemia Net (ELN) has described 3 phases of this disease: chronic phase (CP), acceleration phase (AP), and blast transformation phase (BP), which is described by elevated levels of blasts of myeloid and/or lymphoid lineages [1, 2]. According to International Consensus Classification (ICC), 10–19% of bone marrow or peripheral blood blasts and ≥ 20% peripheral blood basophils are considered AP, while if the bone marrow or peripheral blood myeloid blast count is more than 20% then it is considered as the criteria for BP [3]. Most of the patients of CML are categorized under CP, however, if left untreated, this stage can progress to AP and a small fraction of patients may develop BP. The symptoms of CML are generally unspecific including fever, fatigue and weight loss, due to anemia and splenomegaly. Patients who have developed the blast phase, their symptoms may get severe which may include bleeding and bone pain. However, 50% patients of with CML in the chronic phase are asymptomatic and can only be diagnosed after routine blood tests. Reciprocal translocation between chromosomes 9 and 22 resulting in chimeric BCR-ABL is the major cause of CML in more than 90% of cases of this disease. BCR-ABL fusion gene encodes for an aberrant tyrosine kinase enzyme, which leads towards the increased proliferation of immature granulocytes, increased production of reactive oxygen species and genetic instability [4]. Genetic alterations in various other genes have been reported to correlate with CML, indicating the involvement of various other genetic factors in the development and progression of this disease [5]. The incidence rate of CML is reported to be 1-1.5 cases per 100,000 individuals, annually, without any geographic or racial bias. However, the prevalence of leukaemia is expected to increase and it is estimated that by the year 2040, the prevalence of CML will reach 0.18 million cases per year [6].

Single nucleotide polymorphism (SNP) is the most common type of genetic alteration occurring in DNA. SNPs occur when a single nucleotide in DNA is replaced with another nucleotide, resulting in a change in the genetic sequence. SNPs present in the non-coding or intronic regions can impact the regulation of the expression of the gene [1]. It was shown that intronic SNPs in P53 were associated with poor response and disease progression in CML patients [2]. SNPs located in the coding region of DNA can alter the structure, functions, and interactions of the subsequent protein, which may lead to the development of diseases such as cancers [7]. In the past decade, studies have shown various SNPs to be associated with various patient features such as response to therapy, survival, and prognosis in AML patients. Likewise, in CML, SNPs have been shown to impact the patient’s response to treatment and the overall prognosis of the disease. SNPs in BCR-ABL have been shown to induce resistance in patients against TKIs. A genetic association study has shown that SNPs in genes PSMB10, TNFRSF10D, PSMB2, PPARD and CYP26B1 were associated with CML predisposition [8, 9]. Therefore, it is essential to investigate SNPs in various genes that can result in increased cancer susceptibility to understand the underlying pathogenesis of various cancers.

Adhesion G protein-coupled receptors (GPCRs), commonly referred to as the ADGRE gene family, are a class of cell surface receptors that are vital to many physiological functions [10]. Numerous physiological processes, including cell adhesion, migration, immune response control, and tissue homeostasis are mediated by ADGRE receptors [11]. Their interactions with immune cells and extracellular matrix constituents impact various processes, including inflammation, tissue healing, and leukocyte trafficking [10, 12]. ADGRE2 is located on human chromosome 19 at position 19q13.31. ADGRE2, a member of the ADGRE family, is expressed in various tissues, including immune cells and the nervous system [13]. While its precise functions are still under investigation, the dysregulation of ADGRE2 can cause several diseases, including cancer, autoimmune disorders, and inflammatory ailments [14]. A recent study showed that the ADGRE2 gene has aberrant expression in AML which is associated with poor patient outcomes [15]. Targeting of the ADGRE2 gene with siRNA resulted in anti-leukemic effects in both in vitro and in vivo settings [10]. Furthermore, ADGRE2 has been shown to activate various signalling cascades such as PI3K/AKT, and PKC/MEK/ERK pathways to enhance the maintenance of proteostasis in leukaemia [15]. Furthermore, another study showed that various members of the ADGRE family including ADGRE2 were associated with poor prognosis and short survival in AML patients [16]. ADGRE2 gene has also been used to target AML via the CAR-T therapy approach [17]. Furthermore, ADGRE2 has been shown to activate various signalling cascades such as PI3K/AKT, and PKC/MEK/ERK pathways to enhance the maintenance of proteostasis in leukaemia [15]. This emerging evidence suggests that the ADGRE2 gene might have a role in CML. Moreover, since non-synonymous SNPs can alter the resultant protein’s structural and functional aspects, the presence of these variants in the ADGRE2 gene can have a damaging effect.

