Patients

During our studies, we identified a family in which two members (brothers, RK, MK) displayed symptoms of hemolytic anemia that did not match the characteristics of the known disease [32]. Two brothers with symptoms of hemolytic anemia were admitted to the out-patient Clinic of the Department of Haematology, Blood Neoplasms and Bone Marrow Transplantation, Wroclaw Medical University. This study was approved by the ethics committee (permission KB-541/2011) of the Wroclaw Medical University, and informed consent was obtained from the patients and healthy patients’ family members and unrelated healthy individuals serving as a controls.

The brothers were hospitalized as children at the age of 6 and 4 years (RK and MK, respectively) with an initial diagnosis of hereditary spherocytosis (HS). Further studies ruled out HS, but the conclusive and accurate diagnosis was not possible. Diagnostic criteria for HA were based on features: yellow skin and a high hard palate, splenomegaly, jaundice, increased bilirubin, reticulocytosis, and anemia (Table 1). Both patients had normal values of erythrocyte osmotic fragility and presence of stomatocytes on peripheral blood smears (as well as anisocytosis and azurophilic stipplings) but no spherocytes typical of hereditary spherocytosis in their blood smears. The EMA-binding test, a reliable method for identifying hereditary spherocytosis showed an increase in mean fluorescent intensity (109–112% of reference), whereas for typical spherocytosis a decrease is observed [33, 34]. The direct anti-globulin test was negative, and other causes of hemolytic anemia such as thalassemia syndromes and G6PD deficiency were excluded by appropriate tests (Additional file 1: Table S1.1). SDS–PAGE profile of their erythrocyte membrane protein was not different from normal [35], as was Na+K+ equilibrium of the erythrocytes.

Table 1 Hematological characteristics of studied family membersDNA and RNA isolation

EDTA anticoagulated blood was collected from two patients (brothers, RK and MK) who displayed symptoms of hemolytic anemia and asymptomatic family members (father AK and sister EK) as well as volunteers (healthy individuals of Polish origin), and a patient with HS [29, 32] by venipuncture. The genetic material of a likely asymptomatic mother of the brothers was unavailable. Genomic DNA was isolated from the whole blood of each studied family member using the standard method (QIAamp DNA Blood Mini Kit, Qiagen, Hilden, Germany) and stored at −20 °C until analysis. Total RNA was extracted from whole blood as well as reticulocytes and collected using the miRNeasy Mini Kit (Qiagen, Hilden, Germany). Both procedures were carried out according to the manufacturer’s recommendations. RNA samples were stored at −70 °C until use. DNA and RNA concentrations were calculated on the basis of the absorbance at 260 nm.

Transcriptome analysis (RNA-seq)

Transcriptome analysis was performed by the Heflin Center for Genomic Science Core Laboratories, the University of Alabama at Birmingham, AL, USA. Reticulocyte RNA samples isolated from patients with HA and two controls [one healthy control (CtrlH) and one affected HS [29, 32]] were sequenced and aligned to the human reference genome assembly (human genome19). For contextualization of expression patterns and functional signatures, the RNA-seq results were analyzed using the GeneAnalytics tool [36]. The full procedure has been described by Skulski et al. [32].

Whole-exome sequencing on Illumina platforms

Exome sequencing was performed for both affected patients as well as their father and sister. DNA samples were sequenced at the Heflin Center, UAB (see above). Exome capture was carried out using the Agilent SureSelect all Human exon v3 capture kit (Agilent SureSelect Human All Exon 50 Mb for target enrichment; Agilent Technologies, Santa Clara, CA, USA). The SureSelect Target Enrichment System (Agilent Technologies, CA, USA) was used, followed by 2 × 100 bp paired-end sequencing on the HiSeq 2000 next generation sequencer from Illumina (Illumina, San Diego, CA, USA). Raw sequence reads were aligned to the reference human genome (human genome19/GRCh37.13). The bioinformatics analysis for detecting single nucleotide variants (SNVs) and inserts/deletions was performed at the Heflin Center at UAB and annotated according to dbSNP.

