Background

Adenylate kinase (AK) deficiency (OMIM 103000) is an autosomal recessive disorder associated with moderate to severe congenital hemolytic anemia, with psychomotor impairment observed in few cases [1]. The enzyme involved in this disorder is adenylate kinase type I (AK1), which catalyzes the conversion of adenine nucleotides in the presence of Mg2+ or Mn2+: Mg2++ ATP + AMP = Mg2++ ADP + ADP.

AK1 belongs to the cytosolic enzyme family (EC 2.7.4.3). The AK1 gene is located on chromosome 9 on location 9q34. 11 (NCBI Gene ID: 203) and is highly expressed mainly in tissues with a high turnover rate, such as blood, brain, and muscles[2]. According to the available literature, certain mutations in this gene resulting in a functionally inadequate enzyme; so far, only ten mutations have been reported in the AK1 gene. [1, 3,4,5,6,7,8,9] (details are mentioned in Table 1).

Table 1 Mutations update in AK1 gene
Full size table

This study investigated the molecular basis of erythrocyte AK deficiency in an Indian family and provided prenatal diagnosis to them for subsequent pregnancy.

Methods

Clinical history

A 5-year-old Indian boy presented with severe neonatal jaundice and severe anemia requiring regular blood transfusions was referred to us for a complete hemolytic anemia workup. He had a hepatosplenomegaly with a liver 4 cm in size and spleen up to the umbilicus. The direct and indirect Coombs tests were negative and increased serum lactate dehydrogenase level ( 3400 U/L; reference range 140–280 U/L). The haemoglobin concentration was in the range of 5–7 gm/dl (male reference range 13–16 gm/dl). He has no history of fever and skin rashes. HPLC of the patient and parents indicated the absence of haemoglobinopathies. The Peripheral blood smear suggested dimorphic anemia with predominantly saw hypochromic, normocytic cells at the initial investigation. Bone marrow examination showed erythroid hyperplasia with megaloblasts. The biochemical test for the RBC membrane protein defect, i.e. Hereditary spherocytosis, was performed using eosin 5' maleimide by flow cytometry was within the normal range (980 MCF; reference range 900–1200 MCF). The activity of erythrocyte glucose-6-phosphate dehydrogenase, pyruvate kinase, and glucose phosphate isomerase was normal (details are given in Table 2). Presently, no developmental delay or mental retardation has been observed in the patient.

Table 2 Clinical and hematological data of the proband and parents
Full size table

Molecular studies

We collected peripheral blood from healthy controls, patients, and parents after dually signed informed consent. DNA was extracted using a standard protocol, and a targeted next-generation sequencing (t-NGS) library was generated. We performed library preparation using Illumina's TruSeq Custom Amplicon v1.5 kit (FC 130 1001) using 250 ng genomic DNA, following the manufacturer's instructions. Samples were pooled and loaded at 20 pM on MiSeq using a v3 600 cycle reagent kit sequencing 2 × 301 paired‐end reads (Illumina, San Diego, CA, USA). The library was sequenced to mean > 80-100X coverage on the Illumina MiSeq sequencing platform. The gene panel includes red cell haemoglobinopathies, enzymopathies, membrane disorders, congenital dyserythropoietic anaemias, and bone marrow failure syndrome-related genes. Numbers of genes included in the panels with corresponding accession numbers are obtained from the Single Nucleotide Polymorphism database (dbSNP at www.nchi.nlm.nih.gov/SNP), and the Ensemble Genome Browser (www.ensembl.org) are listed in Additional file 1: Table S1). The sequences obtained were aligned to the human reference genome (GRCh37/hg19) using the BWA program and analyzed using Picard and GATK version 3.6. The clinically relevant variants were annotated with the published literature and databases such as ClinVar, OMIM, GWAS, HGMD, and SwissVar. When sequence changes were found, independent PCR products were sequenced to confirm the mutations. In support that these sequence changes were not polymorphic variations, we verified that none was reported in the 1000 Genomes, https://www.internationalgenome.org/ and Human Gene Mutation Database http://www.hgmd.cf.ac.uk/ac/index.php

Prenatal diagnosis and detection of the familial AK deficiency causative mutation

A gynecologist conducted chronic Villus Sampling (CVS) during the 11th week of pregnancy of the mother. Genomic DNA was isolated using the standard protocol. We used the exon-specific primers mentioned by Dongerdiye 2020 et al. [9] for DNA amplification and Sanger sequencing of mother, father, and fetus samples.

Bioinformatics analysis

The effect of the variant was studied by multiple algorithms, such as MutationTaster, https://www.mutationtaster.org Polyphen-2, https://genetics.bwh.harvard.edu/pph2 SIFT, https://sift.jcvi.org. Mutation Assessor, https://mutationassessor.org M-CAP, http://bejerano.stanford.edu/mcap/, Combined Annotation Dependent Depletion (CADD), https://cadd.gs.washington.edu/. The probability of the mutation affects protein function was evaluated. Therefore the output “low” indicates a neutral variant. For Condel, the score ranges from 0 (neutral) to 1 (damaging). The crystallographic model of recombinant human adenylate kinase (EC 2.7.4.3) was downloaded from the Protein Data Bank (www.rcsb.org/pdb/; PDB-ID: IZ83) [6]. The impact of substitution on the structure and function of the protein was studied using PyMol software (DeLano Scientific, San Carlos, CA, USA) (http://www.pymol.org/) and Swiss Protein databank viewer (https://spdbv.vital-it.ch/).