A previous study indicated that a missense SNP in the ADGRE2 gene resulted in amino acid substitution at residue 492 (p. C492Y) and was associated with vibratory urticaria [18]. The presence of this SNP led to an extension of the degranulation, an increase in the number of responsive mast cells, and a lowering of the activation threshold [19], suggesting that SNPs in ADGRE2 might be associated with diseases. However, most of the genetic variants in ADGRE2 are still uncharacterized in terms of their association with CML. Therefore, the current study aimed to explore the impact of the ADGRE2 non-synonymous variants by employing various computational tools that identified damaging SNPs in the gene and then analyzing the effects of those variants on protein function and structural stability. Furthermore, experimental analysis was performed to investigate the association of the variant exhibiting the most damaging potential with CML through genotyping analysis.

Methods

Data retrieval and mapping

Data on total genetic variants of ADGRE2 was retrieved from Ensembl [20], COSMIC [21], and gnomAD (https://gnomad.broadinstitute.org) [22]. The non-synonymous SNPs from all three databases were filtered out and merged while removing any redundant data. Information regarding exons and their corresponding amino acid residues was obtained from Ensembl and NCBI. The ADGRE2 protein sequence data was obtained from the UniProt database (https://www.uniprot.org/) [23], which is believed to be the most reliable database for protein sequences.

Prediction of pathogenicity

In the present study, various computational tools were utilized to analyze the pathogenic impact of non-synonymous SNPs. These tools evaluated ADGRE2 non-synonymous variants and assigned each variant a specific score based on its potential pathogenicity as determined by the algorithms. PolyPhen tool (http://genetics.bwh.harvard.edu/pph2/) predicts variant damaging effects based on the structural and evolutionary characteristics of the amino acids. This tool provides scores ranging from 0 to 1, with scores from 0.4 to 0.8 and 0.9-1 being possibly damaging and probably damaging, respectively. Variants with PolyPhen scores ≥ 0.999 were chosen [24]. SIFT tool (https://sift.bii.a-star.edu.sg) scores range from 0 to 1, where 0 signifies deleterious and 1 signifies tolerated, and for this study, variants with scores ≤ 0.05 were chosen. This tool predicts the damaging potential of the variant through the physical properties of the amino acid and the sequence homology [25]. REVEL categorizes variants as benign (< 0.5) or likely disease-causing (> 0.5), with variants scoring ≥ 0.5 chosen. REVEL predicts the pathogenicity of a missense variant by combining the scores for a particular variant from 13 individual bioinformatics tools [26]. MetaLR categorizes SNPs as damaging (0.5–0.9) or tolerated (0-0.4), and the variants scoring ≥ 0.5 were selected for this study. MetaLR employs logistic regression analysis to integrate the allele frequency and independent variant damaging score to estimate the pathogenicity of the non-synonymous variants [27]. Lastly, CADD (https://cadd.gs.washington.edu) utilizes sequence conservation and functional information to predict if variants are benign or damaging by scoring them from 0 to 30 and 31 to 35, respectively [28].

Effect of polymorphisms on RNA stability

In this study, the effect of selected non-synonymous SNP on mRNA stability was also determined. For this purpose, the RNAfold web server was used to predict the impact of the genetic variant by determining the minimum free energy and base pair probabilities of the mRNA secondary structure and comparing the results of the variant with wild-type mRNA (https://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi) [29].