Whole-exome sequencing data analysis

Trimmed raw data were aligned to the hg19 human reference genome, and variant identification analysis was performed using the Ingenuity Variant Analysis plugin (IVA; QIAGEN, CA, USA). Additionally, variants that may be responsible for the observed phenotype were also made to select compound polymorphisms present in one of the genes or several genes (connected by common metabolic pathways) using the GeneAnalytics tool (LifeMap Sciences, CA, USA; accessed on 22 December 2016). The clinical significance of selected changes was made according to the 1000 Genome Browser [37], ClinVar [38], HGMD [39], Online Mendelian Inheritance in Man (OMIM) [40], Single Nucleotide Polymorphism (dbSNP) [41], GeneCards [42], and The Universal Protein Resource (UniProt) [43] databases (accessed on 29 December 2020) as well as literature data.

Sanger sequencing validation

Potentially pathogenically significant variants identified by WES were verified by target Sanger sequencing of genomic DNA and cDNA. Total RNA was reversely transcribed into cDNA for sequencing using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s recommendations. Sanger sequencing and all primer synthesis were performed by Genomed SA (Warsaw, Poland). The sequences of all primers and the PCR amplification parameters are available upon request. Additional file 1: Tables S1.2 and S1.3 include the sequences of the primers used for verification of the key data obtained from WES.

SDS–PAGE and western blotting

After freezing and thawing five times in Red Cell Stock Solution [0.9% (w/v) NaCl containing 28% (w/v) glycerol and 2.8% (w/v) sorbitol], erythrocyte lysate was centrifuged at 10,000g for 5 min at 4 °C, and the supernatant was collected. Equal volumes of supernatant and reducing agent [40 mM Tris–HCl, 1.5% (w/v) SDS, 1% (w/v) 2-mercaptoethanol, 1.3 M urea, and 0.005% (w/v) bromophenol blue] were mixed and incubated for 5 min at 95 °C for solubilization and then centrifuged at 10,000g, for 5 min at 4 °C. Supernatants were mixed with reducing agent in a ratio of 1:4 and loaded on (12% separating, 4.5% stacking) polyacrylamide gel. Electrophoresis was carried out at a constant current of 25 mA for 2 h at room temperature in a running buffer [0.025 M Tris, 0.192 M glycine buffer, pH 8.3 containing, and 0.1% (w/v) SDS] until bromophenol blue reached the bottom of the gel. After electrophoresis, a wet transfer onto nitrocellulose paper was performed at a constant current of 250 mA in transfer [Towbin buffer [44]—25 mM Tris buffer, pH 8.3 containing 0.192 M glycine and 20% (v/v) methanol] for 2 h.

Then membranes were stained with Ponceau S solution [0.2% (w/v) Ponceau S and 5% (v/v) acetic acid] to confirm proper protein transfer. Membranes were blocked in 5% nonfat dry milk solution in TBS-T buffer [20 mM Tris–HCl buffer, pH 7.5 containing 0.15 M NaCl and 0.05% (v/v) Tween-20] for 45 min with gentle shaking at room temperature. After washing with TBS buffer (20 mM Tris buffer, pH 7.5 containing 0.15 M NaCl), membranes were incubated with the primary, mouse, monoclonal antibody anti-NT5C3 (sc-390782, Santa Cruz Biotechnology, Dallas, TX, USA), diluted 1:2000 in TBS-T buffer, for 2 h at room temperature with gentle shaking. Then, after washing the membrane in TBS-T buffer, a secondary goat-anti-mouse IgG (H + L) antibody, conjugated with HRP (Jackson ImmunoResearch, Ely, UK), was used to detect the primary antibody. Nitrocellulose membranes were incubated with secondary antibody diluted 1:10,000 in TBS-T buffer for 60 min in the dark at room temperature with gentle shaking. After washing the membrane in TBS-T buffer a chemiluminescence reaction was induced by using ECL Western Blotting Detection Reagent (RPN 2209, Amersham, Marlborough, MA, USA) and visualized in the Azure c500 Imaging System (Azure Biosystems, Dublin, CA, USA). Quantification of WB membrane by ImageJ 1.51n software was performed according to the procedure described by Stael et al. [45]. The area of each peak (without background) for relative density calculation was used.