Results

The complete details of the biochemical and hematological investigation of the patient and family are summarized in Table 2. Genetic analysis performed by the t-NGS panel revealed a single nucleotide substitution in exon 5 (c.301C > A) of AK1 gene, which caused glutamine to lysine (CAA to AAA) substitution at codon 101 (p. Gln101Lys). We observed a homozygous mutation in the proband. Parents were analyzed for the c.301C > A mutation by DNA Sanger sequencing; both parents were heterozygous for the mutation (Fig. 1). The novel variant c.301C > A, p.Gln101Lys, was submitted to the ClinVar database and submitted raw data (Accession No.: PRJNA745516: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA745516.) We measured AK enzyme activity in 50 healthy controls to the established normal range (reference range 297–360 IU/gHb), the proband (38.0 IU/g Hb), and parents' sample (mother 192.0 IU/gHb, father 208.0 IU/gHb). Biochemical findings correlated with molecular results.

Fig. 1
figure 1

Pedigree and electropherogram of the patient and family carrying the AK1 gene c.301C > A mutation

Full size image

We counseled the family for the consequences of severe enzyme deficiency. Therefore, at the time of the second pregnancy, they decided to undergo prenatal screening. The fetal DNA was screened for the complete AK1 gene. DNA Sanger sequencing identified substitution c.301C > A at codon 101, causing a heterozygous change from glutamine to lysine. The pregnancy continued, and the normal healthy child was born after nine months and followed up for one year. There were no symptoms of anemia and jaundice.

The AK1 protein (PDB ID-1Z83) consists of three chains, A, B, and C, spanning 194 residues. Each chain consists of one large central "CORE" domain and two small peripheral domains, the NMP binding and the LID domains. Upon ATP binding, the LID domain closes over the phosphoryl transfer site. The amino acid residue position Q101 is an important AMP binding site and 39,44,138,149 residues. Any changes at these AMP binding sites possibly hamper the catalytic cycle of the enzyme. Figure 2a shows a complete ribbon representation of the protein (PDB ID-1Z83) with chains A, B, and C of the AK1 enzyme, along with an insight into the Q101 position helical structure (Fig. 2b) and amino acid change from wild type (glutamine) to mutant type (lysine) (Fig. 2c, d). Multiple sequence alignment confirmed that the amino acid glutamine at 101 positions is conserved across species (Fig. 2e). Most bioinformatic prediction tools demonstrate the harmful effect of the amino acid change from glutamine to lysine (Table 3 summarizes the prediction results).

Fig. 2
figure 2

a Complete ribbon representation of adenylate kinase protein (PDB ID-1Z83) with chains A, B, and C. b Secondary structure of the protein (PDB ID-1Z83) showing the amino acid residue at position 101 (Q101). c Wild type amino acid residue glutamine 101 (Q101). d Mutant type amino acid residue Lysine101 (K101). e The residue (Q) at position 101 of AK-1 is highly conserved across species

Full size image
Table 3 Bioinformatical prediction data
Full size table

Discussion

In the present study, an Indian patient was diagnosed with rare red cell adenylate kinase deficiency. A novel homozygous (p. Gln101Lys) mutation in the AK1 gene was detected using a disease-targeted NGS panel. Previously, we reported the first case of AK deficiency from India, caused by compound heterozygous (c.71A > G and c.413G > A) mutations in the AK1 gene [9]. Clinically, these patients have incidents of chronic hemolytic anemia but no evidence of mental retardation or psychomotor impairment. Similar to other erythrocyte enzyme deficiencies, the choice of treatment depends upon the severity of the disease. Splenectomy may be recommended for severe transfusion-dependent AK deficient patients [14]. During the study, we excluded all the possible causes of hemolytic anemia, including RBC membrane defects. A recent review on red cell membrane defects mentioned that the flow‐cytometric osmotic fragility test is a new gold standard method for diagnosing HS, HE, and DHS in combination with eosin‐5'‐maleimide testing [13].

In India, only a few research groups have incorporated NGS-based genetic analysis for the routine diagnosis of RBC enzymopathy [9, 10]. However, custom NGS panels or whole-genome sequencing are widely used in various other laboratories to diagnose hemolytic anemia. The use of next-generation sequencing allows the identification of new causative genes, and polygenic conditions, and genetic factors that modify the disease severity of hereditary anemias. Disease-targeted gene panels often have higher diagnostic rates than those of exome sequencing or genome sequencing and are designed to maximize coverage, sensitivity, and specificity for the included genes [11]. The only drawback of using a custom NGS panel is that it involves a limited number of genes. Therefore, a continuous update is required for the best results. We have achieved an approximately 80% diagnostic yield using our custom panel [12]. However, these results may vary depending upon the number of patients, their clinical history, and phenotype-genotype correlations.

We identified three missense substitutions in the AK1 gene, Gln24Arg, Gln101Lys, and Arg138His, in two unrelated Indian families. These mutations could lead to dysfunction of the enzyme molecule by hampering AMP binding capacity. This study focuses on providing prenatal diagnosis to the family and gives accurate genetic advice. The fetal DNA was heterozygous for the substitution c.301C > A; p.Gln101Lys and advised to continue the pregnancy. Next-generation sequencing has many advantages, as it is cost-effective and gives high yield and speed. In contrast, there are certain limitations of NGS that significantly impact the accuracy of the results. The t-NGS panel has proven precise for our study, but its application may vary from lab to lab.

Conclusion

In conclusion, the targeted NGS panel identified a novel causative mutation in the AK1 gene in a 5-year-old male child with severe transfusion-dependent haemolytic anaemia. Identification of the pathogenic mutation helped us to offer a prenatal diagnosis in this family. This study also re-emphasizes the importance of NGS in diagnosing unexplained haemolytic anaemia in severe patients.