Structure prediction and validation

The complete structure of the EMR2 protein encoded by ADGRE2 was unavailable in the protein data bank, so protein structure prediction was performed through I-TASSER (Iterative Threading ASSEmbly Refinement) [30], which employs the threading technique to predict the structure on the basis of the peptide sequence, assigning the confidence score ranging from − 5 to 2 based on the significance of template alignment. For the subsequent analysis, the model exhibiting the highest C-score was chosen, and the 3-D structure was visualized using the PyMOL v.3.0. (The PyMOL Molecular Graphics System, Version 3.0 Schrödinger, LLC) [31]. The predicted structure was validated through ERRAT [32], which determined the overall quality factor of the protein, while PROCHECK was used to determine the stereochemical quality of the protein structure by producing Ramachandran plots [33]. The information regarding the domains of protein and their subsequent amino acids was obtained from InterPro (https://www.ebi.ac.uk/interpro/search/sequence/). The structures were validated through InterPro and highlighted in PyMOL.

Stability analysis

The impact of the non-synonymous variant on protein stability and flexibility was also analyzed. The stability of the protein structure is determined through free energy. Evaluating the impact of a non-synonymous variant on the free energy of the protein can give a clear picture of its structural stability. For this purpose, various bioinformatic tools were applied to investigate any changes in the free energy of the protein induced by the pathogenic non-synonymous SNP in the ADGRE2 gene. I-Mutant 2.0 predicts the alteration in the stability of protein caused by variation in terms of DDG (Kcal/mol) (http://folding.uib.es/i-mutant/i-mutant2.0.html). A DDG value < 0 indicates a decrease, while > 0 indicates an increase in the stability of protein as a result of variation [34]. MUpro was also used to determine the structural stability of the variant protein (http://mupro.proteomics.ics.uci.edu/). This tool works on machine-learning algorithms trained on large mutational datasets. This tool also predicted the protein stability in terms of DDG values [35]. Lastly, the DynaMut2 tool was also used in this study(https://biosig.lab.uq.edu.au/dynamut2/). This tool predicted the impact of non-synonymous SNP on protein structural stability and changes in the intramolecular interactions due to the presence of the variant amino acid residue [36].

Variant effect on molecular characteristics

Genetic variations also affect the structures and functions of proteins. For this purpose, project HOPE was utilized (https://www3.cmbi.umcn.nl/hope/). This tool was designed to describe the role of variations in the human proteins in the molecular basis of disease-associated phenotype. This tool gathers information from various sources to determine protein sequence annotation, predict 3-dimensional coordinates of the protein, predict the impact of the amino acid variant on the protein’s functional and physio-chemical properties, and generate a heterogeneous report comprising of tables, figures, and texts [37].

Evolutionary conservation analysis of ADGRE2

The evolutionary conservation of amino acids in ADGRE2 encoded protein was predicted through the ConSurf tool, which is a web server that utilized the Bayesian method to highlight the protein regions that are important for its structure and functions (https://consurfdb.tau.ac.il). It gives the residue conservation score ranging from 1 to 9, where a score of 1 indicates the least conserved residue, while a score of 9 represents the most highly conserved residues [38].

Sub-cellular localization

The Subcellular localization of the ADGRE2 protein was predicted using Deeploc1.0 (https://services.healthtech.dtu.dk/services/DeepLoc-1.0/). This algorithm predicts the location of protein based on information regarding amino acid sequences by identifying those sequences that are important for the localization of a protein in a particular cellular compartment [39].

IRB and sample collection

Prior to the initiation of the experimental analysis, IRB approval (IRB number; 2024-IRB-A-24/24, approved on 6th may, 2024), was attained from the Ethical Review Committee, ASAB, NUST. Blood samples from the 105 patients suffering from CML were obtained from the Combined Military Hospital (CMH), Rawalpindi, Pakistan. CML patients with other morbidities were excluded from this study. Samples from 102 healthy individuals were taken as controls.

Primer designing

The Primers for ARMS PCR were designed using the tool Primer 1 (https://primer1.soton.ac.uk/primer1.html), and the primer sequences are shown in Table 1. Primers were designed to investigate the genotype of the ADGRE2 gene in the Control and Patient data. The primers were validated through UCSC in silico PCR (https://genome.ucsc.edu/cgi-bin/hgPcr).

Table 1.