Enzymatic activity

The estimation of enzymatic activity of cytosolic pyrimidine 5′-nucleotidase was performed according to Beutler et al. (1984) with slight modifications [46]. The whole blood was collected by venipuncture using EDTA as an anticoagulant, diluted 1:1.5 with cold 0.9% NaCl and passed through a leukofilter as described previously [32]. The flow-through was centrifuged (600g, 15 min, 4 °C) and then the erythrocyte pellet was washed three times (800g, 10 min, 4 °C) with cold 0.9% NaCl. Erythrocyte pellets were suspended in Red Cell Stock Solution (mentioned in the above paragraph) in a ratio of 1:1 (v/v) and lysed via five freezing and thawing cycles. The UMP (U6375, Sigma-Aldrich, Saint Louis, MO, USA) as a substrate for P-5′-nucleotidase was dissolved in a buffer (0.1 M Tris–HCl, 0.1 M MgCl2, 0.1 M EDTA, and 0.7 mM 2-mercaptoethanol) to a final concentration of 4.6 mM. Next, 150 µL of hemolysate (diluted 1.25- to 5-fold) and 150 µL of UMP solution were mixed and incubated at 37 °C for 120 min. Then 120 µL of the mixture was transferred to 1200 µL of iron–TCA solution (10% trichloroacetic acid, 1% thiourea, 3% ammonium iron (II) sulfate hexahydrate) to stop the reaction. The control sample was prepared by taking 120 µL of incubation mixture immediately after mixing 150 µL of hemolysate and 150 µL of UMP solution and diluting in 1200 µL of iron–TCA solution. Samples were centrifuged (10,000g, 5 min, RT), and the supernatants (FeTCA extracts) were collected in clean tubes. Inorganic phosphate (Pi) was quantitated by the Rouser method in a 96-well plate: 120 µL of each FeTCA extract, 20 µL of 70% perchloric acid, 30 µL of dH2O, 20 µL of 2.5% (w/v) ammonium molybdate, and 20 µL of 10% (w/v) ascorbic acid were added [47]. As a blank sample, dH2O was used in place of acid extract. The plate was gently shaken on the orbital shaker and later incubated for 30 min at 37 °C. After incubation, absorbance at 750 nm was read on the Rayto RT-6100 microplate reader (Rayto Life and Analytical Sciences, Shenzhen, China). On the basis of the calibration curve, the Pi nanomolar concentration in each sample was calculated. To convert these data into specific enzymatic activity (expressed as pmol Pi/mg Hb/min), hemoglobin was quantified using absorbance at 512 nm and the molar extinction coefficient for hemoglobin of 26,936.8 cm−1/M [48] (accessed on 4 July 2022). Nonspecific nucleotidase activity tested on 5′-AMP as a substrate showed low values in the range of 0–15% (data not shown). Each sample was accompanied by the zero time control, thus including “background” phosphate derived from the sample and from the 5′-UMP.

Measurement of miRNA and mRNA levels using qPCR

The StepOnePlus Real-Time PCR System was used to run the qPCR reactions. TaqMan One-Step RT-PCR Master Mix Reagents (Thermo Fisher Scientific, Waltham, MA, USA) and the manufacturer’s protocol were used as described previously [49]. The relative expression levels were calculated using the comparative relative standard curve method [50]. We used glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA and small nucleolar RNA (C/D Box 44 (RNU44) RNA) as the relative control for our studies. The relative control stability during experiments was verified with 18S rRNA. TaqMan probes used were the following: GAPDH, Hs99999905_m1; RNU44, 001094; 18 s, Hs99999901_s; 5′-nucleotidase, cytosolic IIIA (NT5C3A, transcript variant 4 expressed only in reticulocytes), Hs01052076_m1 and Hs01052075_m1; zinc finger DHHC-type containing 17 (ZDHHC17), Homo sapiens microRNA 4775, 463622_mat.