ADGRE2 primers for ARMS PCR

D67N (rs765071211)	Common Primer	CCCCATGGAGACTTGTGCCG
	C primer (Wild type)	GCTGCCCTCAAGCCTCTGTCCT
	T Primer (variant)	GCCCAGGCTGGTCTTGAATTCC

Open in a new tab

Genotype analysis

In this study, the genomic DNA from the blood samples was isolated using the phenol-chloroform extraction method. The genotyping primers designed in the previous step were used to detect the presence of the target genetic variant in the ADGRE2 gene. ARMS PCR was used to perform genotyping analysis using the Applied Biosystems™ Veriti™ 96-Well Thermal Cycler. For each sample, 12 µl PCR reaction mixture was made, which was composed of 6 µl Solis Biodyne FIREpol master mix, 1 µl common primer, 1 µl allele-specific primers,1 µl DNA sample, and 3 µl PCR water. Conditions used for PCR are given in Table 2.

Table 2.

PCR conditions

Steps	Temperature	Time
Initial Denaturation	95◦C	10 min
Denaturation	95◦C	45 s
Annealing	60◦C	45 s
Extension	72◦C	45 s
Final extension	72◦C	7 min
Storage	4◦C	∞

Open in a new tab

Statistical analysis

The genotype data were statistically analyzed through Graph Pad Prism 9 (La Jolla, California, United States). Fisher’s exact test was performed on the variant data attained from healthy controls and CML patients. Additionally, odds ratio, relative risk, and 95% confidence intervals were also determined. Statistical significance was determined at a P-value < 0.05.

Results

Variant identification

A total of 17,162 variants were retrieved from all three databases, of which 1,514 were from gnomAD, 802 from COSMIC, and 14,846 from Ensemble, as shown in Fig. 1A. Variants were then grouped into various categories based on their consequences, which include missense variants, frameshift variants, 5`UTR, 3`UTR, and splice variants (Fig. 1B), and from this data, 674 unique non-synonymous ADGRE2 gene variants were determined for further analyses. ADGRE2 gene is comprised of 21 exons that encode their respective amino acids. This study also found that exon 16 of the ADGRE2 gene encodes the highest number of amino acid residues, while in respect of non-synonymous variant frequency, exon number 9 contained the highest number of variants, as shown in Fig. 1C.

Fig. 1 — Identification of ADGRE 2 genetic variants (A) ADGRE2 variants retrieved from Ensembl, COSMIC, gnomAD (B) Total count of various ADGRE2 variants retrieved from the three databases (C) Frequency of amino acids per exon and variants per exon

Variant pathogenicity analysis

The pathogenicity of the filtered non-synonymous variants was evaluated through various computational tools mentioned in the methodology. For that purpose, the pathogenicity scores given by the tools for each variant were determined, and the pathogenicity percentage for each variant was calculated (Fig. 2A). For this study, a pathogenicity percentage of > 80% was determined as the threshold. Out of 674 initial non-synonymous variants, only one variant was found to have a pathogenicity percentage of > 80% (Fig. 2B). A thorough analysis of the selected variants led to the selection of the non-synonymous variant rs765071211 for further in silico analysis due to its highest pathogenicity percentage, i.e., 100%. This variant caused amino acid alteration (D67N) at residue number 67 in the Egfca_6 domain of ADGRE2 encoded protein EMR2.

Fig. 2 — Analysis of pathogenicity of ADGRE2 non-synonymous variants (A) Filtration of SNPs through various algorithms (B) Pathogenicity of non-synonymous SNPs predicted through computational tools (SIFT, PolyPhen, CADD, REVEL, MetaLR)

Effect of rs765071211 polymorphism on RNA secondary structure

The impact of selected SNP on the secondary structure of ADGRE2 mRNA was also predicted. For this purpose, minimum free energy (MFE) for the variant and wild-type mRNA were determined and compared. The secondary structure of variant mRNA exhibited a significant change compared to the wildtype mRNA, revealing a prominent impact of altered allele on the structure and stability of ADGRE2 mRNA. MFE value determined for the wildtype and rs765071211 variant demonstrated that reference allele C was stabilizing in nature with a low MFE value (-37Kcal/mol) while its corresponding variant allele T destabilized the mRNA secondary structure and had a high MFE value of 36Kcal/mol (Fig. 3). Decreased mRNA stability can alter the levels of the translated proteins [40].

Fig. 3 — Effect of ADGRE2 variant on its RNA secondary structure. The lower the value of MFE, the higher the stability of mRNA secondary structure and vice versa