Molecular karyotype (array CGH)

Identification of genome dosage imbalances with the array CGH method (the array-based Comparative Genomic Hybridization) was performed in a certified laboratory cooperating with the Genomed Healthcare Center, Warsaw, Poland (The Institute of Mother and Child, Warsaw, Poland). Material collected from two family members was analyzed: patient RK (analysis of gDNA isolated from saliva, the Oragene kit, DNA Genoteck) and the asymptomatic father AK (analysis of gDNA isolated from peripheral blood collected in an EDTA Vacutainer).

Luciferase reporter assays

Assignment of potentially important microRNAs predicted to change the mRNA expression levels of the NT5C3A gene was performed using miRBase [51]. Human 3′ untranslated region (UTR) of the wild type of the NT5C3A gene (WT -HmiT067108-MT06) and three different mutation variants (M1 with substitution T/C, M2 insertion TCTT, and M1 with both substitution T/C and insertion TCTT; for details, see Additional file 1: Fig. S1.1) of firefly luciferase reporter constructs (HmiT018551-MT06-01/02/03, respectively) and their control vector (Vc) (CmiT000001-MT06) were purchased from GeneCopoeia (Rockville, MD, USA). The post-transcriptional activity of the 3′-UTR regions of the human erythrocyte-specific nucleotidase isoform (P5N-R) was tested by transfecting HEK293T cells with the constructs above described or with control plasmid. Cells were seeded into a 24-well plate and incubated for 24 h at 37 °C until 80% confluency. Cells were transfected with expression plasmids using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. Each well received 300 ng of total plasmid DNA (control vector or WT or mutated sequences M1, M2, M3) as well as vectors: hsa-miR-4775 mimic (MC21381, Thermo Fisher Scientific) or hsa-miR-4775 inhibitor (AM21381, Thermo Fisher Scientific) or hsa-miR-4775 precursor (PM21381, Thermo Fisher Scientific) or the cel-miR-67 scramble control (Thermo Fisher Scientific) at a final concentration of 15 pM. After 48 h of transfection cells were lysed and dual-luciferase reporter assay was performed according to the manufacturer’s protocol (Promega, Madison, WI, USA). Firefly luciferase expression values were normalized to Renilla luciferase expression values. Standard deviations were calculated from independent biological and technical replicates. Results were plotted as a relative decrease in arbitrary light units compared with control cells.

Determination of the pattern of methylation

Peripheral blood samples of four studied family members and five Polish healthy subjects were obtained under the approval of the Ethics Committee of Wroclaw Medical University (study protocol KB-199/2017). DNA was extracted from a 200 µL whole blood sample using QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) following the instructions of the manufacturer. The bisulfite sequencing PCR method (BSP) was used for methylation mapping. EpiJET Bisulfite Conversion Kit (Thermo Fisher Scientific) was used in nine samples using 180–400 ng of gDNA. The CpG-rich amplified fragments of the NT5C3A gene, after the bisulfite conversion of unmethylated cytosines, are then sequenced. PCR amplicons were used for bisulfite conversion and Sanger sequencing. Primer homologous to the converted (NT5C3A_bsDN) and unconverted template (NT5C3A_gDNA) were used for detecting converted amplicons. Primer homologous to the unconverted template (gDNA) were used for amplification of gDNA (Additional file 1: Table S1.4). Primers detecting converted template (NT5C3A_bsDNA) did not include cytosines located within the CpG islands (Additional file 1: Fig. S1.2A). They had homology only to the converted cytosines.

Statistical analysis

GraphPad Prism v. 6.01 and MS Excel software was used to process all data in this work, presented as mean ± standard deviation. Comparisons were performed using Student’s t-test. Statistical significance was accepted as a p-value of < 0.05, used for most analyses besides next-generation sequencing (NGS) analysis where the false discovery rate (FDR)-corrected p-value, i.e., q-value, of < 0.05 was considered as significant [52].